Personnel access multi-dimensional information statistical method applied to checkpoint scene

Abstract

The invention discloses a personnel access multi-dimensional information statistical method applied to a checkpoint scene, relates to the technical field of information statistics, solves the problems of low accuracy of a personnel statistical result and high detection technology cost in the checkpoint scene, and comprises the following steps: issuing a personnel access statistical task of a target personnel corresponding to the checkpoint scene; setting a monitoring area and a counting line position and a counting direction in the monitoring area; collecting video data in the monitoring area, and analyzing the video data to obtain personnel frame information and multi-dimensional personnel data of target personnel in the monitoring area; tracking a target person in the video frames, carrying out person association on the same target person in the plurality of video frames, and identifying a motion track of the target person; and judging the line collision condition of the motion fitting function of the motion track corresponding to the target person and the counting line function, and obtaining the behavior state corresponding to the target person, so that the accuracy of a person detection result is improved, and meanwhile, the cost of a person detection technology is also reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of information statistics, and in particular to a method for counting multi-dimensional information of people entering and leaving a checkpoint. Background Art

[0002] In the era of rapid technological development, artificial intelligence has been applied to all levels and fields of society, such as security, medical care, government affairs, transportation, etc. The extensive application of personnel entry and exit statistics algorithms in social life scenarios has made significant contributions to achieving better management and people's livelihood convenience. For example, in important urban transportation hubs such as stations and subway stations, the application of personnel entry and exit statistics algorithms can more accurately control the flow of people in each time period in real time, so as to adjust the corresponding strategies and avoid more serious consequences caused by passenger flow congestion; At present, when counting the information of the checkpoints, the method of combining artificial intelligence with embedded devices has high consumption and maintenance costs, and false detection and missed detection may occur during the process of personnel detection, which affects the statistical results; To this end, the present invention proposes a method for counting multi-dimensional information of people entering and leaving the checkpoint scene. Summary of the invention

[0003] The purpose of the present invention is to propose a method for counting multi-dimensional information of people entering and leaving the checkpoint scene, so as to solve the problem of inaccurate personnel counting results in the checkpoint scene proposed in the above background technology.

[0004] In order to achieve the above object, the present invention adopts the following technical solutions: A method for counting multi-dimensional information of people entering and leaving a checkpoint, the method comprising: Step S1, issuing a personnel entry and exit counting task corresponding to the target personnel in the checkpoint scene, setting the monitoring area and the counting line position and counting direction in the monitoring area; Step S2, collecting video data in the monitoring area, and analyzing the video data to obtain the personnel frame information and multi-dimensional personnel data of the target personnel in the monitoring area; Step S3, tracking the target person in the video frame, associating the same target person in multiple video frames, and identifying the movement trajectory of the target person; Step S4, determining the collision between the motion fitting function of the target person's corresponding motion trajectory and the counting line function, and obtaining the corresponding behavior state of the target person; Step S5, obtaining the behavior status and multi-dimensional personnel data corresponding to the target personnel whose behavior status has changed and reporting them to the server and the display terminal.

[0005] Furthermore, the monitoring area refers to the area where the entry and exit statistics of personnel need to be executed. When the entry and exit results are counted, only the objects in the area are counted, and the objects outside the area are not included in the statistics; The counting line is set in the detection area. When the target person crosses the counting line, it means that the behavior state of the target person has changed; The counting direction is used to confirm the behavior status of the target person. If the target person's motion trajectory is the same as the set counting direction, it means that the target person is in the entry state. If the target person's corresponding motion trajectory is opposite to the set counting direction, it means that the target person is in the exit state. The person border information includes the border height, border width and border center position corresponding to the person border; The multi-dimensional personnel data includes the age, gender and corresponding confidence level of the target personnel.

