CN109558812B - Face image extraction method and device, training system and storage medium - Google Patents

Abstract

The invention relates to a method and device for extracting a face image, a training system and a storage medium. The method includes acquiring a plurality of images captured by a multi-eye camera, and extracting a first face image from the images according to the skin color feature of the face, Determine the second face image in the first face image according to the Euler number of the standard face image, obtain the overlapping area of the second face image between the images, and extract the third face image in the image according to the overlapping area face image. Based on the first face image extracted by the facial skin color feature, combined with the Euler number of the standard face image to exclude the image background similar to the face skin color in the first face image to obtain the second face image, and then according to the difference between the different images The face image is extracted from the overlapping area of the second face image, which improves the accuracy of extracting the face image, and is applied to the training system of the embedded device, which can accurately extract the trainees from the training scene images. Face image, easy to analyze the training personnel.

Description

Face image extraction method and device, training system and storage medium

technical field

The present invention relates to the technical field of image processing, and in particular, to a method for extracting a face image, a device for extracting a face image, a training system for embedded equipment, and a computer-readable storage medium.

Background technique

With the development of image processing technology, more and more industries need to collect images for data analysis. For example, in teaching training, collecting face images of students in the classroom can conduct statistical analysis on whether the trainees are absent from the training. , and the accurate extraction of face images from the captured images is the basis for data analysis of face images.

The face image extraction method adopted by the traditional technology is easily affected by background noise, so that the image background is mistakenly extracted as a face image, and it is difficult to accurately extract the face image from the captured source image.

SUMMARY OF THE INVENTION

Based on this, it is necessary to provide a method for extracting a face image, a device for extracting a face image, a training system for embedded equipment, and a computer readable storage medium.

A method for extracting a face image, comprising the steps of:

Obtain multiple images captured by a multi-camera;

Extract the first face image from the image according to the skin color feature of the face;

Determine a second face image in the first face image according to the Euler number of the standard face image;

Acquiring an overlapping area of the second face image between each of the images; and extracting a third face image in the images according to the overlapping area.

A device for extracting a face image, comprising:

The acquisition module is used to acquire multiple images captured by the multi-eye camera;

The first extraction module is used for extracting the first face image from the image according to the skin color feature of the face;

A determination module for determining a second human face image in the first human face image according to the Euler number of the standard human face image;

The second extraction module is configured to acquire the overlapping area of the second face images between the images; and extract the third face image in the images according to the overlapping area.

An embedded device training system, comprising: an embedded device, a programming host for programming the embedded device, a plug-in platform for the embedded device, a plurality of expansion modules for the embedded device, and a binocular Cameras and directing devices; of which,

The embedded device is electrically connected to the plurality of expansion modules through the plug-in platform, and is used for training personnel to perform training operations;

The binocular camera is used for collecting training scene images and sending them to the broadcasting director; the training scene images carry the face images of the training personnel;

The broadcasting device is used for extracting the face image of the trainee from the training scene image according to the above-mentioned method for extracting the face image.

A computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, implements the steps of the above-mentioned method for extracting a face image.

The above-mentioned extraction method and device, training system and storage medium of a face image obtain multiple images captured by a multi-eye camera, extract a first face image from the image according to the skin color feature of the face, and extract the first face image from the image according to the characteristics of the skin color of the face. The method determines the second face image in the first face image, obtains the overlapping area of the second face image between each image, and extracts the third face image in the image according to the overlapping area, which can be based on The first face image is obtained by extracting the skin color feature of the face, combined with the Euler number of the standard face image to exclude the image background similar to the skin color of the face in the first face image to obtain the second face image, and then according to the difference between different images The overlapping area of the second face image can extract an accurate face image in the image, which overcomes the problem that the traditional technology is easy to extract the image background as a face image by mistake, and improves the accuracy of extracting the face image. In the training system of the embedded device, the face images of the trainees can be accurately extracted from the training scene images, which is convenient for statistical analysis of the trainees.

Description of drawings

1 is a schematic flowchart of a method for extracting a face image in one embodiment;

2 is a structural block diagram of an apparatus for extracting a face image in one embodiment;

3 is a schematic structural diagram of a training system of an embedded device in one embodiment;

4 is a schematic structural diagram of a socket in one embodiment;

5 is a schematic structural diagram of a socket in another embodiment;

6 is a schematic structural diagram of a pin in one embodiment;

7 is a schematic structural diagram of a jack connecting line in one embodiment;

FIG. 8 is a diagram of the internal structure of a computer device in one embodiment.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention.

When an element is referred to as being "connected" to another element, it can be directly connected to the other element or intervening elements may also be present. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terms used herein in the description of the present invention are for the purpose of describing specific embodiments only, and are not intended to limit the present invention. The "first" and "second" mentioned in the present invention do not represent a specific quantity and order, but are only used for the distinction of names.

In one embodiment, a method for extracting a face image is provided. Referring to FIG. 1, FIG. 1 is a schematic flowchart of a method for extracting a face image in one embodiment. The device is implemented, and the method for extracting the face image may include the following steps:

Step S101, acquiring multiple images captured by a multi-eye camera.

