Monitoring intelligent analysis system and method based on machine vision

Abstract

The invention relates to the technical field of monitoring data analysis, in particular to a monitoring intelligent analysis system and method based on machine vision, and the method comprises the steps: analyzing the people flow distribution dangerous states of straight-through entrances and exits corresponding to elevators in all function subareas and the vehicle entry risk states of the positions where all the straight-through entrances and exits are located; constructing a human-vehicle safety linkage risk assessment model, importing a human flow distribution dangerous state analysis result and a vehicle driving-in risk state analysis result into the human-vehicle safety linkage risk assessment model, and assessing human-vehicle safety linkage risks of all straight-through entrances and exits; a safe parking route planning model is constructed according to the human-vehicle safety linkage risk assessment result of each straight-through entrance and exit, and a safe parking route is planned for a new entering vehicle; therefore, the passing efficiency of the underground parking lot of the shopping mall is optimized, and the safety of people and vehicle passing in the underground parking lot is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of monitoring data analysis, and in particular to a monitoring intelligent analysis system and method based on machine vision. Background Art

[0002] With the rapid development of urban commercial complexes, underground parking lots, as key areas of human-vehicle interaction, face complex environments and dynamic risks in their safety management. Traditional monitoring systems rely on limited data sources and static rules, making it difficult to cope with the dim lighting, blind spots, and unexpected behaviors found in underground parking lots. The introduction of machine vision technology offers a new direction for precise monitoring in the complex environments of commercial complexes: by integrating low-light imaging, behavior recognition, and spatial semantic alignment, machine vision technology enables real-time capture and analysis of vehicle trajectories, pedestrian movements, and environmental characteristics. At the same time, the unique functional zoning of shopping malls leads to significant differences in customer behavior in different areas, especially the active behavior of children, which can pose safety risks. These zones are directly connected to specific entrances and exits of underground parking lots via vertical transportation, making the control of risks associated with human-vehicle safety interactions more complex.

[0003] However, existing technologies are deficient in analyzing the risk transmission characteristics of mall functional zoning, failing to accurately model the behavioral dynamics of children. Furthermore, existing technologies overlook the risk linkage mechanisms between vertical transportation and parking entrances and exits. Existing mall surveillance data analysis technologies fail to fully consider the specific behavioral characteristics of high-risk functional zones, such as infant and toddler areas and youth activity areas, resulting in delayed risk perception when responding to sudden behaviors like children running or stopping. Furthermore, existing technologies process mall and parking data in isolation, failing to establish a dynamic connection from functional zones to elevators and then to entrances and exits through pedestrian path tracking. Consequently, they are unable to predict the potential impact of surges in passenger flow within specific functional zones on underground parking lot safety. The limitations of existing technologies, which rely on static rules and linear weighted assessments, make it difficult to accurately quantify the risk coupling effects in complex scenarios, such as vehicle blind spots and the combined effects of children's active behavior. This ultimately results in insufficient comprehensiveness and timeliness in risk assessments, which in turn impacts the diversion of new vehicles into the mall's underground parking lot.

[0004] In order to solve these problems, this application designs a monitoring intelligent analysis system and method based on machine vision. Summary of the Invention

[0005] The purpose of the present invention is to provide a monitoring intelligent analysis system and method based on machine vision. By constructing a risk assessment model for the linkage of human and vehicle safety, a comprehensive analysis of the dangerous state of human flow distribution and the risk state of vehicle entry is conducted; and then a safe parking route planning model is constructed to plan safe parking routes for newly entered vehicles. This optimizes the traffic efficiency of the shopping mall's underground parking lot and improves the safety of human and vehicle traffic in the underground parking lot.

[0006] The present invention is achieved in that:

[0007] In a first aspect, the present invention provides a monitoring intelligent analysis method based on machine vision, comprising the following steps:

[0008] S1. Obtain vehicle dynamic perception data in the mall's underground parking lot and regional vertical traffic data of elevators in each functional area of ​​the mall through the mall's internal monitoring system; at the same time, obtain spatial topological data of the mall's underground parking lot;

[0009] S2. Each entrance and exit of the underground parking lot that the elevators in each functional zone pass through is used as the direct entrance and exit corresponding to the elevators in each functional zone; the regional vertical traffic data of the elevators in each functional zone of the shopping mall is imported into the pedestrian distribution hazard state analysis model, and the pedestrian distribution hazard state of the direct entrance and exit corresponding to the elevators in each functional zone is analyzed;

[0010] S3. Importing vehicle dynamic perception data and spatial topology data of the mall's underground parking lot into a vehicle entry risk status analysis model to analyze the vehicle entry risk status at each through-entry entrance and exit;

[0011] S4. Construct a pedestrian-vehicle safety linkage risk assessment model. The analysis results obtained from the pedestrian distribution risk state analysis model and the vehicle entry risk state analysis model are imported into the pedestrian-vehicle safety linkage risk assessment model to assess the pedestrian-vehicle safety linkage risk at each through-entry entrance and exit.

[0012] S5. Based on the risk assessment results of the pedestrian-vehicle safety linkage at each through entrance and exit, a safe parking route planning model is constructed to plan safe parking routes for new vehicles entering the site.

[0013] Preferably, in step S2, the dangerous state of the flow of people at the direct entrances and exits corresponding to the elevators in each functional area is analyzed, specifically including:

[0014] S21. Extracting regional vertical traffic data of elevators in each functional area of ​​the shopping mall and the corresponding direct entrances and exits;

[0015] S22. Construct a crowd distribution hazard state analysis model, import the regional vertical traffic data of the elevators in each functional area of ​​the shopping mall into the crowd distribution hazard state analysis model, analyze the crowd distribution hazard state of the direct entrances and exits corresponding to the elevators in each functional area, and obtain the crowd distribution hazard state analysis results of the direct entrances and exits corresponding to the elevators in each functional area.

