People flow analysis system using IoT information collection

Abstract

The present invention discloses a people flow analysis system using Internet of Things (IoT) information collection. The system includes a data collection module, a data storage module, a people flow statistics module, an obstruction analysis module, a combined analysis module, and a predictive analysis module. The data collection module collects all human flow data in the laboratory's machine room through an image collection device and stores the people flow data in the data storage module. The people flow statistics module analyzes crowd densities in different covered spaces and determines the average distance through hierarchical division and cluster analysis of the human head label points, and further determines the human head dimensions using a Gaussian function. When the crowd density is low and not average, the k-nearest neighbor method is used to solve the uncertainty problem. The obstruction analysis module combines the features of obstructions in the covered space with the features of adjacent crowds to obtain a new data set. The predictive analysis module performs overall people flow prediction analysis based on the combined new data set and the people flow data of crowds not adjacent to the obstructions, thereby greatly improving the accuracy of people flow analysis and the efficiency of control.

Description

[Technical Field]

[0001] The present invention relates to the technical field of intelligent control, and more particularly to a people flow analysis system using Internet of Things (IoT) information collection. [Background technology]

[0002] The people flow analysis system analyzes the dynamic changes in people flow at different time slots through pedestrian count statistics and crowd spatial distribution, and then performs data analysis to better understand pedestrian behavioral habits and take appropriate safety measures to reduce pedestrian congestion. In the laboratory machine room, the experimental equipment obstructs pedestrians, reducing the accuracy of people flow statistical analysis. The crowd distribution density in different spaces varies, and the closer the crowd distribution density, the more accurate the model analysis results. However, during peak traffic periods, the people flow density changes rapidly, reducing the accuracy of the model analysis. Crowds at different locations in the same space interact with each other, making congestion highly likely. When analyzing crowds in the entire space, the accuracy of the analysis is affected by obstructions. Accurate analysis based on the distance between crowds is not suitable for all crowd density analyses. To improve traffic efficiency and meet the requirements for accuracy and real-time performance in the people flow analysis process, the present invention proposes a people flow analysis system using Internet of Things (IoT) information collection. Summary of the Invention [Means for solving the problem]

[0003] In order to overcome the shortcomings of the prior art in the above situation, the present invention provides a people flow analysis system using Internet of Things (IoT) information collection. The system's people flow statistics module analyzes crowd density in different covered spaces, determines the average distance through hierarchical division and cluster analysis of the human head label points, and further determines the human head dimensions using a Gaussian function. When the crowd density is low and not average, the uncertainty problem is solved by the k-nearest neighbor method. The obstruction analysis module combines the features of obstructions within the covered space with the features of adjacent crowds to obtain a new data set. The predictive analysis module performs overall people flow predictive analysis based on the combined new data set and the people flow data of crowds not adjacent to obstructions, thereby greatly improving the efficiency of remote people flow analysis and control. Overall people flow analysis can be performed on data collected by different Internet of Things (IoT) collection devices.

[0004] A people flow analysis system using Internet of Things (IoT) information collection, comprising: a data collection module, a data storage module, a people flow statistics module, an obstruction analysis module, a connection analysis module, and a predictive analysis module; The data collection module collects all the human flow data in the laboratory's machine room through an image collection device, and stores the human flow data in a data storage module; The image collection devices are installed at different positions in the machine room of the laboratory, and the machine room is divided into different cover spaces. After that, the people flow statistics module performs crowd density analysis on the people flow data in each different cover space, and obtains the corresponding loss function, crowd density, and people flow rate. The loss function represents the error between the model analysis result of the people flow statistics model and the actual crowd value. The combined analysis module combines the analysis results of the people flow statistics module to analyze the changes in crowd position to obtain the combined degree of crowds in different covered spaces, and calculates the exchange rate of crowd numbers in adjacent covered spaces; The obstruction analysis module analyzes the loss function of the statistical analysis process and the influence of the experimental equipment on the change in the direction of people flow to obtain obstruction factors, and then combines the coupling degree obtained by the analysis of the coupling analysis module to perform location analysis of the obstruction factors for people flow in the machine room of the laboratory; The obstruction analysis module performs feature fusion of crowd features in the laboratory and close to the obstruction location based on the obstruction factor analysis result, and obtains a new dataset by merging the features of the obstruction location and the crowd location; The predictive analysis module performs crowd density analysis based on the analysis result of the obstruction analysis module and the new fused data set to obtain a crowd density map for the entire space, and then performs crowd density prediction based on the density map to obtain a prediction result. Based on the prediction result, the safety of the people flow density within the space is evaluated and management measures for the experimental equipment are formulated.

