Subway car passenger cleaning judgment method and electronic device based on image analysis

Abstract

The present application relates to a kind of metro carriage passenger discharge judgment method and electronic equipment based on image analysis, method includes: according to passenger discharge command, the real-time video stream of metro carriage is acquired, and decoding generation second image sequence;Current frame image in the second image sequence is segmented into seat area and ground area;Difference processing is carried out to reference image and current frame image in the second image sequence, and pixel analysis is carried out to difference processing result to detect whether the seat area and ground area exist anomaly, if at least one of the seat area and ground area exists anomaly, indicate that passenger discharge is not completed, and feedback exception data;If the seat area and ground area do not exist anomaly, indicate that passenger discharge is completed;Wherein, the reference image is obtained by the metro carriage video stream under unmanned condition, and first image sequence is obtained by decoding generation.This application can effectively identify whether there is passenger in metro carriage, and efficiently complete metro carriage passenger discharge.

Description

Technical Field

[0001] This invention relates to the field of target detection technology, and in particular to a method and electronic device for determining passenger clearance in subway cars based on image analysis. Background Technology

[0002] Against the backdrop of information technology development in the rail transit industry, real-time monitoring of passenger behavior is crucial not only for improving operational efficiency and protecting public safety and property, but also for enhancing the ability of rail companies to handle emergencies. Passenger evacuation and clearance are essential procedures during normal or emergency train operations. However, currently, passenger clearance decisions in subway cars are still made manually, which is inefficient and cannot adapt to intelligent development.

[0003] When a subway train arrives at its final stop, a planned passenger evacuation is usually carried out. Subway announcements will remind passengers to disembark as the train approaches the final stop, and station staff will also remind any remaining passengers in the carriages so that plans such as turning the train around or returning it to the depot can be made. In the event of an emergency requiring the subway train to stop at a station, an unplanned passenger evacuation will also be carried out, which will also use the traditional method of announcements and manual evacuation.

[0004] Cameras installed inside train carriages enable remote video monitoring of the carriages, providing a hardware foundation for the automated identification of stranded personnel. Identifying specific objects through target detection technology is a hot topic in video surveillance. Commonly used algorithms include background subtraction, inter-frame subtraction, and optical flow. Background subtraction and inter-frame subtraction are simpler and easier to implement than optical flow, and have lower hardware requirements. Background subtraction calculates the difference between the current frame in an image sequence and a fixed or real-time acquired background reference model, identifying regions where the pixel difference exceeds a certain threshold as moving areas. This allows for the determination of the position, contour, and size of moving objects. This method can accurately extract moving and even stationary targets, but it is easily affected by the external environment. Inter-frame subtraction is similar to background subtraction, obtaining the contour of moving targets by performing a difference operation on adjacent frames in a video image sequence. This method is more sensitive to moving targets and less affected by natural environmental interference. Optical flow is a method that uses the temporal changes of pixels in an image sequence and the correlation between adjacent frames to infer the speed and direction of an object's movement. This method can accurately detect and identify the position of moving targets without needing to know scene information.

[0005] Traditional passenger clearing methods require station staff to inspect each carriage individually, which is time-consuming, unpredictable, and prone to delays. Furthermore, for fully automated subway lines, traditional manual passenger clearing methods are inefficient and reduce the line's automation level. Object detection-based methods also have shortcomings: optical flow methods have high computational complexity and poor timeliness, making them less effective in passenger clearing tasks. Frame difference methods are prone to missing slow-moving or stationary objects. Background difference methods are easily affected by the external environment and require good background modeling. Summary of the Invention

[0006] Therefore, the technical problem to be solved by the present invention is to overcome the problem of low passenger clearance efficiency in subway cars in the prior art.

