CN112836683A - License plate recognition method, device, equipment and medium for portable camera equipment - Google Patents
License plate recognition method, device, equipment and medium for portable camera equipment Download PDFInfo
- Publication number
- CN112836683A CN112836683A CN202110240018.9A CN202110240018A CN112836683A CN 112836683 A CN112836683 A CN 112836683A CN 202110240018 A CN202110240018 A CN 202110240018A CN 112836683 A CN112836683 A CN 112836683A
- Authority
- CN
- China
- Prior art keywords
- license plate
- vehicle
- frame
- image
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000001514 detection method Methods 0.000 claims description 113
- 238000012549 training Methods 0.000 claims description 33
- 230000006870 function Effects 0.000 claims description 32
- 238000000605 extraction Methods 0.000 claims description 27
- 238000004422 calculation algorithm Methods 0.000 claims description 25
- 238000002372 labelling Methods 0.000 claims description 21
- 238000001914 filtration Methods 0.000 claims description 16
- 230000015654 memory Effects 0.000 claims description 16
- 238000004590 computer program Methods 0.000 claims description 13
- 238000003384 imaging method Methods 0.000 claims description 9
- 230000004044 response Effects 0.000 claims description 6
- 238000007689 inspection Methods 0.000 claims description 3
- 230000015572 biosynthetic process Effects 0.000 abstract 1
- 238000003786 synthesis reaction Methods 0.000 abstract 1
- 230000008569 process Effects 0.000 description 19
- 238000010586 diagram Methods 0.000 description 9
- 239000000284 extract Substances 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 7
- 238000013528 artificial neural network Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 5
- 238000013527 convolutional neural network Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 3
- 238000012015 optical character recognition Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000003068 static effect Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000013139 quantization Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 238000013519 translation Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000009432 framing Methods 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/625—License plates
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Traffic Control Systems (AREA)
- Image Analysis (AREA)
Abstract
The application relates to the technical field of machine vision, and provides a license plate recognition method, a license plate recognition device, license plate recognition equipment and a license plate recognition medium applied to portable camera equipment. The vehicle license plate recognition method and device can achieve mobile automatic collection and recognition of the vehicle license plate, improve license plate collection and recognition efficiency, and improve accuracy of license plate recognition of the license plate through synthesis of license plate recognition results of multiple frames of images. The method comprises the following steps: the method comprises the steps of obtaining a video shot by a portable camera device in a moving mode, responding to the fact that a vehicle to be recognized is tracked in the video, extracting a vehicle image of the vehicle to be recognized from a multi-frame image containing the vehicle to be recognized to obtain a multi-frame vehicle image, inputting each frame of vehicle image to a pre-constructed license plate recognition model to obtain a plurality of license plate recognition results, and determining the license plate of the vehicle to be recognized based on statistical information of the license plate recognition results.
Description
Technical Field
The present application relates to the field of machine vision technologies, and in particular, to a license plate recognition method and apparatus for a portable camera device, and a storage medium.
Background
With the development of machine vision technology, the machine vision technology is applied more and more widely in fields such as object detection, OCR (Optical Character Recognition) and the like. The machine vision technology is applied to vehicle information collection such as license plates and the like, and data support can be effectively provided for smart cities.
In the conventional technology, a camera fixedly installed at a road intersection is used for capturing a vehicle to identify the license plate of the vehicle, but the comprehensive coverage of areas such as cities is difficult, so that illegal vehicles can drive by bypassing the camera, and the license plate information of vehicles parked statically at the road side can be recorded manually in a polling mode by polling personnel, but the efficiency is low.
Disclosure of Invention
In view of the above, it is necessary to provide a license plate recognition method and apparatus applied to a portable image pickup device, and a storage medium.
A license plate recognition method applied to a portable camera device comprises the following steps:
acquiring a video shot by the portable camera device;
in response to a vehicle to be identified being tracked in the video, extracting a vehicle image of the vehicle to be identified from a plurality of frames of images of the video containing the vehicle to be identified to obtain a plurality of frames of vehicle images;
inputting the plurality of frames of vehicle images into a pre-constructed license plate recognition model so that the license plate recognition model outputs license plate recognition results corresponding to the frames of vehicle images respectively to obtain a plurality of license plate recognition results;
and determining the license plate of the vehicle to be recognized based on the statistical information of the plurality of license plate recognition results.
A license plate recognition device applied to a portable camera device, comprising:
the video acquisition module is used for acquiring a video shot by the portable camera device;
the image extraction module is used for responding to the fact that a vehicle to be identified is tracked in the video, extracting a vehicle image of the vehicle to be identified from a plurality of frames of images containing the vehicle to be identified in the video, and obtaining a plurality of frames of vehicle images;
the model identification module is used for inputting the plurality of frames of vehicle images into a pre-constructed license plate identification model so that the license plate identification model outputs license plate identification results corresponding to the frames of vehicle images respectively to obtain a plurality of license plate identification results;
and the license plate determining module is used for determining the license plate of the vehicle to be recognized based on the statistical information of the plurality of license plate recognition results.
A portable imaging apparatus comprising a memory storing a computer program and a processor implementing the following steps when executing the computer program:
acquiring a video shot by the portable camera device; in response to a vehicle to be identified being tracked in the video, extracting a vehicle image of the vehicle to be identified from a plurality of frames of images of the video containing the vehicle to be identified to obtain a plurality of frames of vehicle images; inputting the plurality of frames of vehicle images into a pre-constructed license plate recognition model so that the license plate recognition model outputs license plate recognition results corresponding to the frames of vehicle images respectively to obtain a plurality of license plate recognition results; and determining the license plate of the vehicle to be recognized based on the statistical information of the plurality of license plate recognition results.
A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of:
acquiring a video shot by the portable camera device; in response to a vehicle to be identified being tracked in the video, extracting a vehicle image of the vehicle to be identified from a plurality of frames of images of the video containing the vehicle to be identified to obtain a plurality of frames of vehicle images; inputting the plurality of frames of vehicle images into a pre-constructed license plate recognition model so that the license plate recognition model outputs license plate recognition results corresponding to the frames of vehicle images respectively to obtain a plurality of license plate recognition results; and determining the license plate of the vehicle to be recognized based on the statistical information of the plurality of license plate recognition results.
