CN-121392330-B - Method and system for discriminating collapse target of rock high slope

CN121392330BCN 121392330 BCN121392330 BCN 121392330BCN-121392330-B

Abstract

The invention relates to the field of rock high slope safety monitoring, in particular to a method and a system for discriminating a rock high slope collapse target. The center end builds and issues reference environment data according to the space monitoring unit based on the high-resolution image and the three-dimensional point cloud. The method comprises the steps of periodically collecting current images and point clouds by a slope monitoring terminal, registering the current images and the point clouds, calculating image change indexes and terrain change indexes to obtain comprehensive change indexes, screening candidate collapse monitoring units, storing multi-phase data of the candidate units in a preset time window, constructing a multi-time phase feature sequence, and outputting collapse risk grades and collapse target judging results according to collapse judging parameters. The terminal adaptively adjusts the sampling period, the change threshold and the time window control calculated amount according to the resource state, uploads multi-time phase characteristics and monitoring data to the central terminal for rechecking, counts the false alarm rate and the false alarm rate, updates related parameters, and solves the problem of real-time collapse identification of high-resolution image point cloud under the condition of low power consumption.

Inventors

WU TIANWEI
CHEN SUIYANG
GAO QIANG

Assignees

四川省综合地质调查研究所
四川省德阳地质工程勘察院有限公司

Dates

Publication Date: 20260508
Application Date: 20251223

Claims (10)

