CN112200036A - Student behavior remote monitoring method and system - Google Patents

Abstract

The invention provides a remote monitoring method and a remote monitoring system for student behaviors, which can carry out video shooting on a student and extract a plurality of corresponding images, determine a corresponding target image pair according to pixel correlation information of two adjacent preprocessed images, finally judge whether the student behaviors are normal or not according to image similarity between the two images contained in the target image pair, and carry out corresponding warning operation.

Description

Method and system for remote monitoring of student behavior

technical field

The invention relates to the technical field of intelligent education, in particular to a method and system for remote monitoring of student behavior.

Background technique

At present, the behavior monitoring of students in the course of class is usually realized by tracking and photographing the students. Specifically, by shooting a class video about the students, and manually browsing and screening the class video frame by frame, it is judged that the students are in the class process. Whether there is abnormal behavior in . but. The above behavior monitoring methods not only require a lot of manpower and material resources to browse and screen massive video data, but also browse and screen by manual methods are prone to browsing omissions. In addition, the above behavior monitoring methods cannot be effectively and accurately screened. This will seriously affect the automaticity, reliability and accuracy of student behavior monitoring. It can be seen that the existing technology needs to be able to conduct comprehensive, automatic and accurate remote monitoring of the behavior status of students during class, so as to effectively identify various types of abnormal behaviors made by students.

SUMMARY OF THE INVENTION

In view of the defects in the prior art, the present invention provides a method and system for remote monitoring of student behavior, which obtains a video of the student within a preset time period by photographing the student, and performs image extraction processing on the video, thereby After obtaining several images and preprocessing several images, obtain the pixel correlation information between any two adjacent images, and according to the pixel correlation information, determine the two adjacent images as targets image pair, and then obtain the image similarity information between the two images included in the target image pair, and according to the image similarity information, determine whether the student's behavior is normal, and then carry out corresponding warning operations according to the judgment result. ; It can be seen that the method and system for remote monitoring of student behavior can shoot video of students and extract several corresponding images, and determine the corresponding target image pair according to the pixel correlation information of the two adjacent images after preprocessing, and finally according to The image similarity between the two images included in the target image pair judges whether the student's behavior is normal or not, and makes corresponding warning operations. Comprehensive, automatic and accurate remote monitoring of the behavior status in the course of class, so as to effectively identify various types of abnormal behaviors made by students.

The present invention provides a method for remote monitoring of student behavior, characterized in that it comprises the following steps:

Step S1, photographing a student to obtain a video of the student within a preset time period, and performing image extraction processing on the video to obtain several images;

Step S2, after preprocessing several of the images, obtain pixel correlation information between any two adjacent said images, and determine two adjacent said images according to the pixel correlation information as the target image pair;

Step S3, obtain the image similarity information between the two images included in the target image pair, and according to the image similarity information, judge whether the behavior of the student is normal, and then according to the result of the judgment, carry out corresponding steps. warning operation;

Further, in the step S1, the student is photographed to obtain a video of the student within a preset time period, and the video is subjected to image extraction processing, so as to obtain several images including:

Step S101, taking a panoramic shot of the student to obtain a video of the student within the preset time period;

Step S102, performing image extraction processing on the video according to a predetermined time interval and along the forward playback sequence of the video, thereby obtaining several images;

Further, in the step S2, after preprocessing several of the images, obtain pixel correlation information between any two adjacent images, and determine the adjacent pixels according to the pixel correlation information. The two described images as the target image pair specifically include:

Step S201, performing Kalman filtering noise reduction processing and image pixel smoothing processing on several of the images in sequence;

In step S202, each of the images is divided into N rectangular image sub-regions with the same area, and the pixel linear correlation coefficient between any two adjacent images is determined according to the following formula (1):

In the above formula (1), R(a, b) represents the pixel linear correlation coefficient between adjacent image a and image b, S _i represents the pixel chromaticity value of the ith rectangular image sub-region of image b, G _i represents the pixel chromaticity value of the ith rectangular image sub-region of image a, F _i represents the pixel texture value of the ith rectangular image sub-region of image b, and K _i represents the ith rectangular image sub-region of image a The pixel texture value of the The value is [0.7, 0.9];

