Data collection and user behavior analysis system for urban rail transit

Abstract

This invention discloses a data acquisition and user behavior analysis system for urban rail transit, relating to the field of behavioral analysis technology. It includes an image acquisition module that collects image data containing pixel coordinates and numbers in real time; a data preprocessing module that preprocesses the data; and a central computing module that performs comprehensive calculations and standardization on the preprocessed image data to obtain passenger transfer disorder parameters. If the analysis indicates disorder, the prompt module is invoked; otherwise, the system terminates. The prompt module is responsible for issuing polite transfer prompts. Through image recognition and algorithmic analysis, the system can quickly identify transfer disorder and issue prompts, avoiding accidents and protecting vulnerable groups. It can also objectively assess order, reduce delays, and alleviate passenger anxiety and enhance trust through polite prompts.

Description

Technical Field

[0001] The present invention relates to the technical field of behavior analysis, and in particular to a data collection and user behavior analysis system for urban rail transit. Background Art

[0002] Analyzing user behavior in urban rail transit is a key measure for improving operational efficiency and service quality. By analyzing data such as passengers' travel time distribution, transfer route choices, and station traffic changes, operators can accurately adjust train schedules to avoid insufficient capacity during peak hours or wasted resources during off-peak periods. At the same time, ticketing systems and gate settings can be optimized based on passengers' ticket purchase preferences and entry and exit habits to reduce queues and congestion. Furthermore, analyzing abnormal behavior patterns can promptly identify safety hazards and create a more convenient and safe travel environment for passengers. This kind of data-driven, refined management is both a necessary requirement for meeting the challenges of large passenger flows and a key manifestation of the intelligent development of urban transportation.

[0003] For example, China Patent Publication No.: CN109377161B provides a system for office data collection and user behavior analysis in the urban rail transit industry, including a business process diagram and a system architecture, which is characterized by: realizing the collection, storage and analysis of user login data, paying attention to the time period, source, device type (PC, mobile), frequency, etc. of user usage behavior, and then formulating a more scientific and reasonable office operation and maintenance strategy. It realizes the collection, storage and analysis of user access portal system column data, records the column type and access volume visited by users for data structuring and labeling, obtains user demand preferences, and improves user experience. It realizes the collection, storage and analysis of user OA office process approval operation behavior data, solves business needs, optimizes business processes, improves office efficiency and reduces costs. The above-mentioned system for office data collection and user behavior analysis in the rail transit industry meets the second-level protection requirements of the information system security protection level.

[0004] However, the above plan does not take into account that in the rail transit transfer waiting area, when the transfer area lacks clear flow guidance, a large number of passengers will fall into a blind flow state. When the passengers getting off the train are getting off, the passengers getting on the train are eager to get on the train, resulting in pushing and shoving between the passengers getting on and off the train, causing some passengers to lose balance and increasing the risk of trampling. Summary of the Invention

[0005] Technical problems solved

[0006] In response to the shortcomings of the existing technology, the present invention provides a data collection and user behavior analysis system for urban rail transit, which solves the problem of unclear flow guidance in the rail transit transfer waiting area, resulting in blind flow and pushing of passengers getting on and off the train, causing some passengers to lose balance and increase the risk of trampling.

[0007] Technical Solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: a data acquisition and user behavior analysis system for urban rail transit, including the following specific modules: an image acquisition module: starts real-time acquisition of image data, the image data including pixel coordinates and pixel numbers; a data preprocessing module: preprocesses the image data; a central computing and processing module: performs comprehensive calculations and standardization on the preprocessed image data to obtain passenger transfer chaos parameters, analyzes the passenger transfer chaos situation based on the passenger transfer chaos parameters, calls a prompt module and ends the analysis if the analysis shows passenger transfer chaos, and directly ends the analysis if the analysis shows that the passenger transfer is normal; and a prompt module: is used to issue polite transfer prompts.