[0016] Preferably, the process of constructing the crowd flow distribution dangerous state analysis model in step S22 specifically includes:

[0017] S221. Calculate the regional passenger density risk index for elevators in each functional zone based on the regional vertical traffic data for elevators in each functional zone of the shopping mall;

[0018] The calculation formula for the regional crowd density risk index is:

[0019]

[0020] Where Qr is the regional crowd density risk index of the direct entrance and exit corresponding to the elevator in the current functional zone at the current moment, Fp is the risk correction coefficient of the current functional zone at the current moment, nc is the number of stranded children in the direct entrance and exit area corresponding to the elevator in the current functional zone in the regional vertical transportation data at the current moment, Ae is the total number of stranded people in the direct entrance and exit area corresponding to the elevator in the current functional zone in the regional vertical transportation data at the current moment, st is the standard deviation of the historical length of children's detention at the direct entrance and exit corresponding to the elevator in the current functional zone in the regional vertical transportation data, and Ts is the mean length of children's detention at the direct entrance and exit corresponding to the elevator in the current functional zone in the regional vertical transportation data.

[0021] The calculation formula for the risk correction coefficient of the current functional zoning is:

[0022]

[0023] Where A is the area of ​​the current functional zone, Ea is the number of acceleration mutations per unit area of ​​the current functional zone at the current moment in the regional vertical traffic data, n is the number of children in the current functional zone at the current moment in the regional vertical traffic data, Tavg is the average cumulative length of stay of customers in the current functional zone at the current moment in the regional vertical traffic data, vs is the safe moving speed threshold, σ(v) 2 is the variance of the moving speed of all customers in the current functional zone at the current moment in the regional vertical traffic data;

[0024] S222. Calculate the behavioral activity risk index based on the regional vertical traffic data of elevators within each functional area of ​​the shopping mall;

[0025] The calculation formula for the behavioral activity risk index is:

[0026]

[0027] Where Ba is the behavioral activity risk index of the direct entrance and exit corresponding to the elevator in the current functional zone at the current moment, nc is the number of stranded children in the direct entrance and exit area corresponding to the elevator in the current functional zone in the regional vertical traffic data at the current moment, vi is the real-time moving speed of the i-th stranded child in the regional vertical traffic data, ai is the real-time acceleration of the i-th stranded child in the regional vertical traffic data, tr is the driver's braking reaction time, and i is any value between 1 and nc;

[0028] S223. Conduct a comprehensive analysis of the risk status of pedestrian flow distribution at the direct entrances and exits corresponding to elevators in each functional area based on the regional pedestrian density risk index and the behavioral activity risk index.

[0029] The formula for calculating the dangerous state of crowd distribution is:

[0030]

[0031] Where RF is the dangerous state of the passenger flow distribution at the direct entrance and exit corresponding to the elevator in the current functional zone, and min[Qr·ln(1+Ba)] is the minimum value of Qr·ln(1+Ba) among the direct entrances and exits corresponding to the elevators in all functional zones.

[0032] Preferably, in step S3, the risk status of the vehicle when entering each through entrance and exit is analyzed, which specifically includes the following steps:

[0033] S31, extracting vehicle dynamic perception data and spatial topology data of the shopping mall underground parking lot;

[0034] S32. Construct a vehicle entry risk status analysis model, import the vehicle dynamic perception data and spatial topology data of the shopping mall's underground parking lot into the vehicle entry risk status analysis model, analyze the vehicle entry risk status at each direct entrance and exit, and obtain the vehicle entry risk status analysis results at each direct entrance and exit.

[0035] Preferably, the process of constructing the vehicle entry risk state analysis model in step S32 includes the following specific steps:

[0036] S321. Calculate the structural risk index Sr of each through entrance and exit based on the spatial topology data of the shopping mall's underground parking lot;

[0037] S322. Calculate the vehicle accumulation dynamic risk index at each through-entrance and exit based on the vehicle dynamic perception data and spatial topology data of the shopping mall's underground parking lot;

[0038] The calculation formula for the vehicle accumulation dynamic risk index is:

[0039]

[0040] Where Dr is the dynamic risk index of vehicle accumulation at the current through-entry location, Cq is the vehicle accumulation coefficient at the current through-entry location, vc is the real-time speed of the newly entered vehicle, D is the current braking distance of the newly entered vehicle, L is the real-time lighting intensity at the current through-entry location, and Ds is the safe braking distance threshold.

[0041] S323. Analyze the vehicle entry risk status at each through entrance and exit based on the structural risk index and vehicle accumulation dynamic risk index at each through entrance and exit;

[0042] The calculation formula for the vehicle entering the risk state is:

[0043]

[0044] Where Vr is the vehicle entry risk status at the current through-entry / exit location, To obtain the calculated locations of all direct entrances and exits The minimum value in .

[0045] Preferably, the construction of the human-vehicle safety linkage risk assessment model in step S4 includes the following specific steps:

[0046] S41. Obtaining the pedestrian flow distribution hazard status analysis results of the through-entrances and exits corresponding to the elevators in each functional zone and the vehicle entry risk status analysis results at each through-entrance and exit;

[0047] S42. Based on the pedestrian flow distribution hazard analysis results of the through-entrances and exits corresponding to the elevators in each functional zone and the vehicle entry risk analysis results at each through-entrance and exit, the pedestrian and vehicle safety linkage risk of each through-entrance and exit is evaluated to obtain the pedestrian and vehicle safety linkage risk of each through-entrance and exit;

[0048] The calculation formula for the risk of human-vehicle safety linkage is:

[0049]

[0050] Where RT is the pedestrian-vehicle safety linkage risk at the current direct entrance and exit.

[0051] Preferably, building a safe parking route planning model in step S5 specifically includes:

[0052] S51. Real-time acquisition of the analyzed pedestrian-vehicle safety linkage risk assessment results for each through-entrance and exit;

[0053] S52: Obtain in real time the maximum value of the pedestrian-vehicle safety linkage risk assessment results corresponding to all through entrances and exits in the current lane passed by the newly entered vehicle as the parking route change index of the newly entered vehicle;

[0054] S53: A parking route change threshold is preset. When the parking route change index of the newly entered vehicle is less than or equal to the parking route change threshold, the parking route of the newly entered vehicle does not need to be changed.

