Railway platform safety inspection method and system

Abstract

The invention relates to the technical field of rail transit operation safety, in particular to a railway platform safety inspection method and system, and the method comprises the steps: collecting a visual image flow, infrared distance matrix data and a train speed parameter when a train pull-in signal is received; according to the visual image flow and the infrared distance matrix data, the passenger position and behavior are fused and analyzed, a dynamic risk field is generated, and three-level warning area distribution is divided; dynamically calculating a warning trigger threshold based on the dynamic risk field and the train speed parameter and based on a preset calculation rule; monitoring a continuous risk state, and analyzing a risk response strength parameter according to the continuous risk state; and comprehensively deciding and executing a dynamic protection strategy by combining the third-level warning area, the warning trigger threshold and the risk response intensity parameter. According to the invention, full-process automation from risk early warning to active protection is realized, the misjudgment rate is greatly reduced, and the response speed is improved to a millisecond level.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of rail transit operation safety, and in particular to a railway platform safety inspection method and system. BACKGROUND

[0002] Current railway platform safety management mainly relies on manual inspection, fixed position monitoring cameras and laser transmission and other static security devices, which has the following defects:

[0003] Passive response lag: traditional video monitoring requires manual interpretation of abnormal behavior (such as passengers approaching the track, children running), with high response delay and inability to adapt to instantaneous risks when trains enter the station;

[0004] Poor environmental adaptability: in rainy, foggy weather or low light conditions at night, visual detection accuracy decreases significantly, and infrared sensors alone cannot identify behavior intent;

[0005] Risk quantification is missing: existing technologies lack fusion analysis of multi-dimensional data such as passenger position, movement direction, and gathering density, and cannot dynamically build a risk classification model;

[0006] Protection strategy is rigid: warning broadcasts and physical interception are used as fixed trigger thresholds, without considering dynamic risk factors such as train speed and sudden behavior (such as passengers falling and rushing towards the track), which can easily lead to false positives or false negatives.

[0007] Therefore, improvements are needed. SUMMARY

[0008] To solve the above problems, the present application provides a railway platform safety inspection method and system.

[0009] The purpose of the present application is achieved by the following technical solutions:

[0010] A railway platform safety inspection method, comprising:

[0011] When a train enters the station signal is received, a visual image stream, an infrared distance matrix data, and a train speed parameter are collected;

[0012] According to the visual image stream and the infrared distance matrix data, passenger position and behavior are fused and analyzed, a dynamic risk field is generated, and a three-level alert zone distribution is divided;

[0013] Based on the dynamic risk field and the train speed parameter, based on a pre-designed calculation rule, a warning trigger threshold is dynamically calculated;

[0014] A continuous risk state is monitored, and a risk response intensity parameter is analyzed according to the continuous risk state;

[0015] The dynamic protection strategy is comprehensively decided and executed in combination with the three-level alert zones, the alert trigger threshold and the risk response intensity parameter.

[0016] In a preferred embodiment, the generation of the dynamic risk field and the division of the three-level alert zone distribution include:

[0017] According to the passenger contour coordinates identified from the visual image stream and the infrared distance matrix data, spatial position registration is performed to generate fusion positioning data;

[0018] According to the passenger motion vector direction in the fusion positioning data, the included angle relationship with the platform track is judged, and the high-risk mark is activated when the vector direction points to the track;

[0019] According to the density distribution of the high-risk mark and the passenger gathering degree, a dynamic risk field intensity gradient map is generated;

[0020] According to the dynamic risk field intensity gradient map, three concentric belt-shaped alert zones are divided along the direction of the platform edge, wherein the area with the highest field intensity gradient is defined as the first-level alert zone, and the second-level alert zone and the third-level alert zone are sequentially attenuated outward.

