Intelligent management system for urban service operation platform

Abstract

The invention relates to the technical field of urban intelligent evacuation, in particular to an intelligent management system for an urban service operation platform. The system comprises a path planning module, a congestion condition monitoring module, an escalator monitoring module, an evacuation capability evaluation module and a path correction module, and the path correction module obtains space attitude data of a monitored target by constructing an escalator inclined coordinate system and calculates an optimal path score in combination with real-time congestion distribution of an evacuation channel. And carrying out self-adaptive correction on the initial path based on the weighting model. By identifying the abnormal behavior of the escalator area and combining the starting condition of the standby evacuation channel, the evacuation capacity score of the whole evacuation channel is obtained, and dynamic path scheduling and parameter adjustment in high-density crowds are achieved. According to the method, an evacuation strategy optimization mechanism of perception, decision formation and feedback closed loop is constructed, so that the safety evacuation efficiency and the risk control capability under the urban complex scene are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of urban intelligent evacuation, and in particular to an intelligent management system for an urban service operation platform. Background Art

[0002] With the widespread use of urban underground spaces, ensuring evacuation safety has become a crucial component of urban service operations. Research has shown that in emergencies such as fires, using escalators in conjunction with stairs can shorten evacuation time by approximately 23% to 33% compared to relying solely on stairs. Furthermore, installing fire barriers at the bottom of escalators can improve evacuation efficiency by 6% to 7%. However, current research on escalators as emergency evacuation routes in most scenarios is limited to evacuation time simulations. There is a lack of real-time monitoring mechanisms and dynamic evaluation systems for the risks of escalator ridership, and the impact of unexpected behaviors such as climbing and falling on evacuation safety is not fully considered.

[0003] On the other hand, mainstream path planning strategies are mostly based on static shortest paths or simulated paths. Although some studies have attempted to use simulation models, artificial intelligence, or reinforcement learning methods for real-time path optimization and dynamically adjust evacuation strategies based on personnel distribution, most of them lack closed-loop designs integrated with visual recognition technology, and are unable to achieve real-time linkage between congestion, risk, and the opening and closing status of backup channels, nor can they adaptively calibrate path weights based on on-site feedback.

[0004] In addition, the current sanitation operation systems and urban operation platforms rely heavily on manual inspections and static deployment, failing to integrate intelligent visual monitoring with multi-path collaborative scheduling, resulting in insufficient evacuation efficiency and difficulty in early identification of safety risks.

[0005] Therefore, there is an urgent need for a comprehensive underground space evacuation system that integrates real-time visual behavior recognition, escalator risk judgment, dynamic congestion assessment, coordinated activation of backup channels and adaptive adjustment of path scores.

[0006] Chinese patent document CN107241841A discloses a mobile application system for urban lighting operation management. Relying on the operating platform of the urban operation management system, the functions of the handheld terminal include: information query function, control strategy setting function, alarm push function, GIS map positioning function, work order dispatch function, equipment management function, etc.

[0007] This invention completes the operation and management of the urban lighting system through data communication between handheld terminals and central servers. Although it has certain real-time and remote control capabilities, it still has the following defects: the system relies on manual operation and preset strategies, and lacks the ability to perceive and respond to complex dynamic factors in the operation scenario; the lighting strategy is mainly based on time control, light control and longitude and latitude control, with limited intelligence and insufficient multi-source data fusion and scheduling capabilities; work order management mainly relies on personnel's subjective feedback and operating procedures, and cannot achieve intelligent linkage of fault diagnosis, task dispatch and path planning. It can be seen that the existing intelligent management platform is difficult to meet the comprehensive decision-making needs of multiple scenarios, multiple factors and multiple goals in urban service management. Summary of the Invention

[0008] To this end, the present invention provides an intelligent management system for an urban service operation platform to overcome the problem in the prior art of low resource allocation efficiency caused by a single source of path scheduling and unreasonable path selection in the absence of feedback adjustment.

[0009] To achieve the above objectives, the present invention provides an intelligent management system for an urban service operation platform, comprising:

[0010] Path planning module, used to determine the initial evacuation path based on the distribution status of monitored targets and the layout of evacuation channels;

[0011] a congestion monitoring module, connected to the path planning module, for monitoring the target distribution density and distribution position of any evacuation exit in the initial evacuation path to analyze the congestion situation of each evacuation exit;

[0012] An escalator monitoring module, connected to the congestion monitoring module, is used to identify abnormal behavior of the monitoring target through key posture point detection, calculate the safety risk index of the escalator based on the abnormal behavior, and determine the feasibility of the current evacuation;

[0013] An evacuation capacity assessment module, connected to the escalator monitoring module, is used to determine whether to activate a backup evacuation channel based on the current flow state of the monitored target, analyze the opening and closing status of the backup evacuation channel, and obtain the evacuation capacity of all evacuation channels based on the congestion level and the number of backup evacuation channels when the backup evacuation channel is activated;

[0014] A path correction module is connected to the path planning module, the escalator monitoring module and the evacuation capacity evaluation module respectively, and selects the optimal evacuation path based on the optimal weighted calculation result and the initial evacuation path and broadcasts it.