[0005] The image collecting device collects people flow data in the machine room from different angles, and the people flow statistics module uses the head characteristics of the human body to detect and count the crowd through the generated density map, specifically: The head size of each person is inversely proportional to the distance of the image acquisition device, and the label point x i Step 1: The position of the human head in the image is represented using [theta]. The labeled points in the image are divided into different layers according to the distance, and different layers correspond to different weighted averages. The positions of the labeled points in different layers are subjected to cluster analysis to obtain different cluster centers, and the number of cluster centers is the same as the number of layers of the labeled points. Convert the crowd image labels into a crowd density map based on different weighted averages,

[0006]

number

[0007] x i The human head is placed at the position of the Dirichlet function δ(xx i ) where N represents the number of labeled points on the human head. Step 2 performs head simulation using a radial basis function. The radial basis function spreads the human head into a range, and the range value depends on the head dimensions.

[0008]

number

[0009] where:

[0010]

number

[0011] where i is the subscript of the label point, β i is the weighted average of the label points in different layers, n is the number of cluster centers, G σi (x) is the radial basis function, d i - (where " - " is d i Step 3: The overline indicates the average distance from the label point to the cluster center within one layer. Step 4 includes using a deep learning neural network to train all label points and output a density map, and then performing an integral operation on the density map to obtain the number of people.

[0012] The obstruction analysis module analyzes the people flow pattern to determine the obstruction factors that affect the people flow in the machine room. Different obstructions are included in different coverage areas. The specific analysis process is as follows: Based on the spatial distribution of crowds, a biomathematical equation of state is established, and the activity factors of people flow diffusion are extracted. The dynamic equation of crowd diffusion is:

[0013]

number

[0014] where Q is the flow rate, ρ is the crowd density, and V is the average velocity. When the average speed is maximum,

[0015]

number

[0016]

number

[0017] Here, ρ(x,t) is the density of x locations at time t, m is the total number of people in the measurement area, and c(x j (t)-x) is a Gaussian function, x j (t) represents the coordinate vector of pedestrian j at time t, and the measurement area is x j Step 2 depends on the position of (t) and x, Step 3: Analyzing and extracting obstruction factors based on the change of position vectors in the dynamic system of crowd diffusion, then performing feature extraction on the image, extracting object features of the obstruction objects in the image, and determining the location of the obstruction objects in the machine room based on the obstruction factors; The target features of the crowd adjacent to the obstacle and the features of the obstacle are fused, and calculation is performed based on the pixel point position, and the coordinates of the heads of people in the crowd are calculated as x i Let x i and the position of the pixel point of the adjacent obstruction is y i Let,different positions correspond to different feature vectors, and the feature fusion formula is as follows:

[0018]

number

[0019]

number

[0020] where x1 and y2 are feature vectors of the same dimension, y∈Y is the fused feature mapping set, and W(x1+y2)∈R n,Step 4, in which a new dataset is obtained by fusing the ,obstacles with the neighboring crowds through feature fusion.

[0021] The data collection module includes different image collection devices, and the data collection module numbers the image collection devices, and then installs the image collection devices with different numbers at different positions in the laboratory's machine room. The coverage areas of the different image collection devices include overlapping areas and non-overlapping areas. The joint analysis module analyzes the crowd distribution image of the overlapping areas, and the people flow statistics module analyzes the people flow distribution rules of the non-overlapping areas.

[0022] The people flow statistics module extracts crowd features within the coverage space of the image collection device, and then performs data training based on the people flow data and density map to obtain the training loss error between the predicted crowd and the actual crowd. The loss error formula is as follows:

[0023]

number

[0024] where N is the number of images, and F(x;θ) and F represent the points in the density map.

[0025] In the laboratory, obstacles that impede people's walking exert the same pressure level on the adjacent crowds, and the obstruction analysis module analyzes the obstructions and people flow as a whole by performing a fusion analysis on the data set of the crowds adjacent to the obstructions. When the prediction analysis module analyzes the entire crowd within the entire covered space, it takes into account the impact of different parts of the obstructions on people flow.