[0021] In a preferred embodiment, the dynamic calculation of the alert trigger threshold based on the dynamic risk field and the train speed parameter includes:

[0022] According to the real-time field intensity peak value of the dynamic risk field and the train speed parameter, the matching calculation of the alert trigger threshold is performed based on a pre-designed calculation rule;

[0023] The alert trigger threshold includes a regular alert trigger threshold and an enhanced alert trigger threshold;

[0024] The pre-designed calculation rule includes a regular calculation rule and an enhanced calculation rule;

[0025] The regular calculation rule includes:

[0026] When the following conditions are met simultaneously: no real-time field intensity step mutation is detected, and the train speed parameter is less than or equal to a preset high-speed threshold, the adjusted basic threshold weight is adjusted;

[0027] The adjusted basic threshold weight, the real-time field intensity peak value and a preset safety compensation coefficient are weighted and fused to generate the regular alert trigger threshold;

[0028] The enhanced calculation rule includes:

[0029] When any of the following conditions is met: a field intensity step mutation is detected, and the train speed parameter is greater than the preset high-speed threshold, the regular alert trigger threshold is taken as a reference value;

[0030] The preset security compensation coefficient is coupled with the preset emergency response compensation amount to generate an enhanced alert trigger threshold value.

[0031] In a preferred embodiment, the analysis of the risk response intensity parameter includes:

[0032] According to the time series data of the continuous risk state and the historical accident feature library, fusion analysis is performed.

[0033] If the risk duration of the same alert zone exceeds a set value, the risk response intensity parameter is gradually increased.

[0034] If a sudden behavior of a child approaching the track is detected according to the visual image stream, the gradual escalation process is skipped and the highest risk response intensity parameter is directly matched.

[0035] In a preferred embodiment, the combination of the three alert zones, the alert trigger threshold value, and the risk response intensity parameter includes comprehensive decision-making and execution of a dynamic protection strategy, including:

[0036] Based on the infrared distance matrix data and the visual image stream, a platform three-dimensional space model is constructed, the forbidden area is marked by column, seat, and gate coordinates, and the forbidden area coordinates are stored.

[0037] Based on the spatial overlap of the three alert zones and the alert trigger threshold value, the forbidden area, the risk response intensity parameter, and the preset grading rules are combined to perform risk area grading.

[0038] The risk area grading includes an overlapping high-risk area, a non-overlapping medium-risk area, and a low-risk area.

[0039] The preset grading rules include:

[0040] The area where the first alert zone overlaps with the alert trigger threshold value and the risk response intensity parameter is greater than α is the overlapping high-risk area.

[0041] The area where the second alert zone or the alert trigger threshold value is single-covered and the risk response intensity parameter is ∈ [β, α] is the non-overlapping medium-risk area.

[0042] The area where the third alert zone and the risk response intensity parameter are less than β is the low-risk area.

[0043] Wherein, α and β are preset intensity thresholds, and α > β.

[0044] In a preferred embodiment, the combination of the three alert zones, the alert trigger threshold value, and the risk response intensity parameter includes comprehensive decision-making and execution of a dynamic protection strategy, further including:

[0045] Match the corresponding dynamic protection strategy from the preset strategy library based on the risk area classification:

[0046] When the passenger is in the overlapping high-risk area, deploy the physical interception device, and generate a straight-line evacuation path with an included angle greater than a first preset threshold with the track direction;

[0047] When the passenger is in the non-overlapping medium-risk area, start the directional sound and light warning, and generate the shortest evacuation path around the forbidden area with the candidate safe evacuation area as the target point;

[0048] When the passenger is in the low-risk area, enable the broadcast reminder, and mark the low-risk area as a candidate safe evacuation;

[0049] When the passenger moving direction has an included angle greater than a second preset threshold with the track direction, and the passenger is at a distance greater than a third preset threshold from the platform edge and for a preset time, the dynamic protection strategy is released.