[0007] To address the aforementioned technical problems, this invention provides a method for determining passenger clearance in subway cars based on image analysis, comprising:

[0008] Step S1: According to the passenger clearing command, acquire the real-time video stream of the subway car and decode it to generate a second image sequence;

[0009] Step S2: Segment the current frame image in the second image sequence into a seat region and a ground region;

[0010] Step S3: Perform differential processing on the reference image and the current frame image in the second image sequence, and perform pixel analysis on the differential processing result to detect whether there are any anomalies in the seat area and the ground area. If at least one of the seat area and the ground area has an anomaly, it indicates that the passenger clearing is not completed, and the abnormal data is fed back; if neither the seat area nor the ground area has an anomaly, it indicates that the passenger clearing is completed; wherein, the reference image is obtained by acquiring the subway car video stream under unmanned conditions and decoding it to generate the first image sequence.

[0011] In one embodiment of the present invention, step S3, which involves performing differential processing on the reference image and the current frame image in the second image sequence, and performing pixel analysis on the differential processing result to detect whether there is an anomaly in the seat area, includes:

[0012] Step S311: Obtain the pixel points corresponding to the seat region of the current frame image in the second image sequence;

[0013] Step S312: Perform differential processing on the current frame image in the second image sequence and the reference image to obtain the first differential image;

[0014] Step S313: If the pixels in the first difference image are greater than the first background threshold, then the pixels greater than the first background threshold are used as foreground pixels and assigned the value 1; otherwise, they are used as background pixels and assigned the value 0; finally, the first binarized image is generated.

[0015] Step S314: Mesh the first binarized image. When the proportion of foreground pixels in the mesh exceeds a set percentage, calculate the number of meshes where the proportion of foreground pixels exceeds the set percentage.

[0016] Step S315: Determine whether there is an anomaly in the seat area by the number of grids whose foreground pixel ratio exceeds a set percentage.

[0017] In one embodiment of the present invention, in step S312, the current frame image in the second image sequence is differentially processed with the reference image, and the formula is: D i (x,y)=|T i (x,y)-O(x,y)|, where D i (x,y) represents the result of the difference processing, T i (x,y) represents the gray value of the pixel corresponding to the seat area in the i-th frame of the second image sequence, and O(x,y) represents the gray value of the pixel corresponding to the seat area in the reference image.

[0018] In one embodiment of the present invention, step S313 uses an adaptive threshold segmentation method to treat pixels larger than a first background threshold as foreground pixels and assign them a value of 1; otherwise, they are treated as background pixels and assigned a value of 0. The formula is as follows: Among them, D i R represents the first difference image. i Let θ represent the first binarized image, and let θ represent the first background threshold.

[0019] In one embodiment of the present invention, step S315 specifically comprises:

[0020] The subway car is divided into three zones according to the distance from the camera: the close-range zone, the medium-range zone, and the far-range zone. The camera is used to capture video streams from the subway car.

[0021] If the proportion of foreground pixels in the far-distance area of ​​the carriage exceeds a set percentage of the grid, then there is an anomaly in the seating area, and it is considered that the passenger clearing is not completed.

[0022] If the number of grids in the distance area of ​​the carriage where the proportion of foreground pixels exceeds a set percentage is greater than or equal to a first set number, then there is an anomaly in the seating area, and it is considered that the passenger clearance has not been completed.

[0023] If the number of grids in the near-field area of ​​the carriage where the proportion of foreground pixels exceeds a set percentage is less than or equal to a second set number, then there is no abnormality in the seat area; otherwise, there is an abnormality in the seat area.

[0024] In one embodiment of the present invention, step S3, which involves performing differential processing on the reference image and the current frame image in the second image sequence, and performing pixel analysis on the differential processing result to detect whether there are anomalies in the ground region, includes:

[0025] Step S321: Obtain the pixel points corresponding to the ground region of the current frame image in the second image sequence;

[0026] Step S322: Perform differential processing on the current frame image in the second image sequence and the reference image to obtain a second differential image;

[0027] Step S323: If a pixel in the second difference image is greater than the second background threshold, then the pixel greater than the second background threshold is taken as a foreground pixel and assigned a value of 1; otherwise, it is taken as a background pixel and assigned a value of 0; finally, the second binarized image is generated.