[0006] Furthermore, the step S2 specifically includes the following sub-steps: Step S21, obtaining video data in the detection area, and dividing the video data into multiple video frames according to time; Step S22, importing the multiple video frames into the trained target detection model, and the target detection model identifies the multi-dimensional personnel data of the target person in the video frame; Step S23, performing frame selection processing on the target person, obtaining a person frame corresponding to the target person, and then recording the person frame information; Step S24, establishing a regional coordinate system according to the monitoring area, and obtaining a function expression corresponding to the counting line function according to the position of the counting line in the regional coordinate system; Step S25, draw the center position of the person's frame corresponding to the frame in the regional coordinate system to obtain the pedestrian center coordinates.

[0007] Furthermore, the step S3 specifically includes the following sub-steps: Step S301, obtaining the video frame number j corresponding to the current moment, and obtaining the video frame number jc corresponding to the previous c moments of the current moment; wherein j>c, j=1, 2, ..., z, z is a positive integer; Step S302, obtaining the target person's border and pedestrian center coordinates in multiple video frames between time jc and time j; Step S303, creating an association matrix corresponding to adjacent video frames. If the previous video frame contains m person borders and the next video frame contains n person borders, the created association matrix is: ; Among them JL m,n represents the Euclidean distance between the mth person border in the previous video frame and the nth person border in the next video frame, JL m,n The calculation formula is as follows: .

[0008] Furthermore, the step S3 also includes the following sub-steps: Step S304, performing association for row 1, and selecting the largest value in row 1 as the association result for row 1; Step S305, performing association for row 2, selecting the row with the largest value as the association result for row 2, and so on, until row m is associated; Step S306, repeating steps S303 to S305 until all video frames between time jn and time j are associated with people.

[0009] Furthermore, the step S3 further includes the following sub-steps: Step S307, obtaining the pedestrian center coordinates (Xi, Yi) of the target person in multiple video frames; wherein i represents the number of the video frame corresponding to the target person at any time, i=jc,j-c+1, ..., j; Step S308, connecting the pedestrian center coordinates in the order of the video frames to obtain the motion trajectory of the target person; Step S309, calculating the similarity between the target person's motion trajectory and the common function graph; Step S310 , selecting the common function with the highest similarity as the motion fitting function corresponding to the motion trajectory of the target person, and recording the function expression of the motion fitting function as Y=F(X).

[0010] Furthermore, if the function expression of the motion fitting function is Y=F(X)=a1*X, where a1 is the coefficient of the first-order term, then the calculation The value of , and taking the partial derivative with respect to a1, we get: ; Simplify to get , where EX represents the mean of Xi and EY represents the mean of Yi.

[0011] Furthermore, if the function expression of the motion fitting function is Y=F(X)=a0+a1*X, where a0 is a constant term and a1 is a one-time coefficient, then the calculation The value of , and taking partial derivatives with respect to a0 and a1 respectively, we get: Taking partial derivatives with respect to a0 we get: Taking partial derivatives of a1 we get: The combined result is: a0=EY-a1×EX, .

[0012] Furthermore, if the function expression of the motion fitting function is Y=F(X)=a0+a1*X+a3×X 3 ; Among them, a0 is the constant term, a1 is the one-time coefficient, and a3 is the cubic coefficient; Then calculate The value of , and taking partial derivatives of a0, a1 and a3 respectively, we get: Taking partial derivatives with respect to a0 we get: ; Taking partial derivatives of a1 we get: ; Taking partial derivatives of a3 we get: ; Combined, we get: a0=EY-a1*EX-a3*EX 3 ; Among them EX 3 For Xi 3 The mean of ; .

[0013] Furthermore, the step S4 specifically includes the following sub-steps: Step S41, obtaining a function expression of a motion fitting function and a function expression of a counting line function corresponding to a motion trajectory of a target person; Among them, the counting line function is defined in the form of y=ax+b, where a and b are both constants; Step S42, combining the motion fitting function and the counting line function, if there is no real number solution, no operation is performed; Step S43, if there is a real number solution, it is determined that the behavior state of the target person has changed; Step S44, obtaining the center coordinates of the pedestrian at the moment before and the moment after the intersection. If the center coordinates of the pedestrian at the moment before the intersection are below the counting line, and the center coordinates of the pedestrian at the moment after the intersection are above the counting line, it is determined that the behavior state of the target person has changed, and the behavior state of the target person is recorded as leaving state; Step S45, if the pedestrian center coordinates at a moment before the intersection are above the counting line, and the pedestrian center coordinates at a moment after the intersection are below the counting line, it is determined that the behavior state of the target person has changed, and the behavior state of the target person is recorded as the entry state.