The license plate recognition method, the license plate recognition device, the license plate recognition medium and the license plate recognition device applied to the portable camera device are used for acquiring a video shot by the portable camera device in a moving mode, responding to a vehicle to be recognized tracked in the video, extracting a vehicle image of the vehicle to be recognized from a multi-frame image containing the vehicle to be recognized, acquiring a multi-frame vehicle image, inputting each frame of vehicle image to a pre-constructed license plate recognition model to acquire a plurality of license plate recognition results, and determining the license plate of the vehicle to be recognized based on statistical information of the plurality of license plate recognition results. According to the scheme, the movable automatic collection and identification of the vehicle license plate can be realized, the license plate collection and identification efficiency is improved, and the accuracy of license plate number identification can be improved by integrating the license plate identification results of multiple frames of images.
Drawings
Fig. 1 is an application environment diagram of a license plate recognition method applied to a portable camera device in one embodiment;
FIG. 2 is a schematic flowchart illustrating a license plate recognition method applied to a portable camera device according to an embodiment;
FIG. 3 is a schematic flow chart illustrating a license plate extraction model processing a license plate image according to an embodiment;
FIG. 4 is a flowchart illustrating steps of constructing a license plate detection model according to one embodiment;
FIG. 5 is a schematic diagram of a license plate annotation in one embodiment;
FIG. 6 is a schematic flow chart diagram illustrating the steps for tracking a vehicle in video in one embodiment;
FIG. 7 is a schematic flow chart of a method for automatically collecting and recognizing license plates using a forensic instrument in an application example;
FIG. 8 is a block diagram of a license plate recognition apparatus applied to a portable camera device according to an embodiment;
fig. 9 is an internal configuration diagram of the portable image pickup apparatus in one embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The license plate recognition method applied to the portable camera device can be applied to the application environment shown in fig. 1. The application environment may include a portable imaging device 110 and a server 120, wherein the portable imaging device 110 may be communicatively coupled to the server 120 via a network. This portable camera equipment 110 can be the appearance of collecting evidence that the grid person worn at the in-process of patrolling and examining, can be used to shoot the static vehicle that parks in roadside and take the video in real time when the grid person patrols and examines along the direction of patrolling and examining, and portable camera equipment 110 can discern the license plate of vehicle based on the video of recording, uploads server 120 after obtaining license plate information.
For the vehicle configuration identification process, specifically, the portable camera device 110 obtains a video that is shot by the portable camera device 110 in a moving manner, the portable camera device 110 responds to a vehicle to be identified tracked in the video, extracts a vehicle image of the vehicle to be identified from a multi-frame image of the video containing the vehicle to be identified to obtain a multi-frame vehicle image, the portable camera device 110 inputs the multi-frame vehicle image to a pre-constructed license plate identification model, so that the license plate identification model outputs license plate identification results corresponding to the vehicle images of each frame to obtain a plurality of license plate identification results, and the portable camera device 110 determines the license plate of the vehicle to be identified based on statistical information of the license plate identification results.
Through the scheme of the license plate recognition applied to the portable camera device 110, the effects of real-time collection of the license plate of the vehicle, high recognition accuracy and wide coverage can be realized, the recognition mode can be used as effective supplement for capturing the license plate by a camera fixedly installed at the intersection, and the technical problem of low license plate collection and recognition efficiency in the process of inspecting by a gridder can be solved. In the above scenario, the server 120 may be implemented by a stand-alone server or a server cluster composed of a plurality of servers.
The following further describes the license plate recognition method applied to the portable camera device, which is provided by the present application, with reference to the embodiments and the accompanying drawings.
In one embodiment, as shown in fig. 2, a license plate recognition method applied to a portable camera device is provided, which is described by taking the portable camera device 110 in fig. 1 as an example, and includes the following steps:
step S201, acquiring a video shot by the portable camera device;
in this step, the portable camera device 110 may specifically be a forensic instrument, which may be worn on a grid person, and the portable camera device 110 is in a mobile state during the process of patrolling and examining by the grid person, so that the portable camera device 110 may acquire a video by itself in real time during the process of patrolling and examining by the grid person, and the video may include a roadside stationary parked vehicle.
Step S202, in response to the fact that the vehicle to be identified is tracked in the video, extracting the vehicle image of the vehicle to be identified from the multi-frame image containing the vehicle to be identified in the video to obtain a plurality of frames of vehicle images;
in this step, the portable camera device 110 may analyze the captured video, detect whether the video includes a vehicle, if the vehicle in the video is detected and tracked, the tracked vehicle is used as a vehicle to be identified, and then the portable camera device 110 extracts a vehicle image of the vehicle to be identified from a plurality of frames of images of the video including the image to be identified, so as to obtain a plurality of frames of vehicle images. Specifically, the portable imaging device 110 may determine an image area where the vehicle to be recognized is located in a plurality of frames of images including the image to be recognized, extract an image corresponding to the image area where the vehicle to be recognized is located from each frame of image as a vehicle image, and obtain a plurality of frames of vehicle images.
In some embodiments, the portable camera device 110 may recognize whether each frame of image of the video includes a vehicle by using a vehicle detection model that is constructed in advance, for example, the vehicle detection model may be constructed by using deep learning, and in order to implement real-time vehicle detection on the portable camera device 110, specifically, the vehicle detection model may be obtained by training based on a mobilenet0.25-ssdlite model, the size of the obtained vehicle detection model may be 1.3M, and then int8 model quantization is performed on the basis, the size of the vehicle detection model may be converted into 32KB, and then the vehicle detection model may be migrated to an ARM system to be configured in the portable camera device 110 to detect the vehicle.
Step S203, inputting a plurality of frames of vehicle images into a pre-constructed license plate recognition model, so that the license plate recognition model outputs license plate recognition results corresponding to the frames of vehicle images respectively, and a plurality of license plate recognition results are obtained.