[0050] The second object of the application is achieved by the following technical solutions:

[0051] A railway platform safety inspection system, comprising:

[0052] The first module: when receiving a train arrival signal, collect visual image stream, infrared distance matrix data, and train speed parameters;

[0053] The second module: based on the visual image stream and the infrared distance matrix data, fuse and analyze passenger position and behavior, generate a dynamic risk field, and divide three-level alert area distribution;

[0054] The third module: based on the dynamic risk field and the train speed parameters, dynamically calculate the alert trigger threshold based on the pre-designed calculation rule;

[0055] The fourth module: monitor the continuous risk state, and analyze the risk response intensity parameter according to the continuous risk state;

[0056] The fifth module: combine the three-level alert area, the alert trigger threshold, and the risk response intensity parameter, comprehensively decide and execute the dynamic protection strategy.

[0057] The third object of the application is achieved by the following technical solutions:

[0058] A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-mentioned railway platform safety inspection method.

[0059] The fourth object of the application is achieved by the following technical solutions:

[0060] A computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the railway platform safety inspection method.

[0061] In summary, the present application includes at least one of the following beneficial technical effects:

[0062] The method improves the efficiency of platform safety protection through multi-source data fusion and dynamic risk assessment. When the train enters the station, the system synchronously collects visual image stream, infrared distance matrix and train speed (S10), analyzes the passenger position and behavior using visual and infrared data fusion, generates a dynamic risk field and divides three concentric belt-shaped warning zones (S20); based on the risk field intensity gradient and train speed, dynamically calculates the warning trigger threshold (S30), combines the risk duration and sudden behavior (such as children approaching the track) to intelligently adjust the response strength (S40); finally, the warning zone level, trigger threshold and response strength parameters are integrated to execute the hierarchical protection strategy (S50). Effect: realize the whole process automation from risk warning to active protection, greatly reduce the misjudgment rate, and the response speed is improved to milliseconds. BRIEF DESCRIPTION OF DRAWINGS

[0063] Figure 1 is an implementation flowchart of an embodiment of a railway platform safety inspection method of the present application;

[0064] Figure 2 is an implementation flowchart of step S20 in an embodiment of a railway platform safety inspection method of the present application;

[0065] Figure 3 is an implementation flowchart of step S50 in an embodiment of a railway platform safety inspection method of the present application;

[0066] Figure 4 is a principle block diagram of a computer device of the present application. DETAILED DESCRIPTION

[0067] The following will be described in detail below Figures 1-4 The present application will be further described in detail.

[0068] In an embodiment, as shown in Figure 1 The present application discloses a railway platform safety inspection method, specifically including the following steps:

[0069] S10: When receiving the train entering the station signal, collect visual image stream, infrared distance matrix data, and train speed parameters;

[0070] S20: According to the visual image stream and the infrared distance matrix data, the passenger position and behavior are analyzed by fusion, a dynamic risk field is generated, and a three-level alert area distribution is divided;

[0071] S30: Based on the dynamic risk field and the train speed parameter, the alert trigger threshold is dynamically calculated based on the pre-designed calculation rule;

[0072] S40: The continuous risk state is monitored, and the risk response intensity parameter is analyzed according to the continuous risk state;

[0073] S50: The three-level alert area, the alert trigger threshold and the risk response intensity parameter are combined to comprehensively decide and execute the dynamic protection strategy.

[0074] In this embodiment, the method improves the platform safety protection efficiency through multi-source data fusion and dynamic risk assessment. When the train enters the station, the system synchronously collects the visual image stream, the infrared distance matrix and the train speed (S10), analyzes the passenger position and behavior by fusing the visual and infrared data, generates a dynamic risk field and divides a three-level concentric belt-shaped alert area (S20); the alert trigger threshold is dynamically calculated based on the risk field intensity gradient and the train speed (S30), the response intensity is intelligently adjusted in combination with the risk duration and the sudden behavior (such as children approaching the track) (S40); finally, the hierarchical protection strategy is executed in combination with the alert area level, the trigger threshold and the response intensity parameter (S50). Effect: realize the whole process automation from risk warning to active protection, greatly reduce the misjudgment rate, and the response speed is improved to milliseconds.