[0009] Furthermore, the passenger transfer chaos parameter is specifically obtained as follows: image data is calculated using an image contour segmentation algorithm and an image contour tracking algorithm to obtain an image contour, pixel coordinates within the image contour are averaged according to the number of pixels in the image contour to obtain center coordinates, door contours and human contours are screened out according to the image contour, alighting human contours and boarding human contours are screened out according to the pixel coordinates of the door contour, the center coordinates of the door contour, the pixel coordinates of the human contour, and the center coordinates of the human contour, a chaos detection time is set, and within the chaos detection time, the number of boarding human contours and alighting human contours is counted to obtain the number of boarding human contours and the number of alighting human contours, the center coordinates of the boarding human contours are summed and averaged according to the number of boarding human contours to obtain the boarding center coordinates, the center coordinates of the alighting human contours are summed and averaged according to the number of alighting human contours to obtain the alighting center coordinates, the distance between the boarding center coordinates and the alighting center coordinates is calculated to obtain a chaos density, and the chaos density is calculated using a variance method and a time series of the chaos detection time to obtain the passenger transfer chaos parameter.

[0010] Furthermore, the specific method for obtaining the outline of the human body getting off the vehicle is as follows: matching the pixel coordinates of the door outline with the pixel coordinates of the human body outline; if the pixel coordinates of the door outline match the pixel coordinates of the human body outline, it means that the human body outline passes through the door outline; if the pixel coordinates of the door outline do not match the pixel coordinates of the human body outline, it means that the human body outline does not pass through the door outline; setting a passenger getting off detection time period; when the human body outline passes through the door outline, marking this human body outline as the human body outline to be getting off the vehicle, and triggering the passenger getting off detection time period, that is, within the passenger getting off detection time period, calculating the distance between the human body outline to be getting off and the door outline to obtain the getting off distance; comparing the getting off distance of the next moment with the getting off distance of the previous moment according to the time series; if the getting off distance of the next moment is greater than the getting off distance of the previous moment, marking this human body outline to be getting off as the getting off human body outline; if the getting off distance of the next moment is less than or equal to the getting off distance of the previous moment, then not marking this human body outline to be getting off.

[0011] Furthermore, the specific method of matching the pixel coordinates of the door contour with the pixel coordinates of the human body contour is as follows: assuming that the number of pixel points of the door contour is , the number of pixels of a human body contour is , assume that the first The pixel coordinates are , the door profile The pixel coordinates are ,right and Perform vector calculation to get the get-off vector , according to the get-off vector Perform a modulo calculation to obtain the overlap distance, and compare the overlap distance with zero. If the overlap distance is equal to zero, it means that the pixel coordinates of the car door contour match the pixel coordinates of the human body contour. If the overlap distance is greater than zero, continue to match the other pixel coordinates of the human body contour with the other pixel coordinates of the car door contour until the matching is completed. If the overlap distance is greater than zero, it means that the pixel coordinates of the car door contour do not match the pixel coordinates of the human body contour.

[0012] Furthermore, the specific method of obtaining the getting-off distance is as follows: Assume that the coordinates of the center point of the outline of the person to be disembarked are , the center point coordinates of the door contour are , through the Euclidean distance formula and Perform distance calculation to obtain the disembarkation distance; ;in, Indicates the distance to get off. Represents the coordinates of the center point of the silhouette of the person waiting to get off the bus axis, Represents the center point coordinates of the door outline axis, Represents the coordinates of the center point of the silhouette of the person waiting to get off the bus axis, Represents the center point coordinates of the door outline axis.

[0013] Furthermore, the specific method of obtaining the outline of the human body getting on the vehicle is as follows: while triggering the passenger getting off the vehicle detection time period, the moving direction and moving distance of the human body outline toward the vehicle door outline are analyzed within the passenger getting off the vehicle detection time period. If the moving direction of the human body outline is toward the vehicle door outline and the moving distance between the human body outline and the vehicle door outline gradually decreases, then this human body outline is marked as the outline of the human body getting on the vehicle; if the moving direction of the human body outline is not toward the vehicle door outline or the moving distance between the human body outline and the vehicle door outline does not gradually decrease, then this human body outline is not marked as the outline of the human body getting on the vehicle.

