Highway environmental protection construction monitoring method and system based on big data

Abstract

The invention relates to the technical field of big data monitoring, in particular to a highway environmental protection construction monitoring method and system based on big data, and the method comprises the following steps: calculating the spatial distance between sensors based on the geographic position information of a data collection point and the raised dust concentration distribution and particulate matter monitoring value of a road construction environment, and calculating the spatial distance between sensors; and connecting sensor nodes with a direct adjacency relation through difference calculation and normalization processing. According to the invention, through probability density calculation and time sequence analysis, abnormal edge screening, abnormal change detection precision improvement, misjudgment reduction, abnormal edge connection node position and space density calculation, high-density area and cluster center determination, accurate positioning of pollution points is realized, and the method is suitable for the real-time monitoring of the pollution points in combination with monitoring data change trend and diffusion path tracing. The method comprises the following steps: reconstructing a pollution diffusion chain, determining a pollution propagation range, optimizing pollution treatment, accurately identifying a pollution source through weight association degree and gradient analysis between path nodes, reducing monitoring blind areas, and providing comprehensive support for construction monitoring.

Description

Technical Field

[0001] The present invention relates to the field of big data monitoring technology, and in particular to a highway environmental protection construction monitoring method and system based on big data. Background Art

[0002] The field of big data monitoring technology includes the use of big data technology to collect, store, analyze and manage various types of data to achieve dynamic monitoring of target objects or environments. The core content of this technical field is to obtain real-time dynamic information and achieve accurate monitoring and decision support through data analysis and mining. Big data monitoring technology as a whole covers modules such as data collection technology, distributed storage and computing technology, data analysis methods, and visualization presentation. It can be applied to various scenarios such as industry, transportation, and environment, and provide refined management capabilities for complex systems.

[0003] Among them, the highway environmental protection construction monitoring method refers to a specific monitoring technology method for environmental protection issues during highway construction. The patent subject is aimed at the pollution, noise and ecological damage of the surrounding environment caused by construction activities. By deploying a sensor network to collect environmental data in the construction process in real time, such as air quality data, noise intensity data and water quality index data, etc., and using a big data analysis platform to classify, dynamically analyze and detect anomalies of the collected data, and combine with a geographic information system to monitor the construction area. Specific implementation methods include obtaining real-time data through wireless sensor acquisition terminals, using a distributed computing framework for batch processing and stream processing analysis, and using spatiotemporal data models to accurately monitor and analyze the environmental impact of construction activities.

[0004] Existing technologies only deal with simple analysis of data collected in dynamic environments, which makes it difficult to effectively express complex spatial relationships and dynamic pollution patterns, resulting in insufficient positioning accuracy. Anomaly detection mostly relies on a single threshold judgment, which makes it difficult to distinguish interference or random changes, and is prone to misjudgment or omission. Diffusion path analysis lacks the ability to combine data trends and trace chains, making it difficult to clarify the scope of pollution spread and the source of impact. The lack of consistency analysis capabilities for spatiotemporal data makes it impossible to fully respond to the dynamic characteristics of construction activities, resulting in low pertinence and accuracy in pollution monitoring and governance decisions, which may result in waste of resources and expansion of monitoring blind spots. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a highway environmental protection construction monitoring method and system based on big data.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solution: a highway environmental protection construction monitoring method based on big data, comprising the following steps:

[0007] S1: Based on the geographical location information of the data collection points and the dust concentration distribution and particulate matter monitoring values ​​of the road construction environment, the spatial distance between sensors is calculated, and the sensor nodes with direct adjacent relationships are connected through difference calculation and normalization processing to establish a road construction environment weight network;

[0008] S2: Based on the road construction environment weight network, the probability density of the edge weight in the neighborhood is calculated, and the local probability offset value is obtained through difference operation. In combination with the time series data, the prominent edges are screened based on the state change and the associated edges are removed to generate the abnormal edge results of dust and particulate matter;

[0009] S3: Based on the dust and particulate matter abnormal edge results, extract the positions of the nodes connected by the abnormal edges, calculate the spatial distance and local weight density, determine the high-density area, divide the node clusters, calculate the cluster center and weight distribution, and generate the center position of the dust abnormal point cluster;

[0010] S4: Based on the center position of the dust anomaly cluster, extract the associated time series data, analyze the change trend of the particulate matter monitoring value and the atmospheric pollution diffusion coefficient, trace the node source path along the gradient direction, calculate the correlation of the path nodes, and generate the particulate matter diffusion path tracing result;

[0011] S5: Based on the tracing results of the particle diffusion path, compare the weight correlation between the path nodes, update the weight distribution between the nodes, and based on the node with the largest weight change in the path, analyze the weight gradient of the surrounding path of the node and combine it with the spatial density distribution data to generate the dust pollution source positioning result.

[0012] The road construction environment weight network includes data collection points, geographic location information, sensor nodes and dust concentration distribution. The dust and particulate matter abnormal edge results include local probability offset values, time series data, prominent edges and removed associated edges. The dust abnormal point cluster center position includes high-density areas, node clusters, cluster centers and weight distribution. The particulate matter diffusion path tracing results include node source paths, correlation, path nodes and particulate matter diffusion trends. The dust pollution source positioning results include weight correlation between path nodes, path weight gradient and spatial density distribution data.

[0013] As a further solution of the present invention, the steps of acquiring the road construction environment weight network are specifically as follows:

[0014] S111: Calculate the spatial distance between each pair of adjacent sensor nodes and the corresponding particle concentration difference through the geographical location information of the sensor and the dust concentration distribution of the monitoring point in the road construction environment, calculate the spatial coordinates of each sensor based on the location of the monitoring point, calculate the spatial distance between the nodes through the Euclidean distance formula between the coordinates, calculate the concentration difference between the nodes according to the particle concentration values ​​recorded at the monitoring point, and generate an initial adjacency matrix between the sensors;

[0015] S112: normalizing the initial adjacency relationship matrix between the sensors, normalizing the concentration difference to between 0 and 1, removing node connections exceeding the target threshold, and recalculating the weights of the remaining node connections as the ratio of the normalized concentration difference to the spatial distance, and updating the node weights to generate a normalized adjacency weight matrix;

[0016] S113: Based on the normalized adjacency weight matrix, the sensor nodes having direct adjacency relationships are connected, using the formula:

[0017]

[0018] Calculate the weight between each pair of sensor nodes to generate a road construction environment weight network;

[0019] Among them, W ij represents the weight between node i and node j, C i represents the particle concentration at node i, C j represents the particle concentration at node j, D ij represents the spatial distance between node i and node j, γ 1 is the concentration adjustment coefficient, which adjusts the influence of particle concentration on the weight, γ 2 It is the spatial adjustment coefficient, which adjusts the influence of spatial distance on the weight.