[0075] As shown in Figure 2 S20, comprising:

[0076] S201: According to the passenger contour coordinates recognized by the visual image stream and the infrared distance matrix data, spatial position registration is performed to generate fusion positioning data;

[0077] S202: According to the passenger motion vector direction in the fusion positioning data, the included angle relationship between the vector direction and the platform track is judged, and the high-risk mark is activated when the vector direction points to the track;

[0078] S203: According to the density distribution of the high-risk mark and the passenger aggregation degree, a dynamic risk field intensity gradient map is generated;

[0079] S204: According to the dynamic risk field intensity gradient map, a three-level concentric belt-shaped alert area is divided along the direction of the platform edge, wherein the area with the highest field intensity gradient is defined as a first-level alert area, and the second-level alert area and the third-level alert area are attenuated outward in turn.

[0080] In the embodiment, the fusion positioning data is generated by spatial registration of visual contour coordinates and infrared distance matrix (S201), the high-risk label is activated in combination with the angle relationship between passenger motion vector and platform track (S202); the dynamic risk field strength gradient graph is constructed based on label density and aggregation, and the three-level concentric belt-shaped warning areas are divided along the platform edge (the highest field strength area is the first level, and the attenuation is the second and third levels) (S203-S204). Effects: accurately positioning potential dangerous sources, improving the identification accuracy of abnormal behaviors pointing to the track, and focusing resources on high-risk areas through three-level zoning.

[0081] S30, comprising:

[0082] S301: based on the real-time peak value of the dynamic risk field and the train speed parameter, a matching calculation warning trigger threshold is performed based on a pre-designed calculation rule;

[0083] S302: the warning trigger threshold includes a regular warning trigger threshold and an enhanced warning trigger threshold;

[0084] S303: the pre-designed calculation rule includes a regular calculation rule and an enhanced calculation rule;

[0085] The regular calculation rule includes:

[0086] S304: when the following conditions are met simultaneously: no real-time field strength step mutation is detected, and the train speed parameter is less than or equal to a preset high-speed threshold, the basic threshold weight is adjusted;

[0087] S305: the adjusted basic threshold weight, the real-time field strength peak value, and a preset safety compensation coefficient are weighted and fused to generate a regular warning trigger threshold;

[0088] The enhanced calculation rule includes:

[0089] S306: when any of the following conditions is met: a field strength step mutation is detected, and the train speed parameter is greater than a preset high-speed threshold, the regular warning trigger threshold is taken as a reference value;

[0090] S307: a preset safety compensation coefficient and a preset emergency response compensation amount are coupled and calculated, and then combined with the reference value to generate an enhanced warning trigger threshold.

[0091] In the embodiment, a double-layer trigger threshold is matched and calculated according to the real-time peak value of the risk field and the train speed: the regular threshold is generated by weighting the basic weight, the field strength peak value, and the safety compensation coefficient (S304-S305); when the field strength is suddenly changed or the train is overspeed, the enhanced threshold is generated based on the regular threshold and the emergency compensation amount (S306-S307). Effects: the threshold is dynamically adjusted according to the risk, the triggering sensitivity is greatly improved in the overspeed or sudden risk scenario, and false negatives are avoided.

[0092] S40, comprising:

[0093] S401: performing fusion analysis according to time series data of the continuous risk state and a historical accident feature library;

[0094] S402: if the risk duration of the same warning area exceeds a set value, gradually increasing the risk response intensity parameter;

[0095] S403: if a sudden behavior of a child approaching a track is detected according to the visual image stream, skipping the gradual escalation process and directly matching the highest risk response intensity parameter.

[0096] In the embodiment, the continuous risk state is fused and analyzed with the historical accident feature library: when the risk of the same warning area exceeds the time, the response intensity is gradually increased (S402); when a sudden behavior of a child approaching a track is detected, the highest response intensity is directly enabled (S403). Effect: historical data-driven response optimization, reduced response delay of children in special scenarios, and gradual escalation of system resources.