Step S203, compare the pixel linear correlation coefficient R(a, b) between the adjacent image a and the image b with the preset pixel linear correlation threshold, if the pixel linear correlation coefficient R(a, b) is greater than or equal to the preset pixel linear correlation threshold, then the current adjacent image a and image b are used as the target image pair, otherwise, repeat the above step S202 to calculate the pixel linearity between the next group of adjacent two images correlation coefficient, until the calculated pixel linear correlation coefficient is greater than or equal to the preset pixel linear correlation threshold;

Further, in the step S3, the image similarity information between the two images included in the target image pair is obtained, and according to the image similarity information, it is judged whether the behavior of the student is normal, and then according to the image similarity information As a result of the judgment, corresponding warning operations are performed, including:

Step S301, according to the following formula (2), determine the image similarity value between the two images included in the target image pair:

In the above formula (2), sim represents the image similarity value between two images included in the target image pair, and M _d represents the N rectangular image sub-regions corresponding to one image in the target image pair. The number of rectangular image sub-regions with the student body-related pixels, Q _c represents the rectangular image sub-regions containing the student body-related pixels in the N rectangular image sub-regions corresponding to the other image in the target image pair , U _j represents the pixel texture value of the j-th rectangular image sub-region of the rectangular image sub-region where one image in the target image pair contains pixels related to the student body, and T _v represents the target image pair The other image contains the pixel texture value of the vth rectangular image sub-region of the rectangular image sub-region of the student body-related pixels, ε represents the first texture compensation coefficient, and its value is [0.1, 0.15], β represents the second texture compensation coefficient, and its value is [0.05, 0.15];

Step S302, compare the image similarity value sim with a preset image similarity threshold, and if the image similarity value sim is greater than or equal to the preset image similarity threshold, determine that the student's behavior is abnormal , otherwise, judge that the student's behavior is normal;

Step S303, when it is judged that the student's behavior is abnormal, a voice warning message is issued to the student.

The present invention also provides a remote monitoring system for student behavior, which is characterized in that it includes a video shooting module, an image extraction module, a target image pair determination module, a student behavior state judgment module, and a warning operation module; wherein,

The video shooting module is used for shooting students, so as to obtain a video of the students in a preset time period;

The image extraction module is configured to perform image extraction processing on the video to obtain several images;

The target image pair determination module is used to obtain pixel correlation information between any two adjacent said images after preprocessing a number of said images, and according to said pixel correlation information, determine the adjacent ones. The two described images are used as the target image pair;

The student behavior state judging module is used to obtain image similarity information between two images included in the target image pair, and determine whether the student's behavior is normal according to the image similarity information;

The warning operation module is configured to perform a corresponding warning operation according to the judgment result;

Further, the video shooting module shoots the student, so as to obtain the video of the student within the preset time period, specifically:

Taking a panoramic shot of the student to obtain a video of the student within the preset time period;

as well as,

The image extraction module performs image extraction processing on the video, so as to obtain several images including:

Perform image extraction processing on the video according to a predetermined time interval and along the forward playback sequence of the video, thereby obtaining several images;

Further, after the target image pair determination module preprocesses several of the images, obtains pixel correlation information between any two adjacent said images, and determines, according to the pixel correlation information, among the adjacent images. The two described images as the target image pair specifically include:

Performing Kalman filter noise reduction processing and image pixel smoothing processing on several of the images in sequence;

And each of the images is divided into N rectangular image sub-regions with the same area, and the pixel linear correlation coefficient between any two adjacent images is determined according to the following formula (1):

In the above formula (1), R(a, b) represents the pixel linear correlation coefficient between adjacent image a and image b, S _i represents the pixel chromaticity value of the ith rectangular image sub-region of image b, G _i represents the pixel chromaticity value of the ith rectangular image sub-region of image a, F _i represents the pixel texture value of the ith rectangular image sub-region of image b, and K _i represents the ith rectangular image sub-region of image a The pixel texture value of the The value is [0.7, 0.9];