[0020] As a further solution of the present invention, the steps for obtaining the abnormal edge results of dust and particulate matter are specifically as follows:

[0021] S211: Calculate the weight probability density of each edge through the road construction environment weight network, perform weight statistics based on all edges in the neighborhood, use the weight distribution frequency as a reference to calculate the kernel density estimation, perform continuous fitting on the probability distribution of the kernel density, and generate the weight probability density distribution of the edge;

[0022] S212: Perform a difference operation on the weight probability density distribution of the edge to calculate the local probability offset value using the formula:

[0023]

[0024] Calculate the degree of deviation of each edge from the average probability in its neighborhood and generate the local probability offset value of each edge;

[0025] Where ΔP ij Represents the local probability offset value, P ij represents the weight probability density of edges i and j, represents the average weight probability density of the edges in the neighborhood of node i, σ N(i) represents the standard deviation of the edges in the neighborhood, γ 1 To avoid the adjustment parameter whose denominator tends to zero, the local probability offset value is used to measure the degree of change in the edge weight probability;

[0026] S213: Filter the significant edges of the local probability offset value of each edge in combination with the time series data, calculate the fluctuation difference of the associated edges by jointly comparing the offset value with the time series state change, eliminate the associated edges with abnormal fluctuations, and generate dust and particulate matter abnormal edge results.

[0027] As a further solution of the present invention, the step of obtaining the center position of the dust abnormal point cluster is specifically:

[0028] S311: extract all abnormal nodes from the dust and particulate matter abnormal edge results, identify the spatial position of each node, determine the spatial coordinates of the abnormal node through the node connection relationship, determine the complete position of the abnormal edge node based on the weight distribution of the node connection, and generate a node position list;

[0029] S312: performing spatial distance calculation on the node position list, calling the Euclidean distance formula based on the node spatial coordinates, calculating the spatial distance between each pair of nodes in turn, calculating the local weight density of multiple nodes based on the node weight distribution, and obtaining the spatial and weight comprehensive characteristics of each node;

[0030] S313: Based on the spatial and weight comprehensive feature data of each node, determine the high-density area of ​​the node cluster, determine the density peak of the local area based on the node weight density, identify the core node through spatial position aggregation and weight peak, and generate a node cluster;

[0031] S314: Calculate the center position and weight distribution of the node cluster using the formula:

[0032]

[0033] Calculate and generate the center position of dust anomaly cluster;

[0034] Among them, C center Represents the cluster center position, indicating the coordinates of the weighted center point of the node cluster, w i represents the weight of node i, which is used to measure the influence of the node in the cluster center calculation, xi represents the spatial position of node i, represents the coordinates of the node, and γ is an adjustment parameter used to balance the contribution of nodes with smaller weights within the cluster.

[0035] As a further solution of the present invention, the steps for obtaining the particle diffusion path tracing result are specifically as follows:

[0036] S411: based on the center position of the dust anomaly cluster, extract the time series data associated with the cluster center position, screen and analyze the trend of the particulate matter monitoring values ​​in the time series, extract the key time period of particulate matter monitoring according to the fluctuation range and change direction of the particulate matter concentration, and generate the key change segment in the time series data;

[0037] S412: combining the key change segments in the time series data, comparing the time change trends of the particulate matter monitoring value and the atmospheric pollution diffusion coefficient, establishing a gradient change feature through the growth gradient of the particulate matter concentration and the fluctuation trend of the diffusion coefficient within the time period, calculating the change rate and direction based on the growth amount of the particulate matter concentration and the diffusion coefficient in the target period, and obtaining the gradient direction of the particulate matter concentration change;

[0038] S413: Tracing back the node source path along the gradient direction of the particle concentration change, based on the correlation between the weights between nodes and the time series, using the formula:

[0039]

[0040] Calculate the correlation of path nodes and generate particle diffusion path tracing results;

[0041] Among them, R ij represents the correlation between path nodes i and j, indicating the degree of traceability of particle diffusion, C i represents the particle concentration value of node i, indicating the particle concentration monitored during the target period, and is used to calculate the impact of the concentration difference between nodes, C j represents the particle concentration value at node j, and C i Similarly, for the calculation of the concentration difference of paired nodes, D ij Represents the spatial distance between nodes i and j, indicating the physical position relationship between nodes, W ij Represents the weight between nodes i and j, indicating the strength of association between the two nodes in the diffusion path. α is the exponential adjustment coefficient of the weight, which is used to amplify or weaken the contribution of the weight in the calculation. β is the balance parameter, which is used to avoid the situation where the denominator becomes unstable when the spatial distance is too small.

[0042] As a further solution of the present invention, the steps for obtaining the dust pollution source positioning result are specifically as follows:

[0043] S511: Based on the particle diffusion path tracing result, the weight correlation between the path nodes is compared, the key path nodes are determined by the strength of the correlation between the nodes, the weight of each node is re-evaluated, and the weight is redistributed based on the quantified result of the weight change, thereby generating an updated weight distribution between the nodes;

[0044] S512: Selecting a node with the largest weight change amplitude from the updated inter-node weight distribution, taking the node with the largest weight change amplitude as a key node for iterative analysis, determining the key node by calculating the weight change amount and relative weight change rate of each node, and obtaining the weight change amplitude of the key node;

[0045] S513: Analyze the weight gradient of the path around the key node weight change amplitude, and combine the spatial density distribution data of the node, using the formula:

[0046]

[0047] Calculate the probability of dust pollution source location and generate dust pollution source location results;

[0048] Among them, S src It represents the pollution source location score obtained by weighted calculation of weight gradient and spatial density. Represents the weight change gradient of node i, characterizing the rate of change of node weight, ρ i It represents the spatial distribution density of node i, which indicates the proportion and aggregation of nodes in the spatial distribution. α is the weight adjustment coefficient of density, which is used to amplify the influence of nodes with higher spatial density. β is the balance parameter, which is used to avoid the dominant effect of a single high-density node on the overall positioning.