[0015] Furthermore, the congestion monitoring module includes an underground area congestion calculation unit and an underground area congestion analysis unit, wherein:

[0016] The underground area congestion calculation unit is used to monitor the real-time congestion of each upward evacuation exit in the underground area;

[0017] The underground area congestion analysis unit is used to compare the upward standard congestion threshold with the upward real-time congestion, and determine the congestion situation of the upward evacuation exit based on the comparison result.

[0018] Furthermore, the escalator monitoring module includes an escalator state perception unit, an escalator behavior recognition unit and a risk prediction unit, wherein:

[0019] The escalator state sensing unit is used to obtain the escalator usage density when the escalator is in operation;

[0020] The elevator behavior recognition unit is used to identify the category of dangerous behaviors and the frequency of dangerous behaviors during the elevator ride;

[0021] The risk prediction unit is used to determine the category to which the current evacuation state of the escalator belongs based on the identification results of the escalator usage density and the dangerous behavior of escalators and calculate the escalator safety risk value.

[0022] Furthermore, the elevator behavior recognition unit includes a key point construction subunit and a key point detection subunit, wherein:

[0023] The key point construction subunit is used to establish a three-dimensional coordinate system and extract the key points of the joints of the monitoring target;

[0024] The key point detection subunit identifies the category of dangerous behavior of the monitored target according to the joint key points, including climbing behavior identification and falling behavior identification.

[0025] Furthermore, the risk prediction unit includes an escalator safety risk value calculation subunit and an analysis subunit, wherein:

[0026] The escalator safety risk value calculation subunit calculates the escalator safety risk value based on the detected dangerous behavior category and frequency and escalator usage;

[0027] The analysis subunit compares the calculated escalator safety risk value with a preset escalator safety risk value threshold, and determines the category to which the current evacuation state of the escalator belongs based on the comparison result.

[0028] Furthermore, climbing behavior recognition includes,

[0029] Calculate the real-time relative height between the monitored target's hand and the top of the escalator;

[0030] Compare the real-time relative altitude with the standard relative altitude range;

[0031] If the real-time relative height is not within the standard relative height range and the duration exceeds the preset duration of the dangerous behavior, it is determined to be a climbing behavior.

[0032] Furthermore, fall behavior recognition includes:

[0033] Calculate the angle between the spine line and the horizontal plane to obtain the real-time angle;

[0034] Compare the real-time angle with the standard angle.

[0035] If the real-time angle is smaller than the standard angle, obtain the center of gravity descent speed of several consecutive frames and compare the center of gravity descent speed with the descent speed threshold.

[0036] If the center of gravity descending speed is greater than the descending speed threshold, it is determined to be a fall behavior.

[0037] Furthermore, the evacuation capacity assessment module includes a ground area congestion analysis unit and a full evacuation channel score calculation unit, wherein:

[0038] The ground area congestion analysis unit is used to continuously monitor the real-time ground congestion of each ground evacuation exit in the ground area, obtain the real-time ground congestion and compare it with the ground standard congestion threshold, and determine whether to activate the backup evacuation channel and the number of backup evacuation channels corresponding to each main evacuation channel based on the comparison result;

[0039] The full evacuation channel score calculation unit is used to obtain the real-time up-going congestion of each main evacuation channel, the real-time ground congestion of the escalator, and the corresponding number of backup evacuation channels when the backup evacuation channel is activated, and calculate the evacuation capacity corresponding to the comprehensive congestion score of the full evacuation channel.

[0040] Furthermore, the path correction module includes a quantization comparison unit and a path correction unit, wherein:

[0041] The quantitative comparison unit is used to obtain the ranking result of the comprehensive congestion score of all evacuation channels;

[0042] The full evacuation passage is a collection of the main evacuation passage and its backup evacuation passage formed by the one-to-one correspondence of the upward evacuation exit, the escalator, and the ground evacuation exit;

[0043] The path correction unit is used to determine the initial evacuation path based on the priority ranking result of the comprehensive congestion scores of all evacuation channels.

[0044] Furthermore, the path correction unit includes a first correction unit and a second correction unit, wherein:

[0045] The first correction unit is configured to determine, when the backup evacuation channel is not enabled, a comprehensive congestion score of all evacuation channels based on the number of the second backup evacuation channels, so as to select and announce the optimal evacuation path based on a priority ranking result of the comprehensive congestion scores of all evacuation channels;

[0046] The second correction unit is used to determine the comprehensive congestion score of all evacuation channels based on the number of first backup evacuation channels when the backup evacuation channels are activated, so as to select and broadcast the optimal evacuation path based on the priority ranking result of the comprehensive congestion score of all evacuation channels.

[0047] Compared with the existing technology, the beneficial effect of the present invention is that the path planning module performs an initial path score calculation based on the Euclidean distance from the monitoring target to each evacuation exit and the evacuation capacity score, ensuring that the primary selected path is both short and feasible. The evacuation capacity score is obtained by integrating multiple parameters such as crowding density and frequency of dangerous behaviors, avoiding the path selection bias caused by using distance as a single indicator; using visual recognition and coordinate mapping technology, the two-dimensional coordinates of the monitoring target in the image space are accurately converted into actual site coordinates, improving the accuracy of path planning; the congestion monitoring module is linked with the escalator monitoring module to achieve real-time identification of the congestion status and risk level of any channel; the evacuation capacity assessment module further makes decisions on whether to enable the backup channel and assesses the overall load capacity of the system's global evacuation channel; the path correction module selects the optimal path in real time based on the dynamic weighted results and broadcasts it in multiple ways, effectively improving the system's adaptive evacuation capability in emergency situations and enhancing the intelligent response capability of the urban service operation platform in complex pedestrian flow scenarios.