[0028] Step S324: Mesh the second binarized image. When the proportion of foreground pixels in the mesh exceeds a set percentage, calculate the number of meshes where the proportion of foreground pixels exceeds the set percentage.

[0029] Step S325: If there are no grids in the ground area with a foreground pixel ratio exceeding a set percentage, then there is no anomaly in the ground area; if there are grids in the ground area with a foreground pixel ratio exceeding a set percentage, then the RGB histogram of the current frame image in the second image sequence is calculated, and it is determined whether there is an anomaly in the ground area.

[0030] In one embodiment of the present invention, step S325, which involves calculating the RGB histogram of the current frame image in the second image sequence and determining whether there are anomalies in the ground region, includes:

[0031] Obtain the RGB histogram of the reference image and count the number of peaks;

[0032] Obtain the RGB histogram of the current frame image in the second image sequence and count the number of peaks;

[0033] If the number of peaks in the current frame of the second image sequence is greater than the number of peaks in the reference image, then there is an anomaly in the ground region; otherwise, there is no anomaly in the ground region.

[0034] In one embodiment of the present invention, in step S2, the current frame image in the second image sequence is segmented into a seat region and a ground region using an image masking method.

[0035] The present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the above-mentioned image analysis-based subway car passenger clearance judgment method.

[0036] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of the above-described method for determining passenger clearance in a subway car based on image analysis are implemented.

[0037] The technical solution of the present invention has the following advantages compared with the prior art:

[0038] The image analysis-based subway car passenger clearance judgment method described in this invention improves detection efficiency by extracting regions of interest (i.e., ground and seat areas) in areas where stranded passengers are likely to appear through masking. Furthermore, by preprocessing and differentially analyzing the pixels within the regions of interest, and based on the three-stage principle and RGB histogram statistical analysis, it effectively overcomes light spot interference caused by different lighting conditions and train vibrations, thus improving the accuracy of the method. This invention, based on pixel-level image processing, exhibits high robustness and reliability. Attached Figure Description

[0039] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein...

[0040] Figure 1 This is a flowchart of the method of the present invention;

[0041] Figure 2 This is a flowchart of the seat area detection process of the present invention;

[0042] Figure 3 This is a flowchart of the ground area detection process of the present invention;

[0043] Figure 4 This is a schematic diagram illustrating the division of the subway car into three areas in this invention;

[0044] Figure 5 This is a schematic diagram of the grid detection results of the present invention;

[0045] Figure 6 This is a peak comparison result diagram based on RGB histogram statistics of the present invention. Detailed Implementation

[0046] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0047] Example 1

[0048] Reference Figure 1 As shown, the method of the present invention includes:

[0049] Step S1: According to the passenger clearing command, acquire the real-time video stream of the subway car and decode it to generate the second image sequence T;

[0050] Step S2: Divide each frame of the second image sequence T into seat region A and ground region B;

[0051] Step S3: Perform differential processing on the reference image and the frame images in the second image sequence, and perform pixel analysis on the differential processing results to detect whether there are any anomalies in the seat area A and the ground area B. If at least one of the seat area A and the ground area B has an anomaly, it indicates that the passenger clearing is not completed, and the abnormal data is fed back (the abnormal data can be fed back to the control center or reminded to the subway driver through warning lights); if neither the seat area A nor the ground area B has any anomalies, it indicates that the passenger clearing is completed; wherein, the reference image is obtained by: obtaining the subway car video stream under unmanned conditions according to the RTSP address of the camera installed in the car, and decoding it to generate a first image sequence, and using the first image sequence as the reference image O for image differential processing.

[0052] Preferably, in order to improve efficiency, in this embodiment, step S3 first detects whether there is an abnormality in the seating area A. If there is an abnormality in the seating area A, the ground area B is no longer detected and is regarded as not having been cleared. If there is no abnormality in the seating area A, the ground area B is detected. If there is an abnormality in the ground area B, it is regarded as not having been cleared. If there is no abnormality in the ground area B, it is regarded as having been cleared.