In this step, the target area can be photographed by using multiple cameras, and each camera can photograph to obtain an image, so that multiple images photographed by each camera can be obtained. Taking a binocular camera as an example, an image of the same target area can be captured by the binocular camera, and two images can be obtained respectively, that is, the image captured by the left-eye camera and the image captured by the right-eye camera, and when there is a person in the target area. , the multiple images captured by the multi-camera to capture the target area carry face images, and the face images in the multiple images can be extracted based on the multiple images. Specifically, the target area may be a training scene image of the training personnel in the training classroom or the training room.

Step S102, extracting a first face image from the image according to the skin color feature of the face.

This step mainly extracts the first face image from multiple images based on the skin color feature of the face. Among them, according to the characteristics of the skin color of the face, it can be determined that the range of the skin color of the face is 100≤B≤120, 140≤R≤160, so the gray value of the pixel points in the color channel in the image within this range can be set to 1, and the remaining pixels The gray value of the point is set to 0, where B represents the blue channel (B) in the RGB color standard, and R represents the red channel (R) in the RGB color standard. The first face image is extracted respectively, and the gray value of the pixel point in the area of the suspected face in the first face image may be marked as 1, or may be marked as 0.

Step S103, determining a second face image in the first face image according to the Euler number of the standard face image.

In this step, the Euler number refers to a constant calculated according to the number of fragments in an area of the image in the image, and the standard face image has a relatively stable Euler number. The Euler number is used as the main feature to identify the face, and the Euler number is easier to judge for the static image of the face. Since the captured image is usually a static image, the Euler number based on the standard face image can be obtained from the first person. The image background that is similar to the skin color of the human face but not the human face is eliminated from the face image, that is, the background area that is suspected to be a human face but is not a human face in the first human face image is excluded, and the second face image of each image can be obtained respectively.

Step S104, acquiring the overlapping area of the second face image between the images; extracting the third face image in the images according to the overlapping area.

Wherein, each captured image corresponds to a second face image, and this step is mainly to obtain the overlapping area of the second face image between the images, and then according to the overlapping area, each second face image The position in each image is used to determine the third face image in each image, so that the overlapping area of the second face image of different images can be used to identify the face image in each image, which further eliminates the interference of noise. It can improve the accuracy of face image recognition.

The extraction method of the above-mentioned face image, obtains multiple images captured by a multi-eye camera, extracts a first face image from the image according to the skin color feature of the face, and extracts the first face image according to the Euler number of the standard face image in the first face image. Determine the second face image, obtain the overlapping area of the second face image between each image, and extract the third face image in the image according to the overlapping area, which can be based on the third face image obtained by extracting the skin color feature of the face. A face image, combined with the Euler number of the standard face image to exclude the image background similar to the skin color of the face in the first face image to obtain the second face image, and then according to the mutual overlap of the second face images between different images The accurate face image is extracted from the image, which overcomes the problem that the traditional technology easily extracts the image background as a face image by mistake, and improves the accuracy of extracting the face image.

In one embodiment, the step of extracting the first face image from the image according to the skin color feature of the face may include:

Step S201, binarize the image according to the skin color feature of the face to obtain a fourth face image.

In this step, the value range of the color channel of each pixel of the image can be set according to the skin color feature of the face, so as to binarize the image according to the value range to obtain a fourth face image. Specifically, the skin color range of the face is 100≤B≤120, 140≤R≤160, so in each image, the gray value of the pixel whose color channel is within this range can be set to 1, and the remaining The gray value of the pixel can be set to 0, so that a fourth face image can be obtained, wherein B represents the blue channel (B) in the RGB color standard, and R represents the red channel (R) in the RGB color standard.

Step S202, determining multiple target regions in the fourth face image.

In this step, there are multiple target areas in the fourth face image, and the target area here refers to the area that is suspected to be a face in the fourth face image. In the process of binarizing the image with the skin color feature of the face, the The gray value of the pixel point in the suspected face area is set to 1, and the gray value of other pixels is set to 0, then the target area corresponds to the whitening area in the fourth face image (that is, the gray value of the pixel point is 1). area).

Step S203, acquiring the number of pixels of each target area.

In this step, the number of pixels in each target area in the fourth face image can be counted, that is, the number of pixels in the suspected face area in the fourth face image. If the pixel gray value of the target area is set to 1, If the gray value of other pixels is set to 0, the number of pixels in the whitening area in the fourth face image can be counted.

Step S204, if the number of pixels is less than the set threshold, remove the target area in the fourth face image to obtain the first face image.

The number of pixels in the target area can be compared with a set threshold, and if the number of pixels in the target area is less than the threshold, the target area can be removed from the fourth face image, where the threshold can be based on the standard human face. The number of pixels of the image is set. If the number of pixels in the target area is too small (for example, less than 1000), it can be considered that the target area is not the area where the face image is located, so the target area can be removed. If the gray value of the pixel point of the target area is set to 1, the way to remove the target area may be to set the gray value of the pixel point of the target area to 0, so as to obtain the first face image. Further, before comparing the number of pixels in the target area with the set threshold, the fourth face image can be subjected to spherical erosion image processing, and then the fourth face image processed by the spherical erosion image can be adaptively processed. Median filtering achieves a smoothing effect and further improves the accuracy of face image recognition.

In this embodiment, the number of pixels in the suspected face area in the fourth face image can be counted, and the non-face area can be excluded from the fourth face image to obtain the first face image, which further improves the extraction efficiency of the face image. Accuracy, more effectively eliminate the interference of background noise on the extracted face image.