[0026] The predictive analysis module performs overall crowd density analysis based on the data collected by all image collection devices within the space, and the obstruction analysis module and the crowd flow statistics module obtain a complete data set by analyzing the crowd density under different cover spaces. The predictive analysis module further performs crowd density prediction at different times for the data set, and analyzes the pedestrian flow movement path within the entire space based on the prediction result.

[0027] The joint analysis module analyzes the transformation of crowd density in different cover spaces, obtains the changes in the number of crowds at different times through the people flow statistics module, and determines the position changes of the people flow based on the start and end cover spaces of the people flow. [Effects of the Invention]

[0028] By adopting the above technical means, the present invention has the following advantages over the prior art: First, the people flow statistics module of the present invention allocates layers to the head label points, then performs cluster analysis on the label points of different layers, calculates the average distance between the head label point at the center of the cluster and the adjacent head label points, avoids incompatibility with other data sets when the distance coefficient is selected, and solves the uncertainty problem of the k-nearest neighbor method when the crowd density is low and not average, thereby improving the accuracy of head size estimation. Secondly, the obstruction analysis module of the present invention determines obstruction factors that obstruct people flow movement based on the dynamic analysis process of people flow movement in different covered spaces, determines the location of obstructions that obstruct people flow movement within the covered space based on the obstruction factors, and performs feature fusion between the crowd feature vectors adjacent to the obstructions and the feature vectors of the obstructions to obtain a new crowd data set. The obstruction analysis module analyzes the obstruction pressure that the obstructions bring to people flow movement through feature fusion, thereby making the predictive analysis module have higher accuracy and real-time performance in the entire people flow analysis process. Third, the predictive analysis module of the present invention combines people flow data collected at different locations to perform overall predictive analysis. The crowd density distribution in crowd images collected at different locations is different, and all people's movement trajectories in the same space will influence each other. The predictive analysis module combines the analysis results of the obstruction analysis module to greatly improve the efficiency of people flow prediction, and take timely management measures for safety risks that arise when people flow is densely packed. [Brief explanation of the drawings]

[0029] [Figure 1] 1 is an overall analysis flowchart according to the present invention. [Figure 2] FIG. 1 is an overall module diagram according to the present invention. [Figure 3] 3 is an analysis flowchart of a people flow statistics module according to the present invention; [Figure 4] 1 is an analysis flowchart of an inhibition analysis module according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0030] 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 4. The embodiments and features of 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 art of the present invention.

[0031] In order to meet the requirements of real-time performance and strengthen the requirements for real-time control of people flow, the people flow analysis system must process large amounts of data in a short period of time and provide analysis results in a timely manner. It uses data analysis, statistics, and visual investigation methods to predict the flow of people within the machine room space and carry out more real-time statistical analysis and management of people flow. The distribution of people flow within the space is of great significance to the spatial distribution of people flow within the machine room. The experimental equipment in the machine room of a school's laboratory provides shielding for people flow. When students leave the machine room, analyzing the people flow movement rules can optimize the students' movement routes, improve traffic efficiency within the machine room, and avoid traffic congestion. At the same time, analyzing different people flow movement statistics within the machine room can help machine room managers improve the installation efficiency of machine room facilities.

[0032] The crowd density distribution in different cover spaces is different, and when the crowd densities are close, the model analysis process is more accurate. However, when the flow of people within the entire space changes and an unexpected situation occurs, the degree of change in crowd density will change, which will greatly reduce the accuracy of the model analysis process.

[0033] Within the same space, connections occur between people flow data from different covered spaces, and the degree of data connection under different analysis models can be obtained through calculations. When predicting and analyzing people flow within the entire space, obstacles exert pressure on the direction of people flow, changing the movement trajectory of the people flow. To analyze the impact of obstacles, a new dataset is obtained by feature fusion of the characteristics of the obstacles and the characteristics of the adjacent crowd. This weakens the analysis of obstacles when predicting the entire people flow, but takes into account the impact of obstacles, and models are trained on different training datasets, so the accuracy of the analysis results will also differ. The predictive analysis module makes predictions based on the fused dataset and obtains an analysis result of the people flow trajectory that fits the boundary of the obstacle. Based on the results of people flow analysis, the manager can optimize allocation and more accurately grasp the demands and behavioral characteristics of different people, thereby improving resource utilization efficiency. By analyzing the movement patterns and behavioral habits of pedestrians within the laboratory's machine room, future pedestrian flow rates and traffic conditions can be predicted, providing decision-making support for the laboratory manager and allowing corresponding management measures to be formulated in advance.