[0053] Specifically, such as Figure 2 As shown, step S3 involves performing differential processing on the reference image and the current frame image in the second image sequence, and performing pixel analysis on the differential processing result to detect whether there is an anomaly in the seat region A, including:

[0054] Step S311: Obtain the pixel points corresponding to the seat area A, specifically: extract the region of interest from the obtained second image sequence T according to the mask of seat area A, and replace the pixels outside the mask area with black pixels;

[0055] Step S312: Perform Gaussian filtering and normalization preprocessing on the pixels corresponding to seat region A, and then perform difference processing on the current frame image in the second image sequence T (which here corresponds to the pixels corresponding to seat region A) and the reference image O to obtain the first difference image D. i ;

[0056] Further, in step S312, the current frame image in the second image sequence T is differentially processed with the reference image O, using the following formula:

[0057] D i (x,y)=|T i (x,y)-O(x,y)|

[0058] Among them, D i (x,y) represents the result of the difference processing, T i (x,y) represents the gray value of the pixel corresponding to the seat area in the i-th frame of the second image sequence, and O(x,y) represents the gray value of the pixel corresponding to the seat area in the reference image.

[0059] Step S313: If the first difference image D i If a pixel in the image is greater than the first background threshold θ, then that pixel is treated as a foreground pixel and assigned a value of 1; otherwise, it is treated as a background pixel and assigned a value of 0. Finally, the first binarized image R is generated. i ;

[0060] Further, in step S313, a first background threshold θ is calculated using the Adaptive Thresholding (OTSU) method, and pixels larger than the first background threshold θ are designated as foreground pixels and assigned a value of 1; otherwise, they are designated as background pixels and assigned a value of 0, thereby generating a first binarized image R. i The formula is:

[0061]

[0062] Among them, D i R represents the first difference image. i Let θ represent the first binarized image, and let θ represent the first background threshold.

[0063] Step S314: For the first binarized image R i Meshization is performed on the binarized image R. i A grid is laid on top. In this embodiment, the grid size is set to 30×30, which means the image is divided into 64×36 parts. When the foreground pixels in the grid account for more than 75%, the part will be highlighted and the number of grids with the foreground pixels accounting for more than 75% will be calculated.

[0064] Step S315: Determine whether the clearing of customers is complete by the number of grids where the foreground pixels account for more than 75%.

[0065] Further, step S315 specifically involves: determining the number of grid cells based on the three zones of the subway car, such as... Figure 4As shown, based on the current seating arrangement in the subway car, the subway car is divided into three areas from near to far from the camera: the near distance area L0, the medium distance area L1, and the far distance area L2. The camera is used to capture video streams from the subway car.

[0066] (1) If the far-distance area L2 of the carriage has a grid with a foreground pixel ratio of more than 75%, then there is an object and it is considered that the passenger clearance is not completed.

[0067] (2) If the number of grids in the distance area L1 of the carriage where the foreground pixels account for more than 75% is greater than or equal to 2, then there is an object and it is considered that the passenger clearance has not been completed.

[0068] (3) If the number of grids in the near-field area L0 of the carriage with a foreground pixel ratio of more than 75% is less than or equal to 3, it is considered as a light spot and the passenger clearing is considered to be completed; otherwise, it is considered as the passenger clearing is not completed.

[0069] An anomaly alarm is triggered only if the number of detected grids exceeds the grid threshold for its region. If a stage is determined to be incomplete, the system exits the judgment process directly and an anomaly is reported. The detection results are as follows: Figure 5 As shown.