[0014] Furthermore, the specific method of analyzing the movement direction and movement distance of the human body contour toward the door contour is as follows: according to the time sequence of the passenger getting off detection period, the center point coordinates of the human body contour at the next moment and the center point coordinates of the human body contour at the previous moment are vector-calculated to obtain the movement vector. Let the time sequence of the passenger getting off detection period be , then the number of movement vectors is , according to the number of motion vectors, the motion vectors are averaged to obtain the average motion vector. The center point coordinates of the human body contour and the pixel coordinates of the door contour at the current moment of the time series of the passenger getting off detection period are vector-calculated to obtain the predicted vector. The average motion vector and the predicted vector are analyzed to see if they coincide. If the average motion vector coincides with the predicted vector, it means that the human body contour is moving toward the door contour. If the average motion vector and the predicted vector do not coincide, the coordinates of other pixel points of the door contour are calculated. After all the pixel coordinates of the door contour are involved in the calculation, if the motion vector and the predicted vector still do not coincide, it means that the human body contour is not moving toward the door contour. The distance between the center point coordinates of the human body contour and the center point coordinates of the door contour is calculated to obtain the boarding distance. According to the time series, the boarding distance of the next moment is compared with the boarding distance of the previous moment. If the boarding distance of the next moment is less than the boarding distance of the previous moment, it means that the human body contour is gradually approaching the door contour. If the boarding distance of the next moment is greater than or equal to the boarding distance of the previous moment, it means that the human body contour is not gradually approaching the door contour.

[0015] Furthermore, the specific method of obtaining the motion vector is as follows: ;in, represents the motion vector, Indicates the coordinates of the center point of the human body contour at the next moment axis, Represents the coordinates of the center point of the human body contour at the previous moment axis, Indicates the coordinates of the center point of the human body contour at the next moment axis, Represents the coordinates of the center point of the human body contour at the previous moment axis.

[0016] Furthermore, the specific method of analyzing whether the average moving vector and the predicted vector coincide with each other is: calculating the average moving vector and the predicted vector using the angle cosine formula to obtain a cosine value; ;in, Represents the cosine value, and is between negative one and one. represents the average moving vector, represents the prediction vector, represents the magnitude of the mean moving vector, Represents the modulus of the prediction vector; based on the comparison of the cosine value with one, if the cosine value is equal to one, it means that the average moving vector coincides with the prediction vector; if the cosine value is not equal to one, it means that the average moving vector does not coincide with the prediction vector.

[0017] Furthermore, the specific method of analyzing the passenger transfer chaos situation based on the passenger transfer chaos parameter is: setting a transfer chaos threshold, comparing the passenger transfer chaos parameter with the transfer chaos threshold; if the transfer chaos parameter is greater than the transfer chaos threshold, it indicates that the passenger transfer is chaotic; if the transfer chaos parameter is less than or equal to the transfer chaos threshold, it indicates that the passenger transfer is normal.

[0018] Beneficial effects

[0019] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0020] 1. The image acquisition module captures transfer area image data in real time, and uses image contour segmentation and tracking algorithms to accurately identify door and human outlines. Vector calculation and distance analysis are then used to distinguish between boarding and exiting figures. The variance of the chaos density is then used to calculate the passenger transfer chaos parameter. When the parameter exceeds the threshold, the prompt module promptly issues a polite transfer reminder. This allows for rapid intervention when signs of chaos, such as pushing and jostling, occur in passenger flow, preventing local congestion from escalating into stampedes. This provides effective protection for vulnerable groups, such as the elderly and children, and technically addresses the safety pain point of traditional manual observation, which makes timely warnings difficult.

[0021] 2. Through standardized processing and quantitative analysis, transfer order can be objectively evaluated to avoid boarding and alighting congestion caused by disordered passenger flow. Passengers can enter the carriages in an orderly manner, reducing train stop delays. At the same time, polite reminders replace mandatory management, which not only maintains order but also alleviates passenger anxiety and enhances trust in rail transit.

[0022] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is the flow chart of the present invention: data collection and user behavior analysis for urban rail transit.

[0024] Figure 2 This is the structural diagram of the data collection and user behavior analysis system for urban rail transit. DETAILED DESCRIPTION

[0025] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0026] It should be noted that, in this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include," "comprise," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0027] Example 1:

[0028] like Figure 1-Figure 2 As shown, the embodiment of the present invention provides a data collection and user behavior analysis system for urban rail transit, which includes the following specific modules:

[0029] Image acquisition module: Starts to collect image data in real time through cameras arranged in the waiting area. The image data includes pixel coordinates and pixel numbers.

[0030] Data preprocessing module: performs noise reduction on image data to improve the image quality.