In this step, the portable camera device 110 inputs the multiple frames of vehicle images obtained in step S202 into a pre-constructed license plate recognition model, where the license plate recognition model is used to recognize license plates in the images, and specifically, the license plate recognition model can recognize and output a license plate recognition result corresponding to each frame of input vehicle image, so that the portable camera device 110 can obtain multiple license plate recognition results corresponding to each frame of vehicle image, where the license plate recognition result may include license plate numbers or license plates represented by or corresponding to each result.
And step S204, determining the license plate of the vehicle to be recognized based on the statistical information of the plurality of license plate recognition results.
In this step, the portable camera device 110 may count the license plate recognition results to obtain corresponding statistical information, and determine the license plate of the vehicle to be recognized according to the statistical information, so as to avoid the problem that the recognition error exists when the license plate is recognized only according to a single frame image.
In some embodiments, step S204 may include: counting license plates corresponding to the license plate recognition results to obtain the number of the license plate recognition results corresponding to each license plate; and identifying the license plate with the largest number corresponding to the license plate identification result in each license plate as the license plate of the vehicle to be identified.
In the moving process of the portable camera device 110, shaking of the portable camera device 110 is easily caused to cause imaging blurring of a license plate in an image to cause a recognition error, but a wrongly recognized license plate is randomly changed and has a small probability of being kept on the same license plate, so that a plurality of license plate recognition results of the same vehicle to be recognized can be counted in a video, and license plates corresponding to the plurality of license plate recognition results are counted, for example, ten license plate recognition results are obtained by the portable camera device 110 and respectively correspond to ten frames of vehicle images, and the license plate recognition results may not represent the same license plate number due to the recognition error caused by shaking during moving, so that the portable camera device 110 can count license plates corresponding to the ten license plate recognition results, and the license plates counted by the portable camera device 110 are set to include a license plate number a, The number of license plate recognition results corresponding to each license plate is further obtained by the portable camera device 110, the number of license plate recognition results corresponding to each license plate is set to be 8, the number of license plate recognition results corresponding to license plate a is set to be 1, the number of license plate recognition results corresponding to license plate B and license plate C is set to be 1, then the portable camera device 110 recognizes the license plate with the largest number of license plates corresponding to the license plate recognition results in each license plate as the license plate of the vehicle to be recognized, namely the portable camera device 110 uploads the license plate a serving as the license plate of the vehicle to be recognized to the server 120.
It can be understood that the portable camera device 110 may also calculate, based on the number of the license plate recognition results corresponding to the obtained number plates, a probability that each license plate can serve as the license plate of the vehicle to be recognized, and take the license plate with the highest probability as the license plate of the vehicle to be recognized, for example, the probability of the license plate a is 8/10, and the probabilities of the license plates B and C are both 1/10, so that the portable camera device 110 takes the license plate a with the highest probability as the license plate of the vehicle to be recognized, thereby improving the accuracy of recognizing the license plate in the video.
According to the license plate recognition method applied to the portable camera device, the portable camera device 110 obtains a video shot by the portable camera device in a moving mode, a vehicle to be recognized is responded to be tracked in the video, the portable camera device 110 extracts a vehicle image of the vehicle to be recognized from a multi-frame image containing the vehicle to be recognized to obtain a multi-frame vehicle image, the portable camera device 110 inputs each frame of vehicle image to a pre-constructed license plate recognition model to obtain a plurality of license plate recognition results, and the portable camera device 110 determines the license plate of the vehicle to be recognized based on statistical information of the plurality of license plate recognition results. According to the scheme, the movable automatic collection and identification of the vehicle license plate can be realized, the license plate collection and identification efficiency is improved, and the accuracy of license plate number identification can be improved by integrating the license plate identification results of multiple frames of images.
In an embodiment, the license plate recognition model in step S203 may further include a license plate detection model and a license plate extraction model, where the license plate detection model may be configured to detect an image area where a license plate is located in the vehicle image and extract a corresponding license plate image, and the license plate extraction model may be configured to extract a specific license plate number from the license plate image output by the license plate detection model as a license plate recognition result for output. Specifically, step S203 specifically includes:
the portable camera device 110 inputs a plurality of frames of vehicle images into the license plate detection model, so that the license plate detection model outputs license plate images corresponding to each frame of vehicle image, inputs each license plate image into the license plate extraction model, and outputs corresponding license plate recognition results through the license plate extraction model to obtain a plurality of license plate recognition results.
In this embodiment, the license plate recognition model may include a license plate detection model and a license plate extraction model that are sequentially connected, the portable camera device 110 inputs a plurality of frames of vehicle images into the license plate detection model, the license plate detection model recognizes an image area where a license plate is located in each frame of vehicle image, inputs a license plate image corresponding to the image area where the license plate is located in each frame of vehicle image into the license plate extraction model, recognizes a specific license plate number in each frame of license plate image through the license plate extraction model, and outputs a plurality of corresponding license plate recognition results. For the license plate extraction model, as shown in fig. 3, in the model training or construction stage, end-to-end training is performed by adopting neural network multi-license plate recognition, so that error recognition caused by errors caused by character segmentation is avoided. In the model application stage, the license plate extraction model can perform reasoning calculation on the input license plate image to obtain the license plate number, the license plate image can be firstly corrected when being input to the license plate extraction model, and the corrected license plate image is input to the license plate extraction model to perform license plate number extraction so as to improve the accuracy of license plate number identification. The license plate extraction model can be implemented based on a CRNN (Convolutional Recurrent Neural Network) to extract characters in the license plate image through CRNN general OCR. The specific process may include: the license plate extraction model obtains an input corrected license plate image (the license plate number is set as A.12B 34), and characteristic extraction is carried out by utilizing a CNN (Convolutional Neural Network), the Convolutional layer is a common CNN Network and is used for extracting a characteristic diagram (Convolutional feature maps) of the input license plate image, the characteristic diagram of the license plate image outputs the license plate number through an RNN (Recurrent Neural Network) and a translation layer, wherein the RNN is a deep two-way LSTM Network (Long Short-Term Memory Network), and the translation layer is used for carrying out software function processing on the output of the RNN to obtain corresponding characters to be output as the license plate number corresponding to each frame of license plate image.