1. The method for discriminating the rock high slope collapse target is characterized by comprising the following steps of: The method comprises the steps that firstly, central terminal equipment obtains rock high slope initial high resolution image data and initial three-dimensional point cloud data, generates reference environment data containing static morphological characteristics of a plurality of space monitoring units, and transmits the reference environment data to a slope monitoring terminal; Step two, the slope monitoring terminal acquires current monitoring image data and current monitoring three-dimensional point cloud data according to a preset sampling period, and performs spatial registration on the current monitoring image data and the current monitoring three-dimensional point cloud data with reference environment data in sequence to obtain alignment data; calculating an image change index and a terrain change index of each space monitoring unit by the slope monitoring terminal based on the alignment data and the reference environment data to obtain a comprehensive change index, and determining the space monitoring units with the comprehensive change index larger than a preset change threshold as candidate collapse monitoring units; Storing multi-time alignment data of the candidate collapse monitoring units in a preset time window by the slope monitoring terminal, and obtaining a multi-time phase characteristic sequence according to the stored multi-time alignment data of the candidate collapse monitoring units; step five, outputting a collapse risk level and a collapse target discrimination result of the candidate collapse monitoring unit by the side slope monitoring terminal according to the collapse discrimination parameters and the multi-time phase characteristic sequence, and entering a step seven when the collapse risk level reaches a preset high risk level, otherwise entering a step six; step six, acquiring resource state parameters of the slope monitoring terminal, controlling the number of candidate collapse monitoring units and the length of the multi-time phase characteristic sequence according to the resource state parameters, enabling the calculated amount for judging the collapse target in a single time to be not more than a preset upper limit, returning to the step two, and judging the next period; and seventhly, uploading the multi-time phase characteristic sequence, the alignment data and the collapse target judging result of the candidate collapse monitoring unit to central terminal equipment by the slope monitoring terminal, carrying out rechecking judgment by the central terminal equipment in a preset statistical time period based on the reference environment data and the uploaded data, outputting a rechecking result, counting the false alarm rate and the false alarm rate, updating one or more of collapse judging parameters, a preset sampling period, a preset change threshold value and a preset time window according to the false alarm rate and the false alarm rate, and transmitting the updated collapse judging parameters, the preset sampling period, the preset change threshold value and the preset time window to the slope monitoring terminal.
2. The method for discriminating a rock high slope collapse target according to claim 1, wherein the central side device acquires initial high resolution image data and initial three-dimensional point cloud data of the rock high slope, generates reference environment data including static morphological characteristics of a plurality of space monitoring units, and transmits the reference environment data to the slope monitoring terminal, and comprises the steps of: The central terminal equipment acquires initial high-resolution image data through the remote sensing photographing equipment, acquires initial three-dimensional point cloud data through the laser radar equipment, constructs a multi-level resolution image data set for the initial high-resolution image data, divides a rock high-slope area into a plurality of space monitoring units, performs rasterization processing on the initial three-dimensional point cloud data to obtain terrain raster data corresponding to the space monitoring units, calculates static morphological characteristics according to preset static morphological parameters based on the multi-level resolution image data set and the terrain raster data, wherein the static morphological parameters comprise at least two of average elevation, gradient, slope direction and surface roughness, and forms reference environment data.
3. The method for judging the collapse target of the rock high slope according to claim 2, wherein the slope monitoring terminal collects current monitoring image data and current monitoring three-dimensional point cloud data according to a preset sampling period, and performs spatial registration on the current monitoring image data and the current monitoring three-dimensional point cloud data with reference environment data in sequence to obtain alignment data, and the method comprises the following steps: The slope monitoring terminal acquires current monitoring image data and current monitoring three-dimensional point cloud data according to a preset sampling period, maps the current monitoring image data to a target resolution image level in the reference environment data, performs translational alignment and rotational alignment on the current monitoring image data and the target resolution image data based on feature point matching to obtain registration image data, and projects the current monitoring three-dimensional point cloud data to a coordinate system consistent with terrain grid data in the reference environment data based on spatial reference of the registration image data, and performs gridding processing to obtain alignment data.
4. The method for determining a target of rock high slope collapse according to claim 3, wherein the slope monitoring terminal calculates an image change index and a terrain change index of each spatial monitoring unit based on the alignment data and the reference environmental data to obtain a comprehensive change index, and determines the spatial monitoring units with the comprehensive change index greater than a preset change threshold as candidate collapse monitoring units, comprising: For each space monitoring unit, obtaining an image change index according to an image change parameter based on the aligned current monitoring image data and corresponding image data in the reference environment data, wherein the image change parameter comprises at least one of a gray level difference value, a texture change amount and an edge direction change amount; Obtaining a terrain variation index according to a terrain variation parameter based on the aligned current terrain raster data and the corresponding terrain raster data in the reference environment data, wherein the terrain variation parameter comprises at least one of an elevation difference value, a gradient difference value and a surface roughness difference value; And carrying out normalization processing on the image change index and the terrain change index, carrying out weighted summation according to preset weights to obtain a comprehensive change index, and selecting space monitoring units with the comprehensive change index being larger than a preset change threshold value and the number not larger than a preset upper limit as candidate collapse monitoring units on the basis of sequencing the comprehensive change index from high to low.