[0055] S54. When the parking route change index of the newly entered vehicle is greater than the parking route change threshold, a parking route change strategy is implemented for the newly entered vehicle. The parking route change strategy specifically includes: when the parking route change index of the newly entered vehicle obtained in real time is greater than the parking route change threshold, prompting the newly entered vehicle to detour to the nearest intersection in the current lane;

[0056] S55. Continue to execute steps S52-S54 for the new vehicle in the lane after the detour until the new vehicle stops in the parking area.

[0057] In a second aspect, the present invention provides a monitoring intelligent analysis system based on machine vision, comprising:

[0058] The data acquisition module is used to obtain vehicle dynamic perception data in the mall's underground parking lot and regional vertical traffic data of elevators in each functional area of ​​the mall through the mall's internal monitoring system; at the same time, it obtains the spatial topology data of the mall's underground parking lot;

[0059] The pedestrian flow distribution hazard state analysis module is used to use the underground parking lot entrances and exits that the elevators in each functional zone pass through as the direct entrances and exits corresponding to the elevators in each functional zone; and import the regional vertical traffic data of the elevators in each functional zone of the shopping mall into the pedestrian flow distribution hazard state analysis model to analyze the pedestrian flow distribution hazard state of the direct entrances and exits corresponding to the elevators in each functional zone;

[0060] The vehicle entry risk status analysis module is used to import the vehicle dynamic perception data and spatial topology data of the shopping mall's underground parking lot into the vehicle entry risk status analysis model to analyze the vehicle entry risk status at each direct entrance and exit;

[0061] The pedestrian-vehicle safety linkage risk assessment module is used to build a pedestrian-vehicle safety linkage risk assessment model. The analysis results obtained from the pedestrian distribution risk state analysis model and the vehicle entry risk state analysis model are imported into the pedestrian-vehicle safety linkage risk assessment model to evaluate the pedestrian-vehicle safety linkage risk at each direct entrance and exit.

[0062] The safe parking route planning module is used to build a safe parking route planning model based on the pedestrian-vehicle safety linkage risk assessment results of each direct entrance and exit, and plan safe parking routes for new vehicles entering the park;

[0063] The control module is used to control the operation of the data acquisition module, the crowd distribution dangerous state analysis module, the vehicle entry risk state analysis module, the human-vehicle safety linkage risk assessment module, and the safe parking route planning module.

[0064] In a third aspect, the present invention provides an electronic device comprising: a processor and a memory, wherein the memory stores a computer program that can be called by the processor, and the processor executes a machine vision-based monitoring intelligent analysis method by calling the computer program stored in the memory.

[0065] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0066] The present invention analyzes the dangerous status of pedestrian distribution at the direct entrances and exits corresponding to the elevators in each functional area and the risk status of vehicle entry at each direct entrance and exit; constructs a pedestrian and vehicle safety linkage risk assessment model, imports the pedestrian and vehicle safety linkage risk assessment results into the pedestrian and vehicle safety linkage risk assessment model, and assesses the pedestrian and vehicle safety linkage risks of each direct entrance and exit; and constructs a safe parking route planning model based on the pedestrian and vehicle safety linkage risk assessment results of each direct entrance and exit, and plans safe parking routes for newly entered vehicles; thereby optimizing the traffic efficiency of the shopping mall's underground parking lot and improving the safety of pedestrian and vehicle traffic in the underground parking lot. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0068] Figure 1 Schematic diagram of the overall process of the monitoring intelligent analysis method based on machine vision of the present invention;

[0069] Figure 2 Schematic diagram of the structure of the monitoring intelligent analysis system based on machine vision of the present invention;

[0070] Figure 3 This is an analysis flow chart of step S2 of the machine vision-based monitoring intelligent analysis method of the present invention;

[0071] Figure 4 This is an analysis flow chart of step S3 of the machine vision-based monitoring intelligent analysis method of the present invention. DETAILED DESCRIPTION

[0072] The technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0073] Example 1

[0074] like Figure 1 As shown, this embodiment provides a monitoring intelligent analysis method based on machine vision, which specifically includes the following steps:

[0075] S1. Obtain vehicle dynamic perception data in the mall's underground parking lot and regional vertical traffic data of elevators in each functional area of ​​the mall through the mall's internal monitoring system; at the same time, obtain spatial topological data of the mall's underground parking lot;

[0076] S2. Each entrance and exit of the underground parking lot that the elevators in each functional zone pass through is used as the direct entrance and exit corresponding to the elevators in each functional zone; the regional vertical traffic data of the elevators in each functional zone of the shopping mall is imported into the pedestrian distribution hazard state analysis model, and the pedestrian distribution hazard state of the direct entrance and exit corresponding to the elevators in each functional zone is analyzed;

[0077] S3. Importing vehicle dynamic perception data and spatial topology data of the mall's underground parking lot into a vehicle entry risk status analysis model to analyze the vehicle entry risk status at each through-entry entrance and exit;

[0078] S4. Construct a pedestrian-vehicle safety linkage risk assessment model. The analysis results obtained from the pedestrian distribution risk state analysis model and the vehicle entry risk state analysis model are imported into the pedestrian-vehicle safety linkage risk assessment model to assess the pedestrian-vehicle safety linkage risk at each through-entry entrance and exit.

[0079] S5. Based on the risk assessment results of the pedestrian-vehicle safety linkage at each through entrance and exit, a safe parking route planning model is constructed to plan safe parking routes for new vehicles entering the site.

[0080] In this embodiment, if Figure 3 As shown, in step S2, the dangerous state of the flow distribution of people at the direct entrances and exits corresponding to the elevators in each functional area is analyzed, specifically including:

[0081] S21. Extracting regional vertical traffic data of elevators in each functional area of ​​the shopping mall and the corresponding direct entrances and exits;

[0082] S22. Construct a crowd distribution hazard state analysis model, import the regional vertical traffic data of the elevators in each functional area of ​​the shopping mall into the crowd distribution hazard state analysis model, analyze the crowd distribution hazard state of the direct entrances and exits corresponding to the elevators in each functional area, and obtain the crowd distribution hazard state analysis results of the direct entrances and exits corresponding to the elevators in each functional area.