[0034] An embodiment of the present invention discloses a people flow analysis system using Internet of Things (IoT) information collection, which includes a data collection module, a data storage module, a people flow statistics module, an obstruction analysis module, a connection analysis module, and a prediction analysis module; The data collection module collects all the human traffic data in the laboratory's machine room through an image collection device, and then stores the human traffic data in a data storage module. The image collection device includes a camera, a thermal sensor, and an infrared device. The image collection device at different positions has different distances to the crowd location, and the clarity of the collected images is different. The degree of obstruction caused by the experimental equipment in the laboratory's machine room in different covered spaces is also different. The image collection devices are installed at different positions in the laboratory's machine room, and the machine room is divided into different cover spaces. After that, the people flow statistics module performs crowd density analysis on the people flow data in each different cover space to obtain the corresponding loss function, crowd density, and people flow rate. The loss function represents the error between the model analysis results of the people flow statistics model and the actual crowd values. The crowd density distribution in different cover spaces is different, and the people flow statistics module performs analysis on images of different crowd densities, greatly improving the accuracy and real-timeness of the people flow analysis. The coupling analysis module combines the analysis results of the people flow statistics module to analyze the changes in crowd position to obtain the degree of crowd coupling in different covered spaces, and calculates the exchange rate of the crowd numbers in adjacent covered spaces. The machine room of the experiment room is a fixed space, and students use experimental equipment to carry out experimental learning. The changes in the people flow movement law during non-peak times are gradual, and the crowd densities in different covered spaces are mutually transformed, and the degree of coupling is used to describe the change degree of people flow in different covered spaces. The obstruction analysis module analyzes the loss function of the statistical analysis process and the influence of the experimental equipment on the change in the direction of people flow to obtain obstruction factors, and then combines the coupling degree obtained by the analysis of the coupling analysis module to perform location analysis of the obstruction factors for people flow in the machine room of the laboratory; The obstruction analysis module performs feature fusion of crowd features in the laboratory and close to the obstruction location based on the obstruction factor analysis result, and obtains a new data set by fusion of the obstruction location and the crowd location feature. When the prediction analysis module performs overall analysis based on all the traffic data, although there is no image collection device that can collect all the unobstructed crowd distribution images, within the entire space of the laboratory, the flow direction of people within the entire space during peak periods will follow the prediction analysis result, and the obstruction analysis module processes the pressure on people flow movement caused by obstructions in the overall analysis process. The predictive analysis module performs crowd density analysis based on the analysis result of the obstruction analysis module and the new fused data set to obtain a crowd density map for the entire space, and then performs crowd density prediction based on the density map to obtain a prediction result. Based on the prediction result, the safety of the people flow density within the space is evaluated and management measures for the experimental equipment are formulated.

[0035] In addition, the image collecting device collects people flow data in the machine room from different angles, and the people flow statistics module uses the head characteristics of the human body to detect and count the crowd through the generated density map, specifically: If the k-nearest neighbor method is used to calculate the average distance between all label points and adjacent human head label points in the cover space, and then the distance coefficient is selected based on the average value, the suitability for other datasets will be reduced. When the crowd density is low and not average, the k-nearest neighbor method will cause uncertainty issues, which will further affect the accuracy of head size estimation. Therefore, it is preferable to first perform hierarchical division and cluster analysis on the label points in the cover space in advance, and then calculate the average distance. The head size of each person is inversely proportional to the distance of the image collection device, and the label point x iStep 1: The position of the human head in the image is represented using [theta]. The labeled points in the image are divided into different layers according to the distance, and different layers correspond to different weighted averages. The positions of the labeled points in different layers are subjected to cluster analysis to obtain different cluster centers, and the number of cluster centers is the same as the number of layers of the labeled points. Convert the crowd image labels into a crowd density map based on different weighted averages,

[0036]

number

[0037] x i The human head at the position is calculated by the Dirichlet function δ(xx i ) where N represents the number of labeled points on the human head. Step 2 performs head simulation using a radial basis function. The radial basis function spreads the human head into a range, and the range value depends on the head dimensions.