In one embodiment, after the step of extracting the third face image in the image according to the overlapping area, the method may further include:

Perform binarization processing on the third face image; count the third face image after the binarization processing to determine the number of people in the image.

This embodiment mainly performs binarization processing on the third face image, and the face area in the third face image can be segmented and marked for binarization counting, that is, the gray value of the pixel point of the face area is set. is 1, the pixel gray value of the non-face area is set to 0, and the number of people in each image can be obtained by counting the number of areas with a pixel gray value of 1. In this way, the number of characters in the image can be quickly and accurately counted, which is convenient for statistics on the training personnel present in the training classroom or training room.

In one embodiment, after the step of extracting the third face image in the image according to the overlapping area, the method may further include:

Obtain a face image database; wherein, the face image database records multiple face images and the identity of the person corresponding to the face image; the third face image is matched with the face image in the face image database to determine The identity of the person in the image.

This embodiment mainly performs matching analysis between the third face image and the face image in the face image database, so as to determine the identity of the person in each image. Specifically, the face area in the third face image can be segmented and marked, and the face area can be matched and compared with the face image database pre-stored in the background, such as the face image database of all the trainees, to determine the actual training experience. The identity information of the trainees can also be determined, and the trainees who are absent can also be determined.

In one embodiment, after the step of extracting the third face image in the image according to the overlapping area, the method may further include:

Determine a variety of standard face poses; perform pose analysis on the third face image according to the multiple standard face poses to determine the face pose in the image.

In this embodiment, a variety of standard face poses can be obtained, and the pose comparison analysis can be performed between the multiple standard face poses and the face poses of the face region segmented and marked from the third face image. The face pose area is compared with the face pose area in the third face image, so as to determine the face pose of the person in each image. In the training teaching scenario, this embodiment can recognize the face posture of the trainee, and the trainee who has difficulty in operation usually has a standard face posture, so the solution of this embodiment can identify the difficulty in operation and provide early warning prompts to the training teachers in a timely manner, and send early warning prompts to the training teachers about the training status of the trainees.

In one embodiment, an apparatus for extracting a human face image is provided. Referring to FIG. 2, FIG. 2 is a structural block diagram of an apparatus for extracting a human face image in one embodiment. The apparatus for extracting a human face image may include:

an acquisition module 101, configured to acquire multiple images captured by a multi-eye camera;

The first extraction module 102 is used to extract the first face image from the image according to the skin color feature of the face;

Determining module 103, for determining the second face image in the first face image according to the Euler number of the standard face image;

The second extraction module 104 is configured to acquire the overlapping area of the second face image between the images; extract the third face image in the images according to the overlapping area.

In one embodiment, the first extraction module 102 is further configured to:

Binarize the image according to the skin color feature of the face to obtain a fourth face image; determine multiple target areas in the fourth face image; obtain the number of pixels in each target area; if the number of pixels is less than the set threshold , the target area is removed from the fourth face image to obtain the first face image.

In one embodiment, it can also include:

The number determination unit is used for binarizing the third face image; counting the third face image after the binarization processing to determine the number of people in the image.

In one embodiment, it can also include:

An identity determination unit is used to obtain a face image database; wherein, the face image database records a plurality of face images and the person identities corresponding to the face images; Face images are matched to determine the identity of the person in the image.

In one embodiment, it can also include:

The posture determination unit is used to determine various standard face postures; perform posture analysis on the third human face image according to the various standard human face postures, and determine the human face posture in the image.

The apparatus for extracting a face image of the present invention corresponds to the method for extracting a face image of the present invention. For the specific limitation of the apparatus for extracting a face image, please refer to the limitation on the method for extracting a face image above. The technical features and beneficial effects described in the embodiments of the face image extraction method are all applicable to the embodiments of the face image extraction apparatus, and will not be repeated here. Each module in the above-mentioned apparatus for extracting a face image may be implemented in whole or in part by software, hardware and combinations thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

In one embodiment, a training system for embedded devices is provided. Referring to FIG. 3 , FIG. 3 is a schematic structural diagram of the training system for embedded devices in one embodiment. The training system for embedded devices may include: :

An embedded device 300, a programming host 400 for programming the embedded device 300, a plug-in platform 100 for the embedded device, a plurality of expansion modules 200 for the embedded device, a binocular camera 800, and a broadcast guide device 700; wherein ,

The embedded device 300 is electrically connected with a plurality of expansion modules 200 through the plug-in platform 100, and is used for training personnel to perform training operations;

The binocular camera 800 is used for collecting the training scene image, and sending the training scene image to the broadcasting director 700; wherein, the training scene image carries the face image of the training personnel;

The broadcasting device 700 is configured to extract a face image of a trainee from a training scene image according to the method for extracting a face image according to any one of the above embodiments.