Then compare the pixel linear correlation coefficient R(a, b) between the adjacent image a and the image b with the preset pixel linear correlation threshold, if the pixel linear correlation coefficient R(a, b) is greater than or equal to The preset pixel linear correlation threshold, then the current adjacent image a and image b are used as the target image pair, otherwise, repeat the above step S202 to calculate the pixel linear correlation coefficient between the next group of adjacent two images , until the calculated pixel linear correlation coefficient is greater than or equal to the preset pixel linear correlation threshold;

Further, the student behavior state judging module obtains image similarity information between two images included in the target image pair, and according to the image similarity information, judging whether the student's behavior is normal specifically includes:

According to the following formula (2), determine the image similarity value between the two images included in the target image pair:

In the above formula (2), sim represents the image similarity value between two images included in the target image pair, and M _d represents the N rectangular image sub-regions corresponding to one image in the target image pair. The number of rectangular image sub-regions with the student body-related pixels, Q _c represents the rectangular image sub-regions containing the student body-related pixels in the N rectangular image sub-regions corresponding to the other image in the target image pair , U _j represents the pixel texture value of the j-th rectangular image sub-region of the rectangular image sub-region where one image in the target image pair contains pixels related to the student body, and T _v represents the target image pair The other image contains the pixel texture value of the vth rectangular image sub-region of the rectangular image sub-region of the student body-related pixels, ε represents the first texture compensation coefficient, and its value is [0.1, 0.15], β represents the second texture compensation coefficient, and its value is [0.05, 0.15];

Comparing the image similarity value sim with a preset image similarity threshold, if the image similarity value sim is greater than or equal to the preset image similarity threshold, judge that the student's behavior is abnormal, otherwise , judging that the student's behavior is normal;

as well as,

According to the result of the judgment, the warning operation module performs corresponding warning operations specifically including:

When it is judged that the student's behavior is abnormal, a voice warning message is issued to the student.

Compared with the prior art, the method and system for remote monitoring of student behavior obtain a video of the student within a preset time period by photographing the student, and perform image extraction processing on the video to obtain several images, and After preprocessing several of the images, obtain the pixel correlation information between any two adjacent images, and according to the pixel correlation information, determine the two adjacent images as the target image pair, and then obtain The target image pair contains image similarity information between two images, and according to the image similarity information, it is judged whether the student's behavior is normal, and then according to the judgment result, a corresponding warning operation is performed; it can be seen that the student is The behavior remote monitoring method and system can shoot video of students and extract several corresponding images, and determine the corresponding target image pair according to the pixel correlation information of the preprocessed adjacent two images, and finally according to the target image pair contains the corresponding target image pair. The image similarity between the two images judges whether the student's behavior is normal or not, and makes corresponding warning operations, which can make detailed analysis of the captured images from the pixel level to analyze the students' behavior during class. Comprehensive, automatic and accurate remote monitoring of the status can effectively identify various types of abnormal behaviors made by students.

Other features and advantages of the present invention will be set forth in the description which follows, and in part will be apparent from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention may be realized and attained by the structure particularly pointed out in the written description, claims, and drawings.

The technical solutions of the present invention will be further described in detail below through the accompanying drawings and embodiments.

Description of drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained according to these drawings without creative efforts.

FIG. 1 is a schematic flowchart of a method for remote monitoring of student behavior provided by the present invention.

FIG. 2 is a schematic structural diagram of a remote monitoring system for student behavior provided by the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

Referring to FIG. 1 , it is a schematic flowchart of a method for remotely monitoring student behavior according to an embodiment of the present invention. The method for remote monitoring of student behavior includes the following steps:

Step S1, photographing the student to obtain a video of the student within a preset time period, and performing image extraction processing on the video to obtain several images;

Step S2, after preprocessing several of the images, obtain the pixel correlation information between any two adjacent images, and according to the pixel correlation information, determine the two adjacent images as the target image pair. ;

Step S3: Obtain image similarity information between two images included in the target image pair, and determine whether the student's behavior is normal according to the image similarity information, and then perform corresponding warning operations according to the judgment result.