[0048] Furthermore, the congestion calculation unit accurately identifies the number of monitored targets in each frame of the image and matches it with the area of ​​the visible area to obtain the crowd density at the evacuation exit, and then determines whether it exceeds the set threshold; the analysis unit combines the density and threshold results to determine in real time whether there is a risk of blockage in the channel, providing an objective basis for subsequent path adjustments, and effectively improving the rationality and safety of channel use.

[0049] Furthermore, by combining escalator density and the frequency of dangerous behaviors, an escalator safety risk value is generated. The risk prediction unit integrates multiple data sources and significantly improves the accuracy of abnormal situation responses through model fusion. It can effectively determine whether the current escalator is available, avoid evacuation in potential high-risk channels, and ensure the safety of the crowd.

[0050] Furthermore, by analyzing spatial angle changes and speed sequences, the trend of rapid changes in the center of gravity can be accurately captured; this method can filter out misjudgments and trigger recognition only when both the angle and speed are abnormal, effectively improving the real-time and reliability of escalator accident warnings. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1This is a schematic diagram of the structure of an intelligent management system for a city service operation platform according to an embodiment of the present invention;

[0052] Figure 2 This is a connection diagram of a congestion monitoring module according to an embodiment of the present invention;

[0053] Figure 3 This is a logical decision diagram for climbing behavior recognition according to an embodiment of the present invention;

[0054] Figure 4 This is a logical decision diagram for fall behavior recognition according to an embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention.

[0056] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0057] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside", and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0058] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified or limited, the terms "installed," "connected," and "connected" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to internal connections between two components. For technical monitoring purposes in this field, the specific meanings of the above terms in the present invention can be understood based on specific circumstances.

[0059] See also Figure 1 As shown, it is a structural diagram of an intelligent management system for a city service operation platform according to an embodiment of the present invention. The present invention provides an intelligent management system for a city service operation platform, including:

[0060] Path planning module, used to determine the initial evacuation path based on the distribution status of monitored targets and the layout of evacuation channels;

[0061] a congestion monitoring module, connected to the path planning module, for monitoring the target distribution density and distribution position of any evacuation exit in the initial evacuation path to analyze the congestion situation of each evacuation exit;

[0062] An escalator monitoring module, connected to the congestion monitoring module, is used to identify abnormal behavior of the monitoring target through key posture point detection, calculate the safety risk index of the escalator based on the abnormal behavior, and determine the feasibility of the current evacuation;

[0063] An evacuation capacity assessment module, connected to the escalator monitoring module, is used to determine whether to activate a backup evacuation channel based on the current flow state of the monitored target, analyze the opening and closing status of the backup evacuation channel, and obtain the evacuation capacity of all evacuation channels based on the congestion level and the number of backup evacuation channels when the backup evacuation channel is activated;

[0064] A path correction module is connected to the path planning module, the escalator monitoring module and the evacuation capacity evaluation module respectively, and selects the optimal evacuation path based on the optimal weighted calculation result and the initial evacuation path and broadcasts it.

[0065] In this embodiment, the path planning module is based on the Euclidean distance D from the target group to each evacuation exit. i The current evacuation capacity C of the evacuation exit i , calculate the path score of each evacuation channel when the backup channel is not opened according to the following formula:

[0066]

[0067] Among them, α is the weight coefficient of distance;

[0068] β is the weight coefficient of evacuation capacity;

[0069] α+β=1;

[0070] The initial settings are α = 0.9, β = 0.1 and the evacuation exit with the highest score is selected as the initial evacuation path;

[0071] Among them, the spatial position of the monitoring target is obtained based on the visual recognition algorithm and the Euclidean distance from it to each evacuation exit is calculated;

[0072] A deep learning-based target detection network is used to identify the two-dimensional pixel coordinates of the monitoring target in the image frame, and the coordinates are mapped to the physical coordinates (x, y) of the actual site based on the camera installation angle.

[0073] Using the known position of the evacuation exit in the coordinate system (x i ,y i ), calculate the Euclidean distance from the target to the i-th evacuation exit

[0074] Visual density detection devices are deployed at each evacuation exit on the ground and underground. The dangerous behavior category, frequency, and congestion score fed back by the behavior recognition module are integrated to calculate the evacuation capacity score to determine the evacuation capacity of the current evacuation channel;

[0075] Evacuation capability score C i The larger the value, the stronger the evacuation capacity and the smoother the evacuation passage;

[0076] When planning a route, distance is prioritized in route selection, and evacuation capacity parameters are set based on historical evacuation channel data.