[0049] A highway environmental protection construction monitoring system based on big data, the highway environmental protection construction monitoring system based on big data is used to execute the above-mentioned highway environmental protection construction monitoring method based on big data, and the system includes:

[0050] The road construction environment weight calculation module calculates the spatial distance between sensor nodes based on the geographical location information of the data collection points and the dust concentration distribution value, normalizes the measurement difference, connects the adjacent sensor nodes, and generates a road construction environment weight network;

[0051] The local abnormal edge screening module calculates the neighborhood probability density of edge weights based on the road construction environment weight network, compares the differences in adjacent edge weights, extracts probability offset values, analyzes state changes through time series data, screens and processes associated edges, removes irrelevant edges, and generates abnormal edge results for dust and particulate matter;

[0052] The dust anomaly point positioning module extracts the geographical location of the node based on the dust and particulate matter anomaly edge results, calculates the spatial distance and weight density between nodes, clusters the nodes, determines the cluster center and weight distribution, and generates the center position of the dust anomaly point cluster;

[0053] The particle diffusion path tracing module extracts the particle monitoring value and the atmospheric pollution diffusion coefficient based on the center position of the dust anomaly cluster, analyzes the data change trend, analyzes the node connection path according to the gradient direction, tracks the weight association between the path nodes, and generates the particle diffusion path tracing result;

[0054] The dust pollution source positioning module is based on the tracing results of the particle diffusion path, compares the changes in node weights in the path, analyzes the weight gradients around the nodes, and locates the coordinates of the pollution source nodes in combination with the spatial density distribution to generate the dust pollution source positioning results.

[0055] Compared with the prior art, the advantages and positive effects of the present invention are:

[0056] In the present invention, a sensor network reflecting spatial relationships is constructed through the difference and normalization calculation of the geographical location information of the data collection points and the monitoring data, so as to achieve accurate expression of the construction environment information. Through probability density calculation and time series analysis, abnormal edges are screened, the accuracy of abnormal change detection is improved, and misjudgment is reduced. The position and spatial density of the nodes connected by the abnormal edges are calculated to clarify the high-density areas and cluster centers, so as to achieve accurate positioning of the pollution points. Combined with the monitoring data change trend and diffusion path tracing, the pollution diffusion chain is reconstructed, the pollution propagation range is clarified, and pollution control is optimized. Through the weight correlation and gradient analysis between the path nodes, the pollution source is accurately identified, the monitoring blind spots are reduced, and comprehensive support is provided for construction monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0058] Figure 2 A flowchart of the steps for obtaining the road construction environment weight network of the present invention;

[0059] Figure 3 A flow chart of the steps for obtaining abnormal edge results of dust and particulate matter according to the present invention;

[0060] Figure 4 A flow chart of the steps for obtaining the center position of the dust abnormality cluster of the present invention;

[0061] Figure 5 A flowchart of the steps for obtaining the particle diffusion path tracing result of the present invention;

[0062] Figure 6 This is a flow chart of the steps for obtaining the dust pollution source positioning results of the present invention. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0064] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0065] Embodiment 1

[0066] See also Figure 1 The present invention provides a technical solution: a highway environmental protection construction monitoring method based on big data, comprising the following steps:

[0067] S1: Based on the geographical location information of the data collection points and the dust concentration distribution and particulate matter monitoring values ​​of the road construction environment, the spatial distance between sensors is calculated, and the sensor nodes with direct adjacent relationships are connected through difference calculation and normalization processing to establish a road construction environment weight network;

[0068] S2: Based on the road construction environment weight network, the probability density of the edge weight in the neighborhood is calculated, and the local probability offset value is obtained through difference operation. Combined with time series data, the prominent edges are screened based on state changes and the associated edges are removed to generate dust and particulate matter abnormal edge results;

[0069] S3: Based on the results of dust and particulate matter abnormal edges, the positions of nodes connected by abnormal edges are extracted, spatial distances and local weight density are calculated, high-density areas are determined, node clusters are divided, cluster centers and weight distributions are calculated, and the center positions of dust abnormal point clusters are generated;

[0070] S4: Based on the center position of the dust anomaly cluster, extract the associated time series data, analyze the change trend of the particulate matter monitoring value and the atmospheric pollution diffusion coefficient, trace the node source path along the gradient direction, calculate the correlation of the path nodes, and generate the particulate matter diffusion path tracing results;

[0071] S5: Based on the tracing results of the particle diffusion path, compare the weight correlation between the path nodes, update the weight distribution between the nodes, and based on the node with the largest weight change in the path, analyze the weight gradient of the surrounding paths of the node and combine it with the spatial density distribution data to generate the dust pollution source positioning results.

[0072] The road construction environment weight network includes data collection points, geographic location information, sensor nodes and dust concentration distribution. The dust and particulate matter abnormal edge results include local probability offset values, time series data, prominent edges and removed associated edges. The center position of the dust anomaly point cluster includes high-density areas, node clusters, cluster centers and weight distribution. The particulate matter diffusion path tracing results include node source paths, correlation, path nodes and particulate matter diffusion trends. The dust pollution source positioning results include weight correlation between path nodes, path weight gradient and spatial density distribution data.

[0073] See also Figure 2 ,The specific steps for obtaining the road construction environment weight network are:

[0074] S111: Calculate the spatial distance between each pair of adjacent sensor nodes and the corresponding particle concentration difference through the geographical location information of the sensor and the dust concentration distribution of the monitoring point in the road construction environment, calculate the spatial coordinates of each sensor based on the location of the monitoring point, calculate the spatial distance between the nodes through the Euclidean distance formula between the coordinates, calculate the concentration difference between the nodes according to the particle concentration values ​​recorded at the monitoring point, and generate an initial adjacency matrix between the sensors;

[0075] The spatial coordinates of the sensor nodes are determined by collecting GPS signals from the sensor locations within the road construction area. Based on the collected longitude and latitude coordinate data, the Euclidean distance formula is used The spatial distance between sensor nodes is calculated. The dust concentration data is obtained by sampling the particle concentration at each sensor location at the same time point through the monitoring equipment. The concentration difference of each sensor node is calculated by calculating the absolute difference of the particle concentration between the two nodes |C i -C j |Get, form the initial adjacency relationship of the matrix structure with the two sets of data of concentration difference and spatial distance, where the matrix elements are composed of the spatial distance and concentration difference of the node pairs, and finally store the initial adjacency relationship matrix between the sensors in the form of matrix.

[0076] S112: normalizing the initial adjacency relationship matrix between sensors, normalizing the concentration difference to between 0 and 1, removing node connections that exceed the target threshold, and recalculating the weights of the remaining node connections as the ratio of the normalized concentration difference to the spatial distance, and updating the node weights to generate a normalized adjacency weight matrix;

[0077] For the concentration difference column data in the matrix, the formula (XX min ) / (X max -X min ) calculates the normalized value, removes the node connections that exceed a specific threshold after normalization, and the method of screening nodes above the threshold is based on setting a specific ratio range. For example, if the normalized value of the concentration difference exceeds 0.8, the node connection is removed, and the weight of the remaining node connection is determined by the ratio of the normalized concentration difference to the spatial distance. The matrix form of the weights is recalculated by replacing the elements in the initial adjacency matrix with the normalized concentration difference and spatial distance ratio to generate a normalized adjacency weight matrix.

