CN117830960A

CN117830960A - Smart community-based risk identification method, device, equipment and medium

Info

Publication number: CN117830960A
Application number: CN202410245318.XA
Authority: CN
Inventors: 赵昕
Original assignee: Bi Shengyun Wuhan Information Technology Co ltd
Current assignee: Bi Shengyun Wuhan Information Technology Co ltd
Priority date: 2024-03-05
Filing date: 2024-03-05
Publication date: 2024-04-05
Anticipated expiration: 2044-03-05
Also published as: CN117830960B

Abstract

The application provides a risk identification method, device, equipment and medium based on an intelligent community, which relate to the technical field of image identification, wherein the method comprises the following steps: according to a guard monitoring video at a community gate, if the fact that the outside of the community gate contains both an owner and a non-owner is determined, prompting the non-owner to provide visitor information; identifying the identity of the owner and determining residence information of the owner; judging whether the access destination of the non-owner person is the same as the residence of the owner person according to the residence information and the destination information; if the access destination is different from the residence, determining target appearance characteristics of non-owner personnel in the guard monitoring video; determining an actual destination to which a non-owner person actually goes through a plurality of monitoring videos; and if the actual destination is inconsistent with the access destination, carrying out early warning prompt on non-owner personnel through the contact information. The method and the system can continuously track and analyze the behaviors of the trailing personnel entering the community.

Description

Smart community-based risk identification method, device, equipment and medium

Technical Field

The application relates to the technical field of image recognition, in particular to a risk recognition method, device, equipment and medium based on an intelligent community.

Background

The intelligent community is a modern community mode for realizing intelligent management and optimization of infrastructure, services and resident life in the community by using advanced information technology and an intelligent system and technical means such as the Internet, sensors, big data and the like. The intelligent community improves the running efficiency of the community through digital and networked means, strengthens interaction between residents and the community, comprises the aspects of intelligent security protection, intelligent energy management, intelligent traffic, convenience service and the like, and improves the life quality of the residents and the sustainable development level of cities.

At present, security facilities of intelligent communities have some problems which are worth focusing, particularly in the aspect of access control systems. Generally, an entrance guard system of a community performs entrance control by adopting a face recognition or card swiping mode. However, there is a case that: when the owner swipes the card, the trailing person can directly enter behind the owner without swiping the card. In this case, if a stranger follows the community after a resident or other authorized person, it will pose a threat to the overall security of the community.

Thus, there is a need for a method to continuously track and analyze the behavior of the trailing people entering the community.

Disclosure of Invention

The application provides a risk identification method, device, equipment and medium based on an intelligent community, which can continuously track and analyze the behaviors of trailing persons entering the community.

In a first aspect of the present application, there is provided a risk identification method based on a smart community, the method comprising:

according to a guard monitoring video at a community gate, if it is determined that the community gate comprises both an owner and a non-owner, prompting the non-owner to provide visitor information, wherein the visitor information comprises contact information and destination information of the non-owner;

identifying the identity of the owner and determining residence information of the owner;

judging whether the access destination of the non-owner person is the same as the residence of the owner person according to the residence information and the destination information;

if the access destination is different from the residence, determining target appearance characteristics of the non-owner in the guard monitoring video;

after the non-owner personnel enter the community gate, determining an actual destination to which the non-owner personnel actually go through a plurality of monitoring videos;

And if the actual destination is inconsistent with the access destination, carrying out early warning prompt on the non-owner personnel through the contact information.

By adopting the technical scheme, the continuous tracking and analysis of the behavior of the trailing personnel entering the community are realized by introducing a continuous monitoring and analysis system into the community. First, the guard monitoring video can provide real-time entrance monitoring to distinguish between the simultaneous presence of owner and non-owner outside the large gate of the community. When this is monitored, indicating that there may be non-owner personnel trailing the owner personnel, the non-owner personnel is prompted to provide visitor information, including contact information and destination information. And then, carrying out identity recognition on the owner personnel to acquire residence information. By comparing the residence information of the owner and the non-owner with the destination information, it is possible to determine whether the access destination of the non-owner coincides with the residence of the owner. If the two are not identical, it is indicated that the owner and non-owner may not recognize, further identifying the non-owner's target appearance characteristics, such as clothing, appearance, etc. Once the non-owner personnel enter the community gate, the behavior of the non-owner personnel is tracked through a plurality of monitoring videos, and particularly the actual destination to which the non-owner personnel actually travel is monitored, so that the action track and the final arriving position of the non-owner personnel are determined. If the actual destination is not consistent with the access destination provided initially, carrying out early warning prompt on non-owner personnel through the contact information acquired before. In the whole process, through continuous monitoring and real-time analysis, behaviors of non-owners in communities are automatically identified, tracked and analyzed, abnormal behaviors are timely found, and corresponding measures are taken.

Optionally, the determining the target appearance characteristic of the non-owner person in the guard monitoring video specifically includes:

determining a plurality of first appearance features of the non-owner personnel according to the entrance guard monitoring video;

acquiring a plurality of preset monitoring videos in a preset time period, wherein the preset time period is a time period of preset duration before the current moment;

determining second appearance characteristics of a plurality of community personnel in the community according to a plurality of preset monitoring videos;

determining the target appearance feature of the plurality of first appearance features, the target appearance feature being different from any one of the plurality of second appearance features.

Through adopting above-mentioned technical scheme, through multi-level appearance characteristic comparison, the target appearance characteristic of non-owner personnel in the accurate assurance door guard monitoring video. First, in a guard monitoring video, a plurality of first appearance features of non-owner personnel, such as clothes, body states and the like, are extracted. And then, acquiring a plurality of preset monitoring videos within a preset time period, wherein the videos cover different appearance characteristics of a plurality of people in the community within a period of time. Through analysis of the preset monitoring videos, second appearance characteristics of a plurality of people in the community can be determined, and a comprehensive appearance characteristic library is constructed. And then, screening out target appearance characteristics from the extracted first appearance characteristics, so as to ensure that the target appearance characteristics are obviously different from the second appearance characteristics of other people in the community, thereby improving the accuracy and uniqueness of the target appearance characteristics of non-owner people. By comparing the characteristic with the appearance characteristics of other people in the community, the characteristic similar to the resident of the normal community can be eliminated, so that the identity characteristics of the trailing people can be determined more reliably. The differentiated feature extraction method is beneficial to reducing false recognition and improving sensitivity to abnormal behaviors, and provides a more accurate and reliable tool for community security management.

Optionally, after the non-owner person enters the community gate, before determining the actual destination to which the non-owner person actually goes through the plurality of monitoring videos, the method further includes:

if the target appearance characteristic is identified in a first monitoring video of a first camera, determining a second camera directly connected with the first camera according to a preset camera topology map, wherein the preset camera topology map comprises a link of a community road and branch nodes of all monitoring cameras, the first camera is any one monitoring camera of a plurality of monitoring cameras in the community, and the second camera is any one monitoring camera of the plurality of monitoring cameras directly connected with the first camera according to the preset camera topology map;

acquiring a second monitoring video shot by the second camera;

identifying the target appearance characteristics of the second monitoring video, and judging whether the target appearance characteristics can be identified;

if the target appearance characteristics are identified from the second monitoring video, acquiring the position information of the second camera;

and determining the position information as the position corresponding to the non-owner personnel.