[0077] The path planning module calculates an initial path score based on the Euclidean distance from the monitoring target to each evacuation exit and the evacuation capacity score, ensuring that the preselected path is both short and feasible. The evacuation capacity score is obtained by integrating multiple parameters such as crowding density and the frequency of dangerous behaviors, avoiding path selection bias caused by using distance as a single indicator. Visual recognition and coordinate mapping technology are used to accurately convert the two-dimensional coordinates of the monitoring target in image space into actual site coordinates, improving the accuracy of path planning. The congestion monitoring module is linked with the escalator monitoring module to achieve real-time identification of the congestion status and risk level of any channel. The evacuation capacity assessment module further makes decisions on whether to activate backup channels and evaluates the overall load capacity of the system's global evacuation channels. The path correction module selects the optimal path in real time based on dynamic weighting results and broadcasts it in multiple ways, effectively improving the system's adaptive evacuation capabilities in emergency situations and enhancing the intelligent response capabilities of the urban service operation platform in complex pedestrian flow scenarios.

[0078] See Figure 2 , which is a connection diagram of a congestion monitoring module according to an embodiment of the present invention;

[0079] Specifically, the congestion monitoring module includes an underground area congestion calculation unit and an underground area congestion analysis unit, wherein:

[0080] The underground area congestion calculation unit is used to monitor the real-time congestion of each upward evacuation exit in the underground area;

[0081] The underground area congestion analysis unit is used to compare the upward standard congestion threshold with the upward real-time congestion, and determine the congestion situation of the upward evacuation exit based on the comparison result.

[0082] In this embodiment, an intelligent visual acquisition camera with a set angle is installed above each up-bound evacuation exit. The camera is connected to the computing subunit of the congestion monitoring module and continuously obtains video data of the evacuation exit coverage area;

[0083] The underground area congestion calculation unit uses a visual algorithm that combines target detection and behavior recognition to identify and number monitoring targets in the image, and performs continuous tracking in the image sequence to eliminate duplicate counting;

[0084] Assume that the detection area covered by the visual range of the upward evacuation exit is A D , unit is m 2 , the number of human bodies recognized by the system in the video frames acquired per second is N D , then the uplink real-time congestion ρ D Calculated by the following formula,

[0085]

[0086] Among them, N D is the deduplication count of all detected targets in the current frame, A D It is the actual area obtained after conversion of the physical distance measurement within the field of view;

[0087] The underground area congestion analysis unit is used to compare the upward standard congestion threshold with the upward real-time congestion, and the upward standard congestion threshold is taken as 4 people / m 2 ;

[0088] If the real-time congestion level of the upward movement is greater than the standard congestion level threshold of the upward movement, the escalator will be monitored;

[0089] If the uplink real-time congestion is less than or equal to the uplink standard congestion threshold, evacuation is performed along the initial evacuation path.

[0090] The congestion calculation unit accurately identifies the number of monitored targets in each frame of the image and matches it with the area of ​​the visible area to obtain the crowd density at the evacuation exit, and then determines whether it exceeds the set threshold. The analysis unit combines this density with the threshold result to determine in real time whether there is a risk of blockage in the channel, providing an objective basis for subsequent path adjustments, and effectively improving the rationality and safety of channel use.

[0091] Specifically, the escalator monitoring module includes an escalator state perception unit, an escalator behavior recognition unit and a risk prediction unit, wherein:

[0092] The escalator state sensing unit is used to obtain the escalator usage density when the escalator is in operation;

[0093] The elevator behavior recognition unit is used to identify the category of dangerous behaviors and the frequency of dangerous behaviors during the elevator ride;

[0094] The risk prediction unit is used to determine the category to which the current evacuation state of the escalator belongs based on the identification results of the escalator usage density and the dangerous behavior of escalators and calculate the escalator safety risk value.

[0095] In this embodiment, the escalator state sensing unit uses a binocular stereo camera device installed above and on the side of the escalator, combined with an infrared counting module, to obtain the total number of escalator usage targets N in each time period in real time. E , and at the same time obtain the effective area A of all manned platform sections of the escalator E , unit is m 2 , and then the escalator usage density ρ is calculated by the following formula E ,

[0096]

[0097] The escalator behavior recognition unit is based on the visual behavior recognition cameras configured at the upper and lower ends of the escalator. It collects dynamic posture information of escalator passengers, calls the OpenPose skeleton recognition network to extract key nodes, constructs a continuous time series coordinate sequence throughout the escalator process, and obtains the frequency f of dangerous behaviors per unit time. risk ;

[0098] The risk prediction unit is responsible for summarizing the escalator usage density μ and the frequency of dangerous behaviors f risk , use the following formula to calculate the escalator safety risk value R,

[0099]

[0100] Among them, ω1 is the weight of escalator usage density in the escalator safety risk value;

[0101] ω2 is the weight of the frequency of dangerous behavior in the escalator safety risk value;

[0102] μ0 is the standard escalator usage density threshold, which represents the maximum number of passengers that can be safely carried per unit area. In this embodiment, the value is 3.5 people / m 2 ;

[0103] f norm The threshold value for the frequency of implementation of standard dangerous behaviors refers to the number of dangerous behaviors allowed per person per minute. In this embodiment, the value is 0.002.

[0104] By combining escalator density and the frequency of dangerous behaviors, an escalator safety risk value is generated. The risk prediction unit integrates multiple data sources and significantly improves the accuracy of abnormal situation responses through model fusion. It can effectively determine whether the current escalator is available, avoid evacuation arrangements in potential high-risk channels, and ensure the safety of the crowd.