[0078] S113: Based on the normalized adjacency weight matrix, the sensor nodes with direct adjacency are connected using the formula:

[0079]

[0080] Calculate the weight between each pair of sensor nodes to generate a road construction environment weight network;

[0081] Among them, W ij represents the weight between node i and node j, C i represents the particle concentration at node i, C j represents the particle concentration at node j, D ij represents the spatial distance between node i and node j, γ 1 is the concentration adjustment coefficient, which adjusts the influence of particle concentration on the weight, γ 2 It is the spatial adjustment coefficient, which adjusts the influence of spatial distance on the weight.

[0082] formula:

[0083]

[0084] The benefit of the formula is that by introducing concentration adjustment coefficients and distance adjustment coefficients, the weight calculation is made more flexible, and the contribution of both particle concentration difference and distance between sensors to the network weights can be comprehensively considered, thereby improving the ability of the road construction environment weight network to reflect the characteristic distribution of particle pollution.

[0085] Detailed explanation of the formula and the process of formula calculation and derivation:

[0086] First, take a set of sampling data and assume that the particle concentrations at nodes 1 and 2 are C 1 =120 μg / m 3 and C 2 =150μg / m 3 , the spatial distance between the two nodes is D 12=50m, concentration adjustment factor γ 1 and the distance adjustment factor γ 2 The actual value of is set according to the monitoring range and concentration fluctuation range. Assuming that γ 1 = 0.02 and γ 2 =0.05.

[0087] Substituting into the formula:

[0088]

[0089] The calculation process is as follows:

[0090] Concentration difference:

[0091] |120-150|=30;

[0092] Adjustment Value Section:

[0093] 0.02 120 = 2.4;

[0094] 0.05 50 = 2.5;

[0095] Total denominator: 50+2.4+2.5=54.9;

[0096]

[0097] The result shows that the weight between node 1 and node 2 is 0.546. This weight is used to represent the connection strength between two nodes in the network. The larger the value, the greater the contribution of the concentration difference relative to the distance, generating a road construction environment weight network.

[0098] See also Figure 3 The specific steps for obtaining the abnormal edge results of dust and particulate matter are as follows:

[0099] S211: Calculate the weight probability density of each edge through the road construction environment weight network, perform weight statistics based on all edges in the neighborhood, use the weight distribution frequency as a reference to calculate the kernel density estimation, perform continuous fitting on the probability distribution of the kernel density, and generate the weight probability density distribution of the edge;

[0100] Firstly, the edge weight frequency distribution is extracted according to the weight of each edge and the weight set of edges in the neighborhood. The weight range is divided into fixed intervals by using the statistical histogram method. The frequency of each interval is obtained by calculating the ratio of the number of edges whose weights fall into this interval to the total number of edges in the neighborhood. The weight frequency is smoothed by the Gaussian kernel function. The bandwidth parameter of the kernel function is set according to the square root of the difference between the number of neighborhood edges and the weight range. The fitting process of the weight probability density is completed, and the probability density distribution function of the edge is output for subsequent calculations.

[0101] S212: Perform a difference operation on the edge weight probability density distribution and calculate the local probability offset value using the formula:

[0102]

[0103] Calculate the degree of deviation of each edge from the average probability in its neighborhood and generate the local probability offset value of each edge;

[0104] Where ΔP ij Represents the local probability offset value, P ij represents the weight probability density of edges i and j, represents the average weight probability density of the edges in the neighborhood of node i, σ N(i) represents the standard deviation of the edges in the neighborhood, γ 1 To avoid the adjustment parameter whose denominator tends to zero, the local probability offset value is used to measure the degree of change in the edge weight probability;

[0105] formula:

[0106]

[0107] The benefit of the formula is that by introducing the joint operation of the neighborhood mean and standard deviation and combining the adjustment parameters, it can standardize the local weight fluctuations and enhance the ability to identify outliers;

[0108] Detailed explanation of the formula and the process of formula calculation and derivation:

[0109] Among them, P ij is the weight probability density of edges i and j, which is obtained by kernel density estimation in the probability density distribution function of the edge. For example, the weight probability density P of an edge is ij =0.45;

[0110] is the average weight probability density of the neighboring edges of node i, which is obtained by subtracting the P of all neighboring edges from the ij The values ​​are added and divided by the number of neighboring edges. Assuming the probability density of the neighboring edge weights is 0.30, 0.35, 0.50, and 0.60, then:

[0111]

[0112] σ N(i) is the standard deviation of the edges in the neighborhood, which is obtained by taking the square root of the variance calculated by the weighted probability density set. Assuming that the variance of the above neighborhood edges is 0.0156, the standard deviation is:

[0113]

[0114] γ 1To adjust the parameters to prevent the standard deviation from being too small and causing the amplification of outliers, it is set by the inverse of the number of neighborhood edges. For example, if the neighborhood contains 4 edges, then:

[0115] γ 1 =1 / 4=0.25;

[0116] Substituting the above values ​​into the formula, calculate ΔP ij as follows:

[0117]

[0118] The result shows that the local probability offset value of edges i and j is 0.0333, which indicates the degree of deviation of the weight probability of the edge relative to the neighborhood average value, and is used for subsequent screening of significantly abnormal edges.

[0119] S213: Filter the significant edges of the local probability offset value of each edge in combination with the time series data, calculate the fluctuation difference of the associated edges by jointly comparing the offset value with the time series state change, eliminate the associated edges with abnormal fluctuations, and generate the abnormal edge results of dust and particulate matter.

[0120] First, the significance of the local probability offset value of each edge is determined. The upper and lower limits are set according to the median and interquartile range of the weight offset value set. The edges below the lower limit or above the upper limit are marked as significant offset edges. For example, if the offset value range of a set is [0.01, 0.08, 0.12, 0.15, 0.22], the median is 0.12, the interquartile range is 0.15-0.08=0.07, the upper and lower limits are 0.12±1.5·0.07=[0.015, 0.225], and the edges below 0.015 or above 0.225 are significant offset edges. Secondly, the edges with abnormal fluctuations are detected in combination with the changes in the state of the time series. The fluctuation detection formula is used Calculate the state fluctuation value of each edge in the time series, where S i (t) is the state value at time t. Assuming the state value sequence is [0.1, 0.15, 0.25, 0.20, 0.05], the fluctuation value is:

[0121] B i =|0.15-0.1|+|0.25-0.15|+|0.20-0.25|+|0.05-0.20|=0.05+0.10+0.05+0.15=0.35;

[0122] Combine the significant offset edges with the high fluctuation value edges, remove the associated edges, and generate the abnormal edge results of dust and particulate matter.