By adopting the technical scheme, through intelligent analysis of a plurality of monitoring videos, comprehensive tracking of the track of non-owner personnel entering the community is realized, and the actually-going destination is accurately determined. In addition, through presetting the camera topology map, only the monitoring videos of part of cameras are analyzed and identified according to the community roads and the camera layout in the map, but not all the monitoring videos of the community are analyzed and identified. Thus, the calculation amount brought by video recognition is greatly reduced, and the recognition efficiency is improved.

Optionally, after the target appearance feature is identified from the second surveillance video, the method further includes:

when the fact that the non-owner person passes through the second camera by the first camera is determined, a third camera is determined according to the preset camera topological graph, wherein the third camera is any one monitoring camera except the first camera among a plurality of monitoring cameras directly connected with the second camera;

determining a first end camera directly connected or indirectly connected with the third camera according to the preset camera topology map, wherein the preset camera topology map further comprises a terminal camera serving as a terminal node and a residential building of a community serving as a leaf node, the terminal node is directly connected with the leaf node, and the first end camera is any one of a plurality of terminal cameras;

Determining a first residential building directly connected with the first terminal camera from a plurality of residential buildings contained in the community according to the preset camera topology map;

according to visitor information, if the first residential building is determined to be a destination residential building corresponding to the destination information, calculating first probability that the non-owner personnel pass through the shooting area of the third camera from the second camera, wherein the first probability is calculated specifically by the following formula:

；

wherein P (X) ₂ |X ₃ ) For the first probability that the non-owner person passes through the third camera shooting area from the second camera, lambda is a preset attenuation coefficient, d (3, a) is the distance between the third camera and the first terminal camera, n is the number of the monitoring cameras directly connected with the second camera, d (i, a) is the distance between the ith monitoring camera directly connected with the second camera and the first terminal camera, and P ₀ Target probability of shooting areas for the non-owner personnel without passing through the plurality of cameras directly connected with the second camera;

determining a first number of cameras between the community gate and the second camera according to the preset camera topology map;

Determining the second number of the cameras corresponding to the monitoring video of the target appearance feature detected in the cameras between the community gate and the second camera;

and determining the target probability according to the quotient of the second quantity and the first quantity.

By adopting the technical scheme, the above formula is that by using an exponential decay function, when the distance between the third camera and the first terminal camera is longer, the distance from the current position of the non-owner person to the provided terminal residential building is longer, and the probability that the non-owner person arrives there is smaller. Because the farther this distance is, the less likely it is that a non-owner will pass the corresponding third camera to go to his own provided endpoint.

Optionally, after the calculating the first probability that the non-owner person passes from the second camera through the third camera shooting area, the method further includes:

when the fact that the non-owner person passes through the second camera by the first camera is determined, a fourth camera is determined according to the preset camera topology map, wherein the fourth camera is any one monitoring camera except the first camera and the third camera among a plurality of monitoring cameras directly connected with the second camera;

Determining a second end camera which is directly connected or indirectly connected with the third camera according to the preset camera topology map, wherein the second end camera is any one of a plurality of end cameras;

determining a second residential building connected with the second terminal camera from a plurality of residential buildings according to the preset camera topology map;

if the second residential building is determined to be the terminal residential building, calculating a second probability that the non-owner personnel pass through the fourth camera shooting area from the second camera;

judging the magnitude relation between the first probability and the second probability, and if the first probability is determined to be larger than the second probability, preferentially identifying the target appearance characteristics of the monitoring video shot by the third camera.

By adopting the technical scheme, the magnitude relation between the first probability and the second probability is compared, so that the magnitude relation between the probability that a non-owner person possibly passes through the third camera and the probability that the non-owner person passes through the fourth camera is predicted. And then determining the monitoring video shot by the monitoring camera with higher priority selection probability to carry out target appearance feature recognition. According to the intelligent camera selection strategy, the non-owner personnel can be further predicted in efficiency by comparing the probabilities that the non-owner personnel pass through different monitoring cameras.

Optionally, determining, by using a plurality of monitoring videos, an actual destination to which the non-owner personnel actually goes specifically includes:

acquiring the time of each terminal monitoring video containing the appearance characteristics of the target, wherein the terminal monitoring video is a video shot by the terminal camera;

sequencing the terminal monitoring videos according to a time sequence, and determining a third monitoring video in the terminal monitoring videos, wherein the third monitoring video is the monitoring video with the latest time in the terminal monitoring videos;

determining the distance between each residential building and the third terminal camera, wherein the third monitoring video is terminal monitoring video shot by the third terminal camera;

sequencing the distances between a plurality of buildings and the third-end camera according to the order of the sizes, and determining a second residential building with the shortest distance between the second residential building and the third-end camera;

and determining the second residential building as the actual destination.

By adopting the technical scheme, firstly, the time information of each terminal monitoring video containing the target appearance characteristics is acquired. By ordering these end monitoring videos in time order, it is possible to quickly locate time periods of non-owner personnel activity. After sequencing, determining a third monitoring video with the latest time in the plurality of terminal monitoring videos, selecting a third residential building directly connected with the third terminal camera as an actual destination according to a third terminal camera corresponding to the third monitoring video and a preset camera topology map. This choice is to ensure that the latest location information is obtained to reflect the actual destination to which non-owner personnel may ultimately be going.

Optionally, according to the guard monitoring video, if it is determined that the outside of the large gate of the community contains both an owner and a non-owner, the non-owner is prompted to provide visitor information, and before the method further includes:

detecting human body targets of the guard monitoring video to obtain a plurality of human body targets;

respectively identifying face areas corresponding to all human targets, and judging whether owner information can be identified;

if the owner information can be identified, determining the corresponding human body target as the owner personnel;

and if the owner information can not be identified, determining the corresponding human body target as the non-owner person.

By adopting the technical scheme, firstly, a guard monitoring video is acquired, human body target detection is carried out on the video, and a plurality of human body targets are obtained. Then, the face area corresponding to each human target is respectively identified, so as to judge whether the information of the owners can be identified. If the owner information is successfully identified, namely that the identity corresponding to the face is confirmed to be the community owner, the human body target is confirmed to be the owner person. In contrast, if the owner information cannot be identified in the face area, it is determined that the human target is a non-owner person. The technical scheme has the effect that people outside the community gate are rapidly and accurately classified into two categories of owners and non-owners through the face recognition technology. By effectively identifying the identity information of the owner, community management is able to more accurately perform subsequent security management measures on non-owner personnel, such as requiring the non-owner to provide visitor information. Overall, this scheme helps improving the real-time monitoring and the discernment ability of community gate, reinforcing community security.