In this embodiment, the embedded device 300 can be placed on the patch cord platform 100, and the embedded device 300 can be electrically connected to a plurality of expansion modules inserted on the patch cord platform 100 through the patch cord platform. The device can be an Arduino embedded device. The Arduino embedded device can be used in programming teaching. The embedded device can be equipped with different expansion modules. The expansion modules can include but are not limited to: LED light modules, traffic light modules, and laser head sensor modules. , temperature and humidity sensor module, PS2 joystick module, relay module, obstacle avoidance photoelectric sensor module, finger detection heartbeat module, high-sensitivity microphone sensor module, touch sensor module, flame sensor module, full-color LED module, line-hunting sensor module , Hall sensor module, rotary encoder module, buzzer module, soil sensor module, moisture sensor module, mercury switch module, gas sensor module, photoresistor module and vibration switch module, so that the embedded The device implements corresponding functions based on different expansion modules.

The programming host 400 can be signal-connected to the embedded device 300 through a connecting line such as a USB data cable. The programming host 400 has the functions of programming and debugging programs. After the programming host 400 is successfully connected to the embedded device 300, the trainee can program The host 400 writes a program to the embedded device 300 to debug the hardware and software of the embedded device 300 .

The binocular camera 800 can be used to collect the training scene image carrying the face image of the training personnel, and send the training scene image to the broadcasting director 700, so that the broadcasting director 700 can accurately identify the training scene image from the training scene image. The face image of the trainee, and analyze the trainee based on the face image, and the broadcast director 700 can perform face recognition analysis and facial posture analysis on the trainee, and timely report to the teacher for students who have difficulty in operation. Provide early warning prompts so that teachers can provide timely guidance and answers, so that the embedded equipment training system integrates embedded equipment teaching, teaching video recording, directing and classroom monitoring and analysis, which is conducive to improving the training effect of embedded equipment. It is convenient for teaching and student training of embedded devices.

In one embodiment, the plug-in platform is provided with a plurality of sockets; the inner surface of the sockets is provided with a first magnetic area and a first conductive area; the expansion module is provided with a plurality of pins matching the sockets The outer surface of the pin is provided with a second magnetic area and a second conductive area; wherein, when the extension module is inserted into the jack of the plug-in platform through the pin, the second magnetic area and the first magnetic area pass through the magnetic Attraction, the second conductive area is in conductive contact with the first conductive area.

In this embodiment, the plug-in platform 100 is provided with a plurality of jacks 110, and the jacks are matched with the pins of the expansion module, that is to say, the shape, size and other attribute parameters of the jacks are consistent with the pins. Matching, so that the pins of the expansion module can be inserted into the jack. The inner surface of the socket 110 is provided with a first magnetic area and a first conductive area, and the first magnetic area is used for adsorbing corresponding magnetic components. For example, when pins with corresponding magnetic properties are inserted into the socket, the The first magnetic area firmly adsorbs the pin, and the first conductive area is mainly used for conducting conductive contact with other conductive components. For example, when a conductive pin is inserted into the jack, the first conductive area and the pin Conducting electrical contact, which can conduct electric current to each other. There are various ways of opening the jack 110 on the patching platform 100 . The jack 110 can be directly opened on the patching platform 100 , or can be opened through the socket 120 installed on the patching platform 100 .

Taking the socket 120 installed on the plug-in platform 100 as an example, the opening method of the socket 110 will be described. Referring to FIG. 4 , FIG. 4 is a schematic structural diagram of the socket in one embodiment. The socket 120 can be connected by welding, integral molding, etc. The socket 120 can be installed on the plug-in platform 100 in a way, and a through hole can be opened on the socket 120 as the jack 110. The inner surface of the jack 110 is a side 111, and a first conductive area and a first magnetic area can be set on the side 111. , and there are many ways to set the conductive area and the magnetic area on the inner surface of the jack. For example, you can coat the inner surface of the jack with magnetic materials and conductive materials, or you can attach a magnetic patch on the inner surface of the jack. and conductive patches. Specifically, a magnetic conductive metal patch can be attached to the inner surface of the jack. Since the magnetic conductive metal has both magnetism and conductivity, the magnetic conductive metal will simultaneously form a first magnetic field at the position attached to the inner surface. area and the first conductive area, so that when the pin is inserted into the jack, the corresponding magnetic area of the pin will attract the magnetic conductive metal, and the magnetic conductive metal will interact with the corresponding conductive area of the pin. The conductive contact makes the pins and sockets fit more closely, not easy to fall off, and the reliability of conduction is also higher.

In one embodiment, the number of sockets provided on the socket may be multiple, and the first conductive areas on the inner surfaces of the sockets are communicated with each other. The socket body may be provided with a plurality of sockets, and the first conductive areas on the inner surfaces of the sockets communicate with each other. Specifically, referring to FIG. 5 , FIG. 5 is a schematic structural diagram of a socket in another embodiment. The socket 130 is provided with a first socket 131 , a second socket 132 and a third socket 133 , wherein the first socket 131 , The inner surfaces of the second jack 132 and the third jack 133 can communicate with each other, and the inner surfaces of the first jack 131, the second jack 132 and the third jack 133 are respectively provided with a first conductive area, and each jack The first conductive areas of the holes communicate with each other, so that the first pin inserted into the first insertion hole 131 can be in conductive contact with the second pin inserted into the second insertion hole 132 or the third insertion hole 133, so that each insertion hole The conducted current can communicate with other sockets, and the socket has a more compact structure, which facilitates the simultaneous conductive connection of pins of multiple expansion modules.

In one embodiment, further, the shape of the socket may be a cube.