[0083] In this embodiment, the process of constructing the crowd flow distribution dangerous state analysis model in step S22 specifically includes:

[0084] S221. Calculate the regional passenger density risk index for elevators in each functional zone based on the regional vertical traffic data for elevators in each functional zone of the shopping mall;

[0085] The calculation formula for the regional crowd density risk index is:

[0086]

[0087] Wherein, Qr is the regional crowd density risk index of the through-entrance and exit corresponding to the elevator in the current functional zone at the current moment, Fp is the risk correction coefficient of the current functional zone at the current moment, nc is the number of stranded children in the through-entrance and exit area corresponding to the elevator in the current functional zone in the regional vertical traffic data at the current moment, Ae is the total number of stranded people in the through-entrance and exit area corresponding to the elevator in the current functional zone in the regional vertical traffic data at the current moment, st is the standard deviation of the historical length of children's detention at the through-entrance and exit corresponding to the elevator in the current functional zone in the regional vertical traffic data, Ts is the mean length of the historical length of children's detention at the through-entrance and exit corresponding to the elevator in the current functional zone in the regional vertical traffic data; in this embodiment, by calculating the proportion of children to the total number of people, it is possible to reflect the degree of control that parents can have over their children. When the proportion of children increases, it is more difficult for parents to control children's sudden running and other behaviors; when the proportion of children is small, it means that it is easier for parents to control children's sudden running and other behaviors;

[0088] The calculation formula for the risk correction coefficient of the current functional zoning is:

[0089]

[0090] Where A is the area of ​​the current functional zone, Ea is the number of acceleration mutations per unit area of ​​the current functional zone at the current moment in the regional vertical traffic data, n is the number of children in the current functional zone at the current moment in the regional vertical traffic data, Tavg is the average cumulative length of stay of customers in the current functional zone at the current moment in the regional vertical traffic data, vs is the safe moving speed threshold, σ(v) 2is the variance of the moving speed of all customers in the current functional zone at the current moment in the regional vertical traffic data. In this embodiment, σ(v) 2 The calculation is performed by extracting the thermal imaging trajectory tracking speed sequence from the surveillance video of the current functional partition; performing inter-frame analysis on the surveillance video of the current functional partition using the optical flow method to obtain the number of acceleration mutations per unit area of ​​the current functional partition; extracting the thermal imaging trajectory tracking speed sequence from the surveillance video and calculating the variance of the movement speed based on the thermal imaging trajectory tracking speed sequence are all existing technologies in this technical field and will not be described in detail.

[0091] For example, this embodiment constructs the molecular term by dynamically associating the number of acceleration mutations per unit area of ​​the current functional partition with the number of children gathering. The optical flow method is used to calculate the change of customer movement acceleration in the current functional area in real time, and normalize it by the functional area and average stay time. This embodiment also introduces a nonlinear risk amplification mechanism to construct a logarithmic term. The variance of customer movement speed is used as an indicator to reflect the customer's running and sudden stop behaviors. The ratio of the variance of customer movement speed to the safe movement speed threshold is used to quantify the movement chaos, realizing real-time dynamic adjustment of the risk correction coefficient of the current functional zoning.

[0092] Specifically, this embodiment introduces a risk correction coefficient Fp to quantify the inherent risk differences of different functional zones, which can accurately reflect the special risks of areas with high-frequency child activity in shopping malls. By integrating the static spatial properties of different functional zones with dynamic changes in passenger flow, a calculation formula for the regional passenger flow density risk index is constructed. Through monitoring and analysis of real-time surveillance videos, real-time response to changes in passenger flow within functional zones is achieved. When children are detected running (i.e., the number of acceleration mutations per unit area of ​​the current functional zone) or gathering (i.e., the number of children in the current functional zone increases), the risk correction coefficient Fp is amplified nonlinearly. Compared with the traditional fixed risk correction coefficient, the risk correction coefficient Fp provided by this embodiment can significantly improve the timeliness of risk judgment.

[0093] S222. Calculate the behavioral activity risk index based on the regional vertical traffic data of elevators within each functional area of ​​the shopping mall;

[0094] The calculation formula for the behavioral activity risk index is:

[0095]

[0096] Wherein, Ba is the behavioral activity risk index of the through-entrance / exit corresponding to the elevator in the current functional zone at the current moment, nc is the number of stranded children in the through-entrance / exit area corresponding to the elevator in the current functional zone in the regional vertical traffic data at the current moment, vi is the real-time moving speed of the i-th stranded child in the regional vertical traffic data, ai is the real-time acceleration of the i-th stranded child in the regional vertical traffic data, tr is the driver's braking reaction time, and i is any value from 1 to nc. In this embodiment, the driver's braking reaction time is the time it takes for the driver to perceive the danger and complete the action (such as stepping on the brakes). For example, the normal range of the driver's braking reaction time is generally 2-3 seconds. Therefore, in this embodiment, the driver's braking reaction time defaults to the median of the normal range of 2.5 seconds.

[0097] For example, this embodiment quantifies the child's kinetic energy by introducing a squared velocity term, reflecting the potential collision risk with passing vehicles posed by a stranded child suddenly running in a through-entrance or exit area. Secondly, the product of acceleration and the driver's braking reaction time can be used to predict the stranded child's sudden displacement distance. Furthermore, this embodiment also dynamically assesses the driver's ability to effectively brake in an emergency using the driver's braking reaction time.