[0123] See also Figure 4 , the specific steps for obtaining the center position of the dust anomaly cluster are:

[0124] S311: extract all abnormal nodes from the dust and particulate matter abnormal edge results, identify the spatial position of each node, determine the spatial coordinates of the abnormal node through the node connection relationship, determine the complete position of the abnormal edge node based on the weight distribution of the node connection, and generate a node position list;

[0125] By identifying abnormal edge nodes, the spatial coordinates of each node are extracted one by one. The spatial coordinates obtain the geographic location information collected by the sensor, and the sensor's installation location information is matched and confirmed with the coordinate system to generate the two-dimensional or three-dimensional coordinates of each abnormal node. The relevance of these nodes is further confirmed according to the node connection relationship. By analyzing the weight distribution of the abnormal edges, the concentration or dispersion of the node weights is verified, and the spatial distribution pattern of the abnormal nodes is further refined. The geographic location and connection relationship of each node are recorded as a weighted point set, which are integrated into a complete set of node spatial coordinates and weights to generate a node location list.

[0126] S312: Perform spatial distance calculation on the node position list, call the Euclidean distance formula based on the node spatial coordinates, calculate the spatial distance between each pair of nodes in turn, calculate the local weight density of multiple nodes based on the node weight distribution, and obtain the spatial and weight comprehensive characteristics of each node;

[0127] Based on the node spatial coordinates, the Euclidean distance formula is called to calculate the spatial distance between each pair of nodes in turn. Among them, d ij represents the spatial distance between node i and node j, x i ,y i ,z i are the spatial coordinates of node i respectively. By substituting the coordinates of the nodes one by one, the distance matrix between all nodes is calculated. Combined with the node weights, the local weight density of each node is calculated through the correlation between the weights and spatial distribution. The local weight density is calculated by calculating the ratio of the sum of the weights in the neighborhood of each node to the number of nodes in the neighborhood. The neighborhood range is limited by the preset threshold in the distance matrix. The screening of neighborhood nodes is completed to obtain the comprehensive spatial and weight characteristics of each node.

[0128] S313: Based on the spatial and weight comprehensive feature data of each node, determine the high-density area of ​​the node cluster, determine the density peak of the local area based on the node weight density, identify the core node through spatial position aggregation and weight peak, and generate a node cluster;

[0129] The node screening threshold in the high-density area is calculated by the local weight density. The threshold is set according to the median and standard deviation of the local weight density. The high-density node set is identified by node-by-node screening. The density peak identification method is further applied to calculate the density peak of each high-density node through the spatial aggregation of these high-density nodes. The density peak is obtained by subtracting the variance of the neighborhood weight density from the average weight in the local neighborhood. The peak node is the core node in the local area. The core node set of each cluster is generated. Through the connection relationship between the core node and its neighborhood nodes, the neighborhood of each core node is merged to form a complete node cluster.

[0130] S314: Calculate the center position and weight distribution of the node cluster using the formula:

[0131]

[0132] Calculate and generate the center position of dust anomaly cluster;

[0133] Among them, C center Represents the cluster center position, indicating the coordinates of the weighted center point of the node cluster, w i represents the weight of node i, which is used to measure the influence of the node in the cluster center calculation, x i represents the spatial position of node i, represents the coordinates of the node, and γ is an adjustment parameter used to balance the contribution of nodes with smaller weights within the cluster.

[0134] formula:

[0135]

[0136] The benefit of the formula is that by introducing square root and absolute value operations and adding the adjustment parameter γ, it ensures that the weight contribution of each node can be more balanced in the calculation of the node cluster center and prevents the minimum weight from affecting the precise positioning of the cluster center;

[0137] Detailed explanation of the formula and the process of formula calculation and derivation:

[0138] Through the node clusters generated in paragraph 3, extract the node space coordinates and weight sets in each node cluster, and convert the weight w i With coordinate x i Substitute into the formula one by one, assuming that there are 5 nodes in a node cluster, the corresponding weights are [3.2, 4.1, 2.5, 3.8, 4.0], the corresponding coordinates are [10, 12, 11, 13, 14], the adjustment parameter γ is 2.0, and the calculation is performed step by step through the formula:

[0139] First, calculate the absolute value of the product of the weight and coordinate of each node and take the square root:

[0140]

[0141] Then calculate the numerator:

[0142]

[0143] Calculate the denominator again:

[0144]

[0145] Finally, enter the formula for calculation:

[0146]

[0147] The result shows that the weighted coordinate value of the cluster center is 7.32, indicating that the center of the cluster is mainly affected by the nodes with larger weights. At the same time, adjusting the parameter γ balances the offset contribution of the nodes with smaller weights to the cluster center, further verifying the rationality of the cluster center position.

[0148] See also Figure 5 , the specific steps for obtaining the particle diffusion path tracing results are as follows:

[0149] S411: Based on the center position of the dust anomaly cluster, extract the time series data associated with the cluster center position, screen and analyze the trend of the particulate matter monitoring values ​​in the time series, extract the key time period of particulate matter monitoring through the fluctuation range and change direction of the particulate matter concentration, and generate the key change fragment in the time series data;

[0150] First, for the time series of particulate matter monitoring values, a preliminary judgment is made based on the fluctuation range of the monitoring data at each time point to extract the time period with obvious fluctuations. Specifically, the change in the monitoring value can be calculated by the formula |C t+1 -C t |Judgement, where C t and C t+1 are the monitored concentration values ​​at adjacent time points. If the change exceeds the set threshold (for example, the average fluctuation value of the local time series plus the standard deviation is taken as the threshold standard), the time period is determined to be a critical time period. Secondly, combined with the particle concentration data in the time period, the specific starting point and end point of the large-scale change (such as the rate of change) are recorded and compared with the time correlation data of the cluster center. For example, the dynamic particle concentration distribution change in the node coverage area in the geographical location is used to confirm the time period in the time series that is most directly related to the central cluster. Furthermore, for the particle monitoring values ​​in the time period, a detailed analysis of the change trend is performed, and the trend slope (C t+n -C t) / n, where n represents the number of time points in the time period. The point with the larger absolute value of the slope is defined as the significant change point, and the significant change point segment is used as the key segment for particle diffusion monitoring. The key segments of the time series are extracted to generate the key change segments in the time series data.