[0031] Central computing processing module: Processes the denoised image data through image contour segmentation algorithm and image contour tracking algorithm. Image contour segmentation algorithm: For example, the Sobel algorithm first converts the image data into a grayscale image to simplify the calculation, and then convolves with 3×3 convolution kernels in the horizontal and vertical directions to obtain horizontal and vertical gradient matrices respectively; then calculates the gradient amplitude and direction, the former determines the edge, and the latter determines the edge extension direction; finally, through threshold binarization, the pixels above the threshold are marked as contours, and finally the image contour is obtained. Image contour tracking algorithm: For example, the Lucas-Kanade optical flow method, establishes the optical flow constraint equation based on the assumption of constant brightness of adjacent frames, calculates the spatial and temporal gradients through the Sobel algorithm and frame difference, assumes that the neighborhood motion is consistent, constructs an overdetermined equation group, and solves the optical flow velocity by the least squares method, and finally performs tracking and position update, thereby tracking the image contour;

[0032] The door contours and human body contours are screened out based on the image contours, and are subjected to comprehensive calculation and standardization processing to obtain the passenger transfer chaos parameters. The dimensional differences of the passenger transfer chaos parameters are eliminated, and the values ​​of different orders of magnitude are converted into a unified numerical range. The passenger transfer chaos situation is analyzed according to the passenger transfer chaos parameters. If the analysis shows that the passenger transfer is chaotic, the prompt module is called and the analysis ends. If the analysis shows that the passenger transfer is normal, the analysis ends directly.

[0033] Prompt module: used to issue polite transfer prompts.

[0034] Example 2 differs from Example 1 in that:

[0035] The specific method of filtering out the door outline and human body outline based on the image outline is as follows:

[0036] A door template contour area threshold and a human body template contour area threshold are preset, and the number of pixels within the image contour is summed to obtain the image contour area. The image contour area is compared with the door template contour area threshold. If the image contour area is within the door template contour area threshold, the image contour is marked as a door contour. If the image contour area is outside the door template contour area threshold, the image contour is not marked as a door contour. The image contour area is compared with the human body template contour area threshold. If the image contour area is within the human body template contour area threshold, the image contour is marked as a human body contour. If the image contour area is outside the human body template contour area threshold, the image contour is not marked as a human body contour.

[0037] The specific method of obtaining passenger transfer chaos parameters is as follows:

[0038] According to the number of pixels of the image contour, the pixel coordinates within the image contour are averaged to obtain the center point coordinates, which are used to represent the overall image contour for subsequent calculations. According to the pixel coordinates of the door contour, the center point coordinates of the door contour, the pixel coordinates of the human body contour and the center point coordinates of the human body contour, the contours of the getting-off-the-car and the getting-on-the-car are screened out, and a confusion detection time is set, that is, the timing starts from the time the contours of the getting-off-the-car are obtained, and the timing ends when the distance between all the contours of the getting-off-the-car and the door contour is greater than the distance between all the contours of the getting-on-the-car and the door contour, indicating that the contours of the getting-off-the-car have all gotten rid of the obstruction of the contours of the getting-on-the-car. During this process, the contours of the getting-on-the-car stand still or move toward the door contour. During the confusion detection time, the number of the contours of the getting-on-the-car and the contours of the getting-on-the-car are counted to obtain the distance between the contours of the getting-on-the-car and the human body. The number of silhouettes and the number of silhouettes of people getting off the bus are calculated. The coordinates of the center points of the silhouettes of people getting on the bus are summed up and averaged according to the number of silhouettes of people getting on the bus to obtain the center coordinates of the boarding bus. The coordinates of the center points of the silhouettes of people getting off the bus are summed up and averaged according to the number of silhouettes of people getting off the bus to obtain the center coordinates of the boarding bus. The distance between the center coordinates of the boarding bus and the center coordinates of the getting off bus is calculated to obtain the chaos density. The specific method for obtaining the chaos density is the same as the specific method for obtaining the distance from the bus. The chaos density is calculated by using the variance method and the time series of the chaos detection time to obtain the passenger transfer chaos parameter, that is, the passenger transfer chaos parameter reflects the degree of crowding and pushing between the passengers getting off the bus and the passengers getting on the bus. The higher the degree of crowding and pushing, the greater the density fluctuation between the passengers getting off the bus and the passengers getting on the bus, that is, the greater the fluctuation of the chaos density, and the greater the passenger transfer chaos parameter.

[0039] ;

[0040] in, represents the passenger transfer chaos parameter, Indicates the A chaotic density, A time series representing the time of chaos detection.