[0098] Specifically, this embodiment uses surveillance cameras and the optical flow method to capture the movement speed and acceleration of stranded children in each direct entrance and exit area in real time, and dynamically calculates the instantaneous risk contribution value of each stranded child to the behavioral activity risk index. For example, when a stranded child suddenly accelerates, its risk value shows a nonlinear growth trend; further, this embodiment summarizes the instantaneous risk contribution values ​​of all stranded children in each direct entrance and exit area to the behavioral activity risk index through a summation term, and normalizes them in combination with the safe speed threshold; further, by calculating the behavioral activity risk index, this embodiment improves the warning response speed to sudden running incidents of children, which can effectively reduce the probability of accidents.

[0099] S223. Conduct a comprehensive analysis of the risk status of pedestrian flow distribution at the direct entrances and exits corresponding to elevators in each functional area based on the regional pedestrian density risk index and the behavioral activity risk index.

[0100] The formula for calculating the dangerous state of crowd distribution is:

[0101]

[0102] Where RF is the dangerous state of the passenger flow distribution at the direct entrance and exit corresponding to the elevator in the current functional zone, and min[Qr·ln(1+Ba)] is the minimum value of Qr·ln(1+Ba) among the direct entrances and exits corresponding to the elevators in all functional zones.

[0103] For example, first, this embodiment introduces the natural logarithm term ln(1+Ba) to perform a nonlinear transformation on the behavioral activity risk index, which can amplify the contribution of a high behavioral activity risk index to the dangerous state of crowd distribution, such as: sudden behaviors related to children running or stopping suddenly; secondly, this embodiment adopts the product form of the regional crowd density risk index and the behavioral activity risk index to achieve a comprehensive evaluation of the static density of the functional zone and the dynamic behavior of children leaving the functional zone; finally, this embodiment normalizes the formula through min[Qr·ln(1+Ba)] to eliminate the absolute dimension difference, so that the dangerous state of crowd distribution of each direct entrance and exit corresponding to the elevator in each functional zone is comparable. Compared with the traditional linear weighted method, this embodiment can more accurately describe the dangerous state of crowd distribution of each direct entrance and exit through the nonlinear coupling mechanism.

[0104] In this embodiment, if Figure 4 As shown, in step S3, the risk status of the vehicle when entering the location of each through entrance and exit is analyzed, which specifically includes the following steps:

[0105] S31, extracting vehicle dynamic perception data and spatial topology data of the shopping mall underground parking lot;

[0106] S32. Construct a vehicle entry risk status analysis model, import the vehicle dynamic perception data and spatial topology data of the shopping mall's underground parking lot into the vehicle entry risk status analysis model, analyze the vehicle entry risk status at each direct entrance and exit, and obtain the vehicle entry risk status analysis results at each direct entrance and exit.

[0107] In this embodiment, the process of constructing the vehicle entry risk state analysis model in step S32 includes the following specific steps:

[0108] S321. Calculate the structural risk index of each through entrance and exit based on the spatial topology data of the shopping mall's underground parking lot;

[0109] The calculation formula of the structural risk index is:

[0110]

[0111] Where Sr is the structural risk index of the current through-entry location, Sb is the blind spot area of ​​the driver at the current through-entry location in the spatial topology data, St is the total area of ​​the monitoring area at the current through-entry location in the spatial topology data, Rmin is the minimum turning radius corresponding to the new vehicle model, Rnom is the minimum turning radius of a standard small car, and θ is the slope angle of the current lane of the new vehicle.

[0112] For example, the blind spot ratio is positively correlated with the accident rate. In this embodiment, the blind spot area is introduced to quantify the blocking effect of structures such as pillars and walls on the driver's vision. Secondly, this embodiment also introduces the sine term of the slope angle to amplify the impact of steep slopes on the change in the driver's blind spot and the vehicle's braking performance.

[0113] S322. Calculate the vehicle accumulation dynamic risk index at each through-entrance and exit based on the vehicle dynamic perception data and spatial topology data of the mall's underground parking lot;

[0114] The calculation formula for the vehicle accumulation dynamic risk index is:

[0115]

[0116] Where Dr is the dynamic risk index of vehicle accumulation at the current through-entry location, Cq is the vehicle accumulation coefficient at the current through-entry location, vc is the real-time speed of the newly entered vehicle, D is the current braking distance of the newly entered vehicle, L is the real-time lighting intensity at the current through-entry location, and Ds is the safe braking distance threshold.

[0117] The calculation formula for the vehicle stacking coefficient is:

[0118]

[0119] Where Nv is the number of vehicles queuing at the current through-entrance / exit at the current moment, and Nmax is the lane capacity.

[0120] For example, first, this embodiment analyzes the linear relationship between lane capacity and congestion risk at the through-entry location by introducing a vehicle accumulation coefficient, and quantifies the compression effect of queuing vehicles on the braking space of newly entering vehicles; secondly, this embodiment reflects the dynamic balance between vehicle kinetic energy and braking capacity by multiplying the speed by the braking distance; finally, this embodiment reflects the risk perception ability in the dim environment inside the underground parking lot by the inverse of the lighting intensity at the through-entry location, and realizes adaptive assessment of the dynamic risk of vehicle accumulation at the through-entry location by coupling dynamic parameters with static thresholds.

[0121] S323. Analyze the vehicle entry risk status at each through entrance and exit based on the structural risk index and vehicle accumulation dynamic risk index at each through entrance and exit;

[0122] The calculation formula for the vehicle entering the risk state is:

[0123]

[0124] Where Vr is the vehicle entry risk status at the current through-entry / exit location, To obtain the calculated locations of all direct entrances and exits The minimum value in ;

[0125] For example, this embodiment analyzes the vehicle entry risk status by nonlinearly fusing the vehicle accumulation dynamic risk index and the structural risk index. The interaction and joint influence of the vehicle accumulation dynamic risk index and the structural risk index extend beyond the impact of a single risk alone, allowing the vehicle entry risk status to exhibit an amplified effect and avoid the one-sidedness of a single risk factor. This embodiment also introduces a global minimum for normalization, making the vehicle entry risk status at each entrance and exit relatively comparable. A square root operation is used to suppress extreme fluctuations in vehicle accumulation risk, such as instantaneous congestion. This allows for a precise analysis of the risk transmission mechanism of the vehicle entry risk status.