In an embodiment, the portable camera device 110 or the server 120 may construct the license plate detection model in the license plate recognition model by the following steps, as shown in fig. 4, the specific steps include:
step S401, acquiring a training image sample;
step S402, acquiring a real license plate label, a real license plate labeling frame and real license plate key points obtained by labeling the training image sample;
the server 120 can obtain and label the training image sample to obtain a real license plate label, a real license plate labeling frame and a real license plate key point. The real license plate label can be used for distinguishing whether a certain area is a license plate, the real license plate marking frame is used for marking the real position of the license plate in the training image sample, and the real license plate key points are used for marking four key points of the license plate in the training image sample. Specifically, as shown in fig. 5, a license plate "car a.12b 34" may be labeled, a region where the license plate is located is labeled as a license plate by a real license plate label, an actual frame of the license plate is labeled as a real license plate labeling frame 510, and four corner points (P11 to P14) of the real license plate labeling frame 510 are labeled as real license plate key points.
Step S403, training the license plate detection model to be trained based on the training image sample, the real license plate label, the real license plate labeling frame and the real license plate key point, and constructing to obtain the license plate detection model.
On the basis of training an image sample, training a license plate detection model to be trained by taking four key points of classification recognition of a license plate or a non-license plate, a labeling frame of the license plate and the license plate as supervision information, and constructing to obtain the license plate detection model. Therefore, in the model application stage, the problem of inaccurate license plate recognition caused by shaking or inclined parking of some vehicles in the moving process of the portable camera device 110 can be overcome to a certain extent, so that the recognition accuracy is improved, and in the model application stage, the license plate image can be corrected by using four key points on the premise of detecting the rectangular frame of the license plate, so that the inclined license plate can be corrected to the average normal position, and the license plate recognition accuracy in the subsequent license plate recognition process can be improved.
In an embodiment, the step S403 may specifically include: inputting the training image sample into a license plate detection model to be trained so that the license plate detection model to be trained outputs a predicted license plate label, a predicted license plate marking frame and predicted license plate key points; constructing a license plate classifier loss function according to the predicted license plate label and the real license plate label; constructing a regression loss function of a marking frame according to the predicted license plate marking frame and the real license plate marking frame; constructing a key point positioning loss function according to the predicted license plate key points and the real license plate key points; and training a license plate detection model to be trained based on the license plate classifier loss function, the mark frame regression loss function and the key point positioning loss function.
In this embodiment, the license plate detection model may adopt a light-weight MTCNN detection network (Multi-task convolutional neural network), and a loss function of the license plate detection model may be divided into three parts, where the first part is a license plate or non-license plate recognition part, the second part is a license plate labeling frame regression part, and the third part is a license plate key point location part, where loss functions of the three parts are a license plate classifier loss function, a labeling frame regression loss function, and a key point location loss function, and the license plate detection model may be obtained based on the three loss functions through training. As shown in fig. 5, for the prediction of the license plate by the license plate detection model, the license plate detection model can predict a predicted license plate labeling frame 520, an image area where the predicted license plate labeling frame 520 is located can be provided with a predicted license plate label, and four corner points (P21 to P24) corresponding to the predicted license plate labeling frame 520 are key points of the predicted license plate.
In particular, the license plate classifier penalty function
Can be expressed as:
wherein, PiIs the probability that the ith target belongs to the license plate or the predicted license plate label,
labeling the marked real license plate label;
mark box regression loss function
Can be expressed as:
wherein,
a label box for indicating the predicted license plate,
representing a real license plate marking frame;
key point localization loss function
Can be expressed as:
wherein,
representing the key points of the predicted license plate,
representing the real license plate key points.
Based on the above three loss functions, a total loss function L can be constructed as:
wherein, the training process of the license plate detection model is the function L for minimizing the total loss, wherein N represents the number of training samples, and alphajWhich indicates the importance of the task or tasks,
the weight of the sample of the label is represented,
as above three loss functions.
The embodiment provides a method for constructing a license plate detection model, and the identification accuracy can be improved in the application stage of the model. Specifically, when the portable camera device 110 uses the license plate detection model, the cutout of the vehicle image obtained by the pre-constructed vehicle detection model can be used as an input, so as to improve the efficiency of license plate detection and recognition.
In some embodiments, as shown in fig. 6, the portable camera device 110 may track the vehicle through the following steps, specifically, after the video captured by the movement of the portable camera device is acquired in step S201, the method may further include:
step S601, responding to a pre-constructed vehicle detection model to detect that a current frame image of a video contains a vehicle, acquiring a corresponding current frame vehicle detection frame, and acquiring a vehicle prediction frame of the vehicle contained in the current frame image in a next frame image by using a Kalman filtering algorithm to obtain a next frame vehicle prediction frame;
the method mainly includes the steps that the portable camera device 110 can detect whether a vehicle is included in a current frame image in real time by using a vehicle detection model which is constructed in advance, when the current frame image is detected to include the vehicle, the portable camera device 110 obtains a corresponding current frame vehicle detection frame, the current frame vehicle detection frame refers to a detection frame for framing the vehicle in the current frame image, and the detection frame is obtained by outputting the vehicle detection model. The portable camera device 110 further predicts a vehicle prediction frame corresponding to the vehicle in the next frame image included in the current frame image by using a kalman filtering algorithm, so as to obtain the next frame vehicle prediction frame.
Step S602, obtaining a next frame vehicle detection frame obtained by detecting a next frame image by a vehicle detection model;
in this step, the portable camera device 110 further detects the vehicle in the next frame of image by using the vehicle detection model, so as to obtain a next frame of vehicle detection frame.