The beneficial effects of the above-mentioned technical solutions are as follows: the method for remote monitoring of student behavior can shoot video of students and extract several corresponding images, and determine the corresponding target image pair according to the pixel correlation information of the preprocessed adjacent two images. Finally, according to the image similarity between the two images included in the target image pair, determine whether the student's behavior is normal or not, and make corresponding warning operations. Comprehensive, automatic and accurate remote monitoring of students' behavior during class, so as to effectively identify various types of abnormal behaviors made by students.

Preferably, in this step S1, the student is photographed to obtain a video of the student within a preset time period, and the video is subjected to image extraction processing, thereby obtaining several images including:

Step S101, taking a panoramic shot of the student to obtain a video of the student within the preset time period;

Step S102, performing image extraction processing on the video according to a predetermined time interval and along the forward playing sequence of the video, thereby obtaining several images.

The beneficial effects of the above technical solutions are: by sampling and extracting several images from the video obtained by shooting, the calculation amount of subsequent image processing can be greatly reduced, and it is also convenient to select appropriate images for processing according to actual needs, thereby improving student behavior. Flexibility and controllability of monitoring.

Preferably, in this step S2, after preprocessing several of the images, obtain pixel correlation information between any two adjacent images, and determine the two adjacent ones according to the pixel correlation information. The image as the target image pair specifically includes:

Step S201, performing Kalman filtering noise reduction processing and image pixel smoothing processing on several of the images in sequence;

In step S202, each of the images is divided into N rectangular image sub-regions with the same area, and the pixel linear correlation coefficient between any two adjacent images is determined according to the following formula (1):

In the above formula (1), R(a, b) represents the pixel linear correlation coefficient between adjacent image a and image b, S _i represents the pixel chromaticity value of the ith rectangular image sub-region of image b, G _i represents the pixel chromaticity value of the ith rectangular image sub-region of image a, F _i represents the pixel texture value of the ith rectangular image sub-region of image b, and K _i represents the ith rectangular image sub-region of image a The pixel texture value of the The value is [0.7, 0.9];

Step S203, compare the pixel linear correlation coefficient R(a, b) between the adjacent image a and the image b with the preset pixel linear correlation threshold, if the pixel linear correlation coefficient R(a, b) is greater than or is equal to the preset pixel linear correlation threshold, then the current adjacent image a and image b are used as the target image pair, otherwise, the above step S202 is repeated to calculate the pixel linear correlation coefficient between the next group of adjacent two images, Until the calculated pixel linear correlation coefficient is greater than or equal to the preset pixel linear correlation threshold.

The beneficial effects of the above technical solutions are: by performing Kalman filter noise reduction processing and image pixel smoothing processing on the image, the interference information in the image can be effectively removed, thereby improving the accuracy of subsequent image processing; In addition, by the above formula ( 1) Calculate the pixel linear correlation coefficient between any two adjacent images, which can determine the correlation of image pixel features for two adjacent images from the level of image pixel chromaticity and image pixel texture, thereby improving Determine the reliability and objectivity of the target image pair.

Preferably, in this step S3, the image similarity information between the two images included in the target image pair is obtained, and according to the image similarity information, it is judged whether the student's behavior is normal, and then according to the result of the judgment, The corresponding warning actions include:

Step S301, according to the following formula (2), determine the image similarity value between the two images included in the target image pair:

In the above formula (2), sim represents the image similarity value between the two images included in the target image pair, and M _d represents that the N rectangular image sub-regions corresponding to one image in the target image pair contain the The number of rectangular image sub-regions of the student's body-related pixels, Q _c represents the number of rectangular image sub-regions containing the student's body-related pixels in the N rectangular image sub-regions corresponding to the other image in the target image pair, U _j Indicates that one image in the target image pair contains the pixel texture value of the jth rectangular image sub-region of the rectangular image sub-region of the student's body-related pixels, and T _v indicates that the other image in the target image pair contains the student's body. The pixel texture value of the vth rectangular image sub-region of the rectangular image sub-region of the body-related pixels, ε represents the first texture compensation coefficient, and its value is [0.1, 0.15], β represents the second texture compensation coefficient, and its The value is [0.05, 0.15];

Step S302, compare the image similarity value sim with a preset image similarity threshold, if the image similarity value sim is greater than or equal to the preset image similarity threshold, then judge that the student's behavior is abnormal, otherwise, judge the student is behaving normally;

Step S303, when it is judged that the student's behavior is abnormal, a voice warning message is issued to the student.