In the second aspect of the application, a risk recognition device based on an intelligent community is provided, which comprises a judgment module, a recognition module, a detection module and an early warning module, wherein:

the judging module is used for prompting the non-owner personnel to provide visitor information according to a guard monitoring video at a community gate, and if the fact that the community gate is outside and simultaneously contains the owner personnel and the non-owner personnel is determined, the visitor information comprises contact information and destination information of the non-owner personnel;

the identification module is used for carrying out identity identification on the owner personnel and determining residence information of the owner personnel;

the judging module is used for judging whether the access destination of the non-owner person is the same as the residence of the owner person according to the residence information and the destination information;

the detection module is used for determining target appearance characteristics of the non-owner personnel in the guard monitoring video if the access destination is determined to be different from the residence;

the judging module is used for determining an actual destination to which the non-owner personnel actually go through a plurality of monitoring videos after the non-owner personnel enter the community gate;

And the early warning module is used for carrying out early warning prompt on the non-owner personnel through the contact information if the actual destination is not consistent with the access destination.

Optionally, the identification module is configured to determine a plurality of first appearance features of the non-owner person according to the guard monitoring video;

the judging module is used for acquiring a plurality of preset monitoring videos in a preset time period, wherein the preset time period is a time period of preset duration before the current moment;

the identification module is used for determining second appearance characteristics of a plurality of community personnel in the community according to a plurality of preset monitoring videos;

the detection module is used for determining the target appearance characteristics in the first appearance characteristics, wherein the target appearance characteristics are different from any one of the second appearance characteristics.

Optionally, the determining module is configured to determine, if the target appearance feature is identified in a first surveillance video of a first camera, a second camera directly connected to the first camera according to a preset camera topology map, where the preset camera topology map includes links with a community road and each surveillance camera as a branch node, the first camera is any one surveillance camera of a plurality of surveillance cameras in the community, and the second camera is any one surveillance camera of a plurality of surveillance cameras directly connected to the first camera according to the preset camera topology map;

The judging module is used for acquiring a second monitoring video shot by the second camera;

the identification module is used for identifying the target appearance characteristics of the second monitoring video and judging whether the target appearance characteristics can be identified;

the identification module is used for acquiring the position information of the second camera if the target appearance characteristic is identified from the second monitoring video;

and the judging module is used for determining the position information as the position corresponding to the non-owner personnel.

Optionally, the detection module is configured to determine, when it is determined that the non-owner person passes through the second camera by the first camera, a third camera according to the preset camera topology map, where the third camera is any one of a plurality of monitoring cameras directly connected to the second camera, except for the first camera;

the judging module is configured to determine a first end camera directly connected or indirectly connected to the third camera according to the preset camera topology map, where the preset camera topology map further includes an end camera serving as an end node and a residential building of a community serving as a leaf node, the end node is directly connected to the leaf node, and the first end camera is any one of the end cameras;

The judging module is used for determining a first residential building directly connected with the first terminal camera from a plurality of residential buildings contained in the community according to the preset camera topology map;

the judging module is configured to calculate, according to visitor information, a first probability that the non-owner person passes through the third camera shooting area from the second camera if it is determined that the first residential building is a destination residential building corresponding to the destination information, and specifically calculate by the following formula:

；

the judging module is used for determining first numbers of a plurality of cameras between the community gate and the second camera according to the preset camera topology map;

The judging module is used for determining the second number of the cameras corresponding to the monitoring video of the target appearance characteristic among the cameras between the community gate and the second camera;

the judging module is used for determining the target probability according to the quotient of the second quantity and the first quantity.

Optionally, the detection module is configured to determine, when it is determined that the non-owner person passes through the second camera by the first camera, a fourth camera according to the preset camera topology map, where the fourth camera is any one of a plurality of monitoring cameras directly connected to the second camera, except for the first camera and the third camera;

the judging module is used for determining a second end camera which is directly connected or indirectly connected with the third camera according to the preset camera topology graph, wherein the second end camera is any one end camera of a plurality of end cameras;

the judging module is used for determining a second residential building connected with the second terminal camera from a plurality of residential buildings according to the preset camera topology map;

The judging module is used for calculating a second probability that the non-owner personnel pass through the fourth camera shooting area from the second camera if the second residential building is determined to be the terminal residential building;

the judging module is configured to judge a magnitude relation between the first probability and the second probability, and if it is determined that the first probability is greater than the second probability, preferentially identify the target appearance feature for the monitoring video shot by the third camera.

Optionally, the judging module is configured to obtain time of each terminal monitoring video including the appearance feature of the target, where the terminal monitoring video is a video shot by the terminal camera;

the detection module is used for sequencing the terminal monitoring videos according to a time sequence, and determining a third monitoring video in the terminal monitoring videos, wherein the third monitoring video is the monitoring video with the latest time in the terminal monitoring videos;

the detection module is used for determining a third terminal camera corresponding to the third monitoring video;

the detection module is used for determining a third residential building connected with the third terminal camera from a plurality of residential buildings according to the preset camera topology map;

And the early warning module is used for determining the third residential building as the actual destination.

Optionally, the detection module is configured to perform human body target detection on the guard monitoring video to obtain a plurality of human body targets;

the detection module is used for respectively identifying the face areas corresponding to the human targets and judging whether the owner information can be identified;

the judging module is used for determining that the corresponding human body target is the owner person if the owner information can be identified;

and the judging module is used for determining that the corresponding human body target is the non-owner person if the owner information cannot be identified.

In summary, one or more technical solutions provided in the embodiments of the present application at least have the following technical effects or advantages:

continuous tracking and analysis of behavior of trailing persons entering the community is achieved by introducing a continuous monitoring and analysis system in the community. First, the guard monitoring video can provide real-time entrance monitoring to distinguish between the simultaneous presence of owner and non-owner outside the large gate of the community. When this is monitored, indicating that there may be non-owner personnel trailing the owner personnel, the non-owner personnel is prompted to provide visitor information, including contact information and destination information. And then, carrying out identity recognition on the owner personnel to acquire residence information.

By comparing the residence information of the owner and the non-owner with the destination information, it is possible to determine whether the access destination of the non-owner coincides with the residence of the owner. If the two are not identical, it is indicated that the owner and non-owner may not recognize, further identifying the non-owner's target appearance characteristics, such as clothing, appearance, etc. Once the non-owner personnel enter the community gate, the behavior of the non-owner personnel is tracked through a plurality of monitoring videos, and particularly the actual destination to which the non-owner personnel actually travel is monitored, so that the action track and the final arriving position of the non-owner personnel are determined. If the actual destination is not consistent with the access destination provided initially, carrying out early warning prompt on non-owner personnel through the contact information acquired before. In the whole process, through continuous monitoring and real-time analysis, behaviors of non-owners in communities are automatically identified, tracked and analyzed, abnormal behaviors are timely found, and corresponding measures are taken.