In this embodiment, the shape of the socket body is a cube, and the cube is used as the shape of the socket, which is conducive to making the socket more stably installed on the plug-in platform. The first conductive areas on the inner surface of each socket are connected with each other, and the sockets provided on the six surfaces of the socket body can be connected to pins of different expansion modules, which is also beneficial to improve the utilization rate of the plug-in platform.

The expansion module 200 is provided with a plurality of pins 210 that match the jacks 110 of the plug-in platform 100 , and the pins 210 match the jacks 110 , that is, the shape, size and other attribute parameters of the pins 210 match the jacks. Matching, so that the expansion module 200 can be inserted into the socket 110 through its pins 210, and the outer surface of the pins 210 is provided with a second magnetic area and a second conductive area, and the second magnetic area is used to connect with the socket. The corresponding magnetic area of the hole is adsorbed, for example, when the pin 210 is inserted into the jack 110 with the corresponding magnetic area, the second magnetic area and the jack can be firmly adsorbed; the second conductive area is mainly used to connect with The corresponding conductive area of the jack 110 is in conductive contact. For example, when the pin 210 is inserted into the jack 110 with the corresponding conductive area, the second conductive area is in conductive contact with the jack, and can conduct current with each other. Disposing the first magnetic region on the inner surface of the socket can also prevent the miscellaneous magnetic substances from contaminating the socket and affecting the firmness when contacting the plug.

There are various ways to set the conductive area and the magnetic area on the outer surface of the pin 210. For example, the outer surface of the pin 210 can be coated with magnetic material and conductive material, or the outer surface of the pin 210 can be pasted. Set up magnetic patch and conductive patch.

Specifically, referring to FIG. 6, FIG. 6 is a schematic diagram of the structure of a pin in an embodiment. A magnetic conductive metal patch can be attached to the outer surface 211 of the pin 210. Since the magnetic conductive metal has both magnetism and conductivity, Therefore, the magnetic conductive metal will simultaneously form a second magnetic region and a second conductive region at the position where the outer surface is attached. In this way, when the pin 210 is inserted into the socket 110 , the magnetic conductive metal will interact with the socket 110 . The first magnetic area on the inner surface is attracted, and the magnetic conductive metal will conduct conductive contact with the first conductive area on the inner surface of the jack, so that the pin 210 and the jack 110 can be more closely matched, not easy to fall off, and reliable in conduction Sex is also higher.

The jack 110 of the patching platform 100 can be used in conjunction with the pins 210 of the expansion module 200. When the expansion module 200 is inserted into the jack 110 of the patching platform 100 through its pins 210, the second magnetic area of the pins 210 It is magnetically attracted to the first magnetic area of the jack 110, and the second conductive area of the pin 210 is in conductive contact with the first conductive area of the jack 110, so that the pin 210 and the jack 110 can be in stable magnetic contact, so that the expansion The module 200 is not easy to fall off from the patch platform, and the conductive contact between the first conductive area of the jack 110 and the second conductive area of the pin 210 can ensure that the currents of the expansion module 200 and the patch platform can be connected to each other.

In the above-mentioned embodiment, the plug-in platform of the embedded device and a plurality of expansion modules, the plug-in platform of the embedded device is provided with a plurality of jacks with a first magnetic area and a first conductive area on the inner surface, and the expansion module A plurality of pins are provided, and the outer surface of the pins is provided with a second magnetic area and a second conductive area. The pins are matched with the jacks on the plug-in platform. When the extension module is inserted into the socket through its pins When the socket is inserted, the second magnetic area and the first magnetic area are magnetically attracted, and the second conductive area is in conductive contact with the first conductive area. The area is pulled together, so that the plug-in platform and the expansion module are in stable contact, and it is not easy to fall off, so as to avoid poor circuit contact between the plug-in platform and the expansion module.

In one embodiment, magnetic conductive metal is attached to the inner surface of the jack of the plug-in platform; the pins of the extension module are made of magnetic conductive metal.

In this embodiment, a magnetic conductive metal patch can be attached to the inner surface of the jack 110 of the patching platform 100. Since the magnetic conductive metal has both magnetic properties and electrical conductivity, the magnetic conductive metal patch is included in the The first magnetic area and the first conductive area will be formed at the position where the surface is attached; the pin 210 of the expansion module 200 can be made of magnetic conductive metal. A second magnetic region and a second conductive region are simultaneously formed thereon. In this way, when the pins 210 of the expansion module 200 are inserted into the jacks 110 of the plug-in platform 100 , the outer surfaces of the pins 210 and the inner surfaces of the jacks 110 are magnetically attracted, and the outer surfaces of the pins 210 are magnetically attracted to each other. The inner surface of the jack 110 is in conductive contact, so that the pins of the extension module can be electrically connected to the inner surface of the jack of the plug-in platform, and the pins and the inner surface of the jack can also be pulled together through the corresponding magnetic area. , so that the expansion module is in stable contact with the socket, not easy to fall off, and avoids poor circuit contact.

In one embodiment, further, the shapes of the sockets and the pins are hexagonal prisms.

In this embodiment, the shape of the jack of the plug-in platform may be a hexagonal prism, and correspondingly, the shape of the pins of the expansion module may also be a hexagonal prism. When the pin is matched with the jack, it can fit more closely, and it can also avoid the problem that the pin is easy to rotate relative to the jack when the pin is connected to the jack in the traditional technology, which affects the electrical conductivity. reliability of conduction. Further, a magnetic conductive metal patch is attached to the inner surface of the hexagonal column-shaped socket, so that when the pins of the expansion module are inserted into the socket, the magnetic conductive metal patch is used for six-sided contact. Ensure the reliability of the conductive connection.