[0151] S412: Combine the key change segments in the time series data to compare the time change trends of the particulate matter monitoring values ​​and the atmospheric pollution diffusion coefficient, establish the gradient change characteristics through the growth gradient of the particulate matter concentration and the fluctuation trend of the diffusion coefficient within the time period, calculate the change rate and direction based on the growth amount of the particulate matter concentration and the diffusion coefficient in the target period, and obtain the gradient direction of the particulate matter concentration change;

[0152] First, through the dynamic monitoring value K of the atmospheric pollution diffusion coefficient t The slope of the time trend of the particle concentration (C t+n -C t ) / n for point-by-point correspondence analysis. Further, the formula |K is calculated based on the gradient of the diffusion coefficient. t+1 -K t |, establish a correlation between the dynamic fluctuation of the diffusion coefficient and the fluctuation of the particle concentration; secondly, determine the directional characteristics of the pollutant diffusion within the time period through the dynamic growth trend of the particle concentration, for example, by calculating (C t+n -C t ) / (x t+n -x t ), where x t+n and x t It represents the spatial position corresponding to the time points t+n and t. The positive or negative slope represents the direction of the diffusion trend. It specifically quantifies the change gradient and the dynamic diffusion direction. It also superimposes the spatial comparison data of the atmospheric diffusion coefficient change trend based on the temporal correlation of each node. It further combines the particle diffusion gradient direction with the spatial position information to obtain the overall directionality of the concentration change. Finally, it combines the time series data with the dynamic diffusion directional information to refine the gradient direction relationship of the concentration change and obtain the gradient direction of the particle concentration change.

[0153] S413: Trace the node source path along the gradient direction of the particle concentration change, based on the correlation between the weights between nodes and the time series, using the formula:

[0154]

[0155] Calculate the correlation of path nodes and generate particle diffusion path tracing results;

[0156] Among them, R ij represents the correlation between path nodes i and j, indicating the degree of traceability of particle diffusion, C irepresents the particle concentration value of node i, indicating the particle concentration monitored during the target period, and is used to calculate the impact of the concentration difference between nodes, C j represents the particle concentration value at node j, and C i Similarly, for the calculation of the concentration difference of paired nodes, D ij Represents the spatial distance between nodes i and j, indicating the physical position relationship between nodes, W ij Represents the weight between nodes i and j, indicating the strength of association between the two nodes in the diffusion path. α is the exponential adjustment coefficient of the weight, which is used to amplify or weaken the contribution of the weight in the calculation. β is the balance parameter, which is used to avoid the situation where the denominator becomes unstable when the spatial distance is too small.

[0157] formula:

[0158]

[0159] The benefit of the formula is that it improves the sensitivity of the calculation to concentration changes through the square root form of the particle concentration difference. At the same time, combined with the exponential adjustment of the weights between nodes and the distance balance parameter, it optimizes the comprehensive evaluation of the correlation between multiple nodes in the diffusion path, thereby enhancing the applicability and accuracy of the formula in the particle diffusion tracing scenario.

[0160] Detailed explanation of the formula and the process of formula calculation and derivation:

[0161] First, according to the specific monitoring data, the particle concentration C at node i and node j is set i =150μg / m 3 , C j =

[0162] 80 μg / m 3 , the weight between nodes W ij =0.6, the spatial distance between nodes D ij =2.5km, index adjustment coefficient α=1.2, balance parameter β=0.5km;

[0163] 1. First calculate the square root of the difference in particle concentration:

[0164]

[0165] 2. Calculate the exponential adjustment of weights:

[0166]

[0167] 3. Calculate the denominator (distance plus balance parameter):

[0168] D ij +β=2.5+0.5=3.0;

[0169] 4. Comprehensively input the formula to calculate the correlation between nodes:

[0170]

[0171] The result shows that the correlation value between node i and node j is 1.442, which reflects the concentration difference and weight strength.

[0172] See also Figure 6 , the specific steps for obtaining the dust pollution source location results are:

[0173] S511: Based on the particle diffusion path tracing results, the weight correlation between the path nodes is compared, the key path nodes are determined by the strength of the correlation between the nodes, the weight of each node is re-evaluated, and the weight is redistributed based on the quantitative results of the weight change, thereby generating an updated weight distribution between the nodes;

[0174] First, extract the weight data of each node in the path. The weight data is obtained by normalizing the node flow data of each path. The calculation formula is: where w ij is the normalized weight between path nodes i and j, F ij For path flow, after calculating the normalized weight of each node, the correlation between two nodes in all paths is compared, and the key path nodes are divided according to the strength of the correlation. Subsequently, the weights of the key path nodes are re-evaluated. During the re-evaluation process, the weight changes are compared pairwise. The calculation formula for the weight change is: Where Δw ij is the change in the old and new weights between nodes i and j, and are the new and old weights respectively. By comparing the weight changes, the nodes with larger weight changes are screened out and marked as key nodes. These key nodes are used for further analysis to generate the updated weight distribution between nodes.

[0175] S512: Select the node with the largest weight change from the updated inter-node weight distribution, and use the node with the largest weight change as the key node for iterative analysis. Determine the key node by calculating the weight change amount and relative weight change rate of each node, and obtain the weight change range of the key node.

[0176] The key nodes are determined by calculating the weight change of each node and its relative weight change rate. The weight change calculation formula of the key node is: Δw i =∑ j Δw ij Where Δw i is the total weight change of node i, Δw ijis the weight change between node i and adjacent node j. After calculating the total weight change, the relative weight change rate is calculated. The formula is: where r i is the relative weight change rate of node i, ∑w i is the sum of all weights of node i. By sorting the total weight change and the relative weight change rate, the node with the largest weight change is selected and marked as the key node. Finally, the weight change range of the key node is determined to obtain the weight change range of the key node.

[0177] S513: Analyze the weight gradient of the path around the key node weight change amplitude, and combine the spatial density distribution data of the node, using the formula:

[0178]

[0179] Calculate the probability of dust pollution source location and generate dust pollution source location results;

[0180] Among them, S src It represents the pollution source location score obtained by weighted calculation of weight gradient and spatial density. Represents the weight change gradient of node i, characterizing the rate of change of node weight, ρ i It represents the spatial distribution density of node i, which indicates the proportion and aggregation of nodes in the spatial distribution. α is the weight adjustment coefficient of density, which is used to amplify the influence of nodes with higher spatial density. β is the balance parameter, which is used to avoid the dominant effect of a single high-density node on the overall positioning.

[0181] formula:

[0182]

[0183] The benefit of the formula is that by introducing the comprehensive information of weight gradient and spatial density, the influence of high-density areas is optimized in the form of exponential adjustment, while avoiding the dominant role of a single high-density node in the positioning results, thereby improving the accuracy and robustness of positioning.