Drawings

Fig. 1 is a schematic flow chart of a risk identification method based on an intelligent community according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a community route design and monitoring layout as disclosed in an embodiment of the present application;

FIG. 3 is a schematic diagram of a preset camera topology diagram disclosed in an embodiment of the present application;

FIG. 4 is a schematic block diagram of a risk identification device based on a smart community according to an embodiment of the present application;

fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Reference numerals illustrate: 401. a judging module; 402. an identification module; 403. a detection module; 404. an early warning module; 501. a processor; 502. a communication bus; 503. a user interface; 504. a network interface; 505. a memory.

Detailed Description

In order to make those skilled in the art better understand the technical solutions in the present specification, the technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification, and it is obvious that the described embodiments are only some embodiments of the present application, but not all embodiments.

In the description of embodiments of the present application, words such as "for example" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described herein as "such as" or "for example" should not be construed as preferred or advantageous over other embodiments or designs. Rather, the use of words such as "or" for example "is intended to present related concepts in a concrete fashion.

In the description of the embodiments of the present application, the term "plurality" means two or more. For example, a plurality of systems means two or more systems, and a plurality of screen terminals means two or more screen terminals. Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating an indicated technical feature. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless expressly specified otherwise.

Thus, there is a need for a method to continuously track and analyze the behavior of trailing persons.

The embodiment discloses a risk identification method based on an intelligent community, referring to fig. 1, comprising the following steps S110-S150:

and S110, according to the guard monitoring video at the gate of the community, if the fact that the gate of the community is outside and contains both the owner and the non-owner is determined, prompting the non-owner to provide visitor information.

The risk identification method based on the intelligent community is applied to a server, and the server comprises but is not limited to electronic equipment such as a mobile phone, a tablet personal computer, wearable equipment, a PC (Personal Computer ) and the like, and can also be a background server for running the risk identification method based on the intelligent community. The server may be implemented as a stand-alone server or as a server cluster composed of a plurality of servers.

The monitoring for face recognition, namely the entrance guard monitoring, is arranged at a community gate of the community to recognize the identity of a person about to enter the community, and whether the person is an owner is judged. After the server acquires the guard monitoring video shot by the guard monitoring, the video frame is processed by using a target detection algorithm so as to detect a human body target. A common target detection algorithm includes YOLO (You Only Look Once). For each detected human target, its head region is extracted and then identified using face recognition techniques. And comparing the face image with an owner information database. The database stores the face images of the owners and related information, such as the names, addresses, etc. of the owners. If the matching is successful, namely the owner information is identified, the corresponding human body target is determined to be the owner personnel. If the matching is impossible, namely the owner information can not be identified, the corresponding human body target is determined to be a non-owner person.

And acquiring a guard monitoring video, and detecting human body targets on the video to obtain a plurality of human body targets. And respectively identifying the face area corresponding to each human body target, and judging whether the information of the owner can be identified. If the owner information is successfully identified, namely that the identity corresponding to the face is confirmed to be the community owner, the human body target is confirmed to be the owner person. In contrast, if the owner information cannot be identified in the face area, it is determined that the human target is a non-owner person. People outside the community gate are divided into two categories, namely owners and non-owners, and follow-up safety management measures can be carried out on the non-owners more accurately.

Firstly, acquiring a guard monitoring video, and detecting human body targets on the video to obtain a plurality of human body targets. Then, the face area corresponding to each human target is respectively identified, so as to judge whether the information of the owners can be identified. If the owner information is successfully identified, namely that the identity corresponding to the face is confirmed to be the community owner, the human body target is confirmed to be the owner person. In contrast, if the owner information cannot be identified in the face area, it is determined that the human target is a non-owner person. The technical scheme has the effect that people outside the community gate are rapidly and accurately classified into two categories of owners and non-owners through the face recognition technology. By effectively identifying the identity information of the owner, community management is able to more accurately perform subsequent security management measures on non-owner personnel, such as requiring the non-owner to provide visitor information. Overall, this scheme helps improving the real-time monitoring and the discernment ability of community gate, reinforcing community security.

If the server determines that the community is outside the gate and contains both the owner and the non-owner, in order to prevent the non-owner from trailing into the community, the prompt information may be presented by voice, text or other suitable means, informing the non-owner that the visitor information needs to be provided. After the non-owner personnel are prompted, the guard monitoring system can provide an interface or channel for inputting visitor information including contact information and destination information, namely, the contact information of the non-owner personnel and which unit needs to be reached and which layer.

S120, identity recognition is carried out on the owner personnel, and residence information of the owner personnel is determined.

The identity of the person who is about to enter the community is identified through the entrance guard monitoring, and whether the person is an owner or not is judged. And comparing the identified face image with the owner information database, and if the face image is successfully matched with the owner information database, namely the owner information is identified, determining that the corresponding human body target is an owner person. And searching out residence information of the corresponding owner personnel according to the database, namely determining which unit and layer the owner personnel resides in.

S130, judging whether the access destination of the non-owner person is the same as the residence of the owner person according to the residence information and the destination information.

And judging whether the building is in the same unit and the same floor according to the residence information of the identified owner personnel and the destination information input by the non-owner personnel. A non-owner person may be allowed access if the non-owner person's access destination is the same as the residence of the owner person, indicating that the non-owner person and the owner person may be friends.

After determining that the access destination of the non-owner person is different from the residence, the identity information of the non-owner person is obtained through visitor information input by the non-owner person, wherein the identity information can comprise a name, an identity card number, a license plate number and the like. And judging the blacklist of the acquired identity information, wherein the server can maintain a preset blacklist which contains the identity information of the personnel which are not allowed to enter the community. Judging whether the obtained identity information is in a blacklist, if the obtained identity information is not in the blacklist, inquiring a resident personnel list by the server to check whether the obtained identity information is contained, wherein the resident personnel list may comprise the identity information of personnel such as couriers, takeouts or maintenance personnel. If the identity information is in the resident list, the server confirms that the identity information is a resident of the community, and the server can allow the resident to enter the community.

If the blacklist contains the identity information of the non-owner personnel or the resident personnel list does not contain the identity information of the non-owner personnel, corresponding safety measures are triggered, including prompting the non-owner personnel to refuse to enter the system, and the like, and meanwhile, the system interface, the voice prompt, the short message notification and the like can be performed to remind a guard or related security personnel of paying attention to the existence of the non-owner personnel and taking corresponding measures.

After judging whether the preset blacklist contains non-owner personnel identity information, a real-time safety early warning and information notification mechanism is introduced, so that the safety management capability of communities is enhanced. If it is determined that the identity information of the non-owner person exists in the preset blacklist or it is determined that the corresponding identity information cannot be found in the resident person list, immediately taking a safety precaution measure. The effect of this early warning alert mechanism is to respond quickly to potential security risks. By sending the early warning information to security personnel, timely monitoring and disposal of non-owner personnel can be achieved, so that potential threats are prevented.

And S140, if the access destination is determined to be different from the residence, determining target appearance characteristics of non-owner personnel in the guard monitoring video.