[0105] Specifically, the elevator behavior recognition unit includes a key point construction subunit and a key point detection subunit, wherein:

[0106] The key point construction subunit is used to establish a three-dimensional coordinate system and extract the key points of the joints of the monitoring target;

[0107] The key point detection subunit identifies the category of dangerous behavior of the monitored target according to the joint key points, including climbing behavior identification and falling behavior identification.

[0108] In this embodiment, a three-dimensional coordinate system (X, Y, Z) is defined with the escalator movement direction as the X axis, the escalator vertical ascending direction as the Y axis, and the passenger's side direction as the Z axis.

[0109] Extract the skeleton key points of the monitoring target:

[0110] Call the deep neural network model to extract the three-dimensional coordinates of the key nodes for each ladder target detected in the image:

[0111] Neck center point P1, left shoulder center point P2, right shoulder center point P3, spine center point P4, left hand center point P5, right hand center point P6, pelvis center point P7;

[0112] Build skeleton connection relationships based on key nodes and establish a skeleton model;

[0113] The key point detection subunit performs temporal analysis on the above skeleton key points and determines whether there is any dangerous behavior based on the spatial position changes of consecutive frames.

[0114] By defining a three-axis coordinate system, the coordinates of key nodes in the elevator riding process are extracted in real time, and the behavior category is judged based on the skeleton connection relationship. While enhancing the accuracy of identifying abnormal elevator riding behavior, the stability and robustness of the recognition algorithm are maintained in complex action environments, providing refined support for dangerous behavior detection.

[0115] Specifically, the risk prediction unit includes an escalator safety risk value calculation subunit and an analysis subunit, wherein:

[0116] The escalator safety risk value calculation subunit calculates the escalator safety risk value based on the detected dangerous behavior category and frequency and escalator usage;

[0117] The analysis subunit compares the calculated escalator safety risk value with a preset escalator safety risk value threshold, and determines the category to which the current evacuation state of the escalator belongs based on the comparison result.

[0118] See Figure 3 As shown, it is a logical decision diagram for climbing behavior recognition according to an embodiment of the present invention;

[0119] Specifically, climbing behavior recognition includes:

[0120] Calculate the real-time relative height between the monitored target's hand and the top of the escalator;

[0121] Compare the real-time relative altitude with the standard relative altitude range;

[0122] If the real-time relative height is not within the standard relative height range and the duration exceeds the preset duration of the dangerous behavior, it is determined to be a climbing behavior.

[0123] In this embodiment, the key point construction subunit identifies the left hand center point P5 and the right hand center point P6 of the current passenger through the depth image, and obtains their three-dimensional spatial coordinates (X L , Y L , Z L )、(X R , Y R , Z R ));

[0124] At the same time, the top edge point P of the escalator handrail corresponding to the monitoring target in the same horizontal direction of the current evacuation channel is obtained by escalator parameters and image recognition. top Height value Y top ;

[0125] Calculate the center point P of the elevator rider's hand in the current frame LR Real-time relative height H to the top of the escalator real , specifically:

[0126]

[0127] Among them, the hand center point P LR The midpoint of the line connecting the left-hand center point P5 and the right-hand center point P6;

[0128] Set the standard relative height range H for safety control std , in this embodiment, it is set to -[-0.2, 0.2], the unit is m;

[0129] Compare the real-time relative altitude with the standard relative altitude range.

[0130] If the real-time relative altitude is less than or equal to the minimum value of the standard relative altitude interval, it is determined to be a climbing tendency, and the duration is further obtained;

[0131] If the real-time relative altitude is within the standard relative altitude range, it is determined to be non-abnormal behavior;

[0132] If the real-time relative height is greater than or equal to the maximum value of the standard relative height interval, it is determined to be a climbing tendency, and the duration is further obtained to determine whether it is a climbing behavior among dangerous behaviors;

[0133] Set the preset duration of dangerous behavior to 5 seconds, and compare the duration with the preset duration of dangerous behavior.

[0134] If the duration is longer than the preset duration of dangerous behavior, it is determined to be a climbing behavior among dangerous behaviors;

[0135] If the duration is less than or equal to the preset duration of the dangerous behavior, it is determined to be non-abnormal behavior.

[0136] By calculating the relative height between the passenger's hand and the top edge of the escalator in real time and combining it with duration analysis, a climbing behavior judgment model is constructed. Using only spatial distance and time logic as the basis for judgment, it can effectively identify dangerous actions and distinguish them from ordinary standing behaviors, providing data support for behavior monitoring and risk warning in escalator areas.

[0137] See Figure 4 As shown, it is a logical decision diagram for fall behavior recognition according to an embodiment of the present invention;

[0138] Specifically, fall behavior recognition includes:

[0139] Calculate the angle between the spine line and the horizontal plane to obtain the real-time angle;

[0140] Compare the real-time angle with the standard angle.

[0141] If the real-time angle is smaller than the standard angle, obtain the center of gravity descent speed of several consecutive frames and compare the center of gravity descent speed with the descent speed threshold.

[0142] If the center of gravity descending speed is greater than the descending speed threshold, it is determined to be a fall behavior.