[0184] Detailed explanation of the formula and the process of formula calculation and derivation:

[0185] is the weight gradient of node i, expressed as the absolute value of the weight difference of adjacent nodes, and the calculation formula is:

[0186]

[0187] where w ij is the weight between nodes i and j, w i is the average weight of node i, ρ iis the spatial density of node i, which is obtained by normalizing the number of nodes in the unit area. The calculation formula is:

[0188]

[0189] Where N i is the number of nodes in the neighborhood of node i, A i is the spatial area of ​​the neighborhood of node i, α and β are adjustment coefficients, α is used to amplify the influence of high-density areas, and β is used to balance the impact of density on the overall result.

[0190] Example values:

[0191] Assume that the weight data is w ij =0.3, w i =0.4, calculated Density data is N i =50, A i =100, we get ρ i =0.5, take adjustment coefficient α=2, β=1, substitute into the formula:

[0192]

[0193] The results show that the dust pollution source positioning score is 0.075, reflecting the degree of influence of pollution sources in the current area. The lower score indicates that the area may be an intermediate area of ​​pollution diffusion rather than a pollution source area. By further analyzing the scores of nodes in other areas, the exact location of the pollution source can be comprehensively determined to generate the dust pollution source positioning result.

[0194] A highway environmental protection construction monitoring system based on big data, which is used to execute the above-mentioned highway environmental protection construction monitoring method based on big data, comprises:

[0195] The road construction environment weight calculation module calculates the spatial distance between sensor nodes based on the geographical location information of the data collection points and the dust concentration distribution value, normalizes the measurement difference, connects the adjacent sensor nodes, and generates a road construction environment weight network;

[0196] The local abnormal edge screening module is based on the road construction environment weight network, calculates the neighborhood probability density of edge weights, compares the differences in adjacent edge weights, extracts probability offset values, analyzes state changes through time series data, screens and processes related edges, removes irrelevant edges, and generates dust and particulate matter abnormal edge results;

[0197] The dust anomaly point positioning module extracts the geographical location of the node based on the dust and particulate matter anomaly edge results, calculates the spatial distance and weight density between nodes, clusters the nodes, determines the cluster center and weight distribution, and generates the center location of the dust anomaly point cluster;

[0198] The particle diffusion path tracing module extracts the particle monitoring value and the atmospheric pollution diffusion coefficient based on the center position of the dust anomaly cluster, analyzes the data change trend, analyzes the node connection path according to the gradient direction, tracks the weight association between the path nodes, and generates the particle diffusion path tracing results;

[0199] The dust pollution source positioning module is based on the tracing results of the particle diffusion path, compares the changes in node weights in the path, analyzes the weight gradients around the nodes, and locates the coordinates of the pollution source nodes in combination with the spatial density distribution to generate the dust pollution source positioning results.

[0200] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A highway environmental protection construction monitoring method based on big data, characterized in that: The following steps are involved: S1: Based on the geographical location information of the data collection points and the dust concentration distribution and particulate matter monitoring values ​​of the road construction environment, the spatial distance between sensors is calculated, and the sensor nodes with direct adjacent relationships are connected through difference calculation and normalization processing to establish a road construction environment weight network; S2: Based on the road construction environment weight network, the probability density of the edge weight in the neighborhood is calculated, and the local probability offset value is obtained through difference operation. In combination with the time series data, the prominent edges are screened based on the state change and the associated edges are removed to generate the abnormal edge results of dust and particulate matter; S3: Based on the dust and particulate matter abnormal edge results, extract the positions of the nodes connected by the abnormal edges, calculate the spatial distance and local weight density, determine the high-density area, divide the node clusters, calculate the cluster center and weight distribution, and generate the center position of the dust abnormal point cluster; S4: Based on the center position of the dust anomaly cluster, extract the associated time series data, analyze the change trend of the particulate matter monitoring value and the atmospheric pollution diffusion coefficient, trace the node source path along the gradient direction, calculate the correlation of the path nodes, and generate the particulate matter diffusion path tracing result; S5: Based on the tracing results of the particle diffusion path, compare the weight correlation between the path nodes, update the weight distribution between the nodes, and based on the node with the largest weight change in the path, analyze the weight gradient of the surrounding path of the node and combine it with the spatial density distribution data to generate the dust pollution source positioning result.

2. The highway environmental protection construction monitoring method based on big data according to claim 1 is characterized in that: The road construction environment weight network includes data collection points, geographic location information, sensor nodes and dust concentration distribution. The dust and particulate matter abnormal edge results include local probability offset values, time series data, prominent edges and removed associated edges. The dust abnormal point cluster center position includes high-density areas, node clusters, cluster centers and weight distribution. The particulate matter diffusion path tracing results include node source paths, correlation, path nodes and particulate matter diffusion trends. The dust pollution source positioning results include weight correlation between path nodes, path weight gradient and spatial density distribution data.

3. The highway environmental protection construction monitoring method based on big data according to claim 2 is characterized in that: The steps for obtaining the road construction environment weight network are specifically as follows: S111: Calculate the spatial distance between each pair of adjacent sensor nodes and the corresponding particle concentration difference through the geographical location information of the sensor and the dust concentration distribution of the monitoring point in the road construction environment, calculate the spatial coordinates of each sensor based on the location of the monitoring point, calculate the spatial distance between the nodes through the Euclidean distance formula between the coordinates, calculate the concentration difference between the nodes according to the particle concentration values ​​recorded at the monitoring point, and generate an initial adjacency matrix between the sensors; S112: normalizing the initial adjacency relationship matrix between the sensors, normalizing the concentration difference to between 0 and 1, removing node connections exceeding the target threshold, and recalculating the weights of the remaining node connections as the ratio of the normalized concentration difference to the spatial distance, and updating the node weights to generate a normalized adjacency weight matrix; S113: Based on the normalized adjacency weight matrix, the sensor nodes having direct adjacency relationships are connected, using the formula: Calculate the weight between each pair of sensor nodes to generate a road construction environment weight network; Among them, W ij represents the weight between node i and node j, C i represents the particle concentration at node i, C j represents the particle concentration at node j, D ij represents the spatial distance between node i and node j, γ1 is the concentration adjustment coefficient, which adjusts the influence of particle concentration on the weight, and γ2 is the spatial adjustment coefficient, which adjusts the influence of spatial distance on the weight.

4. The highway environmental protection construction monitoring method based on big data according to claim 3 is characterized in that: The steps for obtaining the abnormal edge results of dust and particulate matter are specifically as follows: S211: Calculate the weight probability density of each edge through the road construction environment weight network, perform weight statistics based on all edges in the neighborhood, use the weight distribution frequency as a reference to calculate the kernel density estimation, perform continuous fitting on the probability distribution of the kernel density, and generate the weight probability density distribution of the edge; S212: Perform a difference operation on the weight probability density distribution of the edge to calculate the local probability offset value using the formula: Calculate the degree of deviation of each edge from the average probability in its neighborhood and generate the local probability offset value of each edge; Among them, ΔP ij Represents the local probability offset value, P ij represents the weight probability density of edges i and j, represents the average weight probability density of the edges in the neighborhood of node i, σ N(i) represents the standard deviation of the edges in the neighborhood, γ1 is an adjustment parameter to avoid the denominator approaching zero, and the local probability offset value is used to measure the degree of change in the edge weight probability; S213: Filter the significant edges of the local probability offset value of each edge in combination with the time series data, calculate the fluctuation difference of the associated edges by jointly comparing the offset value with the time series state change, eliminate the associated edges with abnormal fluctuations, and generate dust and particulate matter abnormal edge results.

