CN-121564667-B - Underground pipe gallery fault data tracing processing method and system based on big data

CN121564667BCN 121564667 BCN121564667 BCN 121564667BCN-121564667-B

Abstract

The invention relates to the technical field of image processing and discloses a method and a system for tracing fault data of an underground pipe gallery based on big data, wherein the method comprises the steps of obtaining multi-view monitoring image data of the underground pipe gallery; the method comprises the steps of performing downsampling on multi-view monitoring image data to obtain a differential resolution image, performing edge feature enhancement on the differential resolution image according to a pyramid structure to obtain an enhancement feature image, identifying an abnormal region of the enhancement feature image, verifying the authenticity of the abnormal region in a space structure of an underground pipe gallery to obtain a multi-dimensional feature descriptor, constructing a space-time relationship feature image according to the space-time feature information in the multi-dimensional feature descriptor and the feature dimension relevance of the abnormal region, performing multi-angle track tracing on the multi-dimensional feature descriptor based on the space-time relationship feature image to obtain abnormal tracing information, and improving the efficiency of underground pipe gallery fault data tracing processing based on big data.

Inventors

Zhao Yindan
Wen Chenying
WU WENYI
ZHU ZUGANG
ZHAN JUNQING
ZHANG XINMIN

Assignees

杭州市地下管道开发有限公司

Dates

Publication Date: 20260505
Application Date: 20260126

Claims (8)