[0143] In this embodiment, the neck center point P1 and the pelvis center point P7 are obtained, and the spine line vector is formed based on these two points.

[0144]

[0145] Taking the horizontal direction as the reference plane, the real-time angle is calculated by vector projection, that is, the angle θ between the spine line and the horizontal plane. The calculation formula is:

[0146]

[0147] in, is the upward vertical unit vector (0, 1, 0);

[0148] Set the standard angle to θ std =40°, compare the real-time angle with the standard angle,

[0149] If the real-time angle is greater than the standard angle, it is determined to be non-abnormal behavior;

[0150] If the real-time angle is less than or equal to the standard angle, it is determined to be a falling tendency, and the center of gravity falling speed of several consecutive frames is further obtained to determine whether it is a falling behavior in a dangerous behavior;

[0151] The set of the center of gravity of the monitored target in the Y-axis direction [y1, y2, ..., y 150 ];

[0152] The speed sequence is calculated by the inter-frame difference method. The calculation formula is:

[0153]

[0154] 149 sets of inter-frame center of gravity points were calculated;

[0155] The center of gravity is the midpoint of the line connecting the neck center point P1 and the pelvic center point P7;

[0156] Get the center of gravity descent speed of several consecutive frames, set the descent speed threshold to 0.6m / s, and compare the center of gravity descent speed with the descent speed threshold.

[0157] If the center of gravity descending speed is greater than the descending speed threshold, and the center of gravity descending speed is greater than the normal level, it is determined to be a fall behavior among dangerous behaviors;

[0158] If the center of gravity descent speed is less than or equal to the descent speed threshold, the center of gravity descent speed is at a normal level and is not an abnormal behavior.

[0159] By analyzing spatial angle changes and speed sequences, the trend of rapid changes in the center of gravity can be accurately captured; this method can filter out misjudgments and trigger recognition only when both the angle and speed are abnormal, effectively improving the real-time and reliability of escalator accident warnings.

[0160] Specifically, the evacuation capacity evaluation module includes a ground area congestion analysis unit and a full evacuation channel score calculation unit, wherein:

[0161] The ground area congestion analysis unit is used to continuously monitor the real-time ground congestion of each ground evacuation exit in the ground area, obtain the real-time ground congestion and compare it with the ground standard congestion threshold, and determine whether to activate the backup evacuation channel and the number of backup evacuation channels corresponding to each main evacuation channel based on the comparison result;

[0162] The full evacuation channel score calculation unit is used to obtain the real-time upward congestion of each main evacuation channel, the escalator safety risk value and the corresponding number of backup evacuation channels when the backup evacuation channel is activated, and calculate the evacuation capacity corresponding to the comprehensive congestion score of the full evacuation channel.

[0163] In this embodiment, the detection area covered by the field of view of the ground evacuation exit is assumed to be A.U , unit is m 2 , the number of human bodies recognized by the system in the video frames acquired per second is N U , then the ground real-time congestion ρ U It can be calculated by the following formula:

[0164]

[0165] Among them, N U is the deduplication count of all detected targets in the current frame, A U It is the actual area obtained after conversion of the physical distance measurement within the field of view;

[0166] Take the ground standard congestion threshold ρ U，std 3.5 people / m 2 ,The ground area congestion analysis unit compares the ground standard congestion threshold with the ground real-time congestion,

[0167] If the real-time ground congestion is greater than the standard ground congestion threshold, the current main evacuation channel is unable to evacuate the current number of people, and the backup channel needs to be activated;

[0168] If the real-time ground congestion is less than or equal to the ground standard congestion threshold, the current main evacuation channel has the capacity to evacuate the current number of people, and the backup channel will not be activated;

[0169] When the backup channel is enabled, the number of backup channels at the time of activation is set as the number of the first backup evacuation channel. The number of backup evacuation channels n required to be activated for the current main evacuation channel is determined by the following relationship: b ,

[0170]

[0171] Among them, δ represents the crowd density that can be alleviated by each backup evacuation channel, and the system is preset to 0.8;

[0172] Indicates rounding up operation to ensure that at least one backup channel is opened as long as the threshold is exceeded;

[0173] For example, when the ρ of a main channel corresponding to the ground evacuation exit is U =4.6, substituting into the equation,

[0174]

[0175] At this time, immediately link the two paired backup evacuation channels and b =2 Input is passed to the full evacuation channel score calculation unit;

[0176] The calculation formula for the comprehensive score of congestion in all evacuation channels is:

[0177]

[0178] C i,adjusted =100*T

[0179] Among them, ω3, ω4, and ω5 are the preset weights corresponding to the three indicators of real-time up-train congestion, escalator safety risk value, and the number of corresponding backup evacuation channels;

[0180] ρ D,max The maximum tolerable value set for ground congestion historical statistics;

[0181] R ,max The maximum tolerable value set for the historical statistics of escalator safety risk value;

[0182] n max The maximum number of backup evacuation channels that can be opened corresponding to the main evacuation channel;

[0183] C i,adjusted The evacuation capacity corresponding to the comprehensive congestion score of all evacuation channels;

[0184] Take ω3=0.3, ω4=0.4, ω5=0.3;

[0185] The threshold for the comprehensive score comparison of all evacuation channel congestion is set to 0.5;

[0186] Compare the comprehensive congestion score of all evacuation channels with the comprehensive congestion score comparison threshold of all evacuation channels.