The beneficial effect of the above technical solution is: because when the student makes an abnormal behavior, a corresponding difference will occur at the pixel level between two adjacent images, and the image between the two images is obtained by calculating the above formula (2). The similarity value can quantitatively evaluate the pixel difference between two adjacent images, so as to facilitate the subsequent quick and accurate identification of various types of abnormal behaviors made by students.

Referring to FIG. 2 , it is a schematic structural diagram of a remote monitoring system for student behavior provided by an embodiment of the present invention. The student behavior remote monitoring system includes a video shooting module, an image extraction module, a target image pair determination module, a student behavior state judgment module, and a warning operation module; wherein,

The video shooting module is used for shooting a student, so as to obtain a video of the student within a preset time period;

The image extraction module is used to perform image extraction processing on the video to obtain several images;

The target image pair determination module is used to obtain pixel correlation information between any two adjacent images after preprocessing a number of the images, and determine the two adjacent ones according to the pixel correlation information. image as the target image pair;

The student behavior state judging module is used to obtain image similarity information between two images included in the target image pair, and determine whether the student's behavior is normal according to the image similarity information;

The warning operation module is used for performing corresponding warning operations according to the judgment result.

The beneficial effects of the above technical solutions are as follows: the student behavior remote monitoring system can shoot video of students and extract a number of corresponding images, and determine the corresponding target image pair according to the pixel correlation information of the preprocessed adjacent two images. Finally, according to the image similarity between the two images included in the target image pair, determine whether the student's behavior is normal or not, and make corresponding warning operations. Comprehensive, automatic and accurate remote monitoring of students' behavior during class, so as to effectively identify various types of abnormal behaviors made by students.

Preferably, the video shooting module shoots the student, so as to obtain the video of the student within the preset time period, which specifically includes:

Take a panoramic shot of the student to obtain a video of the student within the preset time period;

as well as,

The image extraction module performs image extraction processing on the video, thereby obtaining several images including:

Image extraction processing is performed on the video according to predetermined time intervals and along the forward playback sequence of the video, thereby obtaining several images.

The beneficial effects of the above technical solutions are: by sampling and extracting several images from the video obtained by shooting, the calculation amount of subsequent image processing can be greatly reduced, and it is also convenient to select appropriate images for processing according to actual needs, thereby improving student behavior. Flexibility and controllability of monitoring.

Preferably, after the target image pair determination module preprocesses several of the images, obtains pixel correlation information between any two adjacent images, and determines two adjacent ones according to the pixel correlation information. The image as the target image pair specifically includes:

Perform Kalman filter noise reduction processing and image pixel smoothing processing on several of the images in sequence;

Divide each image into N rectangular image sub-regions with the same area, and determine the pixel linear correlation coefficient between any two adjacent images according to the following formula (1):

In the above formula (1), R(a, b) represents the pixel linear correlation coefficient between adjacent image a and image b, S _i represents the pixel chromaticity value of the ith rectangular image sub-region of image b, G _i represents the pixel chromaticity value of the ith rectangular image sub-region of image a, F _i represents the pixel texture value of the ith rectangular image sub-region of image b, and K _i represents the ith rectangular image sub-region of image a The pixel texture value of the The value is [0.7, 0.9];

Then compare the pixel linear correlation coefficient R(a, b) between the adjacent image a and the image b with the preset pixel linear correlation threshold, if the pixel linear correlation coefficient R(a, b) is greater than or equal to the If the pixel linear correlation threshold is preset, the current adjacent image a and image b are used as the target image pair, otherwise, the above step S202 is repeated to calculate the pixel linear correlation coefficient between the next group of adjacent two images, until the calculation The obtained pixel linear correlation coefficient is greater than or equal to the preset pixel linear correlation threshold.