[0070] Specifically, such as Figure 3 As shown, step S3, which involves performing differential processing on the reference image and the current frame image in the second image sequence, and then performing pixel analysis on the differential processing result to detect whether there are anomalies in the ground area, includes:

[0071] Step S321: Obtain the pixel points corresponding to the ground region B, specifically: extract the region of interest from the obtained second image sequence T based on the ground region B mask, and replace the pixel points outside the mask region with black pixels;

[0072] Step S322: Perform Gaussian filtering and normalization preprocessing on the pixels corresponding to ground region B, and perform difference processing on the current frame image in the second image sequence T and the reference image O to obtain the second difference image;

[0073] Step S323: If a pixel in the second difference image is greater than the second background threshold, then the pixel greater than the second background threshold is taken as a foreground pixel and assigned a value of 1; otherwise, it is taken as a background pixel and assigned a value of 0; finally, the second binarized image is generated.

[0074] Step S324: Mesh the second binarized image. When the proportion of foreground pixels in the mesh exceeds 75%, calculate the number of meshes where the proportion of foreground pixels exceeds 75%.

[0075] Step S325: If there are no grids in the ground area with a foreground pixel ratio exceeding 75%, the clearing of customers is considered complete; if there are grids in the ground area with a foreground pixel ratio exceeding 75%, the RGB histogram of the current frame image in the second image sequence is calculated, and it is determined whether the clearing of customers is complete.

[0076] Further, step S325, which involves calculating the RGB histogram of the current frame image in the second image sequence and determining whether the clearing process is complete, includes:

[0077] (1) Obtain the RGB histogram of the reference image and count the number of peaks;

[0078] (2) Obtain the RGB histogram of the current frame image in the second image sequence and count the number of peaks;

[0079] (3) If the number of peaks in the current frame image in the second image sequence is greater than the number of peaks in the reference image, then the clearing is considered incomplete; otherwise, the clearing is considered complete.

[0080] like Figure 6 As shown, although the peak heights change in the two frames, the number of peaks remains the same. Therefore, it can be concluded that there are no anomalies in ground region B in these two frames.

[0081] It should be noted that this embodiment only analyzes the current frame image in the second image sequence T to determine whether there are any anomalies in the ground area and the seat area. In other embodiments, not only the current frame image can be considered, but also several consecutive frames following the current frame image can be analyzed to ensure the reliability of the method.

[0082] Example 2

[0083] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the image analysis-based subway car passenger clearance judgment method described in Embodiment 1.

[0084] Example 3

[0085] This embodiment provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the steps of the image analysis-based subway car passenger clearance judgment method described in Embodiment 1.

[0086] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0087] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0088] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0089] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0090] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0091] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for determining passenger clearance in subway cars based on image analysis, characterized in that: include: Step S1: According to the passenger clearing command, acquire the real-time video stream of the subway car and decode it to generate a second image sequence; Step S2: Segment the current frame image in the second image sequence into a seat region and a ground region; Step S3: Perform differential processing on the reference image and the current frame image in the second image sequence, and perform pixel analysis on the differential processing result to detect whether there are any anomalies in the seat area and the ground area. If at least one of the seat area and the ground area has an anomaly, it indicates that the passenger clearing is not completed, and the abnormal data is fed back; if neither the seat area nor the ground area has an anomaly, it indicates that the passenger clearing is completed; wherein, the reference image is obtained by acquiring the subway car video stream under unmanned conditions and decoding it to generate the first image sequence; Step S3, which involves performing differential processing on the reference image and the current frame image in the second image sequence, and then performing pixel analysis on the differential processing result to detect whether there are any anomalies in the ground area, includes: Step S321: Obtain the pixel points corresponding to the ground region of the current frame image in the second image sequence; Step S322: Perform differential processing on the current frame image in the second image sequence and the reference image to obtain a second differential image; Step S323: If a pixel in the second difference image is greater than the second background threshold, then the pixel greater than the second background threshold is taken as a foreground pixel and assigned a value of 1; otherwise, it is taken as a background pixel and assigned a value of 0; finally, the second binarized image is generated. Step S324: Mesh the second binarized image. When the proportion of foreground pixels in the mesh exceeds a set percentage, calculate the number of meshes where the proportion of foreground pixels exceeds the set percentage. Step S325: If there are no grids in the ground area with a foreground pixel ratio exceeding a set percentage, then there is no anomaly in the ground area; if there are grids in the ground area with a foreground pixel ratio exceeding a set percentage, then the RGB histogram of the current frame image in the second image sequence is calculated, and it is determined whether there is an anomaly in the ground area. Step S325, which involves calculating the RGB histogram of the current frame image in the second image sequence and determining whether there are anomalies in the ground region, includes: Obtain the RGB histogram of the reference image and count the number of peaks; Obtain the RGB histogram of the current frame image in the second image sequence and count the number of peaks; If the number of peaks in the current frame of the second image sequence is greater than the number of peaks in the reference image, then there is an anomaly in the ground region; otherwise, there is no anomaly in the ground region.