[0014] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: The present invention first issues a personnel entry and exit statistics task corresponding to the target personnel in the checkpoint scene, sets a monitoring area and a counting line position and a counting direction in the monitoring area, then collects video data in the monitoring area, and analyzes the video data to obtain the personnel frame information and multi-dimensional personnel data of the target personnel in the monitoring area, then tracks the target personnel in the video frame, associates the same target personnel in multiple video frames, and identifies the motion trajectory of the target personnel, and then determines the collision between the motion fitting function of the motion trajectory corresponding to the target personnel and the counting line function, and obtains the corresponding behavior state of the target personnel, and finally obtains the behavior state and multi-dimensional personnel data corresponding to the target personnel whose behavior state has changed and reports them to a server and a display terminal. The present invention reduces the cost of personnel detection and improves the accuracy of personnel detection results. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0016] Figure 1 is a flow chart of the method of the present invention; Figure 2 It is a flow chart of the analysis of video data in the present invention; Figure 3 It is a flow chart for analyzing the collision situation in the present invention; Figure 4 It is a schematic diagram of the counting line function and the motion fitting function in the present invention; Figure 5 It is a schematic diagram of the structure of the electronic device in the present invention. DETAILED DESCRIPTION

[0017] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Example 1: Please refer to Figure 1-Figure 4 As shown, the technical solution provided by the present invention is: a method for counting multi-dimensional information of people entering and leaving the checkpoint scene, and the method is specifically as follows: Step S1, issuing a personnel entry and exit counting task corresponding to the target personnel in the checkpoint scene, setting the monitoring area and the counting line position and counting direction in the monitoring area; Monitoring area: The area where the statistics of people entering and leaving the area need to be executed. The area is fully covered in the screen of the image acquisition device. When the statistics of entry and exit are performed, only the objects in the area are counted, and the objects outside the area are not included in the statistics; Counting line: It is set in the detection area. When the target person crosses the counting line, it means that the behavior state of the target person has changed; Counting direction: used to confirm the behavior status of the target person. If the target person's motion trajectory is the same as the set counting direction, it means that the target person is in the entering state. If the target person's corresponding motion trajectory is opposite to the set counting direction, it means that the target person is in the leaving state.

[0019] Step S2, collecting video data in the monitoring area, and analyzing the video data to obtain the personnel frame information and multi-dimensional personnel data of the target personnel in the monitoring area; In this embodiment, step S2 specifically includes the following sub-steps: Step S21, obtaining video data in the detection area, and dividing the video data into multiple video frames according to time; Step S22, importing the multiple video frames into the trained target detection model, and the target detection model identifies the multi-dimensional personnel data of the target person in the video frame; Step S23, performing frame selection processing on the target person, obtaining a person frame corresponding to the target person, and then recording the person frame information; The person border information includes the border height, border width and border center position corresponding to the person border; the multi-dimensional person data includes the age, gender and corresponding confidence level of the target person; Step S24, establishing a regional coordinate system according to the monitoring area, and obtaining a function expression corresponding to the counting line function according to the position of the counting line in the regional coordinate system; It should be noted that when establishing the regional coordinate system, the monitoring area below the counting line is recorded as the area where personnel entry statistics are required by default; Step S25, drawing the center position of the border corresponding to the person's border in the regional coordinate system to obtain the pedestrian center coordinates; It should be noted that the training of the target detection model includes preparatory work such as data collection and data labeling. Data collection is mainly carried out through scenes with large flow of people, such as subway stations, large shopping malls, etc. Scenes with large flow of people can increase the robustness and stability of the training model. When labeling the collected data, the target persons are manually selected and their corresponding gender and age information is labeled. The more accurate and detailed the labeled information is, the further the accuracy of the target detection model will be increased.