[0097] S50, comprising:

[0098] S501: constructing a platform three-dimensional space model based on the infrared distance matrix data and the visual image stream, marking the forbidden area formed by the column, the seat, and the gate coordinate, and storing the forbidden area coordinate;

[0099] S502: based on the spatial overlap degree of the three-level warning area and the warning trigger threshold, combining the forbidden area, the risk response intensity parameter, and presetting a grading rule, performing risk area grading;

[0100] S503: the risk area grading includes an overlapping high-risk area, a non-overlapping medium-risk area, and a low-risk area;

[0101] The preset grading rule includes:

[0102] S504: the area where the first-level warning area overlaps the warning trigger threshold and the risk response intensity parameter is greater than a is the overlapping high-risk area;

[0103] S505: the area where the second-level warning area or the warning trigger threshold is single-covered and the risk response intensity parameter is in [β, a] is the non-overlapping medium-risk area;

[0104] S506: the area where the third-level warning area and the risk response intensity parameter are less than β is the low-risk area;

[0105] S507: where a and β are preset intensity thresholds, and a > β.

[0106] In this embodiment, a three-dimensional model of the platform is constructed based on infrared and visual data, and prohibited areas such as pillars / seats are marked (S501). Based on the spatial overlap between the warning area and the trigger threshold, the location of the prohibited area, and the response intensity parameters, the risk area is divided into: overlapping high-risk area (Level 1 area + threshold overlap + response intensity > α); non-overlapping medium-risk area (Level 2 / threshold single coverage + response intensity ∈ [β, α]); and low-risk area (Level 3 area + response intensity < β) (S502-S507).

[0107] like Figure 3 As shown, S50 also includes:

[0108] SB1: Based on the risk area classification, match the corresponding dynamic protection strategy from the preset strategy library:

[0109] SB2: When passengers are in overlapping high-risk areas, deploy physical interception equipment and generate a straight retreat path with an angle greater than a first preset threshold to the track direction;

[0110] SB3: When a passenger is in a non-overlapping medium-risk area, a directional audio-visual warning is activated, and the shortest escape path is generated by bypassing the restricted area with the candidate safe retreat zone as the target point;

[0111] SB4: When a passenger is in a low-risk area, activate a broadcast reminder and mark the low-risk area as a candidate for safe retreat;

[0112] SB5: When the angle between the passenger's movement direction and the track direction is greater than the second preset threshold and the distance between the passenger and the platform edge is greater than the third preset threshold for a preset time, the dynamic protection strategy is deactivated.

[0113] In this embodiment, a dynamic protection strategy is matched according to risk level: Overlapping high-risk areas: deploy physical interception equipment to generate a straight retreat path with an angle greater than the first preset threshold to the track direction (SB2); Non-overlapping medium-risk areas: activate directional audio-visual warnings and generate the shortest retreat path by detouring through the restricted area (SB3); Low-risk areas: broadcast reminders and mark as safe retreat candidate areas (SB4); When the passenger's movement direction has an angle greater than the second preset threshold to the track direction and the distance between the passenger and the platform edge is greater than the third preset threshold for a preset time, the protection is lifted (SB5).

[0114] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0115] In an embodiment, a railway platform safety inspection system is provided, which corresponds to the railway platform safety inspection method in the above-mentioned embodiment. The railway platform safety inspection system comprises:

[0116] The first module: when receiving a train arrival signal, a visual image stream, infrared distance matrix data, and train speed parameters are collected;

[0117] The second module: passenger positions and behaviors are fused and analyzed according to the visual image stream and the infrared distance matrix data, a dynamic risk field is generated, and a three-level alert zone distribution is divided;

[0118] The third module: based on the dynamic risk field and the train speed parameters, an alert trigger threshold is dynamically calculated based on a pre-designed calculation rule;

[0119] The fourth module: a continuous risk state is monitored, and a risk response intensity parameter is analyzed according to the continuous risk state;

[0120] The fifth module: a dynamic protection strategy is comprehensively decided and executed in combination with the three-level alert zone, the alert trigger threshold, and the risk response intensity parameter.