If the access destination of the non-owner person is different from the residence of the owner person, indicating that the non-owner person is not related to the owner person, the non-owner person may be trailing the owner person into the cell, and the server extracts a plurality of first appearance features of the non-owner person from the entrance guard monitoring video. These features may include, but are not limited to: clothing features (color, style, etc.), hair features (color, length, etc.), body features (height, body shape, etc.), other identifiable features (bag, hand carried item, etc.).

The server acquires a plurality of preset monitoring videos in a preset time period, wherein the time period is preset time before the current moment, and the specific time of the preset time can be determined according to specific requirements. The server extracts second appearance characteristics of a plurality of community staff, such as clothes, hairstyles, bodies and the like of the community staff, from a plurality of preset monitoring videos. The server determines a target appearance feature, that is, a feature different from any one of the plurality of second appearance features, from among the plurality of first appearance features. This may be achieved by image processing and feature extraction algorithms. Computer vision techniques, such as feature point extraction and matching, may be used to compare the target appearance feature to the second appearance feature of the community person. For the target appearance feature and the second appearance feature of community personnel, a similarity measure, such as cosine similarity or euclidean distance, may be used to evaluate the degree of difference between them. Setting a threshold, and when the difference degree between the target appearance characteristic and the second appearance characteristic of any community personnel exceeds the threshold, determining that the target appearance characteristic is different from the second appearance characteristic of the community personnel, and finding out target appearance characteristics different from all the second appearance characteristics for subsequent behavior tracking of non-owner personnel.

And through multi-level appearance characteristic comparison, the target appearance characteristics of non-owner personnel in the guard monitoring video are accurately determined. First, in a guard monitoring video, a plurality of first appearance features of non-owner personnel, such as clothes, body states and the like, are extracted. And then, acquiring a plurality of preset monitoring videos within a preset time period, wherein the videos cover different appearance characteristics of a plurality of people in the community within a period of time. Through analysis of the preset monitoring videos, second appearance characteristics of a plurality of people in the community can be determined, and a comprehensive appearance characteristic library is constructed. And then, screening out target appearance characteristics from the extracted first appearance characteristics, so as to ensure that the target appearance characteristics are obviously different from the second appearance characteristics of other people in the community, thereby improving the accuracy and uniqueness of the target appearance characteristics of non-owner people. By comparing the characteristic with the appearance characteristics of other people in the community, the characteristic similar to the resident of the normal community can be eliminated, so that the identity characteristics of the trailing people can be determined more reliably. The differentiated feature extraction method is beneficial to reducing false recognition and improving sensitivity to abnormal behaviors, and provides a more accurate and reliable tool for community security management.

And S150, after the non-owner personnel enter the community gate, determining the actual destination to which the non-owner personnel actually go through a plurality of monitoring videos.

After the unique target appearance characteristics of the non-owner personnel are determined, the server automatically tracks the behaviors of the non-owner personnel through a plurality of follow-up first monitoring videos for monitoring, and identifies the actual destination of the non-owner personnel.

The application also discloses a preset camera topology map, according to the community road design and the camera layout, the position of each camera on the community route is obtained, the camera is used as a branch node, the community road is used as a link, and all cameras are connected according to the position. Referring to fig. 2, according to the community road and the camera layout, the camera a and the camera b are directly connected through a section of road, no other cameras are installed between the camera a and the camera b, the camera a and the camera b are connected as branch nodes through a link, and the connection modes between the other cameras are the same. Therefore, based on the layout and the connection relationship, referring to fig. 3, a preset camera topology map is generated.

The preset camera topology map comprises a connection relation between the cameras and a residential building, and further comprises a connection relation between the cameras and the residential building. For partial cameras, the partial cameras are arranged at the first layer of the doorway of the residential building and are adjacent to the residential building, the cameras are tail end cameras, the tail end cameras are used as tail end nodes in a preset camera topology graph, the residential building is used as branch and leaf nodes, the tail end nodes are in direct connection with the branch and leaf nodes, and the tail end nodes are also connected with the branch nodes.

After non-owner personnel enter the community through a community gate, according to a preset camera topology graph, after the target appearance characteristics of the non-owner personnel are detected by the monitoring video shot by any one monitoring camera in the community, the surface non-owner personnel pass through the shooting area of the monitoring camera. Further, whether the monitoring video shot by other monitoring cameras in the subsequent time can also identify the target appearance characteristics is required to be identified, so that the action track of non-owner personnel is determined according to the positions of the monitoring cameras corresponding to the monitoring video with the identified target appearance characteristics. However, in order to reduce the calculation amount when the target appearance characteristic recognition is performed, because the calculation amount required for simultaneously recognizing the monitoring videos of all the monitoring cameras in the cell is large, the monitoring cameras possibly passed by non-owners can be predicted based on the preset camera topology map, so that only the monitoring videos of part of the monitoring cameras need to be recognized, and the monitoring videos of all the monitoring cameras do not need to be recognized, so that the calculation amount brought by the recognition is reduced.

The first camera is illustrated with any one of a plurality of monitoring cameras. If the target appearance characteristic is identified in the first monitoring video of the first camera, according to a preset monitoring camera topology map, only the monitoring videos of a plurality of monitoring cameras directly connected with the first camera are needed to be identified. Referring to fig. 2, when a non-owner person passes through an area photographed by the monitoring camera d, then necessarily passes through an area photographed by any one of the monitoring cameras b, g, h, and i. According to the preset camera topology map, a non-owner person can directly connect any one of the areas shot by the plurality of monitoring cameras through the monitoring camera d, and then only needs to identify the monitoring video shot by the monitoring camera b, the monitoring camera g, the monitoring camera h and the monitoring camera i, so as to judge whether the target appearance characteristics can be identified.

Any one of a plurality of cameras directly connected with the first camera is used for monitoring the camera, and the second camera is used for illustration. And acquiring a second monitoring video shot by a second camera, then identifying the appearance characteristics of the target for the second monitoring video, and judging whether the appearance characteristics of the target can be identified. When the target appearance characteristics of the non-owner personnel are detected from the second monitoring video of the second camera, the non-owner personnel are indicated to pass through the shooting area of the second monitoring camera, the server acquires the position information of the second camera, and the acquired position information is determined to be the position of the non-owner personnel corresponding to the shooting time.

Through intelligent analysis of a plurality of monitoring videos, comprehensive tracking of the track of non-owner personnel entering the community is realized, and the actual destination of the non-owner personnel is accurately determined. In addition, through presetting the camera topology map, only the monitoring videos of part of cameras are analyzed and identified according to the community roads and the camera layout in the map, but not all the monitoring videos of the community are analyzed and identified. Thus, the calculation amount brought by video recognition is greatly reduced, and the recognition efficiency is improved.

Further, when a non-owner person passes through the shooting areas of the plurality of monitoring cameras and recognizes the target appearance characteristics from the monitoring recognition of the monitoring cameras, the probability that the non-owner person passes through different monitoring cameras can be calculated, so that the monitoring video of the monitoring camera with higher probability is preferentially recognized, and the recognition efficiency of the target appearance characteristics is improved.