[0041] The specific method of obtaining the human body silhouette after getting off the vehicle is as follows:

[0042] When the vehicle arrives at the waiting area, the camera arranged in the waiting area first collects the door outline. When the door is opened, revealing the passengers in the car, the camera arranged in the waiting area collects the human body outline near the door inside the car. Since the passengers will pass by the door when getting off the car and gradually move away from the door after getting off the car, that is, the human body outline passes through the door outline and gradually moves away from the door outline. Such a human body outline represents the human body outline getting off the car. Therefore, the pixel coordinates of the door outline are matched with the pixel coordinates of the human body outline. If the pixel coordinates of the door outline match the pixel coordinates of the human body outline, it means that the human body outline passes through the door outline. If the pixel coordinates of the door outline do not match the pixel coordinates of the human body outline, it means that the human body outline does not pass through the door outline. A passenger getting off detection time period is set, that is, when the human body outline passes through the door outline, the human body outline is detected within the passenger getting off detection time period. Gradually move away from the door, and use the time constraint after the human body contour passes the door contour to avoid subsequent continuous calculation, waste of computing power and reduced efficiency. When the human body contour passes the door contour, this human body contour is marked as the human body contour to be gotten off, and the passenger getting off detection time period is triggered, that is, within the passenger getting off detection time period, the distance between the human body contour to be gotten off and the door contour is calculated to obtain the getting off distance. According to the time series, the getting off distance of the next moment is compared with the getting off distance of the previous moment. If the getting off distance of the next moment is greater than the getting off distance of the previous moment, that is, the human body contour to be gotten off gradually moves away from the door contour, indicating that the passenger gets off, then the human body contour to be gotten off is marked as the getting off human body contour. If the getting off distance of the next moment is less than or equal to the getting off distance of the previous moment, the human body contour to be gotten off is not marked, that is, there is a situation where the passenger has passed the door but returned to the car.

[0043] The specific method for matching the pixel coordinates of the door outline with the pixel coordinates of the human body outline is:

[0044] Assume the number of pixels of the door outline is , the number of pixels of a human body contour is , assume that the first The pixel coordinates are , the door profile The pixel coordinates are ,right and Perform vector calculation to get the get-off vector , according to the get-off vector Perform modulo calculation and get off the vector The modulus represents the get-off vector The length of the human body is used to represent the first The pixel coordinates and the door contour The distance between the pixel coordinates is used to obtain the overlap distance, and the overlap distance is compared with zero. If the overlap distance is equal to zero, the first pixel of the human body contour is obtained. The pixel coordinates and the door contour The pixel coordinates of the door outline and the pixel coordinates of the human body outline match each other. If the overlap distance is greater than zero, the human body outline passes through the door outline. The pixel coordinates and the door contour If the pixel coordinates of the human body contour do not overlap, the other pixel coordinates of the human body contour are matched with the other pixel coordinates of the door contour until the matching is completed. If the overlap distance is greater than zero, it means that the human body contour does not pass through the door contour, that is, the pixel coordinates of the door contour do not match the pixel coordinates of the human body contour, which means that the human body contour is a passenger who does not get off the car or a passenger in the waiting area.

[0045] The specific method of obtaining the disembarkation distance is as follows:

[0046] Since the image captured by the camera is a two-dimensional image, the pixel coordinates also have two-dimensional values. Therefore, the coordinates of the center point of the outline of the person waiting to get off the car are , the center point coordinates of the door contour are , through the Euclidean distance formula and To calculate the distance, the Euclidean distance formula is used to calculate the straight-line distance between two points in a two-dimensional plane to obtain the distance to get off the bus;

[0047] ;

[0048] in, Indicates the distance to get off. Represents the coordinates of the center point of the silhouette of the person waiting to get off the bus axis, Represents the center point coordinates of the door outline axis, Represents the coordinates of the center point of the silhouette of the person waiting to get off the bus axis, Represents the center point coordinates of the door outline axis.

[0049] The specific method of obtaining the outline of the person getting on the bus is as follows:

[0050] While triggering the passenger getting off the vehicle detection time period, the movement direction and movement distance of the human body contour towards the vehicle door contour are analyzed within the passenger getting off the vehicle detection time period to avoid misjudgment of the human body contour that is only close to the vehicle door contour but not getting on the vehicle or the human body contour that moves towards the vehicle door contour but is not close to the vehicle. If the movement direction of the human body contour is towards the vehicle door contour and the movement distance between the human body contour and the vehicle door contour gradually decreases, then this human body contour is marked as a human body contour getting on the vehicle. If the movement direction of the human body contour is not towards the vehicle door contour or the movement distance between the human body contour and the vehicle door contour does not gradually decrease, then this human body contour is not marked as a human body contour getting on the vehicle.