[0121] Optionally, it further comprises:

[0122] The sixth module: spatial position registration is performed according to passenger contour coordinates recognized from the visual image stream and the infrared distance matrix data, and fusion positioning data are generated;

[0123] The seventh module: according to a passenger motion vector direction in the fusion positioning data, an angle relationship between the vector direction and a platform track is judged, and a high-risk mark is activated when the vector direction points to the track;

[0124] The eighth module: a dynamic risk field intensity gradient map is generated according to a density distribution of the high-risk mark and a passenger aggregation degree;

[0125] The ninth module: according to the dynamic risk field intensity gradient map, a three-level concentric belt-shaped alert zone is divided along a platform edge direction, wherein a highest field intensity gradient area is defined as a first-level alert zone, and successively attenuates to a second-level alert zone and a third-level alert zone.

[0126] Optionally, it further comprises:

[0127] The tenth module: according to a real-time field intensity peak value of the dynamic risk field and the train speed parameters, a matching calculation alert trigger threshold is executed based on a pre-designed calculation rule;

[0128] The eleventh module: the alert trigger threshold comprises a regular alert trigger threshold and an enhanced alert trigger threshold;

[0129] Twelve modules: the pre-design calculation rules include a regular calculation rule and an enhanced calculation rule;

[0130] The regular calculation rule includes:

[0131] Thirteen modules: when the following conditions are met simultaneously: no real-time field strength step mutation is detected, and the train speed parameter is less than or equal to a preset high-speed threshold, the basic threshold weight is adjusted;

[0132] Fourteen modules: the adjusted basic threshold weight, the real-time field strength peak value and a preset safety compensation coefficient are weighted and fused to generate a regular alert triggering threshold;

[0133] The enhanced calculation rule includes:

[0134] Fifteen modules: when any of the following conditions is met: a field strength step mutation is detected, and the train speed parameter is greater than a preset high-speed threshold, the regular alert triggering threshold is taken as a reference value;

[0135] Sixteen modules: a preset safety compensation coefficient and a preset emergency response compensation amount are coupled and calculated, and then combined with the reference value to generate an enhanced alert triggering threshold.

[0136] Optionally, the method further includes:

[0137] Seventeen modules: fusion analysis is performed according to time series data of the sustained risk state and a historical accident feature library;

[0138] Eighteen modules: if the risk duration of the same alert zone exceeds a set value, the risk response intensity parameter is gradually increased;

[0139] Nineteen modules: if a sudden behavior of a child approaching a track is detected according to the visual image stream, the gradual escalation process is skipped, and the highest risk response intensity parameter is directly matched.

[0140] Optionally, the method further includes:

[0141] Twenty modules: a station three-dimensional space model is constructed based on the infrared distance matrix data and the visual image stream, a forbidden area is marked by labeling column bodies, seats and gate coordinates, and the forbidden area coordinates are stored;

[0142] Twenty-one modules: based on the spatial overlap degree of the three-level alert zones and the alert triggering thresholds, the forbidden area and the risk response intensity parameter are combined, a preset grading rule is executed, and a risk area grading is performed;

[0143] Twenty-two modules: the risk area grading includes an overlapping high-risk area, a non-overlapping medium-risk area and a low-risk area;

[0144] The preset grading rule includes:

[0145] The region where the first alert zone overlaps with the alert trigger threshold and the risk response intensity parameter > a is an overlapping high-risk region;

[0146] The region where the second alert zone or the alert trigger threshold is singly covered and the risk response intensity parameter ∈ [β, a] is a non-overlapping medium-risk region;

[0147] The region where the third alert zone and the risk response intensity parameter < β is a low-risk region;

[0148] The sixth module II, wherein a and β are preset intensity thresholds, and a > β.