In one embodiment, the jacks can be connected by jack connecting wires.

In this embodiment, the jacks 110 shown in FIG. 3 may be electrically connected through jack connecting wires. Wherein, the back of the plug-in platform 100 may be provided with a plurality of back jacks such as back jacks 141 and 142, and these back jacks are respectively electrically connected to the jacks on the upper surface of the plug-in platform 100. The jacks 110 on the upper surface of the cable platform 100 are in conduction with the back jacks 141 , that is, the pins inserted into the jacks 110 can be electrically connected to other jacks such as the back jacks 142 through their corresponding back jacks 141 , and the jack connecting line of this embodiment can realize conductive connection between different jacks.

Specifically, referring to FIG. 7 , FIG. 7 is a schematic structural diagram of a jack connecting line in an embodiment, and the jack connecting line may include a first pin 221 and a second pin matching with the jack 110 of the plug-in platform 222, the first pin 221 and the second pin 222 are conductively connected through the wire 20, one end of the wire 20 is conductively connected to the second conductive area on the outer surface of the first pin 221, and the other end of the wire 20 is connected to the second conductive area on the outer surface of the first pin 221. The second conductive area on the outer surface of the second pin 222 is conductively connected, so as to realize the conductive connection between the first pin 221 and the second pin 222 through the wire 20, and the jack connecting line can be used to connect different jacks Conduct conductive connections to improve the reliability and flexibility of electrical connections to jacks. It should be noted that, for the specific definition of the first pin 221 and the second pin 222, reference may be made to the content of the definition of the pin 210 of the expansion module in the above-mentioned embodiment, which is not repeated here.

In one embodiment, it may further include: a first image acquisition device 500 and a projection device 600 .

Among them, the first image acquisition device 500 is used to collect the video images of the training personnel operating the plug-in platform and the expansion module of the embedded equipment, and send them to the broadcast director 700. The broadcast director 700 can receive the video images and project them through The apparatus 600 projects the video image.

In this embodiment, the first image capture device 500 may be disposed above the plug-in platform of the embedded device, and the first image capture device 500 may include a camera 510 and a plurality of auxiliary lights 520 disposed around it. The auxiliary lights The 520 can provide auxiliary lighting when the light is insufficient, so that the camera 510 can capture clear video images. The first image acquisition device 500 can collect the video images of the training personnel operating the plug-in platform and the expansion module of the embedded device, such as recording the teaching demonstration operation video images of the teacher or collecting the operation video images of the trainees, and then the video images are recorded. Sent to the broadcasting director 700, the broadcasting director 700 may have built-in intelligent recording and broadcasting guide software, which can be used for recording and broadcasting instructional videos, and the first image capture device 500 can record teachers on the plug-in platform, extension module of the embedded device And the video images such as hardware connection operation and programming operation on the programming host, the recorded video images can be imported into the database through the intelligent director software background data management software, and then realize on-demand. The directing device 700 can be provided with a touch screen, and the training personnel can directly select and play the teaching video and learning courseware through touch operation on the touch screen, and the on-demand video or courseware is projected on the screen through the projection device 600. The projection apparatus 600 can be used for playing and displaying on-demand videos or courseware, including training teaching or learning operations of embedded devices.

In the above-mentioned embodiment, the operation video of the embedded device, the plug-in platform and the expansion module of the training personnel can be collected by the first image acquisition device, and the video-on-demand and projection can be performed through the broadcast director and the projection device, which is convenient for the teaching and training of the embedded device. , to improve the teaching and training quality of embedded equipment.

In one embodiment, it may further include: a workbench.

In this embodiment, the training system of the embedded device may further include a workbench 10 , and the workbench 10 may be used to place the programming host 400 and the broadcast director 700 , so that the training personnel can easily perform the programming on the host 400 and the programming host on the workbench 10 . The operation of the broadcast guide device 700 is also convenient for teachers to teach and monitor the training students in the classroom.

In one embodiment, a first telescopic rod may also be included; one end of the first telescopic rod is connected to the workbench, and the other end of the first telescopic rod is connected to the projection device and the binocular camera.

In this embodiment, the training system of the embedded device may further include a first telescopic rod 910 , one end of the first telescopic rod 910 can be fixedly or detachably connected to the workbench 10 , and the other end can be connected to the projection device 600 and the dual The binocular camera 800 can be arranged on the back of the projection device 600 to realize the connection between the projection device 600 and the binocular camera 800 and the first telescopic rod 910 . In this embodiment, by adjusting the telescopic state of the first telescopic rod 910, the first telescopic rod 910 can be adjusted to the extended state when the projection device 600 or the binocular camera 800 is used, and the specific length of the first telescopic rod 910 can be adjusted according to the projection Or the image capture needs to be adjusted, and the first telescopic rod 910 can be adjusted to a retracted state when it is not needed to save space.

In one embodiment, it may further include: a second telescopic rod; one end of the second telescopic rod is arranged on the first telescopic rod, and the other end of the second telescopic rod is connected to the first image acquisition device.