[0051] The specific method for analyzing the movement direction and movement distance of the human body contour toward the door contour is as follows:

[0052] Since it is already within the passenger getting off detection time period, the center coordinates of the human body contour at the next moment and the center coordinates of the human body contour at the previous moment are vector-calculated according to the time series of the passenger getting off detection time period to obtain the moving vector. Since the number of moving vectors is the interval of the time series, the time series of the passenger getting off detection time period is set to , then the number of movement vectors is , the moving vectors are averaged according to the number of moving vectors to obtain the average moving vector, which increases the accuracy of detecting the moving direction of the human body contour. The vector calculation is performed based on the center point coordinates of the human body contour and the pixel point coordinates of the door contour at the current moment of the time series of the passenger getting off the bus detection period. The reason for calculating by the center point coordinates of the human body contour is that the center point coordinates can represent the overall moving direction of the human body contour, and the pixel point coordinates of the door contour are calculated because the area of ​​the door contour is larger than the area of ​​the human body contour. The center point coordinates of the door contour alone cannot accurately represent the moving direction of the human body contour towards the door contour, that is, the moving direction of the center point coordinates of the human body contour is not towards the center point coordinates of the door contour. It can also indicate that the passenger enters the car, but when passing through the door, he does not enter from the center of the door. Therefore, the predicted vector is obtained. The specific method of obtaining the predicted vector is the same as the specific method of obtaining the moving vector. According to The average moving vector is analyzed to see whether it coincides with the predicted vector. If the average moving vector coincides with the predicted vector, it means that the human body contour is moving toward the door contour. If the average moving vector does not coincide with the predicted vector, the coordinates of other pixel points of the door contour are calculated. After all the pixel points of the door contour are involved in the calculation, if the moving vector and the predicted vector still do not coincide, it means that the human body contour is not moving toward the door contour. The distance between the center point coordinates of the human body contour and the center point coordinates of the door contour is calculated to obtain the boarding distance. The specific method for obtaining the boarding distance is the same as the specific method for obtaining the getting off distance. The boarding distance of the next moment is compared with the boarding distance of the previous moment according to the time series. If the boarding distance of the next moment is less than the boarding distance of the previous moment, it means that the human body contour is gradually approaching the door contour. If the boarding distance of the next moment is greater than or equal to the boarding distance of the previous moment, it means that the human body contour is not gradually approaching the door contour.

[0053] The specific method of obtaining the motion vector is as follows:

[0054] ;

[0055] in, represents the motion vector, Indicates the coordinates of the center point of the human body contour at the next moment axis, Represents the coordinates of the center point of the human body contour at the previous moment axis, Indicates the coordinates of the center point of the human body contour at the next moment axis, Represents the coordinates of the center point of the human body contour at the previous moment axis.

[0056] The specific method of analyzing whether the average moving vector coincides with the predicted vector is:

[0057] The average moving vector and the predicted vector are calculated using the angle cosine formula to obtain the cosine value; ;

[0058] in, Represents the cosine value, and is between negative one and one. represents the average moving vector, represents the prediction vector, represents the magnitude of the mean moving vector, represents the modulus of the prediction vector;

[0059] The cosine value is compared with one. If the cosine value is equal to one, it means that the average moving vector coincides with the predicted vector. If the cosine value is not equal to one, it means that the average moving vector does not coincide with the predicted vector.

[0060] The specific method of analyzing the passenger transfer chaos situation based on the passenger transfer chaos parameters is as follows:

[0061] The transfer chaos parameter data set of historical orderly transfers is averaged and calculated. The transfer chaos parameter data set of historical orderly transfers is composed of transfer chaos parameters calculated when all disembarking passengers get off the bus first and then the boarding passengers get on the bus. Then, the transfer chaos threshold is obtained, and the passenger transfer chaos parameter is compared with the transfer chaos threshold. When there is a push and shove between the outlines of the boarding and disembarking people, the distance between the outlines of the boarding and disembarking people will fluctuate by shrinking and expanding. Although the distance between the outlines of the boarding and disembarking people is large in the transfer chaos threshold, the fluctuation is small. Therefore, if the transfer chaos parameter is greater than the transfer chaos threshold, it indicates that the passenger transfer is chaotic. If the transfer chaos parameter is less than or equal to the transfer chaos threshold, it indicates that the passenger transfer is normal.