The beneficial effects of the above technical solutions are: by performing Kalman filter noise reduction processing and image pixel smoothing processing on the image, the interference information in the image can be effectively removed, thereby improving the accuracy of subsequent image processing; In addition, by the above formula ( 1) Calculate the pixel linear correlation coefficient between any two adjacent images, which can determine the correlation of image pixel features for two adjacent images from the level of image pixel chromaticity and image pixel texture, thereby improving Determine the reliability and objectivity of the target image pair.

Preferably, the student behavior state judging module obtains image similarity information between two images included in the target image pair, and determines whether the student's behavior is normal according to the image similarity information, specifically including:

According to the following formula (2), determine the image similarity value between the two images included in the target image pair:

In the above formula (2), sim represents the image similarity value between the two images included in the target image pair, and M _d represents that the N rectangular image sub-regions corresponding to one image in the target image pair contain the The number of rectangular image sub-regions of the student's body-related pixels, Q _c represents the number of rectangular image sub-regions containing the student's body-related pixels in the N rectangular image sub-regions corresponding to the other image in the target image pair, U _j Indicates that one image in the target image pair contains the pixel texture value of the jth rectangular image sub-region of the rectangular image sub-region of the student's body-related pixels, and T _v indicates that the other image in the target image pair contains the student's body. The pixel texture value of the vth rectangular image sub-region of the rectangular image sub-region of the body-related pixels, ε represents the first texture compensation coefficient, and its value is [0.1, 0.15], β represents the second texture compensation coefficient, and its The value is [0.05, 0.15];

Compare the image similarity value sim with the preset image similarity threshold, if the image similarity value sim is greater than or equal to the preset image similarity threshold, judge that the student's behavior is abnormal, otherwise, judge the student behave normally;

as well as,

According to the result of the judgment, the warning operation module performs corresponding warning operations specifically including:

When it is judged that the student's behavior is abnormal, a voice warning message is issued to the student.

The beneficial effect of the above technical solution is: because when the student makes an abnormal behavior, a corresponding difference will occur at the pixel level between two adjacent images, and the image between the two images is obtained by calculating the above formula (2). The similarity value can quantitatively evaluate the pixel difference between two adjacent images, so as to facilitate the subsequent quick and accurate identification of various types of abnormal behaviors made by students.

It can be seen from the content of the above embodiment that the method and system for remote monitoring of student behavior obtains a video of the student within a preset time period by photographing the student, and performs image extraction processing on the video to obtain several images, After preprocessing several of the images, obtain the pixel correlation information between any two adjacent images, and according to the pixel correlation information, determine the two adjacent images as the target image pair, and then Obtain the image similarity information between the two images included in the target image pair, and according to the image similarity information, determine whether the student's behavior is normal, and then perform a corresponding warning operation according to the judgment result; The method and system for remote monitoring of student behavior can shoot video of students and extract several corresponding images, and determine corresponding target image pairs according to the pixel correlation information of the two adjacent images after preprocessing, and finally according to the target image pair contains The image similarity between the two images can judge whether the student's behavior is normal or not, and make corresponding warning operations. Conduct comprehensive, automatic and accurate remote monitoring of behavior status, so as to effectively identify various types of abnormal behaviors made by students.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit and scope of the invention. Thus, provided that these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims (8)

1. The student behavior remote monitoring method is characterized by comprising the following steps:

step S1, shooting a student to obtain a video of the student in a preset time period, and performing image extraction processing on the video to obtain a plurality of images;

step S2, after preprocessing a plurality of images, acquiring pixel correlation information between any two adjacent images, and determining two adjacent images as a target image pair according to the pixel correlation information;

and step S3, acquiring image similarity information between two images included in the target image pair, judging whether the behavior of the student is normal according to the image similarity information, and performing corresponding warning operation according to the judgment result.