Step S603, the portable imaging device 110 obtains an Intersection-over-unity (IOU) ratio between the vehicle detection frame of the next frame and the vehicle prediction frame of the next frame;
step S604, judging whether the prediction result of the Kalman filtering algorithm on the vehicle is matched with the detection result of the vehicle detection model on the vehicle or not based on the intersection ratio;
in this step, the portable camera device 110 may perform IOU matching on the vehicle prediction frame of the kalman filter algorithm and the vehicle detection frame of the vehicle detection model using the hungarian algorithm to calculate the similarity, and if the cross-over ratio is greater than or equal to the preset threshold, the portable camera device 110 may determine that the prediction result of the kalman filter algorithm on the vehicle matches the detection result of the vehicle detection model on the vehicle.
Step S605, if yes, updating the next frame vehicle prediction frame based on the next frame vehicle detection frame, tracking the vehicle included in the current frame image, and taking the vehicle included in the current frame image as the vehicle to be identified.
If the prediction result of the kalman filter algorithm on the vehicle is matched with the detection result of the vehicle detection model on the vehicle, the portable camera device 110 updates the next frame of vehicle prediction frame obtained by the kalman filter algorithm based on the next frame of vehicle detection frame output by the vehicle detection model, so as to track the vehicle contained in the current frame of image in the subsequent frame of image, and takes the vehicle as the vehicle to be identified, and extracts the license plate image of the vehicle for identification.
In the embodiment, the vehicle is tracked, a gridder generally inspects vehicles parked statically at the roadside, the vehicles are static, and the evidence obtaining instrument worn by the gridder moves, so that the vehicles can be regarded as moving corresponding to the evidence obtaining instrument through a coordinate conversion relation, the evidence obtaining instrument keeps small speed change, the moving distance of the same vehicle between two adjacent frames of images of a video is small, and the portable camera device 110 can form tracking track data of the vehicles by using a Kalman filtering algorithm and a Hungary matching algorithm and detect license plates of the vehicles according to the tracking track data.
In an embodiment, the license plate recognition method applied to the portable camera device provided in this application is applied to a scene where a grid person inspects a still parked vehicle on a roadside, in this application example, the portable camera device 110 is a forensic apparatus worn by the grid person during inspection, and the vehicle to be recognized is a still parked vehicle on the roadside, as shown in fig. 7, the method may include the following steps:
and reading the video stream of the evidence obtaining instrument, and carrying out vehicle detection, tracking, license plate detection and identification based on the video stream.
For vehicle detection, deep learning can be adopted to train a vehicle detection model, and in order to achieve real-time model detection at the embedded end of a evidence obtaining instrument, a mobilenet0.25-ssdlite model is used for training, the size of the model is finally 1.3M, int8 model quantization is carried out, the size of the model is converted into 32KB, and C + + is transplanted to an ARM system for vehicle detection reasoning. The evidence obtaining instrument detects the vehicle, a detected vehicle chain can be established when the vehicle is not detected, whether the tracker vehicle chain is empty or not is further judged, if yes, a follow-up vehicle tracking process is carried out, and if not, the tracker is established and the vehicle is continuously detected.
For vehicle tracking, the vehicle is still and the evidence obtaining instrument moves, so the vehicle can be considered to move corresponding to the evidence obtaining instrument according to the coordinate conversion relation, and under the condition that the speed change of the evidence obtaining instrument is kept to be small, tracking track data of the vehicle is formed by utilizing a Kalman filtering algorithm and a Hungary matching algorithm, and a license plate is detected, identified and output according to the tracking track data in the subsequent steps. Specifically, the flow of the kalman filter algorithm and the hungarian algorithm mainly comprises the following steps: 1. predicting a prediction frame of the vehicle by using a Kalman filtering algorithm; 2. using a Hungarian algorithm to carry out IOU matching on a prediction box of a Kalman filtering algorithm and a detection box output by the vehicle detection model to calculate the similarity; 3. and in the Kalman filtering algorithm, the prediction frame predicted by the Kalman filtering algorithm is updated by using the detection frame, so that the tracking of the vehicle is realized.
For the license plate detection, in order to improve the detection efficiency, the vehicle image cutout obtained by the vehicle detection model detection can be used as the input of the license plate detection model. Because the evidence obtaining instrument shakes in the movement process, and some vehicles park more slantwise, in order to achieve good recognition rate, in the training stage of the license plate detection model, the license plate classifier loss function, the mark frame regression loss function and the key point positioning loss function in the previous embodiment can be adopted to train the license plate detection model, so that in the application stage of the license plate detection model, the rectangular frame of the license plate can be detected, and then the rectangular frame of the license plate can be corrected by using 4 key points to obtain the corrected license plate image.
In the license plate recognition process, a plurality of frames of corrected license plate images output by the license plate detection model can be input into the license plate extraction model, license plate numbers corresponding to the license plate images of all frames are obtained through the output of the license plate extraction model, and the license plate numbers are added into a license plate list of tracking track data of corresponding vehicles.
And finally, filtering the wrongly identified license plate numbers to obtain real license plate numbers. Specifically, in the process of movement of the evidence obtaining instrument, shaking is easily caused to cause license plate blurring to cause error recognition, but the license plates which are wrongly recognized are randomly changed, and the probability kept on the same license plate is small, so that for tracking track data of the same vehicle, a probability maximization algorithm can be adopted, and only the license plate number with the highest corresponding probability in the tracking track data is extracted as the license plate number of the final vehicle, so that the accuracy of license plate recognition can be improved, and further, under the condition that the probability is greater than a preset threshold value, the evidence obtaining instrument can output the license plate information and upload the license plate information to the server 120. The process of filtering and uploading the license plate number can be entered when the prediction frame of the Kalman filtering algorithm is not matched with the IOU of the detection frame output by the vehicle detection model, when the IOU is not matched, the evidence obtaining instrument can delete the created vehicle tracker, extract the license plate with the highest probability in the tracker, and upload the license plate to the server 120 when the probability is greater than a threshold value, and when the probability is less than or equal to the threshold value, ignore the identified license plate and detect the vehicle again.