[0062] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A data collection and user behavior analysis system for urban rail transit, characterized by: Includes the following specific modules: Image acquisition module: starts to collect image data in real time, including pixel coordinates and pixel numbers; Data preprocessing module: preprocess image data; Central computing and processing module: performs comprehensive calculations and standardization on the pre-processed image data to obtain passenger transfer chaos parameters. The module then analyzes the passenger transfer chaos situation based on the passenger transfer chaos parameters. If the analysis indicates that the passenger transfer is chaotic, the prompt module is called and the analysis ends. If the analysis indicates that the passenger transfer is normal, the analysis ends directly. Prompt module: used to issue polite transfer prompts; The specific method of obtaining the passenger transfer chaos parameter is as follows: The image data is calculated by an image contour segmentation algorithm and an image contour tracking algorithm to obtain an image contour. The pixel coordinates within the image contour are averaged according to the number of pixels in the image contour to obtain the center point coordinates. The door contour and the human contour are screened out according to the image contour. The getting-off human contour and the getting-off human contour are screened out according to the pixel coordinates of the door contour, the center point coordinates of the door contour, the pixel coordinates of the human contour and the center point coordinates of the human contour. A chaos detection time is set. Within the chaos detection time, the number of getting-on human contours and the number of getting-off human contours are counted to obtain the number of getting-on human contours and the number of getting-off human contours. The center point coordinates of the getting-on human contours are summed and averaged according to the number of getting-off human contours to obtain the getting-off center coordinates. The center point coordinates of the getting-off human contours are summed and averaged according to the number of getting-off human contours to obtain the getting-off center coordinates. The distance between the getting-off center coordinates and the getting-off center coordinates is calculated to obtain a chaos density. The fluctuation of the chaos density in the time series of the chaos detection time is calculated by a variance method to obtain a passenger transfer chaos parameter.

2. The data collection and user behavior analysis system for urban rail transit according to claim 1 is characterized in that: The specific method for obtaining the outline of the human body getting off the vehicle is as follows: The pixel coordinates of the door contour are matched with the pixel coordinates of the human body contour. If the pixel coordinates of the door contour match the pixel coordinates of the human body contour, it means that the human body contour passes through the door contour. If the pixel coordinates of the door contour do not match the pixel coordinates of the human body contour, it means that the human body contour does not pass through the door contour. A passenger getting off detection time period is set. When the human body contour passes through the door contour, this human body contour is marked as the human body contour to be getting off, and the passenger getting off detection time period is triggered. That is, within the passenger getting off detection time period, the distance between the human body contour to be getting off and the door contour is calculated to obtain the getting off distance. The getting off distance of the next moment is compared with the getting off distance of the previous moment according to the time series. If the getting off distance of the next moment is greater than the getting off distance of the previous moment, the human body contour to be getting off is marked as the getting off human contour. If the getting off distance of the next moment is less than or equal to the getting off distance of the previous moment, the human body contour to be getting off is not marked.

3. The data collection and user behavior analysis system for urban rail transit according to claim 2, characterized in that: The specific method of matching the pixel coordinates of the door contour with the pixel coordinates of the human body contour is as follows: Assume the number of pixels of the door outline is , the number of pixels of a human body contour is , assume that the first The pixel coordinates are , the door profile The pixel coordinates are ,right and Perform vector calculation to get the get-off vector , according to the get-off vector Perform a modulo calculation to obtain the overlap distance, and compare the overlap distance with zero. If the overlap distance is equal to zero, it means that the pixel coordinates of the car door contour match the pixel coordinates of the human body contour. If the overlap distance is greater than zero, continue to match the other pixel coordinates of the human body contour with the other pixel coordinates of the car door contour until the matching is completed. If the overlap distance is greater than zero, it means that the pixel coordinates of the car door contour do not match the pixel coordinates of the human body contour.