[0038]

number

[0039] where:

[0040]

number

[0041] i is the subscript of the label point, β i is the weighted average of the label points in different layers, n is the number of cluster centers, G σi (x) is the radial basis function, d i - (where " - " is d i (showing the overline of ) represents the average distance from the label point to the cluster center in one layer, and σ i Step 3 determines the accuracy of the Gaussian convolution; Step 4 includes using a deep learning neural network to train all label points and output a density map, and then performing an integral operation on the density map to obtain the number of people.

[0042] During peak times of people flow, the people flow in the entire space corresponding to the machine room of the laboratory must be analyzed in order to realize remote control of the entire people flow through the people flow rules. However, the data in the space collected by the image collection device has a high occlusion rate. Therefore, the obstruction analysis module can obtain a new data set that combines obstructions and crowd characteristics. The obstruction analysis module analyzes the people flow rules to determine the obstruction factors that affect the people flow in the machine room. Different obstructions are included in different coverage areas. The specific analysis process is as follows: Based on the spatial distribution of crowds, a biomathematical equation of state is established, and the activity factors of people flow diffusion are extracted. The dynamic equation of crowd diffusion is:

[0043]

number

[0044] where Q is the flow rate, ρ is the crowd density, and V is the average velocity. When the average speed is maximum,

[0045]

number

[0046]

number

[0047] Here, ρ(x,t) is the density of x locations at time t, m is the total number of people in the measurement area, and c(x j (t)-x) is a Gaussian function, x j (t) represents the coordinate vector of pedestrian j at time t, and the measurement area is x j (t) and the position of x,

[0048]

number

[0049]

number

[0050] Here, the range of values ​​for R0 depends on the head dimensions of the label points in step 2. In the dynamic change process of crowd diffusion, the pressure change amount of the obstruction object corresponding to the time of the change in the dynamic factor is recorded as the obstruction factor, and the obstruction degree at different positions of the obstruction object is different, and the experimental equipment in the laboratory room has an obstruction effect on the walking of students. Therefore, the obstruction factor is analyzed and extracted based on the change in the position vector in the dynamic system of crowd diffusion, and then feature extraction is performed on the image to extract the object features of the obstruction object in the image, and the position of the obstruction object in the machine room is determined based on the obstruction factor. Step 3; The target features of the crowd adjacent to the obstacle and the features of the obstacle are fused, and calculation is performed based on the pixel point position, and the coordinates of the heads of people in the crowd are calculated as x i Let x i and the position of the pixel point of the adjacent obstruction is y i Let,different positions correspond to different feature vectors, and the feature fusion formula is as follows:

[0051]

number

[0052]

number

[0053] where x1 and y2 are feature vectors of the same dimension, y∈Y is the fused feature mapping set, and W(x1+y2)∈R n ,A new dataset is obtained by fusing the obstacles and the ,neighboring crowds through feature fusion.

[0054]

number

[0055] w i Step 4 includes,where,is the fusion weight and,i,is a subscript.

[0056] The data collection module includes different image collection devices, and the data collection module numbers the image collection devices, and then installs the image collection devices with different numbers at different positions in the laboratory's machine room. The coverage areas of the different image collection devices include overlapping areas and non-overlapping areas. The joint analysis module analyzes the crowd distribution image of the overlapping areas, and the people flow statistics module analyzes the people flow distribution rules of the non-overlapping areas.

[0057] The people flow statistics module extracts crowd features within the coverage space of the image collection device, and then performs data training based on the people flow data and density map to obtain the training loss error between the predicted crowd and the actual crowd. The loss error formula is as follows:

[0058]

number

[0059] Here, N is the number of images, F(x;θ) and F represent points in the density map, and the density distribution of the crowd varies depending on the cover space, and the loss error generated during the corresponding data analysis is also different.

[0060] In the laboratory, obstacles that obstruct people's walking exert the same pressure level on the adjacent crowd. When the crowd distribution characteristics change, the obstruction pressure exerted by the obstruction position on the crowd does not change, so the impact of the obstacles does not change during the overall crowd flow prediction analysis. The obstruction analysis module performs a fusion analysis on the data set of the crowd adjacent to the obstacle to analyze the obstructions and the crowd flow as a whole. When the prediction analysis module analyzes the entire crowd within the entire covered space, it takes into account the impact of different parts of the obstructions on the crowd flow, and calculates the obstruction pressure using the following formula:

[0061]

number

[0062]

number

[0063] Here, <·> represents the average value at the observation center point x, and Var(V) represents the amount of change in the crowd's velocity.