2. The student behavior remote monitoring method according to claim 1, wherein:

in step S1, capturing a student to obtain a video of the student within a preset time period, and performing image extraction processing on the video to obtain a plurality of images specifically includes:

step S101, carrying out panoramic shooting on the students so as to obtain videos of the students in the preset time period;

step S102, according to a preset time interval and a forward playing time sequence along the video, carrying out image extraction processing on the video, and thus obtaining a plurality of images.

3. The student behavior remote monitoring method according to claim 2, wherein:

in step S2, after preprocessing the images, obtaining pixel correlation information between any two adjacent images, and determining two adjacent images as a target image pair according to the pixel correlation information specifically includes:

step S201, performing Kalman filtering noise reduction processing and image pixel smoothing processing on a plurality of images in sequence;

step S202, dividing each image into N rectangular image sub-regions with the same area, and determining a pixel linear correlation coefficient between any two adjacent images according to the following formula (1):

in the above formula (1), R (a, b) represents a pixel linear correlation coefficient between the adjacent images a and b, S_iRepresenting pixel chrominance values, G, of the ith rectangular image sub-area of image b_iPixel chrominance values, F, of the ith rectangular image sub-region representing image a_iThe texture value of the pixel, K, representing the ith rectangular image sub-region of image b_iThe texture value of the pixel of the ith rectangular image sub-region of the image a is represented, theta represents a preset chroma weight value and takes a value of 0.4, delta represents a preset texture weight value and takes a value of 0.6, and X represents a common correction coefficient of the image a and the image b and takes a value of [0.7, 0.9]；

Step S203, comparing the pixel linear correlation coefficient R (a, b) between the adjacent images a and b with a preset pixel linear correlation threshold, if the pixel linear correlation coefficient R (a, b) is greater than or equal to the preset pixel linear correlation threshold, using the current adjacent image a and image b as the target image pair, otherwise, repeating the step S202 to calculate the pixel linear correlation coefficient between the next group of adjacent two images until the calculated pixel linear correlation coefficient is greater than or equal to the preset pixel linear correlation threshold.

4. A student behaviour remote monitoring method as claimed in claim 3, characterised in that:

in step S3, acquiring image similarity information between two images included in the target image pair, determining whether the behavior of the student is normal according to the image similarity information, and performing corresponding warning operation according to the determination result specifically includes:

step S301, determining an image similarity value between two images included in the target image pair according to the following formula (2):

in the above formula (2), sim represents an image similarity value between two images included in the target image pair, M_dRepresenting the number Q of rectangular image sub-regions containing the student body related pixels in the N rectangular image sub-regions corresponding to one image in the target image pair_cRepresenting the number, U, of the rectangular image sub-regions containing the student body related pixels in the N rectangular image sub-regions corresponding to the other image in the target image pair_jA pixel texture value, T, of a jth rectangular image sub-region of the rectangular image sub-region representing that one of the target image pair contains the student body related pixel_vThe texture value of the pixel of the v-th rectangular image sub-area of the rectangular image sub-area which represents that the other image in the target image pair contains the pixel related to the student body is epsilon, the epsilon represents a first texture compensation coefficient, and the value of epsilon is [0.1, 0.15 ]]And beta represents a second texture compensation coefficient and takes a value of [0.05, 0.15 ]]；

Step S302, comparing the image similarity value sim with a preset image similarity threshold, if the image similarity value sim is greater than or equal to the preset image similarity threshold, judging that the behavior of the student is abnormal, otherwise, judging that the behavior of the student is normal;

step S303, when the behavior of the student is judged to be abnormal, voice warning information is sent to the student.

5. The remote monitoring system for the student behaviors is characterized by comprising a video shooting module, an image extraction module, a target image pair determination module, a student behavior state judgment module and a warning operation module; wherein,

the video shooting module is used for shooting a student so as to obtain a video of the student in a preset time period;

the image extraction module is used for carrying out image extraction processing on the video so as to obtain a plurality of images;

the target image pair determining module is used for preprocessing a plurality of images, acquiring pixel correlation information between any two adjacent images, and determining two adjacent images as a target image pair according to the pixel correlation information;

the student behavior state judging module is used for acquiring image similarity information between two images contained in the target image pair and judging whether the behavior of the student is normal or not according to the image similarity information;

and the warning operation module is used for carrying out corresponding warning operation according to the judgment result.