5. The method for determining a target of rock high slope collapse according to claim 1, wherein the slope monitoring terminal stores multi-temporal alignment data of the candidate collapse monitoring units in a preset time window, and obtains a multi-temporal feature sequence according to the stored multi-temporal alignment data of the candidate collapse monitoring units, and the method comprises the steps of: the method comprises the steps that a preset time window covers the latest M preset sampling periods, M is a preset positive integer, and a slope monitoring terminal stores aligned current monitoring image data and current monitoring three-dimensional point cloud data of a candidate collapse monitoring unit corresponding to each preset sampling period in the preset time window; The method comprises the steps of obtaining current monitoring image data, extracting image change characteristics from the current monitoring image data after alignment to form an image change characteristic sequence, extracting topography change characteristics from the current monitoring three-dimensional point cloud data after alignment to form a topography change characteristic sequence, and obtaining a multi-temporal characteristic sequence based on the image change characteristic sequence and the topography change characteristic sequence.
6. The method for determining a target of collapse of a rocky high slope according to claim 1, wherein the slope monitoring terminal outputs a collapse risk level and a collapse target determination result of the candidate collapse monitoring unit according to the collapse determination parameter and the multi-temporal feature sequence, and the method comprises the steps of: The collapse judging parameters comprise weight coefficients for each characteristic dimension of the multi-time phase characteristic sequence and a multi-level collapse risk level dividing threshold, wherein the slope monitoring terminal weights each characteristic dimension of the multi-time phase characteristic sequence according to the weight coefficients to obtain comprehensive judging quantity, the comprehensive judging quantity is compared with the multi-level collapse risk level dividing threshold, the candidate collapse monitoring unit is judged to be in a normal state when the comprehensive judging quantity is smaller than the lowest collapse risk level dividing threshold, the candidate collapse monitoring unit is judged to be corresponding to the collapse risk level when the comprehensive judging quantity is between the adjacent collapse risk level dividing thresholds, and the candidate collapse monitoring unit is judged to be in a collapse target and to be in a high risk level when the comprehensive judging quantity is larger than the highest collapse risk level dividing threshold.
7. The method for determining a target of collapse of a rocky high slope according to claim 1, wherein the step of obtaining the resource state parameter of the slope monitoring terminal, and controlling the number of candidate collapse monitoring units and the length of the multi-temporal feature sequence according to the resource state parameter, so that the calculated amount of determining the target of collapse in a single time is not more than a preset upper limit of calculation, comprises: The method comprises the steps of periodically reading the occupancy rate of a processor, the electric quantity of a power supply and the temperature of equipment by a side slope monitoring terminal to form a resource state parameter, comparing the resource state parameter with a preset resource threshold, determining that the side slope monitoring terminal is in a high load state when the occupancy rate of the processor is larger than the resource threshold of the processor and/or the electric quantity of the power supply is smaller than the resource threshold of the power supply and/or the temperature of the equipment is larger than the temperature resource threshold, and increasing a preset sampling period and/or increasing a preset change threshold and/or shortening a preset time window under the high load state so that the calculated amount for judging the collapse target in one time is not larger than a preset upper limit.
8. The method for determining a target of collapse of a rocky high slope according to claim 1, wherein the slope monitoring terminal uploads the multi-temporal feature sequence, the alignment data and the target of collapse determination result of the candidate collapse monitoring unit to the central terminal device, and the method comprises the steps of: When the collapse risk level of the candidate collapse monitoring unit reaches a preset high risk level, the slope monitoring terminal combines the multi-temporal feature sequence corresponding to the candidate collapse monitoring unit, the aligned current monitoring image data, the aligned current monitoring three-dimensional point cloud data and the collapse target judging result to form uploading data, and sends the uploading data to the center terminal equipment, and the center terminal equipment carries out rechecking judgment based on the reference environment data and the uploading data to generate a center terminal rechecking judging result and outputs the rechecking result.
9. The method for determining a target of rock high slope collapse according to claim 1, wherein the statistics of the false alarm rate and the false alarm rate, and updating one or more of collapse determining parameters, a preset sampling period, a preset change threshold value and a preset time window according to the false alarm rate and the false alarm rate, and issuing the parameters to the slope monitoring terminal, comprises: The central terminal equipment compares the collapse target discrimination result of the side slope monitoring terminal with the central terminal recheck discrimination result in a preset statistical time period, respectively calculates the false alarm rate and the false alarm rate, respectively compares the false alarm rate and the false alarm rate with a preset threshold, and updates one or more of the collapse discrimination parameter, the preset sampling period, the preset change threshold and the preset time window according to a preset adjustment strategy when the false alarm rate is larger than the preset threshold or the false alarm rate is larger than the preset threshold, so as to form an update parameter and send the update parameter to the side slope monitoring terminal.
10. The rock high slope collapse target judging system is characterized by comprising a central end device and a slope monitoring terminal, wherein the slope monitoring terminal is in communication connection with the central end device; The center terminal equipment comprises a reference environment construction unit, a rechecking and parameter updating unit and a communication device, wherein the reference environment construction unit and the rechecking and parameter updating unit are respectively connected with the communication device; the slope monitoring terminal comprises a data acquisition and registration unit, a multi-source change calculation and candidate screening unit, a multi-time phase characteristic construction unit, a collapse discrimination unit, a resource state sensing and adjustment unit and a result output and uploading unit; The data acquisition and registration unit, the multi-source change calculation and candidate screening unit, the multi-time phase feature construction unit, the collapse judging unit and the resource state sensing and adjusting unit are respectively connected with the result output and uploading unit, and the result output and uploading unit is in communication connection with the communication device.