2. The method for determining passenger clearance in subway cars based on image analysis according to claim 1, characterized in that: Step S3, which involves performing differential processing on the reference image and the current frame image in the second image sequence, and then performing pixel analysis on the differential processing result to detect whether there are any anomalies in the seat area, includes: Step S311: Obtain the pixel points corresponding to the seat region of the current frame image in the second image sequence; Step S312: Perform differential processing on the current frame image in the second image sequence and the reference image to obtain the first differential image; Step S313: If the pixels in the first difference image are greater than the first background threshold, then the pixels greater than the first background threshold are used as foreground pixels and assigned the value 1; otherwise, they are used as background pixels and assigned the value 0; finally, the first binarized image is generated. Step S314: Mesh the first binarized image. When the proportion of foreground pixels in the mesh exceeds a set percentage, calculate the number of meshes where the proportion of foreground pixels exceeds the set percentage. Step S315: Determine whether there is an anomaly in the seat area by the number of grids whose foreground pixel ratio exceeds a set percentage.

3. The method for determining passenger clearance in subway cars based on image analysis according to claim 2, characterized in that: In step S312, the current frame image in the second image sequence is differentially processed with the reference image, using the following formula: ,in, This indicates the result of the difference processing. Indicates the second image sequence The grayscale value of the pixel corresponding to the seat area in the frame image. This represents the grayscale value of the corresponding pixel in the seat area of ​​the reference image.

4. The method for determining passenger clearance in subway cars based on image analysis according to claim 2, characterized in that: In step S313, the adaptive threshold segmentation method is used to identify pixels larger than the first background threshold as foreground pixels and assign them a value of 1. Otherwise, treat it as a background pixel and assign it a value of 0. The formula is: ,in, This represents the first difference image. Represents the first binarized image. This represents the first background threshold.

5. The method for determining passenger clearance in subway cars based on image analysis according to claim 2, characterized in that: Step S315 specifically involves: The subway car is divided into three zones according to the distance from the camera: the close-range zone, the medium-range zone, and the far-range zone. The camera is used to capture video streams from the subway car. If the proportion of foreground pixels in the far-distance area of ​​the carriage exceeds a set percentage of the grid, then there is an anomaly in the seating area, and it is considered that the passenger clearance has not been completed. If the number of grids in the distance area of ​​the carriage where the proportion of foreground pixels exceeds a set percentage is greater than or equal to a first set number, then there is an anomaly in the seating area, and it is considered that the passenger clearance has not been completed. If the number of grids in the near-field area of ​​the carriage where the proportion of foreground pixels exceeds a set percentage is less than or equal to a second set number, then there is no abnormality in the seat area; otherwise, there is an abnormality in the seat area.

6. The method for determining passenger clearance in subway cars based on image analysis according to claim 1, characterized in that: In step S2, the current frame image in the second image sequence is segmented into a seat area and a ground area using an image masking method.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the image analysis-based subway car passenger clearance judgment method as described in any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the image analysis-based subway car passenger clearance judgment method as described in any one of claims 1 to 6.

Images