[0126] Specifically, in this embodiment, when a through entrance or exit on the driving lane where a newly entered vehicle is located has both a high vehicle accumulation and structural defects, the vehicle entry risk status increases significantly, triggering the execution of a parking route change strategy to divert vehicles to low-risk paths in real time, thereby improving vehicle diversion efficiency; through global minimum normalization, this embodiment can also automatically adapt to the baseline risk of different time periods. For example, the baseline value is lowered at night when the traffic volume is low, avoiding risk misjudgment caused by fixed thresholds.

[0127] In this embodiment, step S4 constructs a risk assessment model for human-vehicle safety linkage, including the following specific steps:

[0128] S41. Obtaining the pedestrian flow distribution hazard status analysis results of the through-entrances and exits corresponding to the elevators in each functional zone and the vehicle entry risk status analysis results at each through-entrance and exit;

[0129] S42. Based on the pedestrian flow distribution hazard analysis results of the through-entrances and exits corresponding to the elevators in each functional zone and the vehicle entry risk analysis results at each through-entrance and exit, the pedestrian and vehicle safety linkage risk of each through-entrance and exit is evaluated to obtain the pedestrian and vehicle safety linkage risk of each through-entrance and exit;

[0130] The calculation formula for the risk of human-vehicle safety linkage is:

[0131]

[0132] Where RT is the current pedestrian-vehicle safety linkage risk at the direct entrance and exit;

[0133] Exemplarily, this embodiment reflects the synergistic amplification effect of human-vehicle interaction at through entrances and exits through the product of the dangerous state of pedestrian distribution and the risk state of vehicle entry; 1+exp(-(RF+Vr)) is introduced to dynamically normalize the numerator, thereby suppressing the overflow of extreme risk values. Compared with the traditional linear weighting method, this embodiment dynamically adjusts the risk weight through an exponential function. For example, when the dangerous state of pedestrian distribution and the risk state of vehicle entry are both high, the risk of human-vehicle safety linkage increases nonlinearly, which can more accurately analyze the urgency of changing parking routes in high-risk scenarios.

[0134] In this embodiment, the safe parking route planning model is constructed in step S5, which specifically includes:

[0135] S51. Real-time acquisition of the analyzed pedestrian-vehicle safety linkage risk assessment results for each through-entrance and exit;

[0136] S52: Obtain in real time the maximum value of the pedestrian-vehicle safety linkage risk assessment results corresponding to all through entrances and exits in the current lane passed by the newly entered vehicle as the parking route change index of the newly entered vehicle;

[0137] S53: A parking route change threshold is preset. When the parking route change index of the newly entered vehicle is less than or equal to the parking route change threshold, the parking route of the newly entered vehicle does not need to be changed.

[0138] S54. When the parking route change index of the newly entered vehicle is greater than the parking route change threshold, a parking route change strategy is implemented for the newly entered vehicle. The parking route change strategy specifically includes: when the parking route change index of the newly entered vehicle obtained in real time is greater than the parking route change threshold, prompting the newly entered vehicle to detour to the nearest intersection in the current lane;

[0139] S55. Continue to execute steps S52-S54 for the newly entered vehicle in the driving lane after the detour until the newly entered vehicle stops in the parking area; wherein, the setting parameters (such as weights and thresholds) in this embodiment are obtained in the following manner: obtain vehicle dynamic perception data of the shopping mall's underground parking lot at multiple historical moments and regional vertical traffic data of elevators in each functional area of ​​the shopping mall; simultaneously obtain spatial topology data of the shopping mall's underground parking lot; substitute the vehicle dynamic perception data, regional vertical traffic data, and spatial topology data into the calculation formula of the pedestrian-vehicle safety linkage risk to calculate the pedestrian-vehicle safety linkage risk of each direct entrance and exit at multiple historical moments; based on the pedestrian-vehicle safety linkage risk of each direct entrance and exit at multiple historical moments, extract the parking route change index of multiple newly entered vehicles during driving in history, obtain the corresponding judgment results of whether emergency accidents occurred during driving of multiple newly entered vehicles in history, import the parking route change index of multiple newly entered vehicles during driving in history and the corresponding judgment results of whether emergency accidents occurred into the fitting software, and output the values ​​of the setting parameters (such as weights and thresholds) that meet the parking route change judgment accuracy.

[0140] Example 2

[0141] like Figure 2 As shown, this embodiment provides a monitoring intelligent analysis system based on machine vision, including:

[0142] The data acquisition module is used to obtain vehicle dynamic perception data in the mall's underground parking lot and regional vertical traffic data of elevators in each functional area of ​​the mall through the mall's internal monitoring system; at the same time, it obtains the spatial topology data of the mall's underground parking lot;

[0143] The pedestrian flow distribution hazard state analysis module is used to use the underground parking lot entrances and exits that the elevators in each functional zone pass through as the direct entrances and exits corresponding to the elevators in each functional zone; and import the regional vertical traffic data of the elevators in each functional zone of the shopping mall into the pedestrian flow distribution hazard state analysis model to analyze the pedestrian flow distribution hazard state of the direct entrances and exits corresponding to the elevators in each functional zone;

[0144] The vehicle entry risk status analysis module is used to import the vehicle dynamic perception data and spatial topology data of the shopping mall's underground parking lot into the vehicle entry risk status analysis model to analyze the vehicle entry risk status at each direct entrance and exit;

[0145] The pedestrian-vehicle safety linkage risk assessment module is used to build a pedestrian-vehicle safety linkage risk assessment model. The analysis results obtained from the pedestrian distribution risk state analysis model and the vehicle entry risk state analysis model are imported into the pedestrian-vehicle safety linkage risk assessment model to evaluate the pedestrian-vehicle safety linkage risk at each direct entrance and exit.

[0146] The safe parking route planning module is used to build a safe parking route planning model based on the pedestrian-vehicle safety linkage risk assessment results of each direct entrance and exit, and plan safe parking routes for new vehicles entering the park;

[0147] The control module is used to control the operation of the data acquisition module, the crowd distribution dangerous state analysis module, the vehicle entry risk state analysis module, the human-vehicle safety linkage risk assessment module, and the safe parking route planning module.

[0148] The above-mentioned parameters and steps for each unit module to implement corresponding functions in the machine vision-based monitoring intelligent analysis system of the present invention can refer to the parameters and steps in the embodiment of the machine vision-based monitoring intelligent analysis method above, and will not be repeated here.

[0149] Example 3

[0150] An electronic device according to an embodiment of the present invention includes: a processor and a memory, wherein the memory stores a computer program that can be called by the processor, and the processor executes a machine vision-based intelligent monitoring analysis method by calling the computer program stored in the memory. It should be noted that all computer programs of the machine vision-based intelligent monitoring analysis method are implemented in C language, and the data acquisition module, the pedestrian distribution hazard state analysis module, the vehicle entry risk state analysis module, the pedestrian-vehicle safety linkage risk assessment module, the safe parking route planning module, and the control module are all controlled by a remote server.

[0151] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0152] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A monitoring intelligent analysis method based on machine vision, characterized in that: The steps include: S1. Obtain vehicle dynamic perception data in the mall's underground parking lot and regional vertical traffic data of elevators in each functional area of ​​the mall through the mall's internal monitoring system; at the same time, obtain spatial topological data of the mall's underground parking lot; S2. Use the entrances and exits of the underground parking lot that the elevators in each functional area pass through as the direct entrances and exits corresponding to the elevators in each functional area; Import the regional vertical traffic data of elevators in each functional area of ​​the shopping mall into the pedestrian distribution hazard state analysis model, and analyze the pedestrian distribution hazard state of the direct entrances and exits corresponding to the elevators in each functional area; S3. Importing vehicle dynamic perception data and spatial topology data of the mall's underground parking lot into a vehicle entry risk status analysis model to analyze the vehicle entry risk status at each through-entry entrance and exit; S4. Construct a pedestrian-vehicle safety linkage risk assessment model. The analysis results obtained from the pedestrian distribution risk state analysis model and the vehicle entry risk state analysis model are imported into the pedestrian-vehicle safety linkage risk assessment model to assess the pedestrian-vehicle safety linkage risk at each through-entry entrance and exit. S5. Based on the risk assessment results of the pedestrian-vehicle safety linkage at each through entrance and exit, a safe parking route planning model is constructed to plan safe parking routes for new vehicles entering the site.

2. The monitoring intelligent analysis method based on machine vision according to claim 1 is characterized in that: The step S2 analyzes the dangerous state of the flow of people at the direct entrances and exits corresponding to the elevators in each functional area, specifically including: S21. Extracting regional vertical traffic data of elevators in each functional area of ​​the shopping mall and the corresponding direct entrances and exits; S22. Construct a crowd distribution hazard state analysis model, import the regional vertical traffic data of the elevators in each functional area of ​​the shopping mall into the crowd distribution hazard state analysis model, analyze the crowd distribution hazard state of the direct entrances and exits corresponding to the elevators in each functional area, and obtain the crowd distribution hazard state analysis results of the direct entrances and exits corresponding to the elevators in each functional area.

3. The monitoring intelligent analysis method based on machine vision according to claim 2 is characterized in that: The process of constructing the crowd flow distribution dangerous state analysis model in step S22 specifically includes: S221. Calculate the regional passenger density risk index for elevators in each functional zone based on the regional vertical traffic data for elevators in each functional zone of the shopping mall; The calculation formula for the regional crowd density risk index is: Where Qr is the regional crowd density risk index of the direct entrance and exit corresponding to the elevator in the current functional zone at the current moment, Fp is the risk correction coefficient of the current functional zone at the current moment, nc is the number of stranded children in the direct entrance and exit area corresponding to the elevator in the current functional zone in the regional vertical transportation data at the current moment, Ae is the total number of stranded people in the direct entrance and exit area corresponding to the elevator in the current functional zone in the regional vertical transportation data at the current moment, st is the standard deviation of the historical length of children's detention at the direct entrance and exit corresponding to the elevator in the current functional zone in the regional vertical transportation data, and Ts is the mean length of children's detention at the direct entrance and exit corresponding to the elevator in the current functional zone in the regional vertical transportation data. The calculation formula for the risk correction coefficient of the current functional zoning is: Where A is the area of ​​the current functional zone, Ea is the number of acceleration mutations per unit area of ​​the current functional zone at the current moment in the regional vertical traffic data, n is the number of children in the current functional zone at the current moment in the regional vertical traffic data, Tavg is the average cumulative length of stay of customers in the current functional zone at the current moment in the regional vertical traffic data, vs is the safe moving speed threshold, σ(v) 2 is the variance of the moving speed of all customers in the current functional zone at the current moment in the regional vertical traffic data; S222. Calculate the behavioral activity risk index based on the regional vertical traffic data of elevators within each functional area of ​​the shopping mall; The calculation formula for the behavioral activity risk index is: Where Ba is the behavioral activity risk index of the direct entrance and exit corresponding to the elevator in the current functional zone at the current moment, nc is the number of stranded children in the direct entrance and exit area corresponding to the elevator in the current functional zone in the regional vertical traffic data at the current moment, vi is the real-time moving speed of the i-th stranded child in the regional vertical traffic data, ai is the real-time acceleration of the i-th stranded child in the regional vertical traffic data, tr is the driver's braking reaction time, and i is any value between 1 and nc; S223. Conduct a comprehensive analysis of the risk status of pedestrian flow distribution at the direct entrances and exits corresponding to elevators in each functional area based on the regional pedestrian density risk index and the behavioral activity risk index. The formula for calculating the dangerous state of crowd distribution is: Where RF is the dangerous state of the passenger flow distribution at the direct entrance and exit corresponding to the elevator in the current functional zone, and min[Qr·ln(1+Ba)] is the minimum value of Qr·ln(1+Ba) among the direct entrances and exits corresponding to the elevators in all functional zones.