5. The highway environmental protection construction monitoring method based on big data according to claim 4 is characterized in that: The steps for obtaining the center position of the dust anomaly cluster are specifically as follows: S311: extract all abnormal nodes from the dust and particulate matter abnormal edge results, identify the spatial position of each node, determine the spatial coordinates of the abnormal node through the node connection relationship, determine the complete position of the abnormal edge node based on the weight distribution of the node connection, and generate a node position list; S312: performing spatial distance calculation on the node position list, calling the Euclidean distance formula based on the node spatial coordinates, calculating the spatial distance between each pair of nodes in turn, calculating the local weight density of multiple nodes based on the node weight distribution, and obtaining the spatial and weight comprehensive characteristics of each node; S313: Based on the spatial and weight comprehensive feature data of each node, determine the high-density area of ​​the node cluster, determine the density peak of the local area based on the node weight density, identify the core node through spatial position aggregation and weight peak, and generate a node cluster; S314: Calculate the center position and weight distribution of the node cluster using the formula: Calculate and generate the center position of dust anomaly cluster; Among them, C center Represents the cluster center position, indicating the coordinates of the weighted center point of the node cluster, w i represents the weight of node i, which is used to measure the influence of the node in the cluster center calculation, x i represents the spatial position of node i, represents the coordinates of the node, and γ is an adjustment parameter used to balance the contribution of nodes with smaller weights within the cluster.

6. The highway environmental protection construction monitoring method based on big data according to claim 5 is characterized in that: The specific steps for obtaining the particle diffusion path tracing result are as follows: S411: based on the center position of the dust anomaly cluster, extract the time series data associated with the cluster center position, screen and analyze the trend of the particulate matter monitoring values ​​in the time series, extract the key time period of particulate matter monitoring according to the fluctuation range and change direction of the particulate matter concentration, and generate the key change segment in the time series data; S412: combining the key change segments in the time series data, comparing the time change trends of the particulate matter monitoring value and the atmospheric pollution diffusion coefficient, establishing a gradient change feature through the growth gradient of the particulate matter concentration and the fluctuation trend of the diffusion coefficient within the time period, calculating the change rate and direction based on the growth amount of the particulate matter concentration and the diffusion coefficient in the target period, and obtaining the gradient direction of the particulate matter concentration change; S413: Tracing back the node source path along the gradient direction of the particle concentration change, based on the correlation between the weights between nodes and the time series, using the formula: Calculate the correlation of path nodes and generate particle diffusion path tracing results; Among them, R ij represents the correlation between path nodes i and j, indicating the degree of traceability of particle diffusion, C i represents the particle concentration value of node i, indicating the particle concentration monitored during the target period, and is used to calculate the impact of the concentration difference between nodes, C j represents the particle concentration value at node j, and C i Similarly, for the calculation of the concentration difference of paired nodes, D ij Represents the spatial distance between nodes i and j, indicating the physical position relationship between nodes, W ij Represents the weight between nodes i and j, indicating the strength of association between the two nodes in the diffusion path. α is the exponential adjustment coefficient of the weight, which is used to amplify or weaken the contribution of the weight in the calculation. β is the balance parameter, which is used to avoid the situation where the denominator becomes unstable when the spatial distance is too small.

7. The highway environmental protection construction monitoring method based on big data according to claim 6 is characterized in that: The steps for obtaining the dust pollution source positioning result are specifically as follows: S511: Based on the particle diffusion path tracing result, the weight correlation between the path nodes is compared, the key path nodes are determined by the strength of the correlation between the nodes, the weight of each node is re-evaluated, and the weight is redistributed based on the quantified result of the weight change, thereby generating an updated weight distribution between the nodes; S512: Selecting a node with the largest weight change amplitude from the updated inter-node weight distribution, taking the node with the largest weight change amplitude as a key node for iterative analysis, determining the key node by calculating the weight change amount and relative weight change rate of each node, and obtaining the weight change amplitude of the key node; S513: Analyze the weight gradient of the path around the key node weight change amplitude, and combine the spatial density distribution data of the node, using the formula: Calculate the probability of dust pollution source location and generate dust pollution source location results; Among them, S src It represents the pollution source location score obtained by weighted calculation of weight gradient and spatial density. Represents the weight change gradient of node i, characterizing the rate of change of node weight, ρ i It represents the spatial distribution density of node i, which indicates the proportion and aggregation of nodes in the spatial distribution. α is the weight adjustment coefficient of density, which is used to amplify the influence of nodes with higher spatial density. β is the balance parameter, which is used to avoid the dominant effect of a single high-density node on the overall positioning.

8. Highway environmental protection construction monitoring system based on big data, characterized by: According to the method for monitoring highway environmental protection construction based on big data according to any one of claims 1 to 7, the system comprises: The road construction environment weight calculation module calculates the spatial distance between sensor nodes based on the geographical location information of the data collection points and the dust concentration distribution value, normalizes the measurement difference, connects the adjacent sensor nodes, and generates a road construction environment weight network; The local abnormal edge screening module calculates the neighborhood probability density of edge weights based on the road construction environment weight network, compares the differences in adjacent edge weights, extracts probability offset values, analyzes state changes through time series data, screens and processes associated edges, removes irrelevant edges, and generates abnormal edge results for dust and particulate matter; The dust anomaly point positioning module extracts the geographical location of the node based on the dust and particulate matter anomaly edge results, calculates the spatial distance and weight density between nodes, clusters the nodes, determines the cluster center and weight distribution, and generates the center position of the dust anomaly point cluster; The particle diffusion path tracing module extracts the particle monitoring value and the atmospheric pollution diffusion coefficient based on the center position of the dust anomaly cluster, analyzes the data change trend, analyzes the node connection path according to the gradient direction, tracks the weight association between the path nodes, and generates the particle diffusion path tracing result; The dust pollution source positioning module is based on the tracing results of the particle diffusion path, compares the changes in node weights in the path, analyzes the weight gradients around the nodes, and locates the coordinates of the pollution source nodes in combination with the spatial density distribution to generate the dust pollution source positioning results.

Images