This application example realizes license plate automatic acquisition based on the appearance of collecting evidence, can high-efficient high accuracy ground automatic acquisition curb stop the license plate information of vehicle, for the vehicle management and control of later stage, illegal processing provides effectual data support. Specifically, the application example uses a evidence obtaining instrument worn by a gridder in the process of patrolling to record a video, firstly uses a neural network to identify vehicles which are statically parked at the roadside, then uses a video tracking technology to track the track of the vehicles, uses the neural network to identify the number of the vehicle in the track, and filters the number of the wrong vehicle according to a probability maximization method because of the number of the wrong vehicle which appears in the video jitter, thereby achieving the purposes of real-time acquisition, high accuracy and wide coverage, effectively supplementing the snapshot of the number of the road junction, and solving the problem of low efficiency in the process of patrolling by the gridder.
It should be understood that, although the steps in the above-described flowcharts are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a part of the steps in the above-mentioned flowcharts may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of performing the steps or the stages is not necessarily performed in sequence, but may be performed alternately or alternately with other steps or at least a part of the steps or the stages in other steps.
In one embodiment, as shown in fig. 8, there is provided a license plate recognition apparatus of a portable camera device, the apparatus 800 may include:
a video acquiring module 801, configured to acquire a video that is shot by the portable imaging device;
an image extraction module 802, configured to, in response to a vehicle to be identified being tracked in the video, extract a vehicle image of the vehicle to be identified from a plurality of frames of images of the video that include the vehicle to be identified, so as to obtain a plurality of frames of vehicle images;
the model identification module 803 is configured to input the multiple frames of vehicle images to a pre-constructed license plate identification model, so that the license plate identification model outputs license plate identification results corresponding to each frame of vehicle image, and multiple license plate identification results are obtained;
and the license plate determining module 804 is configured to determine the license plate of the vehicle to be recognized based on the statistical information of the plurality of license plate recognition results.
In an embodiment, the license plate determining module 804 is configured to count license plates corresponding to the plurality of license plate recognition results to obtain the number of license plate recognition results corresponding to each license plate; and identifying the license plate with the largest number corresponding to the license plate identification result in the license plates as the license plate of the vehicle to be identified.
In one embodiment, the license plate recognition model comprises a license plate detection model and a license plate extraction model; the model identification module 803 is configured to input the multiple frames of vehicle images to the license plate detection model, so that the license plate detection model outputs license plate images corresponding to the respective frames of vehicle images, and inputs the license plate images to the license plate extraction model, and the license plate extraction model outputs corresponding license plate identification results to obtain the multiple license plate identification results.
In one embodiment, the apparatus 800 may further include: the model construction unit is used for acquiring a training image sample; acquiring a real license plate label, a real license plate labeling frame and real license plate key points obtained by labeling the training image sample; and training a license plate detection model to be trained based on the training image sample, the real license plate label, the real license plate labeling frame and the real license plate key point, and constructing to obtain the license plate detection model.
In one embodiment, the model construction unit is configured to input the training image sample to the license plate detection model to be trained, so that the license plate detection model to be trained outputs a predicted license plate label, a predicted license plate labeling frame, and a predicted license plate key point; constructing a license plate classifier loss function according to the predicted license plate label and the real license plate label; constructing a mark frame regression loss function according to the predicted license plate mark frame and the real license plate mark frame; constructing a key point positioning loss function according to the predicted license plate key points and the real license plate key points; and training the license plate detection model to be trained based on the license plate classifier loss function, the mark frame regression loss function and the key point positioning loss function.
In one embodiment, the apparatus 800 may further include: the vehicle tracking unit is used for responding to a pre-constructed vehicle detection model to detect that a current frame image of the video contains a vehicle, acquiring a corresponding current frame vehicle detection frame, and acquiring a vehicle prediction frame of the vehicle contained in the current frame image in a next frame image by using a Kalman filtering algorithm to obtain the next frame vehicle prediction frame; acquiring a next frame of vehicle detection frame obtained by detecting the next frame of image by the vehicle detection model; acquiring the intersection ratio of the next frame of vehicle detection frame and the next frame of vehicle prediction frame; judging whether the prediction result of the Kalman filtering algorithm on the vehicle is matched with the detection result of the vehicle detection model on the vehicle or not based on the intersection ratio; if yes, updating the next frame vehicle prediction frame based on the next frame vehicle detection frame, tracking the vehicle contained in the current frame image, and taking the vehicle contained in the current frame image as the vehicle to be identified.
In one embodiment, the portable camera device is a forensic instrument worn by a girderstorm inspection process; the vehicle to be identified is a roadside stationary parked vehicle.
For specific limitations of the license plate recognition device of the portable camera device, reference may be made to the above limitations of the license plate recognition method of the portable camera device, and details thereof are not repeated herein. All or part of the modules in the license plate recognition device of the portable camera equipment can be realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent of a processor in the portable camera device, and can also be stored in a memory in the portable camera device in a software form, so that the processor can call and execute the corresponding operations of the modules.
In one embodiment, a portable image pickup apparatus is provided, an internal structure diagram of which may be as shown in fig. 9. The portable image pickup apparatus includes a processor, a memory, and a communication interface connected by a system bus. Wherein the processor of the portable camera device is used to provide computing and control capabilities. The memory of the portable image pickup apparatus includes a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the portable camera device is used for performing wired or wireless communication with an external device, and the wireless communication can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a license plate recognition method applied to a portable image pickup apparatus.
Those skilled in the art will appreciate that the structure shown in fig. 9 is a block diagram of only a part of the structure related to the present application, and does not constitute a limitation of the portable image pickup apparatus to which the present application is applied, and a specific portable image pickup apparatus may include more or less components than those shown in the figure, or combine some components, or have a different arrangement of components.
In one embodiment, there is also provided a portable imaging apparatus including a memory in which a computer program is stored and a processor that implements the steps in the above method embodiments when the processor executes the computer program.