[0064] The predictive analysis module performs overall crowd density analysis based on the data collected by all image collection devices within the space, and the obstruction analysis module and the crowd flow statistics module obtain a complete data set by analyzing the crowd density under different cover spaces. The predictive analysis module further performs crowd density prediction at different times for the data set, and analyzes the pedestrian flow movement path within the entire space based on the prediction result.

[0065] The coupling analysis module analyzes the transformation of crowd density in different cover spaces, obtains the changes in the number of crowds at different times through the people flow statistics module, and determines the changes in the position of the people flow based on the start and end cover spaces of the people flow. There is also a coupling relationship between the obstacles detected in different cover spaces, and the coupling analysis module analyzes the overlapping and connecting parts between the obstacles.

[0066] Specifically, when the present invention is used, the system includes a data collection module, a data storage module, a people flow statistics module, an obstruction analysis module, a combined analysis module, and a predictive analysis module. The data collection module collects all the human flow data in the laboratory's machine room through an image collection device, and then stores the people flow data in the data storage module. The people flow statistics module of the system analyzes the crowd density in different covered spaces, determines the average distance through hierarchical division and cluster analysis of the human head label points, and further determines the human head dimensions using a Gaussian function. When the crowd density is low and not average, the uncertainty problem is solved by the k-nearest neighbor method. The obstruction analysis module combines the features of the obstructions in the covered space with the features of the adjacent crowds to obtain a new data set. The predictive analysis module performs overall people flow predictive analysis based on the combined new data set and the people flow data of the crowds not adjacent to the obstructions, thereby greatly improving the efficiency of remote people flow analysis and control. This avoids data occlusion when performing overall people flow analysis on the data collected by different Internet of Things (IoT) collection devices.

[0067] Although the present invention has been described in more detail above based on specific embodiments, it is readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent modifications or substitutions to the relevant technical features, and all technical proposals after these modifications or substitutions fall within the scope of protection of the present invention.

Claims

1. The system includes a data collection module, a data storage module, a people flow statistics module, an obstruction analysis module, a connection analysis module, and a prediction analysis module; The data collection module collects all the human flow data in the laboratory's machine room through an image collection device, and stores the human flow data in a data storage module; The image collection devices are installed at different positions in the laboratory's machine room, and the machine room is divided into different cover spaces. Then, the people flow statistics module performs crowd density analysis on the people flow data in the different cover spaces to obtain corresponding loss functions, crowd density, and people flow rate. The loss functions represent the error between the model analysis results of the people flow statistics model and the actual crowd values. The combined analysis module combines the analysis results of the people flow statistics module to analyze the changes in crowd position to obtain the combined degree of crowds in different covered spaces, and calculates the exchange rate of crowd numbers in adjacent covered spaces; The obstruction analysis module analyzes the loss function of the statistical analysis process and the influence of the experimental equipment on the change in the direction of people flow to obtain obstruction factors, and then combines the coupling degree obtained by the analysis of the coupling analysis module to perform location analysis of the obstruction factors for people flow in the machine room of the laboratory; The obstruction analysis module performs feature fusion of crowd features in the laboratory and close to the obstruction location based on the obstruction factor analysis result, and obtains a new dataset by merging the features of the obstruction location and the crowd location; The prediction analysis module performs a crowd density analysis based on the analysis result of the obstruction analysis module and the new fusion data set to obtain a crowd density map for the entire space, and then predicts the crowd density based on the density map to obtain a prediction result. The prediction result is used to evaluate the safety of the crowd flow density in the space and formulate management measures for the experimental equipment. This is a people flow analysis system using Internet of Things (IoT) information collection.