Description

Method and system for discriminating collapse target of rock high slope Technical Field The invention relates to the field of rock high slope safety monitoring, in particular to a method and a system for discriminating a rock high slope collapse target. Background Along with the mass construction of expressways, railways and hydraulic engineering in mountain areas, the number of rock high-grade slopes is rapidly increased, and traffic interruption, facility damage and casualties caused by slope collapse occur. In order to reduce the risk of disasters, it is generally necessary to monitor a rock high slope for a long period of time and to determine the risk of collapse. In the prior art, common slope monitoring means comprise point position monitoring instruments such as total stations, GNSS, crack meters and the like, wherein displacement information of a small number of discrete monitoring points can be obtained only, the space coverage is limited, and the overall stability condition of a rock high slope is difficult to comprehensively reflect; The unmanned aerial vehicle image, the aerial photography and the laser radar scanning are adopted to obtain a high-resolution image and a three-dimensional point cloud, modeling and multi-time phase deformation analysis are carried out in a central end device or a background server, and although finer geometric change information can be obtained, the calculation amount is large and real-time operation on-site low-power-consumption equipment is difficult; Some video intelligent analysis or image recognition methods are deployed on a field terminal, and are generally used for single-frame or short-sequence target detection only based on two-dimensional images, lack of fusion with three-dimensional point clouds, lack of analysis on deformation evolution in a long period of time, and limited in recognition capability on slow-change and mutation processes such as collapse of a rock high slope. In addition, the rock high slope is often in an environment with limited power supply and unstable communication conditions, the on-site monitoring terminal has limited computing resources, and complex algorithms aiming at high-resolution images and three-dimensional point clouds are difficult to directly run on the terminal. If all the data are uploaded to the central terminal equipment for processing, the communication burden is excessive, the timeliness is insufficient, the algorithm parameters of the field terminal are difficult to adjust in time in a targeted manner, and the false alarm rate of long-term operation are difficult to control. Therefore, the prior art still has the following problems: the high-resolution image and the three-dimensional point cloud data are large in quantity, the traditional processing flow mainly depends on central terminal equipment, and real-time collapse judgment is difficult to realize on a slope field low-power consumption terminal; The field terminal is based on a single data source or a simple threshold value, fusion of image change information and three-dimensional point cloud geometric change information is not fully utilized, candidate collapse areas are not accurately screened, and calculation load is heavy; the lack of a multi-temporal feature sequence construction and comprehensive discrimination mechanism aiming at a candidate collapse monitoring unit cannot fully utilize the time evolution rule of morphological change and displacement change; The lack of an adaptive adjustment mechanism based on the terminal resource state and a cloud closed-loop parameter updating mechanism based on the false alarm rate and the missing alarm rate leads to difficulty in considering real-time performance and discrimination precision in long-term operation of a monitoring system. Disclosure of Invention The present invention has been made to solve the above-mentioned problems occurring in the prior art. Specifically, the invention is realized by the following technical scheme: a method for discriminating a rock high slope collapse target comprises the following steps: The method comprises the steps that firstly, central terminal equipment obtains rock high slope initial high resolution image data and initial three-dimensional point cloud data, generates reference environment data containing static morphological characteristics of a plurality of space monitoring units, and transmits the reference environment data to a slope monitoring terminal; Step two, the slope monitoring terminal acquires current monitoring image data and current monitoring three-dimensional point cloud data according to a preset sampling period, and performs spatial registration on the current monitoring image data and the current monitoring three-dimensional point cloud data with reference environment data in sequence to obtain alignment data; calculating an image change index and a terrain change index of each space monitoring unit by the slope monitoring terminal based on the