In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when being executed by a processor, carries out the steps of the above-mentioned method embodiments.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A license plate recognition method applied to a portable camera device is characterized by comprising the following steps:
acquiring a video shot by the portable camera device;
in response to a vehicle to be identified being tracked in the video, extracting a vehicle image of the vehicle to be identified from a plurality of frames of images of the video containing the vehicle to be identified to obtain a plurality of frames of vehicle images;
inputting the plurality of frames of vehicle images into a pre-constructed license plate recognition model so that the license plate recognition model outputs license plate recognition results corresponding to the frames of vehicle images respectively to obtain a plurality of license plate recognition results;
and determining the license plate of the vehicle to be recognized based on the statistical information of the plurality of license plate recognition results.
2. The method of claim 1, wherein the determining the license plate of the vehicle to be recognized based on the statistical information of the plurality of license plate recognition results comprises:
counting license plates corresponding to the license plate recognition results to obtain the number of the license plate recognition results corresponding to each license plate;
and identifying the license plate with the largest number corresponding to the license plate identification result in the license plates as the license plate of the vehicle to be identified.
3. The method of claim 1, wherein the license plate recognition model comprises a license plate detection model and a license plate extraction model; the method for inputting the plurality of frames of vehicle images into a pre-constructed license plate recognition model so that the license plate recognition model outputs license plate recognition results corresponding to each frame of vehicle image to obtain a plurality of license plate recognition results comprises the following steps:
and inputting the plurality of frames of vehicle images into the license plate detection model so that the license plate detection model outputs license plate images corresponding to the frames of vehicle images respectively, inputting the license plate images into the license plate extraction model, and outputting corresponding license plate identification results through the license plate extraction model to obtain a plurality of license plate identification results.
4. The method of claim 3, further comprising:
acquiring a training image sample;
acquiring a real license plate label, a real license plate labeling frame and real license plate key points obtained by labeling the training image sample;
and training a license plate detection model to be trained based on the training image sample, the real license plate label, the real license plate labeling frame and the real license plate key point, and constructing to obtain the license plate detection model.
5. The method of claim 4, wherein training a license plate detection model to be trained based on the training image samples, the real license plate labels, the real license plate labeling frames, and the real license plate key points comprises:
inputting the training image sample into the license plate detection model to be trained so that the license plate detection model to be trained outputs a predicted license plate label, a predicted license plate marking frame and predicted license plate key points;
constructing a license plate classifier loss function according to the predicted license plate label and the real license plate label;
constructing a mark frame regression loss function according to the predicted license plate mark frame and the real license plate mark frame;
constructing a key point positioning loss function according to the predicted license plate key points and the real license plate key points;
and training the license plate detection model to be trained based on the license plate classifier loss function, the mark frame regression loss function and the key point positioning loss function.
6. The method of claim 1, wherein after the video captured by the portable imaging device is captured, the method further comprises:
responding to a pre-constructed vehicle detection model to detect that a current frame image of the video contains a vehicle, acquiring a corresponding current frame vehicle detection frame, and acquiring a vehicle prediction frame of the vehicle contained in the current frame image in a next frame image by using a Kalman filtering algorithm to obtain the next frame vehicle prediction frame;
acquiring a next frame of vehicle detection frame obtained by detecting the next frame of image by the vehicle detection model;
acquiring the intersection ratio of the next frame of vehicle detection frame and the next frame of vehicle prediction frame;
judging whether the prediction result of the Kalman filtering algorithm on the vehicle is matched with the detection result of the vehicle detection model on the vehicle or not based on the intersection ratio;
if yes, updating the next frame vehicle prediction frame based on the next frame vehicle detection frame, tracking the vehicle contained in the current frame image, and taking the vehicle contained in the current frame image as the vehicle to be identified.
7. The method according to any one of claims 1 to 6, wherein the portable camera device is a forensic instrument worn by a grid person during inspection; the vehicle to be identified is a roadside stationary parked vehicle.
8. A license plate recognition device applied to a portable camera device, comprising:
the video acquisition module is used for acquiring a video shot by the portable camera device;
the image extraction module is used for responding to the fact that a vehicle to be identified is tracked in the video, extracting a vehicle image of the vehicle to be identified from a plurality of frames of images containing the vehicle to be identified in the video, and obtaining a plurality of frames of vehicle images;
the model identification module is used for inputting the plurality of frames of vehicle images into a pre-constructed license plate identification model so that the license plate identification model outputs license plate identification results corresponding to the frames of vehicle images respectively to obtain a plurality of license plate identification results;
and the license plate determining module is used for determining the license plate of the vehicle to be recognized based on the statistical information of the plurality of license plate recognition results.