In this embodiment, the training system of the embedded device may further include a second telescopic rod 920. One end of the second telescopic rod 920 can be fixedly or detachably connected to the first telescopic rod 910, and the other end is connected to the first telescopic rod 910. The image acquisition device 500 enables the first image acquisition device 500 to acquire video images of the training personnel operating the plug-in platform and the expansion module of the embedded device. In this embodiment, by adjusting the telescopic state of the second telescopic rod 920, the second telescopic rod 920 can be adjusted to the extended state when the first image acquisition device 500 is used, and the specific length can be determined according to the plug-in platform and the expansion module of the embedded device When the first image capture device 500 is not used, the second telescopic rod 920 can be adjusted to a retracted state, and when the first image capture device 500, the projection device 600 and the binocular camera 800 are not used At the same time, the first telescopic rod 910 and the second telescopic rod 920 are adjusted to a retracted state to save space.

In one embodiment, it may further include: a mobile controller for controlling the broadcasting guide device.

In this embodiment, the training system of the embedded device may further include a mobile controller 30, and the mobile controller 30 may be used to control the broadcast director 700, wherein the mobile control device 30 is connected to the broadcast director 700 through WIFI or Bluetooth, so that , the training personnel such as teachers can use the mobile control device 30 to control the broadcast guide device 700 in the course of moving freely in the classroom or training room. The mobile control device 30 can have built-in broadcast guide control software through WIFI or Bluetooth. It is connected with the broadcasting director 700 to realize the function of selection and on-demand of learning videos.

In one embodiment, a computer device is provided, and the computer device is applied to the broadcasting device according to any of the above embodiments, and its internal structure diagram may be shown in FIG. 8 , which is the computer device in one embodiment. internal structure diagram. The computer equipment includes a processor, memory, a network interface, a display screen, and an input device connected by a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium, an internal memory. The nonvolatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the execution of the operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for extracting a face image is realized. The display screen of the computer equipment may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment may be a touch layer covered on the display screen, or a button, a trackball or a touchpad set on the shell of the computer equipment , or an external keyboard, trackpad, or mouse.

Those skilled in the art can understand that the structure shown in FIG. 8 is only a block diagram of a partial structure related to the solution of the present invention, and does not constitute a limitation on the computer equipment to which the solution of the present invention is applied. Include more or fewer components than shown in the figures, or combine certain components, or have a different arrangement of components.

In one embodiment, a computer device is provided, comprising a memory, a processor, and a computer program stored on the memory and running on the processor, and the processor implements the following steps when executing the computer program:

Obtaining multiple images captured by a multi-eye camera; extracting a first face image from the image according to the skin color feature of the face; determining a second face image in the first face image according to the Euler number of the standard face image; Obtain the overlapping area of the second face image between the images; extract the third face image in the images according to the overlapping area.

In one embodiment, the processor further implements the following steps when executing the computer program:

Binarize the image according to the skin color feature of the face to obtain a fourth face image; determine multiple target areas in the fourth face image; obtain the number of pixels in each target area; if the number of pixels is less than the set threshold , the target area is removed from the fourth face image to obtain the first face image.

In one embodiment, the processor further implements the following steps when executing the computer program:

Perform binarization processing on the third face image; count the third face image after the binarization processing to determine the number of people in the image.

In one embodiment, the processor further implements the following steps when executing the computer program:

Obtain a face image database; wherein, the face image database records multiple face images and the identity of the person corresponding to the face image; the third face image is matched with the face image in the face image database to determine The identity of the person in the image.

In one embodiment, the processor further implements the following steps when executing the computer program:

Determine a variety of standard face poses; perform pose analysis on the third face image according to the multiple standard face poses to determine the face pose in the image.

The above computer equipment, through the computer program running on the processor, improves the accuracy of extracting face images, and is applied to the training system of the embedded equipment, which can accurately extract the training scene images from the training scene images. The face images of the trainees are convenient for statistical analysis of the trainees.

Those of ordinary skill in the art can understand that all or part of the process in the method for extracting a face image described in any of the above embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can store In a non-volatile computer-readable storage medium, when the computer program is executed, it may include the processes of the above-mentioned method embodiments. Wherein, any reference to memory, storage, database or other medium used in the various embodiments provided by the present invention may include non-volatile and/or volatile memory. Nonvolatile memory may include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Road (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

Accordingly, in one embodiment, a computer-readable storage medium is provided on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

Obtaining multiple images captured by a multi-eye camera; extracting a first face image from the image according to the skin color feature of the face; determining a second face image in the first face image according to the Euler number of the standard face image; Obtain the overlapping area of the second face image between the images; extract the third face image in the images according to the overlapping area.

In one embodiment, the computer program further implements the following steps when executed by the processor:

Binarize the image according to the skin color feature of the face to obtain a fourth face image; determine multiple target areas in the fourth face image; obtain the number of pixels in each target area; if the number of pixels is less than the set threshold , the target area is removed from the fourth face image to obtain the first face image.

In one embodiment, the computer program further implements the following steps when executed by the processor:

Perform binarization processing on the third face image; count the third face image after the binarization processing to determine the number of people in the image.

In one embodiment, the computer program further implements the following steps when executed by the processor:

Obtain a face image database; wherein, the face image database records multiple face images and the identity of the person corresponding to the face image; the third face image is matched with the face image in the face image database to determine The identity of the person in the image.