[0187] If the comprehensive congestion score of all evacuation channels is greater than the comprehensive congestion score comparison threshold of all evacuation channels, the evacuation capacity of the initial path meets the evacuation requirements, and the division method of the initial path is not adjusted. The initial evacuation path is the optimal evacuation path;

[0188] If the comprehensive congestion score of all evacuation channels is less than or equal to the comprehensive congestion score comparison threshold, the evacuation capacity of the initial path cannot meet the evacuation requirements. After the evacuation, the weight coefficient of the evacuation capacity is dynamically adjusted to update the initial evacuation path to the optimal evacuation path.

[0189] For example, when ρ D =3.2,ρ D,max =5.0, R=40, R ,max =100,n b =2,n max =4, C i,adjusted=0.29, which is less than the comprehensive congestion score comparison threshold of 0.5 for all evacuation channels. At this time, the weight coefficient of the evacuation capacity needs to be dynamically adjusted to meet the evacuation requirements of the monitoring target. After the evacuation, the weight coefficient of the evacuation capacity is dynamically adjusted with a fixed upward value of 10% of the original weight coefficient of the evacuation capacity.

[0190] A dynamic evacuation capacity adjustment mechanism is established, integrating the real-time congestion density of the ground area, the escalator risk value and the number of activated backup channels to form a comprehensive score for congestion of all evacuation channels, accurately assessing the current path carrying capacity; when the score is lower than the threshold, the weight is automatically increased, and the path scoring model is optimized to make the system evacuation decision more in line with the actual evacuation scenario on site.

[0191] Specifically, the path correction module includes a quantization comparison unit and a path correction unit, wherein:

[0192] The quantitative comparison unit is used to obtain the ranking result of the comprehensive congestion score of all evacuation channels;

[0193] The full evacuation passage is a collection of the main evacuation passage and its backup evacuation passage formed by the one-to-one correspondence of the upward evacuation exit, the escalator, and the ground evacuation exit;

[0194] The path correction unit is used to determine the initial evacuation path based on the priority ranking result of the comprehensive congestion scores of all evacuation channels.

[0195] In this embodiment, the evacuation path corresponding to the highest priority comprehensive evacuation channel congestion score is selected as the optimal evacuation path. The comprehensive congestion scores of the evacuation channels from the monitoring target to each channel are obtained, and the scores are sorted from high to low to form a priority list. The priority is proportional to the sequence number in the priority list.

[0196] The higher the priority, the stronger the evacuation capacity.

[0197] By establishing a sorting mechanism based on a comprehensive congestion score, the priority of each evacuation channel in its current state is clarified. The sorting results are used as the basis for path correction, effectively avoiding the risk of initial path failure; in emergency scenarios, evacuation strategies can be updated in real time to improve response efficiency and channel utilization.

[0198] Specifically, the path correction unit includes a first correction unit and a second correction unit, wherein:

[0199] The first correction unit is configured to determine, when the backup evacuation channel is not enabled, a comprehensive congestion score of all evacuation channels based on the number of the second backup evacuation channels, so as to select and announce the optimal evacuation path based on a priority ranking result of the comprehensive congestion scores of all evacuation channels;

[0200] The second correction unit is used to determine the comprehensive congestion score of all evacuation channels based on the number of first backup evacuation channels when the backup evacuation channels are activated, so as to select and broadcast the optimal evacuation path based on the priority ranking result of the comprehensive congestion score of all evacuation channels.

[0201] In this embodiment, when the backup evacuation channel is not enabled, n b If the value is 0, the number of backup channels when enabled is the number of the second backup evacuation channels;

[0202] Dynamic guide signs display evacuation routes in real time; LED or LCD screens installed at passage entrances and major turning points display evacuation routes and direction arrows simultaneously;

[0203] The optimal evacuation route is announced through the voice broadcast system deployed in the evacuation area via top or wall speakers.

[0204] By building a response mechanism for path correction, dynamic route selection is performed based on the scoring and ranking of all evacuation channels under different conditions, and multi-channel broadcasting is carried out through guiding signs, display screens and voice systems; full coverage and real-time notification of evacuation path information are achieved, enhancing the system's guidance capabilities in high-density crowd areas.

[0205] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

[0206] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Various modifications and variations are possible for the technical monitoring objectives of the art. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. An intelligent management system for an urban service operation platform, characterized in that: include, Path planning module, used to determine the initial evacuation path based on the distribution status of monitored targets and the layout of evacuation channels; a congestion monitoring module, connected to the path planning module, for monitoring the target distribution density and distribution position of any evacuation exit in the initial evacuation path to analyze the congestion situation of each evacuation exit; An escalator monitoring module, connected to the congestion monitoring module, is used to identify abnormal behavior of the monitoring target through key posture point detection, calculate the safety risk index of the escalator based on the abnormal behavior, and determine the feasibility of the current evacuation; An evacuation capacity assessment module, connected to the escalator monitoring module, is used to determine whether to activate a backup evacuation channel based on the current flow state of the monitored target, analyze the opening and closing status of the backup evacuation channel, and obtain the evacuation capacity of all evacuation channels based on the congestion level and the number of backup evacuation channels when the backup evacuation channel is activated; A path correction module is connected to the path planning module, the escalator monitoring module and the evacuation capacity evaluation module respectively, and selects the optimal evacuation path based on the optimal weighted calculation result and the initial evacuation path and broadcasts it.