In the above example, when it is determined that the non-owner person passes through the second camera by the first camera, a plurality of monitoring cameras except the first camera among the plurality of cameras directly connected to the second camera are determined according to the preset camera topology map, and then the probabilities that the non-owner person will pass through the plurality of monitoring cameras are calculated respectively. And among the plurality of monitoring cameras directly connected with the second camera, any monitoring camera except the first camera is exemplified by the third monitoring camera.

According to a preset camera topology diagram, a first end camera which is directly connected or indirectly connected with a third camera is determined, wherein the first end camera is any one end camera of a plurality of end cameras. According to a preset camera topology map, a first residential building which is directly connected with a first terminal camera is determined in a plurality of residential buildings contained in a community, and according to access information filled in by non-owner personnel at a gate of the community, the terminal residential building to which the non-owner personnel will go is the first residential building, for example, the access information of the non-owner personnel shows that the non-owner personnel will visit 3 residential buildings, and the first residential building is 3 east, and the first residential building and the second residential building are the first residential buildings. Then, calculating a first probability of a non-owner person passing through the shooting area of the third camera from the second camera according to the following formula:

；

Wherein P (X) ₂ |X ₃ ) For the first probability that a non-owner passes through a shooting area of the third camera from the second camera, lambda is a preset attenuation coefficient, d (3, a) is the distance between the third camera and the first end camera, n is the number of a plurality of monitoring cameras directly connected with the second camera, d (i, a) is the distance between the ith monitoring camera directly connected with the second camera and the first end camera, and P ₀ Target probability for a plurality of camera shooting areas which are not directly connected by a second camera for non-owner personnel

The numerator portion of the above formula is an exponential decay function, where e is the base of the natural logarithm (approximately equal to 2.71828), which gives a smooth probability estimate based on distance, and the further the distance between the third camera and the first end camera, the further the non-owner is from the current location to the provided end-point residential building, and the less the non-owner has arrived there. Because the farther this distance is, the less likely it is that a non-owner will pass the corresponding third camera to go to his own provided endpoint.

For any monitoring of the distance between two cameras, the definition of the application is as follows, and according to a preset camera topology graph, there may be a plurality of paths between any two cameras, each path including at least one branch node or including at least one link. And respectively calculating the total length of the community roads corresponding to at least one link contained in each path. And then sequencing the total length of the plurality of paths, and selecting the shortest total length as the distance between the two cameras. Similarly, for the distance between the third camera and the first end camera, the total length of the community road corresponding to at least one link included in each of the multiple paths between the third camera and the first end camera needs to be determined. And then sequencing the total lengths of the paths, and selecting the shortest total length as the distance between the third camera and the first end camera. The distance between the ith camera and the first terminal camera which are directly connected with the second camera is calculated in the same way.

For the target probability that non-owner personnel can not pass through a plurality of cameras of second camera direct connection and shoot the region, the calculation mode of this application is as follows: first, according to a preset camera topology map, determining first numbers of a plurality of cameras between a community gate and a second camera. And then determining the second number of the cameras corresponding to the monitoring video with the detected target appearance characteristics among the cameras between the community gate and the second camera. If the user walks on the community road, the first number is the same as the second number, and if the user does not walk on the community road, the second number may be smaller than the first number if a part of the road bypasses the photographing region of the monitoring camera. When the second number is larger, behaviors of non-owner personnel on the surface are more suspicious, the more cameras are avoided, and the larger the target probability is caused, the lower the probability of passing through the third camera is. The target probability of the plurality of camera shooting areas that the non-owner person would not be directly connected through the second camera is calculated by the following formula:

；

wherein P is ₀ For the target probability, N ₁ For a first number, N ₂ A second number.

In the above calculation formula of the first probability, λ is a preset attenuation coefficient for tuning The influence of the whole distance on the probability. If λ is larger, the larger the influence of the distance on the probability. And the value of the preset attenuation coefficient is related to the target probability, and the preset attenuation coefficient is in a proportional relation with the target probability. Because the target probability reflects the accuracy of formula prediction, the attenuation coefficient is affected. In the calculation formula of the first probability, P is divided by the denominator ₀ And the normalization factor is used for calculating the sum of probabilities of a plurality of monitoring cameras directly connected by non-owner personnel through the second camera.

And similarly, determining any one monitoring camera except the first camera and the third camera from a plurality of monitoring cameras directly connected with the second camera according to a preset camera topology map, and a fourth camera. And then determining a second end camera which is directly connected or indirectly connected with the third camera according to a preset camera topology map, wherein the second end camera is any one of a plurality of end cameras. And determining a second residential building connected with the second terminal camera from the plurality of residential buildings according to the preset camera topology map. If it is determined that the second residential building is also the terminal residential building in the access information of the non-owner, calculating a second probability that the non-owner passes through the fourth camera shooting area from the second camera. The calculation method is as above and will not be further described here. And then judging the magnitude relation between the first probability and the second probability, if the first probability is determined to be larger than the second probability, preferentially identifying the target appearance characteristics of the monitoring video shot by the third camera, and then identifying the monitoring video shot by the fourth camera. If the first probability is smaller than the second probability, the monitoring video shot by the fourth camera is preferentially identified in the target appearance characteristic, and then the monitoring video shot by the third camera is identified.

And comparing the magnitude relation between the first probability and the second probability, and predicting the magnitude relation between the probability that a non-owner person possibly passes through the third camera and the probability that the non-owner person passes through the fourth camera. And then determining the monitoring video shot by the monitoring camera with higher priority selection probability to carry out target appearance feature recognition. According to the intelligent camera selection strategy, the non-owner personnel can be further predicted in efficiency by comparing the probabilities that the non-owner personnel pass through different monitoring cameras.

For an end monitoring video shot by an end camera and containing target appearance characteristics of non-owner personnel, time information of the end monitoring video is obtained from a video file or a monitoring system. This may be a time stamp at which the video starts recording, or time information in the video file. And sequencing the end monitoring videos containing the target appearance characteristics according to the time information of the end monitoring videos so as to ensure the chronological order. And selecting the video with the latest time from the sequenced terminal monitoring videos as a third monitoring video, wherein the third monitoring video is the video with the appearance characteristics of the target detected last time. And then determining a third terminal camera corresponding to the third monitoring video, and determining a third residential building connected with the third terminal camera from the plurality of residential buildings according to a preset camera topology map. And finally determining that the third residential building is the actual destination visited by non-owner personnel.

First, time information of each end monitoring video including the appearance characteristics of the target is acquired. By ordering these end monitoring videos in time order, it is possible to quickly locate time periods of non-owner personnel activity. After sequencing, determining a third monitoring video with the latest time in the plurality of terminal monitoring videos, selecting a third residential building directly connected with the third terminal camera as an actual destination according to a third terminal camera corresponding to the third monitoring video and a preset camera topology map. This choice is to ensure that the latest location information is obtained to reflect the actual destination to which non-owner personnel may ultimately be going.