1. The utility model provides a underground pipe gallery fault data traceability processing method based on big data, which is characterized in that the method comprises the following steps: s1, acquiring multi-view monitoring image data of an underground pipe gallery; S2, downsampling the multi-view monitoring image data to obtain a differential resolution image of the underground pipe gallery, and carrying out edge feature enhancement on the differential resolution image according to a pyramid structure to obtain an enhancement feature map of the underground pipe gallery; S3, identifying an abnormal region of the enhanced feature map according to the abnormal dynamic features related to the faults in the enhanced feature map, wherein the method comprises the following steps: extracting a dynamic characteristic sequence related to faults in time sequence change in the enhanced characteristic diagram; Identifying abnormal feature points deviating from normal variation in the enhanced feature map according to abnormal mutation trend of image points in the dynamic feature sequence; Counting the spatial distribution density degree of the abnormal feature points to obtain candidate abnormal areas of the enhanced feature map; performing space consistency verification on the candidate abnormal regions, and screening out the abnormal regions of the enhanced feature map; Identifying abnormal feature points deviating from normal variation in the enhanced feature map according to abnormal mutation trend of image points in the dynamic feature sequence, wherein the method comprises the following steps: Based on the dynamic feature sequence, connecting feature points in the enhanced feature map in time sequence to obtain a feature change track of the dynamic feature sequence; And comparing the deviation of the characteristic change track with a preset normal characteristic change mode to obtain the deviation degree of the characteristic change track, wherein the calculation formula of the deviation degree is as follows: ; In the formula, For the degree of deviation to be stated, The length of the time window for the characteristic change track is, for example, As a point in time at which the current time point is present, For the first order difference of the characteristic change tracks at adjacent time points, For the first order difference of the normal characteristic change pattern at adjacent time points, For the second order difference of the characteristic change trace at adjacent time points, For the second order difference of the normal characteristic change pattern at adjacent time points, For the standard deviation of the first order difference in the normal characteristic change pattern, For the standard deviation of the second order difference in the normal characteristic change pattern, As the weight coefficient of the first order difference term, Is the weight coefficient of the second order difference term, At a point in time for the characteristic change trace Is used for the characteristic value of the (c), At a point in time for the normal characteristic change pattern Is a reference feature value of (a); threshold comparison is carried out on the deviation degree, and abnormal feature points of the dynamic feature sequence are obtained; S4, verifying the authenticity of the abnormal region in the space structure of the underground pipe gallery, and obtaining a multi-dimensional feature descriptor of the enhanced feature map; s5, constructing a space-time relation feature map of the underground pipe gallery according to the association of space-time feature information in the multi-dimensional feature descriptors and feature dimensions of the abnormal region; and S6, performing multi-angle track tracing on the multi-dimensional feature descriptors based on the space-time relation feature graphs to obtain abnormal tracing information of the underground pipe gallery.
2. The method for tracing fault data of an underground pipe gallery based on big data according to claim 1, wherein the step of obtaining multi-view monitoring image data of the underground pipe gallery comprises the following steps: Collecting continuous image sequence data in an underground pipe gallery under different visual angles to obtain original image data of the underground pipe gallery; extracting video frames from the original image data to obtain a multi-view image frame set of the underground pipe gallery; unifying the image format of the multi-view image frame set to generate multi-view monitoring image data of the underground pipe gallery; and verifying the visual angle coverage integrity and the time sequence consistency of the multi-visual angle monitoring image data to obtain the multi-visual angle monitoring image data of the underground pipe gallery.
3. The method for tracing fault data of an underground pipe gallery based on big data according to claim 1, wherein the downsampling the multi-view monitoring image data to obtain a differential resolution image of the underground pipe gallery, and performing edge feature enhancement on the differential resolution image according to a pyramid structure to obtain an enhanced feature map of the underground pipe gallery, comprises: Performing downsampling operation on the multi-view monitoring image data to obtain a differential resolution image of the underground pipe gallery; detecting the image edges of resolution levels in the differential resolution images based on a pyramid structure to obtain a multi-scale edge map of the underground pipe gallery; according to the gray level difference change of the multi-scale edge map, the edge characteristics of the multi-scale edge map are enhanced, and an edge enhanced image of the underground pipe gallery is obtained; and fusing the edge enhancement images to obtain an enhancement feature map of the underground pipe gallery.
4. The method for tracing fault data of an underground pipe gallery based on big data according to claim 1, wherein verifying the authenticity of the abnormal region in the space structure of the underground pipe gallery to obtain the multidimensional feature descriptor of the enhanced feature map comprises: taking the pipe gallery internal components and the space layout information of the underground pipe gallery as space structure data; Carrying out spatial registration on the abnormal region and the spatial structure data to obtain a spatial position relation between the abnormal region and a corresponding pipe gallery component; Based on the spatial position relationship, verifying the consistency of the structural topological relationship between the abnormal region and the pipe gallery component, and screening out a component structure abnormal region conforming to the spatial structure data; And integrating the texture features, the shape features and the dynamic change trend features in the abnormal region of the component structure to obtain the multi-dimensional feature descriptor of the enhanced feature map.
5. The method for tracing fault data of an underground pipe gallery based on big data according to claim 1, wherein the constructing a space-time relation feature map of the underground pipe gallery according to the space-time feature information in the multi-dimensional feature descriptor and the feature dimension association of the abnormal region comprises the following steps: Taking the points representing different time points and different spatial position points in the multi-dimensional feature descriptors as graph nodes; taking a node path connected with a continuous time sequence in the multi-dimensional feature descriptor as a time sequence edge; Taking the propagation paths of adjacent spatial position nodes in the multi-dimensional feature descriptors in the spatial dimension as spatial edges; Performing similarity measurement on feature vectors in the multi-dimensional feature descriptors, and establishing feature edges of the multi-dimensional feature descriptors according to measurement results and node clustering degrees in a feature space; And constructing a space-time relationship characteristic diagram of the underground pipe gallery according to the diagram nodes, the time sequence edges, the space edges and the characteristic edges.
6. The underground pipe gallery fault data tracing processing method based on big data as claimed in claim 5, wherein the calculation formula of the feature edge weight in the multi-dimensional feature descriptor is as follows: ; In the formula, For nodes in the multi-dimensional feature descriptor And node The characteristic edge weight of the space between the two, For nodes in the multi-dimensional feature descriptor And node The cosine distance between the feature vectors, For the mean of cosine distances between pairs of total nodes in the multi-dimensional feature descriptor, For the standard deviation of cosine distances between pairs of total nodes in the multi-dimensional feature descriptor, For nodes in the multi-dimensional feature descriptor And node A spatial euclidean distance between them, For nodes in the multi-dimensional feature descriptor And node The absolute time difference between the two, For a predetermined spatial scale factor, Is a predetermined time scale factor of the time scale, Is a natural constant which is used for the production of the high-temperature-resistant ceramic material, Is a natural constant.
7. The method for tracing the fault data of the underground pipe gallery based on the big data according to claim 1, wherein the tracing the multi-dimensional feature descriptors to the abnormal trace of the underground pipe gallery based on the space-time relation feature map to obtain the abnormal tracing information of the underground pipe gallery comprises the following steps: Taking a space-time association path corresponding to the multidimensional feature descriptor in the space-time relationship feature map as a candidate tracing path set of the underground pipe gallery; Under the time evolution angle, the space propagation angle and the characteristic association angle in the space-time relation feature graph, performing multi-angle track analysis on the candidate tracing path set to obtain a preliminary tracing track of the candidate tracing path set; carrying out track conflict resolution on the preliminary tracing track to obtain a complete tracing link of the underground pipe gallery; and based on the complete traceability link, integrating the abnormal characteristic information in the multidimensional characteristic descriptor to obtain the abnormal traceability information of the underground pipe gallery.
8. The utility model provides a utility tunnel fault data traceability processing system based on big data, which is characterized in that the utility tunnel fault data traceability processing method based on big data is used for realizing the utility tunnel fault data traceability processing method based on big data, and the system comprises: The multi-source data acquisition module is used for acquiring multi-view monitoring image data of the underground pipe gallery; the multi-scale feature enhancement module is used for downsampling the multi-view monitoring image data to obtain a differential resolution image of the underground pipe gallery, and enhancing edge features of the differential resolution image according to a pyramid structure to obtain an enhanced feature map of the underground pipe gallery; The abnormal region identification module is used for identifying an abnormal region of the enhanced feature map according to the abnormal dynamic characteristics related to the faults in the enhanced feature map; The feature verification and description module is used for verifying the authenticity of the abnormal region in the space structure of the underground pipe gallery and obtaining a multi-dimensional feature descriptor of the enhanced feature map; the space-time relation construction module is used for constructing a space-time relation feature map of the underground pipe gallery according to the space-time feature information in the multi-dimensional feature descriptors and the feature dimension relevance of the abnormal region; And the track tracing analysis module is used for carrying out multi-angle track tracing on the multi-dimensional feature descriptors based on the space-time relationship feature map to obtain the abnormal tracing information of the underground pipe gallery.