4. The data collection and user behavior analysis system for urban rail transit according to claim 2, characterized in that: The specific method for obtaining the get-off distance is as follows: Assume that the coordinates of the center point of the silhouette of the person waiting to get off the bus are , the center point coordinates of the door contour are , through the Euclidean distance formula and Perform distance calculation to obtain the disembarkation distance; ； in, Indicates the distance to get off. Represents the coordinates of the center point of the silhouette of the person waiting to get off the bus axis, Represents the center point coordinates of the door outline axis, Represents the coordinates of the center point of the silhouette of the person waiting to get off the bus axis, Represents the center point coordinates of the door outline axis.

5. The data collection and user behavior analysis system for urban rail transit according to claim 1 is characterized in that: The specific method for obtaining the outline of the person getting on the bus is as follows: While triggering the passenger getting off the vehicle detection time period, the movement direction and movement distance of the human body contour toward the vehicle door contour are analyzed within the passenger getting off the vehicle detection time period. If the movement direction of the human body contour is toward the vehicle door contour and the movement distance between the human body contour and the vehicle door contour gradually decreases, then this human body contour is marked as a boarding human body contour. If the movement direction of the human body contour is not toward the vehicle door contour or the movement distance between the human body contour and the vehicle door contour does not gradually decrease, then this human body contour is not marked as a boarding human body contour.

6. The data collection and user behavior analysis system for urban rail transit according to claim 5 is characterized in that: The specific method of analyzing the moving direction and moving distance of the human body contour toward the door contour is as follows: According to the time series of the passenger getting off detection period, the coordinates of the center point of the human body contour at the next moment and the coordinates of the center point of the human body contour at the previous moment are calculated to obtain the moving vector. Let the time series of the passenger getting off detection period be , then the number of movement vectors is , according to the number of motion vectors, the motion vectors are averaged to obtain the average motion vector. The center point coordinates of the human body contour and the pixel coordinates of the door contour at the current moment of the time series of the passenger getting off detection period are vector-calculated to obtain the predicted vector. The average motion vector and the predicted vector are analyzed to see if they coincide. If the average motion vector coincides with the predicted vector, it means that the human body contour is moving toward the door contour. If the average motion vector and the predicted vector do not coincide, the coordinates of other pixel points of the door contour are calculated. After all the pixel coordinates of the door contour are involved in the calculation, if the motion vector and the predicted vector still do not coincide, it means that the human body contour is not moving toward the door contour. The distance between the center point coordinates of the human body contour and the center point coordinates of the door contour is calculated to obtain the boarding distance. According to the time series, the boarding distance of the next moment is compared with the boarding distance of the previous moment. If the boarding distance of the next moment is less than the boarding distance of the previous moment, it means that the human body contour is gradually approaching the door contour. If the boarding distance of the next moment is greater than or equal to the boarding distance of the previous moment, it means that the human body contour is not gradually approaching the door contour.

7. The data collection and user behavior analysis system for urban rail transit according to claim 6, characterized in that: The specific method of obtaining the motion vector is as follows: ； in, represents the motion vector, Indicates the coordinates of the center point of the human body contour at the next moment axis, Represents the coordinates of the center point of the human body contour at the previous moment axis, Indicates the coordinates of the center point of the human body contour at the next moment axis, Represents the coordinates of the center point of the human body contour at the previous moment axis.

8. The data collection and user behavior analysis system for urban rail transit according to claim 6, characterized in that: The specific method of analyzing whether the average moving vector and the predicted vector coincide with each other is as follows: The average moving vector and the predicted vector are calculated using the angle cosine formula to obtain the cosine value; ; in, Represents the cosine value, and is between negative one and one. represents the average moving vector, represents the prediction vector, represents the magnitude of the mean moving vector, represents the modulus of the prediction vector; The cosine value is compared with one. If the cosine value is equal to one, it means that the average moving vector coincides with the predicted vector. If the cosine value is not equal to one, it means that the average moving vector does not coincide with the predicted vector.

9. The data collection and user behavior analysis system for urban rail transit according to claim 1, characterized in that: The specific method of analyzing the passenger transfer chaos situation according to the passenger transfer chaos parameter is as follows: A transfer chaos threshold is set, and the passenger transfer chaos parameter is compared with the transfer chaos threshold. If the transfer chaos parameter is greater than the transfer chaos threshold, it indicates that the passenger transfer is chaotic. If the transfer chaos parameter is less than or equal to the transfer chaos threshold, it indicates that the passenger transfer is normal.

Images