And S160, if the actual destination is not consistent with the access destination, carrying out early warning prompt on non-owner personnel through the contact information.

And comparing the actual destination finally reached by the non-owner personnel with the access destination filled in the gate of the community, judging whether the two are consistent, and if so, indicating that the non-owner personnel act normally, and temporarily eliminating the need for subsequent processing. If the actual destination is inconsistent with the access destination, the fact that the behaviors of the non-owner personnel are suspicious is indicated, the server carries out early warning prompt on the non-owner personnel through contact information filled in by the non-owner personnel, and the server can send a short message to a mobile phone of the non-owner personnel through a filled mobile phone number to prompt, and can further prompt related personnel including community management personnel, security personnel and the like.

The embodiment also discloses a risk identification device based on the smart community, referring to fig. 4, including a judging module 401, an identifying module 402, a detecting module 403 and an early warning module 404, wherein:

the judging module 401 is configured to prompt the non-owner personnel to provide visitor information according to the guard monitoring video at the gate of the community, if it is determined that the gate of the community includes both the owner personnel and the non-owner personnel, the visitor information includes contact information of the non-owner personnel and destination information.

And the identification module 402 is used for carrying out identification on the owner personnel and determining residence information of the owner personnel.

A judging module 401, configured to judge whether the access destination of the non-owner person is the same as the residence of the owner person according to the residence information and the destination information.

And the detection module 403 is configured to determine a target appearance characteristic of a non-owner in the guard monitoring video if it is determined that the access destination is different from the residence.

The judging module 401 is configured to determine, through a plurality of monitoring videos, an actual destination to which the non-owner personnel actually goes after the non-owner personnel enters the community gate.

And the early warning module 404 is used for carrying out early warning prompt on non-owner personnel through the contact information if the actual destination is not consistent with the access destination.

In one possible implementation, the identification module 402 is configured to determine a plurality of first appearance characteristics of non-owner personnel based on the concierge surveillance video.

The judging module 401 is configured to obtain a plurality of preset monitoring videos in a preset time period, where the preset time period is a time period of a preset duration before the current time.

The identification module 402 is configured to determine second appearance characteristics of a plurality of community personnel in the community according to a plurality of preset monitoring videos.

A detection module 403, configured to determine a target appearance feature of the plurality of first appearance features, where the target appearance feature is different from any one of the plurality of second appearance features.

In a possible implementation manner, the determining module 401 is configured to determine, if the target appearance feature is identified in the first surveillance video of the first camera, according to a preset camera topology map, a second camera directly connected to the first camera, where the preset camera topology map includes links with community roads and branch nodes as each surveillance camera, the first camera is any one surveillance camera among multiple surveillance cameras in the community, and according to the preset camera topology map, the second camera is any one surveillance camera among multiple surveillance cameras directly connected to the first camera.

The judging module 401 is configured to obtain a second monitoring video shot by the second camera.

And the identifying module 402 is configured to identify the target appearance feature of the second surveillance video, and determine whether the target appearance feature can be identified.

And the identifying module 402 is configured to acquire the position information of the second camera if the target appearance feature is identified from the second surveillance video.

The judging module 401 is configured to determine that the location information is a location corresponding to a non-owner person.

In a possible implementation manner, the detection module 403 is configured to determine, when it is determined that the non-owner person passes through the second camera by the first camera, a third camera according to a preset camera topology map, where the third camera is any one of the plurality of monitoring cameras directly connected to the second camera, except the first camera.

The judging module 401 is configured to determine, according to a preset camera topology map, a first end camera directly connected or indirectly connected to the third camera, where the preset camera topology map further includes an end camera serving as an end node, and a residential building of a community serving as a leaf node, the end node is directly connected to the leaf node, and the first end camera is any one of the plurality of end cameras.

The judging module 401 is configured to determine, according to a preset camera topology map, a first residential building directly connected to a first terminal camera from a plurality of residential buildings included in the community.

The judging module 401 is configured to calculate, according to the visitor information, a first probability that a non-owner person passes through a third camera shooting area from the second camera if it is determined that the first residential building is a destination residential building corresponding to the destination information, specifically by the following formula:

wherein P (X) ₂ |X ₃ ) For the first probability that a non-owner passes through a shooting area of the third camera from the second camera, lambda is a preset attenuation coefficient, d (3, a) is the distance between the third camera and the first end camera, n is the number of a plurality of monitoring cameras directly connected with the second camera, d (i, a) is the distance between the ith monitoring camera directly connected with the second camera and the first end camera, and P ₀ The target probability of the shooting area of the cameras which are not directly connected by the second camera is determined for non-owner personnel.

The judging module 401 is configured to determine a first number of cameras between the community gate and the second camera according to a preset camera topology map.

The judging module 401 is configured to determine a second number of cameras corresponding to the surveillance video that detects the target appearance feature from the plurality of cameras between the community gate and the second camera.

A judging module 401, configured to determine the target probability according to the quotient of the second number and the first number.

In a possible implementation manner, the detection module 403 is configured to determine, when it is determined that the non-owner person passes through the second camera by the first camera, a fourth camera according to a preset camera topology map, where the fourth camera is any one of the plurality of monitoring cameras directly connected to the second camera except the first camera and the third camera.

The judging module 401 is configured to determine, according to a preset camera topology map, a second end camera directly connected or indirectly connected to the third camera, where the second end camera is any one of the plurality of end cameras.

The judging module 401 is configured to determine, according to a preset camera topology map, a second residential building connected to the second terminal camera from the plurality of residential buildings.

The judging module 401 is configured to calculate a second probability that the non-owner person passes through the fourth camera shooting area from the second camera if the second residential building is determined to be the terminal residential building.

The judging module 401 is configured to judge a magnitude relation between the first probability and the second probability, and if the first probability is determined to be greater than the second probability, preferentially identify the target appearance feature of the monitoring video shot by the third camera.

In a possible implementation manner, the determining module 401 is configured to obtain a time of each end monitoring video including the appearance feature of the target, where the end monitoring video is a video captured by an end camera.

The detection module 403 is configured to sort the end monitoring videos according to a time sequence, determine a third monitoring video in the plurality of end monitoring videos, where the third monitoring video is a monitoring video with the latest time in the plurality of end monitoring videos.

The detection module 403 is configured to determine a third end camera corresponding to the third surveillance video.

The detection module 403 is configured to determine a third residential building connected to the third terminal camera from the plurality of residential buildings according to the preset camera topology map.

And the early warning module 404 is used for determining the third residential building as an actual destination.

In a possible implementation manner, the detection module 403 is configured to perform human target detection on the gate guard monitoring video, so as to obtain a plurality of human targets.

The detection module 403 is configured to identify face regions corresponding to the respective human targets, and determine whether owner information can be identified.

The judging module 401 is configured to determine that the corresponding human target is an owner if it is determined that the owner information can be identified.