[0020] Step S3, tracking the target person in the video frame, associating the same target person in multiple video frames, and identifying the movement trajectory of the target person; Specifically, the step S3 includes the following sub-steps: Step S301, obtaining the video frame number j corresponding to the current moment, and obtaining the video frame number jc corresponding to the previous c moments of the current moment; wherein j>c, j=1, 2, ..., z, z is a positive integer; Step S302, obtaining the target person's border and pedestrian center coordinates in multiple video frames between time jc and time j; Step S303, creating an association matrix corresponding to adjacent video frames. If the previous video frame contains m person borders and the next video frame contains n person borders, the created association matrix is: ; Among them JL m,n represents the Euclidean distance between the mth person border in the previous video frame and the nth person border in the next video frame, JL m,n The calculation formula is as follows: ; Step S304, performing association for row 1, and selecting the largest value in row 1 as the association result for row 1; For example: If JL 1，2 The value of is the largest, then associate the border of the person numbered 1 in the previous video frame with the border of the person numbered 2 in the next video frame, and remove all the matrix elements in the second column of the association matrix; Step S305, performing association for row 2, selecting the row with the largest value as the association result for row 2, and so on, until row m is associated; Step S306, repeating steps S303 to S305 until all video frames between time jn and time j are associated with people; In order to simplify the analysis process, the subsequent sub-steps in step S3 only analyze a single target person, so the target person is not numbered; Step S307, obtaining the pedestrian center coordinates (Xi, Yi) of the target person in multiple video frames; wherein i represents the number of the video frame corresponding to the target person at any time, i=jc,j-c+1, ..., j; Step S308, connecting the pedestrian center coordinates in the order of the video frames to obtain the motion trajectory of the target person; Step S309, calculating the similarity between the target person's motion trajectory and the common function graph; Step S310, selecting the common function with the highest similarity as the motion fitting function corresponding to the motion trajectory of the target person, and recording the function expression of the motion fitting function as Y=F(X); If Y=F(X)=a1*X, where a1 is the coefficient of the first-order term; calculate The value of , and taking the partial derivative with respect to a1, we get: ; Simplify to get , where EX represents the mean of Xi, and EY represents the mean of Yi; If Y = F (X) = a0 + a1 * X, where a0 is a constant term and a1 is a one-time coefficient; calculate The value of , and taking partial derivatives with respect to a0 and a1 respectively, we get: Taking partial derivatives with respect to a0 we get: Taking partial derivatives of a1 we get: The combined result is: a0=EY-a1×EX, ; If Y = F (X) = a0 + a1 * X + a3 × X 3 ; Among them, a0 is the constant term, a1 is the one-time coefficient, and a3 is the cubic coefficient; calculate The value of , and taking partial derivatives of a0, a1 and a3 respectively, we get: Taking partial derivatives with respect to a0 we get: ; Taking partial derivatives of a1 we get: ; Taking partial derivatives of a3 we get: ; Combined, we get: a0=EY-a1*EX-a3*EX 3 ; Among them EX 3 For Xi 3 The mean of ; ; The same goes for other types of functions.

[0021] Step S4, determining the collision between the motion fitting function of the target person's corresponding motion trajectory and the counting line function, and obtaining the corresponding behavior state of the target person; In this embodiment, step S4 specifically includes the following sub-steps: Step S41, obtaining a function expression of a motion fitting function and a function expression of a counting line function corresponding to a motion trajectory of a target person; The counting line function is defined as y=ax+b, where a and b are constants, and their specific values ​​are determined by the position of the counting line in the monitoring area coordinates; Step S42, combining the motion fitting function and the counting line function, if there is no real number solution (that is, the function graphs corresponding to the motion fitting function and the counting line function do not have an intersection), no operation is performed; Step S43: if there is a real number solution (i.e., the function graphs corresponding to the motion fitting function and the counting line function have an intersection), it is determined that the behavior state of the target person has changed; Step S44, obtaining the center coordinates of the pedestrian at the moment before and the moment after the intersection. If the center coordinates of the pedestrian at the moment before the intersection are below the counting line, and the center coordinates of the pedestrian at the moment after the intersection are above the counting line, it is determined that the behavior state of the target person has changed, and the behavior state of the target person is recorded as leaving state; Step S45, if the pedestrian center coordinates at a moment before the intersection are above the counting line, and the pedestrian center coordinates at a moment after the intersection are below the counting line, it is determined that the behavior state of the target person has changed, and the behavior state of the target person is recorded as the entry state.