6. The student behavior remote monitoring system of claim 5, wherein:

the video shooting module shoots students, so that the video of the students in the preset time period is obtained, and the video shooting module specifically comprises the following steps:

carrying out panoramic shooting on the students so as to obtain videos of the students in the preset time period;

and the number of the first and second groups,

the image extraction module performs image extraction processing on the video, so as to obtain a plurality of images specifically comprises:

and carrying out image extraction processing on the video according to a preset time interval and a forward playing time sequence of the video, thereby obtaining a plurality of images.

7. The student behavior remote monitoring system of claim 6, wherein:

the target image pair determining module obtains pixel correlation information between any two adjacent images after preprocessing the plurality of images, and determining two adjacent images as a target image pair according to the pixel correlation information specifically includes:

sequentially carrying out Kalman filtering noise reduction processing and image pixel smoothing processing on a plurality of images; dividing each image into N rectangular image sub-areas with the same area, and determining a pixel linear correlation coefficient between any two adjacent images according to the following formula (1):

in the above formula (1), R (a, b) represents a pixel linear correlation coefficient between the adjacent images a and b, S_iRepresenting pixel chrominance values, G, of the ith rectangular image sub-area of image b_iPixel chrominance values, F, of the ith rectangular image sub-region representing image a_iThe texture value of the pixel, K, representing the ith rectangular image sub-region of image b_iThe texture value of the pixel of the ith rectangular image sub-region of the image a is represented, theta represents a preset chroma weight value and takes a value of 0.4, delta represents a preset texture weight value and takes a value of 0.6, and X represents a common correction coefficient of the image a and the image b and takes a value of [0.7, 0.9]；

And comparing the pixel linear correlation coefficient R (a, b) between the adjacent images a and b with a preset pixel linear correlation threshold, if the pixel linear correlation coefficient R (a, b) is greater than or equal to the preset pixel linear correlation threshold, taking the current adjacent image a and image b as the target image pair, otherwise, repeating the step S202 to calculate the pixel linear correlation coefficient between the next group of adjacent two images until the calculated pixel linear correlation coefficient is greater than or equal to the preset pixel linear correlation threshold.

8. The student behavior remote monitoring system of claim 7, wherein:

the student behavior state judgment module acquires image similarity information between two images included in the target image pair, and judges whether the behavior of the student is normal specifically according to the image similarity information:

determining an image similarity value between two images comprised by the target image pair according to the following formula (2):

in the above formula (2), sim represents an image similarity value between two images included in the target image pair, M_dRepresenting the number Q of rectangular image sub-regions containing the student body related pixels in the N rectangular image sub-regions corresponding to one image in the target image pair_cRepresenting the number, U, of the rectangular image sub-regions containing the student body related pixels in the N rectangular image sub-regions corresponding to the other image in the target image pair_jA pixel texture value, T, of a jth rectangular image sub-region of the rectangular image sub-region representing that one of the target image pair contains the student body related pixel_vThe texture value of the pixel of the v-th rectangular image sub-area of the rectangular image sub-area which represents that the other image in the target image pair contains the pixel related to the student body is epsilon, the epsilon represents a first texture compensation coefficient, and the value of epsilon is [0.1, 0.15 ]]And beta represents a second texture compensation coefficient and takes a value of [0.05, 0.15 ]](ii) a Comparing the image similarity value sim with a preset image similarity threshold, if the image similarity value sim is greater than or equal to the preset image similarity threshold, judging that the behavior of the student is abnormal, otherwise, judging that the behavior of the student is normal;

and the number of the first and second groups,

the warning operation module carries out corresponding warning operation according to the judgment result, and specifically comprises the following steps: and when the behavior of the student is judged to be abnormal, voice warning information is sent to the student.

Images