2. The intelligent management system for the urban service operation platform according to claim 1 is characterized in that: The congestion monitoring module includes an underground area congestion calculation unit and an underground area congestion analysis unit, wherein: The underground area congestion calculation unit is used to monitor the real-time congestion of each upward evacuation exit in the underground area; The underground area congestion analysis unit is used to compare the upward standard congestion threshold with the upward real-time congestion, and determine the congestion situation of the upward evacuation exit based on the comparison result.

3. The intelligent management system for the urban service operation platform according to claim 1 is characterized in that: The escalator monitoring module includes an escalator state perception unit, an escalator behavior recognition unit and a risk prediction unit, wherein: The escalator state sensing unit is used to obtain the escalator usage density when the escalator is in operation; The elevator behavior recognition unit is used to identify the category of dangerous behaviors and the frequency of dangerous behaviors during the elevator ride; The risk prediction unit is used to determine the category to which the current evacuation state of the escalator belongs based on the identification results of the escalator usage density and the dangerous behavior of escalators and calculate the escalator safety risk value.

4. The intelligent management system for the urban service operation platform according to claim 3 is characterized in that: The elevator behavior recognition unit includes a key point construction subunit and a key point detection subunit, wherein: The key point construction subunit is used to establish a three-dimensional coordinate system and extract the key points of the joints of the monitoring target; The key point detection subunit identifies the category of dangerous behavior of the monitored target according to the joint key points, including climbing behavior identification and falling behavior identification.

5. The intelligent management system for the urban service operation platform according to claim 3 is characterized in that: The risk prediction unit includes an escalator safety risk value calculation subunit and an analysis subunit, wherein: The escalator safety risk value calculation subunit calculates the escalator safety risk value based on the detected dangerous behavior category and frequency and escalator usage; The analysis subunit compares the calculated escalator safety risk value with a preset escalator safety risk value threshold, and determines the category to which the current evacuation state of the escalator belongs based on the comparison result.

6. The intelligent management system for the urban service operation platform according to claim 4 is characterized in that: Climbing behavior recognition includes: Calculate the real-time relative height between the monitored target's hand and the top of the escalator; Compare the real-time relative altitude with the standard relative altitude range; If the real-time relative height is not within the standard relative height range and the duration exceeds the preset duration of the dangerous behavior, it is determined to be a climbing behavior.

7. The intelligent management system for the urban service operation platform according to claim 4 is characterized in that: Fall behavior recognition includes: Calculate the angle between the spine line and the horizontal plane to obtain the real-time angle; Compare the real-time angle with the standard angle. If the real-time angle is smaller than the standard angle, obtain the center of gravity descent speed of several consecutive frames and compare the center of gravity descent speed with the descent speed threshold. If the center of gravity descending speed is greater than the descending speed threshold, it is determined to be a fall behavior.

8. The intelligent management system for the urban service operation platform according to claim 1 is characterized in that: The evacuation capacity evaluation module includes a ground area congestion analysis unit and a full evacuation channel score calculation unit, wherein: The ground area congestion analysis unit is used to continuously monitor the real-time ground congestion of each ground evacuation exit in the ground area, obtain the real-time ground congestion and compare it with the ground standard congestion threshold, and determine whether to activate the backup evacuation channel and the number of backup evacuation channels corresponding to each main evacuation channel based on the comparison result; The full evacuation channel score calculation unit is used to obtain the real-time up-going congestion of each main evacuation channel, the real-time ground congestion of the escalator, and the corresponding number of backup evacuation channels when the backup evacuation channel is activated, and calculate the evacuation capacity corresponding to the comprehensive congestion score of the full evacuation channel.

9. The intelligent management system for the urban service operation platform according to claim 8 is characterized in that: The path correction module includes a quantization comparison unit and a path correction unit, wherein: The quantitative comparison unit is used to obtain the ranking result of the comprehensive congestion score of all evacuation channels; The full evacuation passage is a collection of the main evacuation passage and its backup evacuation passage formed by the one-to-one correspondence of the upward evacuation exit, the escalator, and the ground evacuation exit; The path correction unit is used to determine the initial evacuation path based on the priority ranking result of the comprehensive congestion scores of all evacuation channels.

10. The intelligent management system for the urban service operation platform according to claim 9 is characterized in that: The path correction unit includes a first correction unit and a second correction unit, wherein: The first correction unit is configured to determine, when the backup evacuation channel is not enabled, a comprehensive congestion score of all evacuation channels based on the number of the second backup evacuation channels, so as to select and announce the optimal evacuation path based on a priority ranking result of the comprehensive congestion scores of all evacuation channels; The second correction unit is used to determine the comprehensive congestion score of all evacuation channels based on the number of first backup evacuation channels when the backup evacuation channels are activated, so as to select and broadcast the optimal evacuation path based on the priority ranking result of the comprehensive congestion score of all evacuation channels.

Images