The judging module 401 is configured to determine that the corresponding human target is a non-owner person if it is determined that the owner information cannot be identified.

The embodiment also discloses an electronic device, referring to fig. 5, the electronic device may include: at least one processor 501, at least one communication bus 502, a user interface 503, a network interface 504, at least one memory 505.

Wherein a communication bus 502 is used to enable connected communications between these components.

The user interface 503 may include a Display screen (Display) and a Camera (Camera), and the optional user interface 503 may further include a standard wired interface and a standard wireless interface.

The network interface 504 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface), among others.

Wherein the processor 501 may include one or more processing cores. The processor 501 connects various parts throughout the server using various interfaces and lines, performs various functions of the server and processes data by executing or executing instructions, programs, code sets, or instruction sets stored in the memory 505, and invoking data stored in the memory 505. Alternatively, the processor 501 may be implemented in hardware in at least one of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), programmable logic array (Programmable Logic Array, PLA). The processor 501 may integrate one or a combination of several of a central processor 501 (Central Processing Unit, CPU), an image processor 501 (Graphics Processing Unit, GPU), and a modem, etc. The CPU mainly processes an operating system, a user interface, an application program and the like; the GPU is used for rendering and drawing the content required to be displayed by the display screen; the modem is used to handle wireless communications. It will be appreciated that the modem may not be integrated into the processor 501 and may be implemented by a single chip.

The Memory 505 may include a random access Memory 505 (Random Access Memory, RAM), or may include a Read-Only Memory 505. Optionally, the memory 505 comprises a non-transitory computer readable medium (non-transitory computer-readable storage medium). Memory 505 may be used to store instructions, programs, code sets, or instruction sets. The memory 505 may include a stored program area and a stored data area, wherein the stored program area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playing function, an image playing function, etc.), instructions for implementing the above-described various method embodiments, etc.; the storage data area may store data or the like involved in the above respective method embodiments. The memory 505 may also optionally be at least one storage device located remotely from the processor 501. As shown, an operating system, a network communication module, a user interface 503 module, and an application program of a smart community-based risk recognition method may be included in the memory 505 as a computer storage medium.

In the electronic device shown in fig. 5, the user interface 503 is mainly used for providing an input interface for a user, and acquiring data input by the user; and the processor 501 may be configured to invoke an application in the memory 505 that stores a smart community based risk identification method that, when executed by the one or more processors 501, causes the electronic device to perform the method as in one or more of the embodiments described above.

It should be noted that, for simplicity of description, the foregoing method embodiments are all expressed as a series of action combinations, but it should be understood by those skilled in the art that the present application is not limited by the order of actions described, as some steps may be performed in other order or simultaneously in accordance with the present application. Further, those skilled in the art will also appreciate that the embodiments described in the specification are all preferred embodiments, and that the acts and modules referred to are not necessarily required in the present application.

In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to related descriptions of other embodiments.

In the several embodiments provided herein, it should be understood that the disclosed apparatus may be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative, such as a division of units, merely a division of logic functions, and there may be additional divisions in actual implementation, such as multiple units or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some service interface, device or unit indirect coupling or communication connection, electrical or otherwise.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed over a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable memory 505. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a memory 505, including several instructions for causing a computer device (which may be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the methods of the embodiments of the present application. Whereas the aforementioned memory 505 includes: various media capable of storing program codes, such as a U disk, a mobile hard disk, a magnetic disk or an optical disk.

The foregoing is merely exemplary embodiments of the present disclosure and is not intended to limit the scope of the present disclosure. That is, equivalent changes and modifications are contemplated by the teachings of this disclosure, which fall within the scope of the present disclosure. Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure. This application is intended to cover any adaptations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a scope and spirit of the disclosure being indicated by the claims.

Claims

1. A risk identification method based on an intelligent community, the method comprising:

2. The smart community-based risk identification method according to claim 1, wherein the determining the target appearance characteristics of the non-owner in the entrance guard monitoring video specifically comprises:

3. The smart community-based risk identification method of claim 1, wherein after the non-owner enters the community gate, before determining the actual destination to which the non-owner is actually traveling through a plurality of surveillance videos, the method further comprises:

acquiring a second monitoring video shot by the second camera;

4. A method of risk identification based on a smart community according to claim 3, wherein after said identifying said target appearance feature from a second surveillance video, the method further comprises:

；

wherein P (X) ₂ |X ₃ ) For the first probability that the non-owner person passes through the third camera shooting area from the second camera, lambda is a preset attenuation coefficient, d (3, a) is the distance between the third camera and the first terminal camera, n is the number of the monitoring cameras directly connected with the second camera, and d (i, a) is the first cameraThe distance between the ith monitoring camera directly connected with the two cameras and the first terminal camera, P ₀ Target probability of shooting areas for the non-owner personnel without passing through the plurality of cameras directly connected with the second camera;

5. The smart community-based risk identification method of claim 4, wherein after said calculating a first probability that the non-owner person passes from the second camera through the third camera capture area, the method further comprises:

6. The risk identification method based on intelligent communities as in claim 4, wherein determining the actual destination of the actual travel of the non-owner person through a plurality of monitoring videos, specifically comprises:

determining a third terminal camera corresponding to the third monitoring video;

determining a third residential building connected with the third terminal camera from a plurality of residential buildings according to the preset camera topology map;

And determining the third residential building as the actual destination.

7. The smart community-based risk identification method according to claim 1, wherein, according to a guard monitoring video at a gate of a community, if it is determined that the gate of the community includes both an owner and a non-owner, the non-owner is prompted to provide visitor information, the method further comprises:

8. The utility model provides a risk recognition device based on intelligent community which characterized in that includes judging module (401), identification module (402), detection module (403) and early warning module (404), wherein:

the judging module (401) prompts the non-owner personnel to provide visitor information according to a guard monitoring video at a community gate if the fact that the community gate is outside and simultaneously contains the owner personnel and the non-owner personnel is determined, wherein the visitor information comprises contact information and destination information of the non-owner personnel;

The identification module (402) is used for carrying out identity identification on the owner personnel and determining residence information of the owner personnel;

the judging module (401) is configured to judge whether the access destination of the non-owner person is the same as the residence of the owner person according to the residence information and the destination information;

the detection module (403) is configured to determine a target appearance characteristic of the non-owner in the guard monitoring video if the access destination is determined to be different from the residence;

the judging module (401) is used for determining an actual destination to which the non-owner personnel actually go through a plurality of monitoring videos after the non-owner personnel enter the community gate;

and the early warning module (404) is used for carrying out early warning prompt on the non-owner personnel through the contact information if the actual destination is not consistent with the access destination.

9. An electronic device comprising a processor (501), a memory (505), a user interface (503) and a network interface (504), the memory (505) for storing instructions, the user interface (503) and the network interface (504) each for communicating with other devices, the processor (501) for executing the instructions stored in the memory (505) to cause the electronic device to perform the method of any of claims 1-7.

10. A computer readable storage medium storing instructions which, when executed, perform the method of any one of claims 1-7.