CN-121982019-A - Method and system for detecting plume skeleton of emission source of atmospheric particulates

CN121982019ACN 121982019 ACN121982019 ACN 121982019ACN-121982019-A

Abstract

The invention relates to a plume skeleton detection method and a plume skeleton detection system for an atmospheric particulate emission source in the technical field of laser radar detection. The weighted undirected graph G is constructed based on the limited Veno graph generated by the plume profile C, the longest first m paths are selected as candidate path sets S according to the path length of the endpoint pairs in the G as a sequencing basis, and energy calculation is performed according to an energy constraint function epsilon so as to find the global energy minimum path from the S as a skeleton. According to the invention, the problem of skeleton extraction of the plume of the emission source of the atmospheric particulates is converted into the optimal path in the weighted undirected graph G, so that the skeleton extraction of the plume of the emission source of the atmospheric particulates is realized. In the face of complex and changeable geometric shapes and topological characteristics of the plume of the emission source of the atmospheric particulates, epsilon is further optimized, the integral structure of the plume is considered instead of the part, and the extracted plume skeleton is ensured to have smoothness, centrality and continuity.

Inventors

LIU YANPING
WANG TONG
Zhou Maimai
ZHANG NA
ZHANG XINQIAO
LIU JIANPING

Assignees

山东国耀量子雷达科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260401

Claims (10)

1. An atmospheric particulate emission source plume skeleton detection method for detecting the skeleton from an atmospheric particulate emission source plume profile C, characterized by comprising the steps of: generating a limited voronoi diagram according to the plume profile C; Constructing a weighted undirected graph G according to a limited voronoi diagram, wherein G= (V, E, w) is formed by adjacent voronoi vertex coordinates in the limited voronoi diagram, wherein the coordinates of adjacent voronoi vertexes in the limited voronoi diagram form a node set V, E is an undirected edge set formed by adjacent voronoi vertexes, and w is a weight set corresponding to undirected edges; Traversing all endpoint pairs in G, calculating the shortest path in each endpoint pair as a path length, taking the path length as a sorting basis, and selecting the longest m candidate paths as a candidate path set S, S= { P 1 , P 2 , …, P m }, wherein the ith candidate path is P i , i=1, 2, & gt, m; Traversing S, performing energy calculation on the candidate paths P i according to an energy constraint function epsilon, and searching a global energy minimum path from the S as the skeleton.
2. The method for detecting the plume skeleton of the emission source of the atmospheric particulates according to claim 1, the detection method is characterized by further comprising the following steps: compressing the C into a plurality of key points and keeping the shape to generate a contour C', and if the distance between two adjacent key points is larger than a preset distance parameter D, uniformly interpolating new contour points between the two key points to ensure that the distance between any two adjacent contour points after interpolation is smaller than D; generating a voronoi diagram according to the C ', and reserving the voronoi vertexes positioned inside the C' to generate a limited voronoi diagram.
3. The method for detecting the plume skeleton of the emission source of the atmospheric particulates according to claim 2, wherein the profile C' generates a voronoi diagram by a Delaunay triangulation method; and/or the algorithm for compressing the plume profile C is a Douglas-Peucker algorithm or a Teh-Chin algorithm.
4. The method for detecting the plume skeleton of the emission source of the atmospheric particulates according to claim 1, wherein the endpoint in the weighted undirected graph G is a node with a degree of 1, and the Dijkstra algorithm is adopted to calculate the shortest path in each endpoint pair.
5. The method according to claim 1, wherein, to extract the plume skeleton of the emission source of the atmospheric particulates having smoothness, centrality and continuity, the energy constraint function ε is: ; Where ε is the energy constraint function, In order for the term to be of a fidelity, In order to smooth the term(s), Is a length item, Alpha, beta and eta are weights and are all E (0, 1).
6. The method for detecting a plume skeleton of an atmospheric particulate emission source according to claim 5, wherein, The calculation method of (2) is as follows: ; ; ; Wherein, the The number of complement elements for the node set of the candidate path P i and the node set of the weighted undirected graph G, representing the number of missing nodes, The element number of intersection of the node set of the candidate path P i and the node set of the weighted undirected graph G is represented as the effective node number, lambda 1 is a penalty term, lambda 2 is a reward term, and lambda 1 >λ 2 >0 is satisfied; and/or the number of the groups of groups, The calculation method of (2) is as follows: For any three consecutive points (v r-1 , v r , v r+1 ) in the node set of candidate path P i , r=2,..k-1, forming edge a r and edge b r ;a r = v r-1 - v r ,b i = v r+1 - v r with v r as intermediate nodes; The angle θ r between side a r and side b r is: ; In the formula, The method is used for limiting the result to be within the range of [ -1,1] to prevent the result from exceeding the definition domain of the inverse cosine function due to numerical error; The expression is: ; Wherein k represents the total number of nodes in the candidate path P i , and k points form k-2 edges; and/or the number of the groups of groups, The calculation method of (2) is as follows: ; Wherein u, V E V (P i ) is an adjacent node, (u, V) is an undirected edge connecting the node u and the node V, w (u, V) is a weight of the undirected edge (u, V), E (P i ) is an edge set of the candidate path P i , and V (P i ) is a node set of the candidate path P i ; and/or the number of the groups of groups, The calculation method of (2) is as follows: ; In the formula, Indicating an indicating function, wherein the value is 1 when the condition is met, and the value is 0 otherwise; representing the degree of node v in weighted undirected graph G when When the node is established, the node having a degree of more than 3 in the weighted undirected graph G is shown in the candidate path P i .
7. The method for detecting the plume skeleton of the emission source of the atmospheric particulates according to claim 1, wherein the method for obtaining the plume outline C is as follows: scanning and monitoring a region to be detected by adopting an aerosol laser radar; and then solving the time-space distribution of the atmospheric particulates by using echo signals received by the aerosol radar, and acquiring a plume profile C by adopting an image processing method, a statistical model method or a hydrodynamic feature method according to the time-space distribution of the atmospheric particulates.
8. An atmospheric particulate emission source plume skeleton detection system characterized in that it detects a skeleton from an atmospheric particulate emission source plume profile C according to the atmospheric particulate emission source plume skeleton detection method according to any one of claims 1 to 7, monitors the change of the atmospheric particulate emission source plume according to the change of the skeleton.
9. The atmospheric particulate emissions source plume skeleton detection system of claim 8, wherein the atmospheric particulate emissions source plume skeleton detection system comprises: the aerosol laser radar scans and monitors a region to be detected, then solves the time-space distribution of the atmospheric particulate matters through echo signals received by the aerosol laser radar, and acquires the plume profile C of the atmospheric particulate matter emission source in a scanning plane according to the space-time distribution of the atmospheric particulate matters by adopting an image processing, a statistical model and a hydrodynamic feature method; and the processor is used for applying the skeleton detection method.
10. A computer readable storage medium having stored thereon a computer program/instruction which when executed by a processor implements the steps of the atmospheric particulate emission source plume skeleton detection method of any one of claims 1 to 7.