[0022] Step S5, obtaining the behavior status and multi-dimensional personnel data corresponding to the target personnel whose behavior status has changed and reporting them to the server and the display terminal; The server is used to receive and store the behavior status and multi-dimensional personnel data corresponding to the target person, and the display terminal is used to receive and display the behavior status and multi-dimensional personnel data corresponding to the target person; This embodiment is applied to the statistics of people entering and leaving the checkpoint scene. After calculation, the accuracy of this solution reaches 96%, the recall rate of target personnel reaches 99.0%, and the comprehensive evaluation index can reach 99.5%.

[0023] Exemplarily, the present invention first issues a personnel entry and exit statistics task corresponding to the target personnel in the checkpoint scene, sets the monitoring area and the counting line position and counting direction in the monitoring area, and then collects video data in the monitoring area, and analyzes the video data to obtain the personnel border information and multi-dimensional personnel data of the target personnel in the monitoring area, and then tracks the target personnel in the video frame, associates the same target personnel in multiple video frames, and identifies the motion trajectory of the target personnel, and then determines the collision between the motion fitting function and the counting line function of the motion trajectory corresponding to the target personnel, and obtains the corresponding behavior state of the target personnel, and finally obtains the behavior state and multi-dimensional personnel data corresponding to the target personnel whose behavior state has changed and reports them to the server and the display terminal. Due to the adoption of the technical solution in this embodiment, the cost of personnel detection can be reduced, and the accuracy of personnel detection results can be improved.

[0024] In this application, if corresponding calculation formulas appear, the above calculation formulas are all dimensionless and take their numerical calculations. The weight coefficients, proportional coefficients and other coefficients in the formulas are set to a result value obtained by quantifying each parameter. The size of the weight coefficient and the proportional coefficient can be determined as long as it does not affect the proportional relationship between the parameter and the result value.

[0025] Embodiment 2: Figure 5 The structural diagram of an electronic device is illustrated, and the electronic device may include: a processor, a communication interface, a memory, and a system bus, wherein the processor, the communication interface, and the memory communicate with each other through the system bus. The processor may call the logic instructions in the memory to execute a method for counting multi-dimensional information of personnel entering and exiting applied to a checkpoint scene, and the method includes: issuing a personnel entry and exit counting task corresponding to a target person in the checkpoint scene, setting a monitoring area and a counting line position and counting direction in the monitoring area; collecting video data in the monitoring area, and analyzing the video data to obtain the personnel frame information and multi-dimensional personnel data of the target person in the monitoring area; tracking the target person in the video frame, associating the same target person in multiple video frames, and identifying the motion trajectory of the target person; judging the collision between the motion fitting function and the counting line function of the motion trajectory corresponding to the target person, and obtaining the corresponding behavior state of the target person; obtaining the behavior state and multi-dimensional personnel data corresponding to the target person whose behavior state has changed and reporting them to the server and the display terminal.

[0026] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0027] Embodiment 3: The present application also provides a computer program product, which includes a computer program stored on a computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute a method for counting multi-dimensional information of people entering and exiting a checkpoint scene provided by the above methods. The method includes: issuing a task of counting people entering and exiting a target person corresponding to the checkpoint scene, setting a monitoring area and a counting line position and counting direction within the monitoring area; collecting video data within the monitoring area, and analyzing the video data to obtain the person border information and multi-dimensional personnel data of the target person in the monitoring area; tracking the target person in the video frame, associating the same target person in multiple video frames, and identifying the motion trajectory of the target person; judging the collision between the motion fitting function and the counting line function of the motion trajectory corresponding to the target person, and obtaining the corresponding behavior state of the target person; obtaining the behavior state and multi-dimensional personnel data corresponding to the target person whose behavior state has changed and reporting them to the server and display terminal.