[0149] Optionally, further comprising:

[0150] The seventh module II, based on the risk region classification, matching a corresponding dynamic protection strategy from a preset strategy library:

[0151] The eighth module II, when the passenger is in the overlapping high-risk region, deploying a physical interception device, and generating a straight-line retreat path with an included angle greater than a first preset threshold with the track direction;

[0152] The ninth module II, when the passenger is in the non-overlapping medium-risk region, starting a directional sound and light warning, and generating a shortest retreat path around the forbidden area with the candidate safe retreat area as a target point;

[0153] The thirtieth module II, when the passenger is in the low-risk region, enabling a broadcast reminder, and marking the low-risk region as a candidate safe retreat;

[0154] The thirty-first module II, when the passenger moving direction has an included angle greater than a second preset threshold with the track direction, and the passenger is at a distance greater than a third preset threshold from the platform edge and lasts for a preset time, releasing the dynamic protection strategy.

[0155] The specific limitations of the railway platform safety inspection system can be referred to the limitations of the railway platform safety inspection method in the above, which will not be repeated here. Each module in the above railway platform safety inspection system can be realized by software, hardware and their combinations in whole or in part. The above modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so that the processor calls and executes the operations corresponding to each module.

[0156] In one embodiment, a computer device, which can be a server, is provided, and its internal structure diagram can be as shown in Figure 4The computer device includes a processor, a memory, a network interface and a database connected through a system bus. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The database of the computer device is configured to store the alert trigger threshold. The network interface of the computer device is configured to communicate with an external terminal through a network connection. The computer program is configured to be executed by the processor to implement the railway platform safety inspection method.

[0157] In one embodiment, a computer device is provided, which includes a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor is configured to execute the computer program to implement the railway platform safety inspection method.

[0158] In one embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program is configured to be executed by a processor to implement the railway platform safety inspection method.

[0159] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by a computer program instructing related hardware, and the computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. Any reference to memory, storage, database or other medium used in the embodiments provided by the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM) and memory bus dynamic RAM (RDRAM).

Claims

1. A method for safety inspection of railway platforms, characterized in that, include: When a train arrival signal is received, visual image stream, infrared distance matrix data, and train speed parameters are collected. Based on the visual image stream and the infrared distance matrix data, passenger location and behavior are fused and analyzed to generate a dynamic risk field and divide the distribution of three-level warning zones. Based on the dynamic risk field and the train speed parameters, and based on preset calculation rules, the warning trigger threshold is dynamically calculated. Monitor the ongoing risk status and analyze the risk response intensity parameters based on the ongoing risk status; By combining the three-level warning zones, the warning trigger thresholds, and the risk response intensity parameters, a comprehensive decision is made and a dynamic protection strategy is implemented.

2. The railway platform safety inspection method according to claim 1, characterized in that, The generation of a dynamic risk field and the division of three-level warning zones include: Based on the passenger contour coordinates identified from the visual image stream and the infrared distance matrix data, spatial position registration is performed to generate fused positioning data. Based on the direction of passenger movement vector in the fused positioning data, determine the angle between the vector and the platform track. When the vector direction points to the track, activate a high-risk marker. A dynamic risk field intensity gradient map is generated based on the density distribution of the high-risk markers and the passenger aggregation degree. Based on the dynamic risk field strength gradient map, three concentric strip-shaped warning zones are divided along the platform edge direction. The region with the highest field strength gradient is defined as the first-level warning zone, and the second-level and third-level warning zones are successively reduced outward.

3. The railway platform safety inspection method according to claim 2, characterized in that, The dynamic calculation of the warning trigger threshold based on the dynamic risk field and the train speed parameters includes: Based on the real-time peak field strength of the dynamic risk field and the train speed parameters, a matching calculation of the warning trigger threshold is performed according to a preset calculation rule. The alert trigger thresholds include regular alert trigger thresholds and enhanced alert trigger thresholds; The preset calculation rules include regular calculation rules and enhanced calculation rules; The general calculation rules are as follows: When the following conditions are met simultaneously: no real-time field strength step change is detected and the train speed parameter is ≤ preset high-speed threshold, the basic threshold weight is adjusted. The adjusted base threshold weights, real-time field strength peak values, and preset safety compensation coefficients are weighted and fused to generate a regular warning trigger threshold. The enhanced calculation rules: When any of the following conditions are met: a sudden change in field strength is detected, or the train speed parameter is greater than the preset high-speed threshold, the conventional warning trigger threshold is used as the reference value. The preset safety compensation coefficient and the preset emergency response compensation amount are coupled and calculated, and then combined with the benchmark value to generate the enhanced alert trigger threshold.