9. A portable camera device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method according to any of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110240018.9A CN112836683B (en) | 2021-03-04 | 2021-03-04 | License plate recognition method, device, equipment and medium for portable camera equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110240018.9A CN112836683B (en) | 2021-03-04 | 2021-03-04 | License plate recognition method, device, equipment and medium for portable camera equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112836683A true CN112836683A (en) | 2021-05-25 |
| CN112836683B CN112836683B (en) | 2024-07-09 |
Family
ID=75934568
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110240018.9A Active CN112836683B (en) | 2021-03-04 | 2021-03-04 | License plate recognition method, device, equipment and medium for portable camera equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112836683B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113221894A (en) * | 2021-05-31 | 2021-08-06 | 中邮信息科技(北京)有限公司 | License plate number identification method and device of vehicle, electronic equipment and storage medium |
| CN114023076A (en) * | 2021-10-22 | 2022-02-08 | 北京商海文天科技发展有限公司 | Specific vehicle tracking method based on multi-source heterogeneous data |
| CN116977949A (en) * | 2023-08-24 | 2023-10-31 | 北京唯行科技有限公司 | Vehicle parking inspection method, device and equipment |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201503639U (en) * | 2009-09-10 | 2010-06-09 | 上海第二工业大学 | Handheld intelligent vehicle inspection system |
| CN103824066A (en) * | 2014-03-18 | 2014-05-28 | 厦门翼歌软件科技有限公司 | Video stream-based license plate recognition method |
| CN106778648A (en) * | 2016-12-26 | 2017-05-31 | 江苏盛世华安智能科技股份有限公司 | Vehicle tracing and Vehicle License Plate Recognition System and recognition methods |
| CN108470381A (en) * | 2018-03-08 | 2018-08-31 | 沈志强 | Roadside Parking management method |
| CN110033002A (en) * | 2019-04-19 | 2019-07-19 | 福州大学 | Detection method of license plate based on multitask concatenated convolutional neural network |
| CN110675637A (en) * | 2019-10-15 | 2020-01-10 | 上海眼控科技股份有限公司 | Vehicle illegal video processing method and device, computer equipment and storage medium |
| CN111145545A (en) * | 2019-12-25 | 2020-05-12 | 西安交通大学 | Road traffic behavior unmanned aerial vehicle monitoring system and method based on deep learning |
| US20200160700A1 (en) * | 2018-11-20 | 2020-05-21 | Laser Technology, Inc. | Handheld laser-based vehicle speed measurement device incorporating an automatic number plate recognition (anpr) function |
| CN111368830A (en) * | 2020-03-03 | 2020-07-03 | 西北工业大学 | License plate detection and identification method based on multi-video frame information and nuclear phase light filtering algorithm |
-
2021
- 2021-03-04 CN CN202110240018.9A patent/CN112836683B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201503639U (en) * | 2009-09-10 | 2010-06-09 | 上海第二工业大学 | Handheld intelligent vehicle inspection system |
| CN103824066A (en) * | 2014-03-18 | 2014-05-28 | 厦门翼歌软件科技有限公司 | Video stream-based license plate recognition method |
| CN106778648A (en) * | 2016-12-26 | 2017-05-31 | 江苏盛世华安智能科技股份有限公司 | Vehicle tracing and Vehicle License Plate Recognition System and recognition methods |
| CN108470381A (en) * | 2018-03-08 | 2018-08-31 | 沈志强 | Roadside Parking management method |
| US20200160700A1 (en) * | 2018-11-20 | 2020-05-21 | Laser Technology, Inc. | Handheld laser-based vehicle speed measurement device incorporating an automatic number plate recognition (anpr) function |
| CN110033002A (en) * | 2019-04-19 | 2019-07-19 | 福州大学 | Detection method of license plate based on multitask concatenated convolutional neural network |
| CN110675637A (en) * | 2019-10-15 | 2020-01-10 | 上海眼控科技股份有限公司 | Vehicle illegal video processing method and device, computer equipment and storage medium |
| CN111145545A (en) * | 2019-12-25 | 2020-05-12 | 西安交通大学 | Road traffic behavior unmanned aerial vehicle monitoring system and method based on deep learning |
| CN111368830A (en) * | 2020-03-03 | 2020-07-03 | 西北工业大学 | License plate detection and identification method based on multi-video frame information and nuclear phase light filtering algorithm |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113221894A (en) * | 2021-05-31 | 2021-08-06 | 中邮信息科技(北京)有限公司 | License plate number identification method and device of vehicle, electronic equipment and storage medium |
| CN114023076A (en) * | 2021-10-22 | 2022-02-08 | 北京商海文天科技发展有限公司 | Specific vehicle tracking method based on multi-source heterogeneous data |
| CN114023076B (en) * | 2021-10-22 | 2022-10-04 | 北京商海文天科技发展有限公司 | Specific vehicle tracking method based on multi-source heterogeneous data |
| CN116977949A (en) * | 2023-08-24 | 2023-10-31 | 北京唯行科技有限公司 | Vehicle parking inspection method, device and equipment |
| CN116977949B (en) * | 2023-08-24 | 2024-09-20 | 北京唯行科技有限公司 | Vehicle parking inspection method, device and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112836683B (en) | 2024-07-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107305627B (en) | Vehicle video monitoring method, server and system | |
| CN112836683B (en) | License plate recognition method, device, equipment and medium for portable camera equipment | |
| CN106951889A (en) | Underground high risk zone moving target monitoring and management system | |
| CN103986910A (en) | Method and system for passenger flow statistics based on cameras with intelligent analysis function | |
| CN109214280A (en) | Shop recognition methods, device, electronic equipment and storage medium based on streetscape | |
| CN104239386A (en) | Method and system for prioritizion of facial recognition matches | |
| CN112507760B (en) | Method, device and equipment for detecting violent sorting behaviors | |
| WO2021043074A1 (en) | Urban pet motion trajectory monitoring method based on image recognition, and related devices | |
| CN114360261B (en) | Vehicle reverse running identification method and device, big data analysis platform and medium | |
| CN111967396A (en) | Processing method, device and equipment for obstacle detection and storage medium | |
| CN112487894B (en) | Automatic track traffic protection area inspection method and device based on artificial intelligence | |
| CN112434566A (en) | Passenger flow statistical method and device, electronic equipment and storage medium | |
| CN111881322B (en) | Target searching method and device, electronic equipment and storage medium | |
| CN116311166A (en) | Traffic obstacle recognition method and device and electronic equipment | |
| CN111881984A (en) | Target detection method and device based on deep learning | |
| CN113033443B (en) | Unmanned aerial vehicle-based automatic pedestrian crossing facility whole road network checking method | |
| Kiew et al. | Vehicle route tracking system based on vehicle registration number recognition using template matching algorithm | |
| EP3244344A1 (en) | Ground object tracking system | |
| CN113963293A (en) | Intelligent operation and maintenance algorithm and system for charging station | |
| Kamenetsky et al. | Aerial car detection and urban understanding | |
| Patel et al. | Vehicle tracking and monitoring in surveillance video | |
| CN111627224A (en) | Vehicle speed abnormality detection method, device, equipment and storage medium | |
| CN116597482A (en) | Image clustering-based peer identification method, device, equipment and medium | |
| TWI617997B (en) | Intelligent object detection assistance system and method | |
| CN111626992B (en) | Intelligent crossing box inspection method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |