Remote Intelligence People Flow Statistics Control System Based on Internet of Things (IoT)

Abstract

The present invention discloses a remote intelligent people flow statistical control system based on Internet of Things (IoT), the system includes: a data collection module, a data storage module, a people flow control module, an image analysis module, a target recognition module, a model analysis module and a linkage analysis module, the image analysis module, the target recognition module and the model analysis module perform model analysis on the pedestrian target movement trajectory of the people flow in different passage spaces, and then the linkage analysis module performs overall dynamic analysis on the evacuation network consisting of all passage spaces, and based on the model analysis results of different nodes, analyzes the correlation between nodes to obtain a correlation coefficient, and further analyzes the influence between evacuation routes to obtain the linkage value between different evacuation routes, and finally establishes a dynamic observation equation, the people flow control module performs remote control on the intelligent people flow, greatly improving the efficiency of the intelligent people flow remote control, and making the analysis result of the people flow movement conform to the actual people flow movement regulation.

Description

[Technical field]

[0001] The present invention relates to the technical field of intelligent control, in particular to a remote intelligence people flow statistical control system based on the Internet of Things (IoT). [Background technology]

[0002] When an emergency occurs on a campus where people are densely distributed and evacuation is required, many students will evacuate from multiple teaching experiment centers and machine rooms, and crowd accidents are likely to occur in the crowd flow. In order to ensure that pedestrians in the building's corridors can evacuate safely, it is important to implement intelligent statistical monitoring detection and diversion command for the crowd flow in the building. In the conventional technology, the crowd flow control system realizes crowd flow monitoring within the monitoring range by analyzing the crowd flow movement characteristics in a space through image recognition and pedestrian detection. The crowd flow movement characteristics in different types of corridors in a building are different, and the models and model judgment criteria adopted for crowd flow analysis in different corridors are not unique. Under different types of corridor monitoring cameras, People flow movement routes are interconnected, and when people need to evacuate in an emergency, different types of passages form the entire evacuation route. However, it is difficult for remote monitoring personnel to monitor the entire people flow movement through multiple monitoring screens, and the cameras in different passages are limited to covering the space within the passage. When there are location requirements in the evacuation process, people flow analysis under a single monitoring alone cannot dynamically monitor the entire evacuation process. In order to increase the degree of cascade matching of different types of passages, realize remote intelligent control of people flow, and improve the accuracy of evacuation route installation, the present invention puts forward a remote intelligent people flow statistics control system based on the Internet of Things (IoT). Summary of the Invention [Means for solving the problem]

[0003] In the above situation, in order to overcome the shortcomings of the prior art, the objective of the present invention proposes a remote intelligent people flow statistics control system based on the Internet of Things (IoT), in which a model analysis module performs micro-analysis based on a model matching the people flow movement structures of different monitoring and collection ports, and then a linkage analysis module combines the analysis results of the micro-model to analyze the people flow movement rules of all evacuation lines, thereby improving the accuracy of the people flow movement analysis, and through the linkage analysis module, a cascade analysis of the mutual influences between different monitoring and collection ports can be performed, thereby improving the dynamic monitoring efficiency of remote people flow statistics and control.

[0004] The remote intelligence people flow statistical control system based on the Internet of Things (IoT) includes a data collection module, a data storage module, a people flow control module, an image analysis module, a target recognition module, a model analysis module, and a linkage analysis module; The people flow control module collects statistics on all evacuation routes in the building and labels the evacuation routes. i Let i ∈ [1,n], n is the number of evacuation routes in the building. The n evacuation routes intersect to form an evacuation net. The data collection module collects people flow data in different passage spaces through N monitoring collection ports, stores the collected people flow data in a data storage module, and calculates a route r i The number of monitoring collection ports is R i , R i ∈[1,N], covering the different numbers of inlets of the pedestrian flow in the pedestrian movement space by the monitoring collection ports; Collecting image information of people flow movement through a monitoring collection port, the image analysis module performs image frame processing on the image information from different monitoring collection ports, and the target recognition module performs target tracking on the moving target in the people flow, and the image analysis module and the target recognition module map the pixel coordinates of the image into a multi-dimensional spatial coordinate system; The model analysis module combines the analysis process of the model analysis module and the target recognition module to perform data statistics on the changes in people flow in the passage space, and then constructs a corresponding micro model based on the speed, density and flow rate structure of people flow movement. For people flow movement spaces with different numbers of entrances, different micro models are used to analyze the movement rules of people flow in space-time, and the speed-time distribution characteristics and density-time distribution characteristics of people flow movement in the passage are obtained; The above-mentioned linkage analysis module is coupled to the evacuation end position of the people flow movement to establish an objective function, and performs analysis on the feedback relationship between N nodes. The linkage analysis module further performs macro-linkage analysis on the people flow movement path characteristics of all passages in the evacuation network, and the people flow movement space covered by one monitoring collection port is regarded as one node in the people flow movement path, and R i nodes are included, and the feature matrix of the entire evacuation route is obtained by extracting the spatiotemporal distribution features of the speed and the spatiotemporal distribution features of the density of different nodes. Then, dynamic analysis is performed on the connections between different evacuation routes to obtain the influence value between the evacuation routes. Finally, the linkage analysis module predicts the evacuation process of people flow through the established dynamic observation equation, The people flow control module is combined with the prediction result of the linkage analysis module to direct the evacuation of people in the building, so that the people flow movement during emergency evacuation follows the optimal evacuation route.

[0005] In addition, the interlocking analysis module first determines the end point of the pedestrian flow, and then the pedestrians at different positions perform a self-organized pedestrian evacuation movement, and then performs dynamic observation through a holistic analysis of the evacuation network consisting of all nodes. Specifically, Let X be the spatial distance between the end point of the evacuation and each node. i , i ∈ [1,N], the linkage analysis module determines the objective function f(t) by combining the node data based on the end point location of the people flow movement. The node data includes spatial distance, people flow rate and time, and the end point coordinate is (x * ,y * ,z * ), the spatial coordinates of the node are (x, y, z),

[0006]

number

[0007] Step 1: determining the process of people flow movement along different evacuation routes based on an objective function; Let the merging node in the pedestrian flow path be the current node p, and calculate the correlation coefficient between the current node p and the adjacent node q. S1 and S2 are the feature vectors between the current node p and the adjacent node q, respectively. The calculation formula is expressed as follows:

[0008]

number

[0009] Step 2, where ρ represents the correlation coefficient, Cov represents the covariance, and E represents the mathematical expectation value. Based on the value of the correlation coefficient, the correlation between adjacent nodes is judged. If ρ>0, the people flow movement rule between the nodes is positive feedback, and if ρ<0, the people flow movement rule between the nodes is negative feedback. The node where two evacuation routes intersect is defined as an overlap node, and the pedestrian flows of different evacuation routes are merged at the overlap node. The model analysis module then performs a micro-analysis of the pedestrian movement rules at the overlap node. At the overlap node of evacuation routes, the pedestrian movements are mutually influenced, and the degree of change in the pedestrian movement at the overlap node of evacuation routes is expressed as a linkage value. The linkage value of the pedestrian movement between different evacuation routes is calculated to obtain the W ij where i and j represent the subscripts of the path, i, j ∈ [1, n], R i and R j are the numbers of nodes in the two evacuation routes, i and j, respectively,

[0010]

number

[0011] Here, U i (k) and U kj (k) is the flow rate of a node adjacent to the stack node and having the same inflow direction of people flow. The two nodes belong to two evacuation routes, respectively. k is the number of intersecting routes. X k , X ik , X jk Step 3 is the spatial distance, and The interlocking analysis module combines the analysis results of the model analysis module to analyze the spatiotemporal distribution characteristics of the speed and density of different nodes on all evacuation routes to obtain the corresponding feature matrix A nRi is obtained, and R i ∈[1,N], the linkage analysis module combines the linkage degree between different evacuation routes to dynamically observe the dynamic states of different nodes in the process of people flow, and makes judgments on the observed values;

[0012]

number

[0013] Where:

[0014]

number

[0015] η1 is the velocity, η2 is the density, η3 is the flow rate, η4 is the covered space, t i The dynamic prediction model includes a step 4 of predicting parameter fluctuations in the evacuation process based on the changes in the predicted values ​​at different times, and combining the predicted values ​​with the set objective function to determine the predicted values.

[0016] The target recognition module recognizes multiple pedestrians in the monitored video, and then converts the coordinates of the successfully recognized pedestrians and corrects their positions based on the recognition effect judgment. After the image analysis module and the target recognition module convert the coordinates of the collected data, the model analysis module analyzes and calculates the pedestrian flow density and pedestrian flow speed in the spatial area, and obtains a basic map of the pedestrian flow speed and space-time distribution on density through the construction of an analytical model of the space-time distribution on the position of the pedestrian target, and further analyzes the movement rules of the pedestrian flow through the basic map.

[0017] The model analysis module constructs different models based on the type of aisle space to analyze the people flow movement rules, and classifies the aisles including Y-shaped aisles, straight aisles, T-shaped aisles and cross aisles based on the number of people flow junctions in the area covered by the current monitoring collection port.

[0018] The relationship between the bottom and center coordinates of the detection frame of the target analysis module is expressed by the following formula:

[0019]

number

[0020] Here, u1 is the abscissa of the center of the detection frame in the pixel coordinate system, v1 is the ordinate of the center of the detection frame in the pixel coordinate system, h is the height of the detection frame in the pixel coordinate system, u is the abscissa of the bottom center of the detection frame in the pixel coordinate system, and v is the ordinate of the bottom center of the detection frame in the pixel coordinate system.

[0021] The image analysis module converts the coordinates of the pixel coordinate system of the pedestrian in the image, and the coordinates are expressed by the following equation:

[0022]

number

[0023] where M is the transformation matrix, x w is the abscissa in the spatial coordinate system, and yw is the ordinate in the spatial coordinate system, and z w is the vertical coordinate in the spatial coordinate system, and z is the distance for human distance monitoring in the camera coordinate system.

[0024] The pedestrian flow velocity analysis relationship is expressed by the following equation:

[0025]

number

[0026] where v i (t) represents the instantaneous speed of pedestrian i at time t, and w i (t) represents the walking distance of pedestrian i from time t to (t+T), and N represents the total number of pedestrians.

[0027]

number

[0028] where v - (Here, " - " indicates the overline of v) represents the time-averaged velocity, and X i,t represents the coordinates of the t-th frame of pedestrian i.

[0029] The density in space-time is ρ, which is given by:

[0030]

number

[0031] Here, the space covered by surveillance is divided into areas, and A m represents the area of ​​the mth area, and N t represents the number of pedestrians in the study area at time t. Effect of the Invention

[0032] By adopting the above technical means, the present invention has the following advantages over the prior art: First, in this invention, the image analysis module and the target recognition module are used to analyze the video information of pedestrian flow movement collected at the monitoring collection port, and the position coordinates of pedestrians in the pedestrian flow are converted through image frame processing and target recognition, and the movement data of pedestrians in the pedestrian flow is quickly determined through the video information, so as to realize pedestrian flow statistics and remote monitoring, and to improve the efficiency of pedestrian flow statistics and control. Second, in the present invention, the model analysis module combines the coordinate information converted by the image analysis module and the target identification module to construct a corresponding analysis model, and different models are adopted to analyze the people flow in different types of passage spaces in the evacuation route, and the model analysis module matches the model with the passage space type to reduce the error caused by the model, and makes the model analysis results conform to the actual people flow movement rules, so as to construct different people flow analysis models, thereby improving the accuracy of people flow data analysis. Third, in the present invention, the linkage analysis module is combined with the model analysis module to perform linkage analysis for all nodes; when there is a target in the people flow movement, the linkage analysis module dynamically analyzes the overall people flow movement rules through the interaction relationship between different routes and the analysis results of the micro-models of different nodes; the people flow control module controls and directs the people flow movement based on the prediction results of the linkage analysis module, thereby improving the accuracy of dynamic monitoring in the people flow evacuation process, enhancing the degree of cascading between different nodes, and avoiding human monitoring errors when remote monitoring personnel are looking at multiple monitoring screens; and when an emergency evacuation occurs, remotely controls and directs the people flow and supports the establishment of evacuation routes. [Brief description of the drawings]

[0033] [Figure 1] FIG. 2 is an overall module diagram according to the present invention. [Diagram 2] 1 is an overall analysis flowchart according to the present invention. [Diagram 3] 4 is an analysis flowchart of a linkage analysis module according to the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0034] The above and other technical contents, features and effects of the present invention can be clearly shown in the detailed description of the examples in the following formulation reference figures 1 to 3. The embodiments and features in the embodiments of the present application may be combined with each other, and the terms used in the specification have the meanings commonly understood by those skilled in the technical field of the present invention.

[0035] An embodiment of the present invention discloses a remote intelligence people flow statistical control system based on the Internet of Things (IoT), which includes a data collection module, a data storage module, a people flow control module, an image analysis module, a target recognition module, a model analysis module, and a linkage analysis module.

[0036] Campuses are places where people are crowded together. When an emergency evacuation is required due to an unexpected event, a large number of people will gather in the space. Take the statistics and application of people flow in and out of the school's teaching experimental center and machine room as an example. When the people in multiple experimental centers need to flow out quickly in an emergency, and students in different locations need to retreat quickly, it is very important to analyze the people flow movement rules and people flow statistics. In order to prevent crowd accidents from occurring when the people flow is crowded in the same space, it is very important to set up evacuation passages, and remote people flow statistics and control management. Image recognition is used to analyze the movement characteristics of pedestrian flow on the passage of the teaching building in the actual scene, and further analyze the people flow movement characteristics, and through the people flow division command, different people flows in the evacuation passages are evacuated in an orderly manner. Different people's evacuation routes cross and pedestrians interfere with each other, so the remote control staff needs to coordinate the analysis of people flow at multiple monitoring points. The people flow control module compiles statistics on all evacuation routes in the building, labels the evacuation routes, and performs r iLet i ∈ [1,n], n is the number of evacuation routes in the building, and the n evacuation routes intersect to form an evacuation net. The people flow in the building flows into the evacuation end point from all directions, and the evacuation people flow passes through different evacuation routes. Different evacuation routes weave an evacuation net, the number of monitoring collection ports on each evacuation route is different, the spatial type of people flow movement is different, and the probability of people flow crowd accidents occurring is also different. In the process of people evacuation, the pedestrian movement characteristics change constantly, the people flows between different evacuation routes influence each other, and different evacuation directions change.

[0037] The data collection module collects people flow data in different passage spaces through N monitoring collection ports, stores the collected people flow data in a data storage module, and calculates a route r i The number of monitoring collection ports is R i , R i ∈[1,N], the monitoring collection port covers a corresponding number of people flow inlet entrances of different people flow movement spaces, the monitoring collection port covers a corresponding number of people flow inlet entrances of different people flow movement spaces, the data collection module includes a monitoring device, a heat detection device, and an infrared detector at the monitoring collection port, and all monitoring devices installed in the building monitor and cover different people flow movement spaces.

[0038] Video information of people flow movement is collected through monitoring collection ports, the image analysis module performs image frame processing on the video information from different monitoring collection ports, and the target recognition module performs target tracking on moving targets in the people flow. The image analysis module and the target recognition module map the pixel coordinates of the image into a multidimensional space coordinate system. During the process of people flow movement, the image analysis module and the target analysis module perform dynamic target analysis and detection on pedestrians in the collected people flow data, and then obtain statistics on the people flow volume within an area through a polygon detection method.

[0039] The model analysis module combines the analysis process of the model analysis module and the target recognition module to perform data statistics on the changes in people flow in the passage space, and then constructs a corresponding micro model based on the speed, density and flow rate structure of people flow movement. For people flow movement spaces with different numbers of entrances, different micro models are used to analyze the movement rules of people flow in space-time, and the space-time distribution characteristics of people flow speed and space-time distribution characteristics of people flow density in the passage are obtained. The model structures in different types of passage spaces will bring corresponding analysis errors, and optimal models are constructed for different passage spaces, making the analysis results closer to the actual intelligent people flow movement rules.

[0040] The above-mentioned linkage analysis module is coupled to the evacuation end position of the people flow movement to establish an objective function, and performs analysis on the feedback relationship between N nodes. The linkage analysis module further performs macro-linkage analysis on the people flow movement path characteristics of all passages in the evacuation network, and the people flow movement space covered by one monitoring collection port is regarded as one node in the people flow movement path, and R i nodes are included, and the feature matrix of the entire evacuation route is obtained by extracting the spatiotemporal distribution features of the speed and the spatiotemporal distribution features of the density of different nodes. Then, dynamic analysis is performed on the connections between different evacuation routes to obtain the influence value between the evacuation routes. Finally, the linkage analysis module predicts the evacuation process of people flow through the established dynamic observation equation, The people flow control module is combined with the prediction result of the linkage analysis module to direct the evacuation of people in the building, so that the people flow movement during emergency evacuation follows the optimal evacuation route.

[0041] In addition, the linkage analysis module first determines the end position of the people flow movement, and the pedestrians at different positions perform the people flow evacuation movement in a self-organized manner, and then performs dynamic observation through the overall analysis of the evacuation network consisting of all nodes. The linkage analysis module analyzes the overall people flow movement rules of all evacuation routes, and the specific meaning of the same alphabet in different formulas refers to the specific interpretation of the alphabet in the formula, specifically, Let X be the spatial distance between the end point of the evacuation and each node. i, i ∈ [1,N], where N is the number of nodes. The linkage analysis module combines the node data based on the end position of the people flow movement to determine the objective function f(t), where ?? represents time. The objective function is set according to the end position of the evacuation and the limiting conditions required in the evacuation process. The limiting conditions include time limiting conditions, people flow load threshold, evacuation distance, and people flow analysis parameters during the evacuation route. The node data includes spatial distance, people flow rate and time. The end coordinate is (x ? ,y ? ,z ? ), the spatial coordinates of the node are (x, y, z),

[0042]

number

[0043] Step 1: determining the process of people flow movement along different evacuation routes based on an objective function; Let the merging node in the pedestrian flow path be the current node p, and calculate the correlation coefficient between the current node p and the adjacent node q. S1 and S2 are the feature vectors between the current node p and the adjacent node q, respectively. The calculation formula is expressed as follows:

[0044]

number

[0045] Step 2, where ρ represents the correlation coefficient, Cov represents the covariance, and E represents the mathematical expectation value. Based on the value of the correlation coefficient, the correlation between adjacent nodes is judged. If ρ>0, the people flow movement rule between the nodes is positive feedback, and if ρ<0, the people flow movement rule between the nodes is negative feedback. The node where two evacuation routes intersect is defined as an overlap node, and the pedestrian flows of different evacuation routes are merged at the overlap node. The model analysis module then performs a micro-analysis of the pedestrian movement rules at the overlap node. At the overlap node of evacuation routes, the pedestrian movements are mutually influenced, and the degree of change in the pedestrian movement at the overlap node of evacuation routes is expressed as a linkage value. The linkage value of the pedestrian movement between different evacuation routes is calculated to obtain the W ij where i and j represent the subscripts of the path, i, j ∈ [1, n], R i and R j are the number of nodes in the two evacuation routes i and j respectively. The movement speed, direction, density and flow parameters between different evacuation routes affect each other. The generation transformation of the pedestrian movement direction and the linkage value are used to represent the pedestrian flow transformation rate between different pedestrian evacuation routes.

[0046]

number

[0047] Here, U i (k) and U k j (k) is the flow rate of a node adjacent to the stack node and having the same inflow direction of people flow. The two nodes belong to two evacuation routes, respectively. k is the number of intersecting routes. X k , X ik , X jk Step 3 is the spatial distance, The interlocking analysis module combines the analysis results of the model analysis module to analyze the spatiotemporal distribution characteristics of the speed and density of different nodes on all evacuation routes to obtain the corresponding feature matrix A nRi is obtained, and R i ∈[1,N], the linkage analysis module combines the linkage degree between different evacuation routes to dynamically observe the dynamic states of different nodes in the process of people flow, and makes judgments on the observed values;

[0048]

number

[0049] Where:

[0050]

number

[0051] η1 is the velocity, η2 is the density, η3 is the flow rate, η4 is the covered space, t i Step 4 includes a people flow evacuation time, t representing different time values, and the dynamic prediction model predicts the parameter fluctuations of the evacuation process based on the changes of the predicted values ​​at different times, and judges the predicted values ​​in combination with the set objective function. If the predicted value of the interlocking analysis module is abnormal, the people flow control module sends a people flow diversion command.

[0052] The target recognition module recognizes multiple pedestrians in the monitored video, and then converts the coordinates of the successfully recognized pedestrians and corrects their positions based on the recognition effect judgment. After the image analysis module and the target recognition module convert the coordinates of the collected data, the model analysis module analyzes and calculates the pedestrian flow density and pedestrian flow speed within the spatial area, and constructs an analysis model based on the space-time distribution of pedestrian target positions to obtain basic diagrams of the space-time distribution of the pedestrian flow speed and the space-time distribution of the density, and further analyzes the movement rules of the pedestrian flow through the basic diagram.

[0053] In addition, the model analysis module builds different models based on the type of passage space to analyze the people flow movement rules, the model analysis module builds models based on different people flow movement spaces, and judges the models using model judgment criteria to ensure that the built models are consistent with the actual people flow movement regulations. According to the number of people flow junctions in the area covered by the current monitoring collection port, the passages are divided into Y-shaped passages, straight passages, T-shaped passages and cross passages, and the number of people flow junction entrances of the Y-shaped passages and T-shaped passages is 3, and the number of entrances of the straight passages is 2.

[0054] In addition, in order to ensure that the letters in different formulas represent unique mathematical meanings and are not confused with the same letters in other formulas, the relationship equation between the bottom and center coordinates of the detection frame of the target analysis module is expressed by the following formula:

[0055]

number

[0056] Here, u1 is the abscissa of the center of the detection frame in the pixel coordinate system, v1 is the ordinate of the center of the detection frame in the pixel coordinate system, h is the height of the detection frame in the pixel coordinate system, u is the abscissa of the bottom center of the detection frame in the pixel coordinate system, and v is the ordinate of the bottom center of the detection frame in the pixel coordinate system.

[0057] The image recognition technology of the image analysis module can obtain more objective people flow data and provide multi-dimensional verification criteria. Image recognition is an important means of acquiring data. When inspectors monitor people flow through multiple surveillance cameras, the people flow movement space collected by one surveillance camera can be judged by analyzing the data. However, when people flow in multiple locations needs to be monitored at the same time, the human naked eye cannot make an objective evaluation and calculation, and there exists an information cascade between different people flow data. In addition, the image analysis module converts the coordinates of the pixel coordinate system of the pedestrian in the image, which can be expressed by the following formula:

[0058]

number

[0059] where M is the transformation matrix, x w is the abscissa in the spatial coordinate system, and y w is the ordinate in the spatial coordinate system, and z w is the vertical coordinate in the spatial coordinate system, z is the distance for human distance monitoring in the camera coordinate system, The matrix M is actually a multiplication of two matrices, namely M=M1M2, where M1 is the camera matrix parameter and M2 is the camera extrinsic parameter matrix.

[0060]

number

[0061] R is a matrix related to the tilt angle of the camera, t is a matrix related to the position of the camera in the world coordinate system, u0 is the abscissa of the image center point, v0 is the ordinate of the image center point, and α x is the focal length and the x-direction pixel conversion ratio, and α y are the focal length and y-direction pixel conversion rate values, and f is the camera focal length.

[0062] Furthermore, the pedestrian flow velocity analysis relationship is expressed by the following equation:

[0063]

number

[0064] where v i (t) represents the instantaneous speed of pedestrian i at time t, and w i (t) represents the walking distance of pedestrian i from time t to (t+T), and N represents the total number of pedestrians.

[0065]

number

[0066] where v - (where " - " indicates the overline of v) represents the time-averaged velocity, and X i,trepresents the coordinate of the t-th frame of pedestrian i. The speed of the pedestrian flow directly affects the accuracy of image collection, and the spatial distribution characteristics of speed are related to the movement situation of each individual in the pedestrian flow. When individuals stand or gather in a crowd in the pedestrian flow, it affects the analysis process of the speed.

[0067] Furthermore, the density in space-time is ρ, which is given by:

[0068]

number

[0069] Here, the space covered by surveillance is divided into areas, and A m represents the area of ​​the mth area, and N t represents the number of pedestrians in the study area at time t. Different passageways have different methods for pedestrian flow density statistics. The density and speed of pedestrians are analyzed in the polygon formed by pedestrians, and the average density and average speed in the polygon are calculated. The polygon statistical method is basically consistent with the statistical results of traditional statistical methods, but it is very time-consuming. Polynomials are used to perform pedestrian coordinates, two-dimensional kernel density estimation, relative density distribution map, and pedestrian position preference feature fitting, and density spatial change characteristics are obtained from pedestrian coordinates, density statistics, maximum density change trend, and density spatial distribution. Special behaviors such as waiting, loitering, turning around, hugging, and running are removed separately. Pedestrian coordinate space-time data, spatial distribution of speed, and linear interpolation function simulating average nearest neighbor distance analysis are used to obtain the speed spatial component.

[0070] In a specific use of the present invention, the system includes a data collection module, a data storage module, a people flow control module, an image analysis module, a target recognition module, a model analysis module, and a linkage analysis module. The data collection module collects people flow data of different passage spaces of people flow movement through monitoring collection ports, and records the collection processes of different monitoring collection ports as a node. The image analysis module and the target recognition module dynamically analyze the movement trajectories of pedestrian targets in the people flow. Further, the model analysis module analyzes a model matching the people flow movement rule structure in different nodes, and analyzes the people flow movement characteristics in the corresponding space through different monitored image information, and the people flow movement characteristics and research models of different passage spaces are different. The interlocking analysis module performs overall dynamic analysis on the evacuation network consisting of all nodes, firstly combines with the evacuation end position of the people flow movement to establish an objective function, then analyzes the correlation between the interactions between different nodes to obtain a correlation coefficient, then analyzes the influence between the mutually influencing evacuation routes to obtain the linkage value between different evacuation routes, finally establishes a dynamic observation equation, combines and analyzes the objective function, the people flow control module performs remote control on the people flow according to the monitoring result of the interlocking analysis module, performs dynamic analysis on the entire node through the interlocking analysis module, greatly improves the efficiency of intelligent people flow remote control, the people flow control module sends corresponding people flow control commands according to the dynamic changes of the people flow, and combines the micro model analysis of people flow with the overall cascade analysis process through the combination of the model analysis module and the interlocking analysis module, reduces the error of people flow statistical control, and makes the analysis result of people flow movement more consistent with the actual people flow movement regulation.

[0071] Although the present invention has been described in more detail based on specific embodiments, those skilled in the art can easily understand that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent modifications or substitutions to the relevant technical features, and all technical suggestions after such modifications or substitutions fall within the scope of protection of the present invention.

Claims

1. It includes a data collection module, a data storage module, a people flow control module, an image analysis module, a target recognition module, a model analysis module, and a linkage analysis module. The people flow control module collects statistics on all evacuation routes in the building and labels the evacuation routes. i Let i ∈ [1, n], n is the number of evacuation routes in the building, and n evacuation routes intersect to form an evacuation net. The data collection module collects people flow data in different passage spaces through N monitoring collection ports, stores the collected people flow data in a data storage module, and stores the collected people flow data in a route r i The number of monitoring collection ports is R i and R i ∈[1, N], covering the different numbers of inlets of the people flow movement space by the monitoring collection ports; Collecting image information of people flow movement through a monitoring collection port, the image analysis module performs image frame processing on the image information from different monitoring collection ports, and the target recognition module performs target tracking on the moving target in the people flow, and the image analysis module and the target recognition module map the pixel coordinates of the image into a multi-dimensional spatial coordinate system; The model analysis module combines the analysis process of the model analysis module and the target recognition module to perform data statistics on the changes in people flow in the passage space, and then constructs a corresponding micro model based on the speed, density and flow rate structure of people flow movement. For people flow movement spaces with different numbers of entrances, different micro models are used to analyze the movement rules of people flow in space-time, and the speed-time distribution characteristics and density-time distribution characteristics of people flow movement in the passage are obtained; The above-mentioned linkage analysis module is coupled to the evacuation end position of the people flow movement to establish an objective function, and performs analysis on the feedback relationship between N nodes. The linkage analysis module further performs macro-linkage analysis on the people flow movement path characteristics of all passages in the evacuation network, and the people flow movement space covered by one monitoring collection port is regarded as one node in the people flow movement path, and R i nodes are included, and the feature matrix of the entire evacuation route is obtained by extracting the spatiotemporal distribution features of the speed and the spatiotemporal distribution features of the density of different nodes. Then, dynamic analysis is performed on the connections between different evacuation routes to obtain the influence value between the evacuation routes. Finally, the linkage analysis module predicts the evacuation process of people flow through the established dynamic observation equation, The people flow control module is combined with the prediction results of the linkage analysis module to direct the evacuation of people within the building, so that people flow movement during emergency evacuation follows the optimal evacuation route. This is a remote intelligence people flow statistical control system based on the Internet of Things (IoT).

2. The interlocking analysis module first determines the end point of the pedestrian flow, and then performs a self-organized evacuation movement of pedestrians at different positions. Then, the interlocking analysis module performs dynamic observation through a comprehensive analysis of the evacuation network consisting of all nodes. Specifically, Let X be the spatial distance between the end point of the evacuation and each node. i , i ∈ [1, N], the linkage analysis module determines the objective function f(t) by combining the node data based on the end point position of the people flow movement. The node data includes spatial distance, people flow rate and time, and the end point coordinate is (x * , y * , z * ), the spatial coordinates of the node are (x, y, z), [0022] Step 1: determining the process of people flow movement of different evacuation routes based on an objective function; The people flow confluence node in the people flow movement path is the current node p, and the correlation coefficient between the current node p and the adjacent node q is calculated. S1 and S2 are the feature vectors between the current node p and the adjacent node q, respectively. The calculation formula is expressed as follows: [0023] Step 2, where ρ represents correlation coefficient, Cov represents covariance, and E represents mathematical expectation. Based on the value of the correlation coefficient, the correlation between adjacent nodes is judged. If ρ>0, the people flow movement rule between the nodes is positive feedback, and if ρ<0, the people flow movement rule between the nodes is negative feedback. The node where two evacuation routes intersect is defined as an overlap node, and the pedestrian flows of different evacuation routes are merged at the overlap node. The model analysis module then performs a micro-analysis of the pedestrian movement rules at the overlap node, expresses the degree of change in the pedestrian movement at the overlap node of the evacuation routes as a linkage value, and calculates the linkage value of the pedestrian movement between different evacuation routes to obtain the W ij where i and j represent the subscripts of the path, i, j ∈ [1, n], R i and R j are the numbers of nodes in the two evacuation routes, i and j, respectively, ##EQU00024## Here, U i (k) and U ｋ j (k) is the flow rate of a node adjacent to the overlapping node and having the same inflow direction of people flow, and the two nodes belong to two evacuation routes, respectively. k is the number of intersecting routes, and X k , X iｋ , X jｋ Step 3 is the spatial distance, The interlocking analysis module combines the analysis results of the model analysis module to analyze the spatiotemporal distribution characteristics of the speed and density of different nodes on all evacuation routes to obtain the corresponding feature matrix A nRi is obtained, and R i ∈[1, N], the linkage analysis module combines the linkage degree between different evacuation routes to dynamically observe the dynamic states of different nodes in the process of people flow, and makes judgments on the observed values; [0025] Where: [0026] η １ is the speed, η ２ is the density, η ３ is the flow rate, η ４ is the covered space, t i The remote intelligence people flow statistical control system based on the Internet of Things (IoT) according to claim 1, further comprising: step 4, wherein the people flow evacuation time, t, represents different time values, and the dynamic prediction model predicts the parameter fluctuations of the evacuation process based on the changes of the predicted values ​​at different times, and combines with the set objective function to determine the predicted values.

3. 2. The remote intelligent people flow statistics and control system based on Internet of Things (IoT) according to claim 1, wherein the target recognition module recognizes pedestrians in the monitored video, and then converts the coordinates of the successfully recognized pedestrians and corrects their positions based on the recognition effect judgment; the image analysis module and the target recognition module convert the coordinates of the collected data, and then the model analysis module analyzes and calculates the people flow density and the people flow speed in the spatial area, and obtains a basic map of the space-time distribution of the pedestrian flow speed and the space-time distribution of the pedestrian flow density through an analysis model of the space-time distribution of the pedestrian target position, and further analyzes the movement rules of the people flow through the basic map.

4. The remote intelligence people flow statistical control system based on the Internet of Things (IoT) as claimed in claim 1, characterized in that the model analysis module builds different models based on the types of passage spaces to analyze people flow movement rules, and classifies passages including Y-shaped passages, straight passages, T-shaped passages and cross passages according to the number of people flow junctions in the area covered by the current monitoring collection port.

5. The relationship between the bottom and center coordinates of the detection frame of the target analysis module is expressed by the following formula: [0027] Here, u １ is the horizontal coordinate of the center of the detection frame in the pixel coordinate system, and v １ The remote intelligence people flow statistical control system based on the Internet of Things (IoT) according to claim 3, characterized in that x is the ordinate of the center of the detection frame in the pixel coordinate system, h is the height of the detection frame in the pixel coordinate system, u is the abscissa of the center of the bottom of the detection frame in the pixel coordinate system, and v is the ordinate of the center of the bottom of the detection frame in the pixel coordinate system.

6. The image analysis module converts the coordinates of the pixel coordinate system of the pedestrian in the image, and the coordinates are expressed by the following equation: [0028] where M is the transformation matrix, x w is the abscissa in the spatial coordinate system, and y w is the ordinate in the spatial coordinate system, and z w The remote intelligence people flow statistical control system based on the Internet of Things (IoT) according to claim 3, characterized in that z is the vertical coordinate under the spatial coordinate system, and z is the distance of people distance monitoring under the camera coordinate system.

7. The pedestrian flow velocity analysis relationship is expressed by the following equation: [0029] where v i (t) represents the instantaneous speed of pedestrian i at time t, and w i (t) represents the walking distance of pedestrian i from time t to (t+T), N represents the total number of pedestrians, [0030] where v － (where " - " indicates the overline of v) represents the time-averaged velocity, and X i，t The remote intelligence people flow statistical control system based on the Internet of Things (IoT) according to claim 3, characterized in that, x represents the coordinates of the t-th frame of pedestrian i.

8. The density in space-time is ρ, which is given by: [0031] Here, the space covered by surveillance is divided into areas, A m represents the area of ​​the mth area, and N t The remote intelligence people flow statistical control system based on the Internet of Things (IoT) according to claim 3, characterized in that: m represents the number of pedestrians in the study area at time t.

Images