4. A railway platform safety inspection method according to claim 2, characterized in that, The parameters for analyzing the intensity of risk response include: A fusion analysis is performed based on the time-series data of the ongoing risk status and the historical accident feature database. If the duration of risk in the same warning zone exceeds a set value, the risk response intensity parameter will be increased step by step. If a child's sudden behavior of approaching the track is detected based on the visual image stream, the progressive escalation process is skipped, and the highest risk response intensity parameter is directly matched.

5. A railway platform safety inspection method according to claim 4, characterized in that, The process of combining the three-level alert zones, the alert trigger threshold, and the risk response intensity parameters to comprehensively decide and execute a dynamic protection strategy includes: Based on the infrared distance matrix data and the visual image stream, a three-dimensional spatial model of the platform is constructed, and the coordinates of the columns, seats, and turnstiles are marked to form a restricted area, and the coordinates of the restricted area are stored. Based on the spatial overlap between the three-level warning zones and the warning trigger threshold, and combined with the restricted areas and the risk response intensity parameters, a pre-defined classification rule is used to classify risk areas. The risk area classification includes overlapping high-risk areas, non-overlapping medium-risk areas, and low-risk areas. The preset grading rules include: The area where the Level 1 alert zone overlaps with the alert trigger threshold and where the risk response intensity parameter is greater than α is an overlapping high-risk area. The area covered by the Level 2 alert zone or the alert trigger threshold and the area where the risk response intensity parameter ∈ [β, α] is a non-overlapping medium-risk area; The area within the Level 3 alert zone and where the risk response intensity parameter < β is a low-risk area; Where α and β are preset intensity thresholds, and α > β.

6. A railway platform safety inspection method according to claim 5, characterized in that, The method of combining the three-level alert zones, the alert trigger threshold, and the risk response intensity parameter to make a comprehensive decision and execute a dynamic protection strategy also includes: Based on the risk area classification, a corresponding dynamic protection strategy is matched from the preset strategy library: When passengers are in overlapping high-risk areas, physical interception equipment is deployed and a straight retreat path is generated with an angle greater than a first preset threshold to the track direction. When a passenger is in a non-overlapping medium-risk area, a directional audio-visual warning is activated, and the shortest escape path is generated by bypassing the restricted area with the candidate safe retreat zone as the target point. When passengers are in a low-risk area, broadcast reminders are activated, and the low-risk area is marked as a candidate for safe retreat. The dynamic protection strategy is deactivated when the angle between the passenger's direction of movement and the track direction is greater than the second preset threshold and the distance between the passenger and the edge of the platform is greater than the third preset threshold for a preset time.

7. A railway platform safety inspection system, characterized in that, include: Module 1: When a train arrival signal is received, collect visual image stream, infrared distance matrix data, and train speed parameters; The second module: Based on the visual image stream and the infrared distance matrix data, the module fuses and analyzes the passenger's location and behavior, generates a dynamic risk field, and divides the distribution of three-level warning zones. The third module: Based on the dynamic risk field and the train speed parameters, and based on the preset calculation rules, dynamically calculate the warning trigger threshold; Module 4: Monitor the ongoing risk status and analyze the risk response intensity parameters based on the ongoing risk status; The fifth module combines the three-level warning zones, the warning trigger thresholds, and the risk response intensity parameters to make comprehensive decisions and execute dynamic protection strategies.

8. A computer device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of a railway platform safety inspection method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of a railway platform safety inspection method as described in any one of claims 1 to 6.

Images