[0028] Embodiment 4: The present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the above-mentioned method for counting multi-dimensional information of people entering and exiting a checkpoint scene, the method comprising: issuing a task of counting people entering and exiting a target person corresponding to the checkpoint scene, setting a monitoring area and a counting line position and counting direction within the monitoring area; collecting video data within the monitoring area, and analyzing the video data to obtain the person border information and multi-dimensional personnel data of the target person within the monitoring area; tracking the target person in the video frame, associating the same target person in multiple video frames, and identifying the motion trajectory of the target person; judging the collision between the motion fitting function of the motion trajectory corresponding to the target person and the counting line function, and obtaining the corresponding behavior state of the target person; obtaining the behavior state and multi-dimensional personnel data corresponding to the target person whose behavior state has changed, and reporting them to the server and display terminal.

[0029] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0030] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0031] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for counting multi-dimensional information of people entering and leaving a checkpoint, characterized in that: Methods include: Step S1, issuing a personnel entry and exit counting task corresponding to the target personnel in the checkpoint scene, setting the monitoring area and the counting line position and counting direction in the monitoring area; Step S2, collecting video data in the monitoring area, and analyzing the video data to obtain the personnel frame information and multi-dimensional personnel data of the target personnel in the monitoring area; Step S3, tracking the target person in the video frame, associating the same target person in multiple video frames, and identifying the movement trajectory of the target person; Step S4, determining the collision between the motion fitting function of the target person's corresponding motion trajectory and the counting line function, and obtaining the corresponding behavior state of the target person; Step S5, obtaining the behavior status and multi-dimensional personnel data corresponding to the target personnel whose behavior status has changed and reporting them to the server and the display terminal.

2. According to claim 1, a method for counting multi-dimensional information of people entering and leaving a checkpoint, characterized in that: The monitoring area refers to the area where the entry and exit statistics of personnel need to be executed. When the entry and exit results are counted, only the objects within the area are counted, and the objects outside the area are not included in the statistics; The counting line is set in the detection area. When the target person crosses the counting line, it means that the behavior state of the target person has changed; The counting direction is used to confirm the behavior status of the target person. If the target person's motion trajectory is the same as the set counting direction, it means that the target person is in the entry state. If the target person's corresponding motion trajectory is opposite to the set counting direction, it means that the target person is in the exit state. The person border information includes the border height, border width and border center position corresponding to the person border; The multi-dimensional personnel data includes the age, gender and corresponding confidence level of the target personnel.

3. According to claim 1, a method for counting multi-dimensional information of people entering and leaving a checkpoint, characterized in that: The step S2 includes the following sub-steps: Step S21, obtaining video data in the detection area, and dividing the video data into multiple video frames according to time; Step S22, importing the multiple video frames into the trained target detection model, and the target detection model identifies the multi-dimensional personnel data of the target person in the video frame; Step S23, performing frame selection processing on the target person, obtaining a person frame corresponding to the target person, and then recording the person frame information; Step S24, establishing a regional coordinate system according to the monitoring area, and obtaining a function expression corresponding to the counting line function according to the position of the counting line in the regional coordinate system; Step S25, draw the center position of the person's frame corresponding to the frame in the regional coordinate system to obtain the pedestrian center coordinates.