Description

Method and system for detecting plume skeleton of emission source of atmospheric particulates Technical Field The invention relates to the technical field of laser radar detection, in particular to a plume skeleton detection method of an atmospheric particulate emission source and a plume skeleton detection system of the atmospheric particulate emission source by adopting the skeleton detection method. Background The remote sensing task of the particulate matter based on the scanning aerosol laser radar is to monitor the emission of particulate matters or the variation of emission source plume in areas such as cities, grasslands, forests and the like under the conditions of high space-time precision and long distance range. An atmospheric particulate emission source plume is defined as a region in a radar scan image where the concentration of particulate matter due to emission of a particulate matter pollution source is above an ambient background concentration level. Since the aerosol radar-detected plume of atmospheric particulate matter is generated by the emission source, there is a significant variability in the emission amount and emission pattern of the emission source over time. The particle plumes are highly dynamic and various under the action of complex meteorological factors, and the formation, diffusion and sedimentation of atmospheric particulate pollutants are affected by interleaving of a plurality of factors such as space, time and meteorological conditions, so that the characteristics of space-time mutation, expansion, contraction, rotation, recombination and the like are shown on morphological structures. The extraction of the plume skeleton has wide application prospect in aspects such as the analysis of the shape, the pattern recognition, the characteristic representation and the like of the plume of the emission source of the atmospheric particulate matters. The method comprises the steps of obtaining a plume skeleton of an emission source of the atmospheric particulates, quantitatively analyzing indexes such as length, curvature, trend, bifurcation, thickness change and the like of the skeleton while keeping key information of diffusion dynamics and spatial structures, describing diffusion forms of the skeleton, detecting various modes such as emission source types, atmospheric stability, diffusion rules and the like, and providing characteristic representation for inversion of the emission source, pollution tracking and regional atmospheric environment assessment. However, the existing skeleton analysis research on the plume of the emission source of the atmospheric particulates is less, the geometric shape and the topological characteristics of the plume of the emission source of the atmospheric particulates are complex and changeable, and the skeleton of the plume of the emission source of the atmospheric particulates with smoothness, centrality and continuity is difficult to extract. Disclosure of Invention The invention provides an atmospheric particulate emission source plume skeleton detection method and an atmospheric particulate emission source plume skeleton detection system adopting the skeleton detection method, and aims to solve the technical problem that an existing complex and changeable atmospheric particulate emission source plume skeleton is difficult to extract. The first aspect of the present invention provides a method for detecting an atmospheric particulate emission source plume skeleton from an atmospheric particulate emission source plume profile C, the method comprising the steps of: generating a limited voronoi diagram according to the plume profile C; Constructing a weighted undirected graph G according to a limited voronoi diagram, wherein G= (V, E, w) is formed by adjacent voronoi vertex coordinates in the limited voronoi diagram, wherein the coordinates of adjacent voronoi vertexes in the limited voronoi diagram form a node set V, E is an undirected edge set formed by adjacent voronoi vertexes, and w is a weight set corresponding to undirected edges; Traversing all endpoint pairs in the weighted undirected graph G, calculating the shortest path in each endpoint pair as a path length, taking the path length as a sorting basis, and selecting the longest m candidate paths as a candidate path set S, s= { P 1, P2, …, Pm }, wherein the i-th candidate path is P i, i=1, 2, m; Traversing the candidate path set S, performing energy calculation on the candidate paths P i according to an energy constraint function epsilon, and searching a global energy minimum path from the candidate path set S as the skeleton. As a further improvement of the above scheme, the plume profile C is compressed into a plurality of key points and kept in shape to generate a profile C', if the distance between two adjacent key points is greater than the preset distance parameter D, new profile points are uniformly interpolated between the two key points, so that the distance between any tw