4. The monitoring intelligent analysis method based on machine vision according to claim 3 is characterized in that: The risk status of the vehicle when entering each through entrance and exit is analyzed in step S3, which specifically includes the following steps: S31, extracting vehicle dynamic perception data and spatial topology data of the shopping mall underground parking lot; S32. Construct a vehicle entry risk status analysis model, import the vehicle dynamic perception data and spatial topology data of the shopping mall's underground parking lot into the vehicle entry risk status analysis model, analyze the vehicle entry risk status at each direct entrance and exit, and obtain the vehicle entry risk status analysis results at each direct entrance and exit.

5. The monitoring intelligent analysis method based on machine vision according to claim 4 is characterized in that: The process of constructing the vehicle entry risk state analysis model in step S32 includes the following specific steps: S321. Calculate the structural risk index Sr of each through entrance and exit based on the spatial topology data of the shopping mall's underground parking lot; S322. Calculate the vehicle accumulation dynamic risk index at each through-entrance and exit based on the vehicle dynamic perception data and spatial topology data of the mall's underground parking lot; The calculation formula for the vehicle accumulation dynamic risk index is: Where Dr is the dynamic risk index of vehicle accumulation at the current through-entry location, Cq is the vehicle accumulation coefficient at the current through-entry location, vc is the real-time speed of the newly entered vehicle, D is the current braking distance of the newly entered vehicle, L is the real-time lighting intensity at the current through-entry location, and Ds is the safe braking distance threshold. S323. Analyze the vehicle entry risk status at each through entrance and exit based on the structural risk index and vehicle accumulation dynamic risk index at each through entrance and exit; The calculation formula for the vehicle entering the risk state is: Where Vr is the vehicle entry risk status at the current through-entry / exit location, To obtain the calculated locations of all direct entrances and exits The minimum value in .

6. The monitoring intelligent analysis method based on machine vision according to claim 5 is characterized in that: The step S4 constructs a risk assessment model for the linkage of human and vehicle safety, including the following specific steps: S41. Obtaining the pedestrian flow distribution hazard status analysis results of the through-entrances and exits corresponding to the elevators in each functional zone and the vehicle entry risk status analysis results at the locations of each through-entrance and exit; S42. Based on the pedestrian flow distribution hazard analysis results of the through-entrances and exits corresponding to the elevators in each functional zone and the vehicle entry risk analysis results at each through-entrance and exit, the pedestrian and vehicle safety linkage risk of each through-entrance and exit is evaluated to obtain the pedestrian and vehicle safety linkage risk of each through-entrance and exit; The calculation formula for the risk of human-vehicle safety linkage is: Where RT is the pedestrian-vehicle safety linkage risk at the current direct entrance and exit.

7. The monitoring intelligent analysis method based on machine vision according to claim 6, characterized in that: The step S5 constructs a safe parking route planning model, specifically including: S51. Real-time acquisition of the analyzed pedestrian-vehicle safety linkage risk assessment results for each through-entrance and exit; S52: Obtain in real time the maximum value of the pedestrian-vehicle safety linkage risk assessment results corresponding to all through entrances and exits in the current lane passed by the newly entered vehicle as the parking route change index of the newly entered vehicle; S53: A parking route change threshold is preset. When the parking route change index of the newly entered vehicle is less than or equal to the parking route change threshold, the parking route of the newly entered vehicle does not need to be changed. S54. When the parking route change index of the newly entered vehicle is greater than the parking route change threshold, a parking route change strategy is implemented for the newly entered vehicle. The parking route change strategy specifically includes: when the parking route change index of the newly entered vehicle obtained in real time is greater than the parking route change threshold, prompting the newly entered vehicle to detour to the nearest intersection in the current lane; S55. Continue to execute steps S52-S54 for the new vehicle in the lane after the detour until the new vehicle stops in the parking area.

8. A machine vision-based monitoring intelligent analysis system, used to implement the machine vision-based monitoring intelligent analysis method according to any one of claims 1 to 7, characterized in that: The system comprises: The data acquisition module is used to obtain vehicle dynamic perception data in the mall's underground parking lot and regional vertical traffic data of elevators in each functional area of ​​the mall through the mall's internal monitoring system; at the same time, it obtains the spatial topology data of the mall's underground parking lot; The pedestrian flow distribution hazard state analysis module is used to use the underground parking lot entrances and exits that the elevators in each functional zone pass through as the direct entrances and exits corresponding to the elevators in each functional zone; and import the regional vertical traffic data of the elevators in each functional zone of the shopping mall into the pedestrian flow distribution hazard state analysis model to analyze the pedestrian flow distribution hazard state of the direct entrances and exits corresponding to the elevators in each functional zone; The vehicle entry risk status analysis module is used to import the vehicle dynamic perception data and spatial topology data of the shopping mall's underground parking lot into the vehicle entry risk status analysis model to analyze the vehicle entry risk status at each direct entrance and exit; The pedestrian-vehicle safety linkage risk assessment module is used to build a pedestrian-vehicle safety linkage risk assessment model. The analysis results obtained from the pedestrian distribution risk state analysis model and the vehicle entry risk state analysis model are imported into the pedestrian-vehicle safety linkage risk assessment model to evaluate the pedestrian-vehicle safety linkage risk at each direct entrance and exit. The safe parking route planning module is used to build a safe parking route planning model based on the pedestrian-vehicle safety linkage risk assessment results of each direct entrance and exit, and plan safe parking routes for new vehicles entering the park; A control module is used to control the operation of the data acquisition module, the crowd distribution dangerous state analysis module, the vehicle entry risk state analysis module, the pedestrian and vehicle safety linkage risk assessment module, and the safe parking route planning module.

9. An electronic device comprising: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; characterized in that the processor executes the machine vision-based monitoring intelligent analysis method as described in any one of claims 1 to 7 by calling the computer program stored in the memory.

Images