4. The method for counting multi-dimensional information of people entering and leaving a checkpoint according to claim 1, characterized in that: The step S3 includes the following sub-steps: Step S301, obtaining the video frame number j corresponding to the current moment, and obtaining the video frame number jc corresponding to the previous c moments of the current moment; wherein j>c, j=1, 2, ..., z, z is a positive integer; Step S302, obtaining the target person's border and pedestrian center coordinates in multiple video frames between time jc and time j; Step S303, creating an association matrix corresponding to adjacent video frames. If the previous video frame contains m person borders and the next video frame contains n person borders, the created association matrix is: ; Among them JL m,n represents the Euclidean distance between the mth person border in the previous video frame and the nth person border in the next video frame, JL m,n The calculation formula is as follows: 。 5. The method for counting multi-dimensional information of people entering and leaving a checkpoint according to claim 4, characterized in that: The step S3 also includes the following sub-steps: Step S304, performing association for row 1, and selecting the largest value in row 1 as the association result for row 1; Step S305, performing association for row 2, selecting the row with the largest value as the association result for row 2, and so on, until the association of row m is completed; Step S306, repeating steps S303 to S305 until all video frames between time jn and time j are associated with people.

6. The method for counting multi-dimensional information of people entering and leaving a checkpoint according to claim 5, characterized in that: The step S3 also includes the following sub-steps: Step S307, obtaining the pedestrian center coordinates (Xi, Yi) of the target person in multiple video frames; wherein i represents the number of the video frame corresponding to the target person at any time, i=jc,j-c+1, ..., j; Step S308, connecting the pedestrian center coordinates in the order of the video frames to obtain the motion trajectory of the target person; Step S309, calculating the similarity between the target person's motion trajectory and the common function graph; Step S310 , selecting the common function with the highest similarity as the motion fitting function corresponding to the motion trajectory of the target person, and recording the function expression of the motion fitting function as Y=F(X).

7. The method for counting multi-dimensional information of people entering and leaving a checkpoint according to claim 6, characterized in that: If the function expression of the motion fitting function is Y=F(X)=a1*X, where a1 is the coefficient of the first-order term, then calculate The value of , and taking the partial derivative with respect to a1, we get: ； Simplify to get , where EX represents the mean of Xi and EY represents the mean of Yi.

8. The method for counting multi-dimensional information of people entering and leaving a checkpoint according to claim 7, characterized in that: If the function expression of the motion fitting function is Y=F(X)=a0+a1*X, where a0 is a constant term and a1 is a one-time coefficient, then calculate The value of , and taking partial derivatives with respect to a0 and a1 respectively, we get: Taking partial derivatives with respect to a0 we get: Taking partial derivatives of a1 we get: The combined result is: a0=EY-a1×EX, .

9. The method for counting multi-dimensional information of people entering and leaving a checkpoint according to claim 8, characterized in that: If the function expression of the motion fitting function is Y=F(X)=a0+a1*X+a3×X 3 ; Among them, a0 is the constant term, a1 is the one-time coefficient, and a3 is the cubic coefficient; Then calculate The value of , and taking partial derivatives of a0, a1 and a3 respectively, we get: Taking partial derivatives with respect to a0 we get: ; Taking partial derivatives of a1 we get: ; Taking partial derivatives of a3 we get: ; Combined, we get: a0=EY-a1*EX-a3*EX 3 ; Among them EX 3 For Xi 3 The mean of ； 。 10. The method for counting multi-dimensional information of people entering and leaving a checkpoint according to claim 1, characterized in that: The step S4 includes the following sub-steps: Step S41, obtaining a function expression of a motion fitting function and a function expression of a counting line function corresponding to a motion trajectory of a target person; Among them, the counting line function is defined in the form of y=ax+b, where a and b are both constants; Step S42, combining the motion fitting function and the counting line function, if there is no real number solution, no operation is performed; Step S43, if there is a real number solution, it is determined that the behavior state of the target person has changed; Step S44, obtaining the center coordinates of the pedestrian at the moment before and the moment after the intersection. If the center coordinates of the pedestrian at the moment before the intersection are below the counting line, and the center coordinates of the pedestrian at the moment after the intersection are above the counting line, it is determined that the behavior state of the target person has changed, and the behavior state of the target person is recorded as leaving state; Step S45, if the pedestrian center coordinates at a moment before the intersection are above the counting line, and the pedestrian center coordinates at a moment after the intersection are below the counting line, it is determined that the behavior state of the target person has changed, and the behavior state of the target person is recorded as the entry state.

Images