Description

Underground pipe gallery fault data tracing processing method and system based on big data Technical Field The invention relates to the technical field of image processing, in particular to a method and a system for tracing fault data of an underground pipe gallery based on big data. Background The underground pipe gallery is used as a key carrier of urban infrastructure, and the accuracy and efficiency of fault data tracing directly influence the timeliness of operation and maintenance guarantee. In the prior art, when multi-view monitoring image data are processed, the problems of poor data resolution and complicated characteristic information cannot be effectively solved, the characteristic enhancement means lack of pertinence, key characteristics related to faults are difficult to fully mine, the recognition accuracy of abnormal areas is insufficient, false abnormal misjudgment or true abnormal omission exists, and hidden hazards are buried for subsequent tracing work. In the fault tracing analysis link, the prior art often ignores the internal association of space-time characteristics and abnormal region characteristic dimensions, lacks a space-time relation construction mechanism of a system, is limited in the tracing process to track tracing of a single dimension, and cannot realize multi-angle and all-dimensional characteristic association analysis. The method makes the tracing link difficult to completely present the evolution process and the propagation path of the fault, the positioning of the root cause of the fault is fuzzy, the tracing efficiency is low, and the actual operation and maintenance requirements of the rapid investigation and the accurate treatment of the underground pipe gallery fault are difficult to be met. Disclosure of Invention The invention provides a method and a system for tracing fault data of an underground pipe gallery based on big data, which are used for solving the problems in the background technology. In order to achieve the above purpose, the invention provides a method for tracing fault data of an underground pipe gallery based on big data, which comprises the following steps: s1, acquiring multi-view monitoring image data of an underground pipe gallery; S2, downsampling the multi-view monitoring image data to obtain a differential resolution image of the underground pipe gallery, and carrying out edge feature enhancement on the differential resolution image according to a pyramid structure to obtain an enhancement feature map of the underground pipe gallery; s3, identifying an abnormal region of the enhanced feature map according to the abnormal dynamic features related to the faults in the enhanced feature map; S4, verifying the authenticity of the abnormal region in the space structure of the underground pipe gallery, and obtaining a multi-dimensional feature descriptor of the enhanced feature map; s5, constructing a space-time relation feature map of the underground pipe gallery according to the association of space-time feature information in the multi-dimensional feature descriptors and feature dimensions of the abnormal region; and S6, performing multi-angle track tracing on the multi-dimensional feature descriptors based on the space-time relation feature graphs to obtain abnormal tracing information of the underground pipe gallery. In a preferred embodiment, the acquiring multi-view monitoring image data of the underground pipe gallery includes: Collecting continuous image sequence data in an underground pipe gallery under different visual angles to obtain original image data of the underground pipe gallery; extracting video frames from the original image data to obtain a multi-view image frame set of the underground pipe gallery; unifying the image format of the multi-view image frame set to generate multi-view monitoring image data of the underground pipe gallery; and verifying the visual angle coverage integrity and the time sequence consistency of the multi-visual angle monitoring image data to obtain the multi-visual angle monitoring image data of the underground pipe gallery. In a preferred embodiment, the downsampling the multi-view monitoring image data to obtain a differential resolution image of the underground pipe gallery, and edge feature enhancement is performed on the differential resolution image according to a pyramid structure to obtain an enhanced feature map of the underground pipe gallery, which includes: Performing downsampling operation on the multi-view monitoring image data to obtain a differential resolution image of the underground pipe gallery; detecting the image edges of resolution levels in the differential resolution images based on a pyramid structure to obtain a multi-scale edge map of the underground pipe gallery; according to the gray level difference change of the multi-scale edge map, the edge characteristics of the multi-scale edge map are enhanced, and an edge enhanced image of the underground pipe gallery is obtai