In one embodiment, the computer program further implements the following steps when executed by the processor:

Determine a variety of standard face poses; perform pose analysis on the third face image according to the multiple standard face poses to determine the face pose in the image.

The above-mentioned computer-readable storage medium, through the computer program stored therein, improves the accuracy of extracting face images, and is applied to the training system of embedded devices, so that the training students can be accurately extracted from the training scene images. face images, which is convenient for statistical analysis of trainees.

The technical features of the above embodiments can be combined arbitrarily. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features It is considered to be the range described in this specification.

The above-mentioned embodiments only represent several embodiments of the present invention, and the descriptions thereof are specific and detailed, but should not be construed as a limitation on the scope of the invention patent. It should be pointed out that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can also be made, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention should be subject to the appended claims.

Claims (10)

1. A method for extracting a face image is characterized by comprising the following steps:

acquiring a plurality of images shot by a multi-view camera; the multiple images all comprise face images; the multi-camera shoots the same target area through a plurality of cameras to obtain the image; the target area comprises a training scene with training personnel in a training classroom;

extracting a first face image from the image according to the face skin color characteristics;

determining a second face image in the first face image according to the Euler number of the standard face image; the second face image corresponds to each image;

acquiring an overlapping area of the second face image among the images; extracting a third face image in the images according to the positions of the overlapping areas in the second face images;

and carrying out attitude analysis on the third face image according to a plurality of preset standard face attitudes to determine the face attitude of the third face image, and if the face attitude is a target face attitude, determining that a practical training student corresponding to the third face image is a practical training student with difficulty in operation.

2. The method for extracting a face image according to claim 1, wherein the step of extracting a first face image from the image according to the face skin color feature comprises:

carrying out binarization on the image according to the skin color characteristics of the human face to obtain a fourth human face image;

determining a plurality of target regions in the fourth face image;

acquiring the number of pixel points of each target area;

and if the number of the pixel points is less than a set threshold value, removing the target area from the fourth face image to obtain the first face image.

3. The method for extracting a face image according to claim 1 or 2, wherein after the step of extracting a third face image from the images according to the position of the overlapping area in each of the second face images, the method further comprises:

carrying out binarization processing on the third face image;

and counting the third face image after binarization processing, and determining the number of people in the image.

4. The method for extracting a face image according to claim 1 or 2, wherein after the step of extracting a third face image from the images according to the position of the overlapping area in each of the second face images, the method further comprises:

acquiring a face image database; the face image database records a plurality of face images and the identity of a person corresponding to the face images;

and matching the third face image with the face image in the face image database to determine the identity of the person in the image.

5. The method for extracting a face image according to claim 1 or 2, wherein after the step of extracting a third face image from the images according to the position of the overlapping area in each of the second face images, the method further comprises:

determining various standard human face postures;

and carrying out attitude analysis on the third face image according to the plurality of standard face attitudes to determine the face attitude in the image.

6. An extraction apparatus for a face image, comprising:

the acquisition module is used for acquiring a plurality of images shot by the multi-view camera; the multiple images all comprise face images; the multi-camera shoots the same target area through a plurality of cameras to obtain the image; the target area comprises a training scene with training personnel in a training classroom;

the first extraction module is used for extracting a first face image from the image according to the face complexion characteristics;

the determining module is used for determining a second face image in the first face image according to the Euler number of a standard face image; the second face image corresponds to each image;

the second extraction module is used for acquiring an overlapping area of the second face image among the images; extracting a third face image in the images according to the positions of the overlapping areas in the second face images; and carrying out attitude analysis on the third face image according to a plurality of preset standard face attitudes to determine the face attitude of the third face image, and if the face attitude is a target face attitude, determining that a practical training student corresponding to the third face image is a practical training student with difficulty in operation.

7. A practical training system of embedded equipment is characterized by comprising: the system comprises embedded equipment, a programming host used for programming the embedded equipment, a plug-in platform of the embedded equipment, a plurality of extension modules of the embedded equipment, a binocular camera and a director; wherein,

the embedded equipment is electrically connected with the plurality of extension modules through the plug wire platform and is used for training personnel to perform training operation;

the binocular camera is used for acquiring practical training scene images and sending the practical training scene images to the broadcasting guide device; the training scene image carries a face image of the training personnel;

the director is used for extracting the facial image of the practical training personnel from the practical training scene image according to the facial image extraction method of any one of claims 1 to 5.

8. The practical training system of the embedded device according to claim 7, wherein the plug wire platform is provided with a plurality of jacks; a first magnetic region and a first conductive region are arranged on the inner surface of the jack; the expansion module is provided with a plurality of pins matched with the jacks; a second magnetic region and a second conductive region are arranged on the outer surface of the pin; wherein,

when the expansion module is inserted into the jack of the plug wire platform through the pins, the second magnetic area and the first magnetic area are attracted through magnetism, and the second conductive area is in conductive contact with the first conductive area.

9. The practical training system of the embedded device according to claim 8, wherein a magnetic conductive metal is attached to the inner surface of the jack; the pins are made of magnetic conductive metal; the shapes of the jacks and the pins are hexagonal prisms.

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 for extracting a face image according to any one of claims 1 to 5.

Images