2. The image collecting device collects people flow data in the machine room from different angles, and the people flow statistics module uses the head characteristics of the human body to detect and count the crowd through the generated density map, specifically: The head size of each person is inversely proportional to the distance of the image acquisition device, and the label point x i Step 1: Represent the position of the human head in the image using [mathematical formula - see original document], divide the labeled points in the image into different layers according to the distance, and different layers correspond to different weighted averages. Then, perform cluster analysis on the positions of the labeled points in different layers to obtain different cluster centers, and the number of cluster centers is the same as the number of layers of the labeled points. Convert the crowd image labels into a crowd density map based on different weighted averages, [0000] x i The human head at the position is calculated by the Dirichlet function δ(xx i ) where N represents the number of labeled points on the human head. Step 2 performs head simulation using a radial basis function. The radial basis function spreads the human head into a range, and the range value depends on the head dimensions. [Equation 25] where: [Equation 26] i is the subscript of the label point, β i is the weighted average of the label points of different layers, n is the number of cluster centers, G σi (x) is the radial basis function, d i - (where " - " is d i Step 3: The overline indicates the average distance from the label point to the cluster center within one layer. and step 4 of using a deep learning neural network to train all label points to output a density map, and then performing an integral operation on the density map to obtain the number of people.

3. The obstruction analysis module analyzes the people flow pattern to determine the obstruction factors that affect the people flow in the machine room. Different obstructions are included in different coverage areas. The specific analysis process is as follows: Based on the spatial distribution of crowds, a biomathematical equation of state is established, and the activity factors of people flow diffusion are extracted. The dynamic equation of crowd diffusion is: [0000] Here, Q is the flow rate, ρ is the crowd density, and V is the average speed. When the average speed is maximum, [0000] [0000] Here, ρ(x,t) is the density of x locations at time t, m is the total number of people in the measurement area, and c(x j (t)-x) is a Gaussian function, x j (t) represents the coordinate vector of pedestrian j at time t, and the measurement area is x j Step 2 depends on the position of (t) and x, Step 3: Analyzing and extracting obstruction factors based on the change of position vectors in the dynamic system of crowd diffusion, then performing feature extraction on the image, extracting object features of obstructions in the image, and determining the position of the obstructions in the machine room based on the obstruction factors; The target features of the crowd adjacent to the obstacle and the features of the obstacle are fused, and calculation is performed based on the pixel point position, and the coordinates of the heads of people in the crowd are calculated as x i Let x i and the position of the pixel point of the adjacent obstruction is y i Let,different positions correspond to different feature vectors, and the feature fusion formula is as follows: [Equation 30] [Equation 31] where x １ and y ２ are feature vectors of the same dimension, y∈Y is the feature mapping set after fusion, W(x 1 +y 2 )∈R n and step 4, in which a new data set is obtained by fusing the obstructions and the nearby crowds through feature fusion.

4. 2. The people flow analysis system using Internet of Things (IoT) information collection according to claim 1, wherein the data collection module comprises different image collection devices, the data collection module numbers the image collection devices, and then installs the image collection devices with different numbers at different positions in the laboratory's machine room, the areas covered by the different image collection devices include overlapping parts and non-overlapping parts, the joint analysis module analyzes the crowd distribution images of the overlapping parts, and the people flow statistics module analyzes the people flow distribution rules of the non-overlapping parts.

5. The people flow statistics module extracts crowd features within the coverage space of the image collection device, and then performs data training based on the people flow data and density map to obtain the training loss error between the predicted crowd and the actual crowd. The loss error formula is as follows: [Equation 32] The people flow analysis system using Internet of Things (IoT) information collection according to claim 1, wherein N is the number of images, and F(x;θ) and F represent points in the density map.

6. 4. The people flow analysis system using Internet of Things (IoT) information collection according to claim 3, wherein obstacles that impede people's walking in the laboratory exert the same pressure level on the adjacent crowds, the obstruction analysis module analyzes the obstacles and people flow as a whole by performing a fusion analysis on the data sets of the crowds adjacent to the obstacles, and the predictive analysis module takes into account the impact of different parts of the obstacles on the people flow when analyzing the entire crowd within the entire covered space.

7. 2. The people flow analysis system using Internet of Things (IoT) information collection according to claim 1, wherein the predictive analysis module performs an overall people flow density analysis based on data collected by all image collection devices within the space; the obstruction analysis module and the people flow statistics module obtain a complete data set by analyzing crowd densities under different cover spaces; the predictive analysis module further performs crowd density predictions at different times for the data set, and analyzes people flow movement paths within the entire space based on the prediction results.

8. 2. The people flow analysis system using Internet of Things (IoT) information collection according to claim 1, wherein the combined analysis module analyzes the transformation of crowd density in different cover spaces, obtains the changes in the number of crowds at different times through the people flow statistics module, and determines the changes in the location of the people flow based on the start and end cover spaces of the people flow.

Images