Research On Monitoring Method Of Surface Movement And Vegetation Parameters Based On RTK-UAV

In the current "double carbon" context,the high-quality development of coal industry has entered a brand new stage in the new era,and it is still the main theme of the future coal industry to restore the mining ecosystem in a timely and effective manner,to rejuvenate the traditional coal mines,and to maintain the green and sustainable development.In order to obtain these key parameters in the coal mining process timely and accurately and to analyze the interaction between them,this thesis proposes a method to extract surface movement information without manual targeting based on UAV photogrammetry technology,and to identify typical vegetation on the surface and its key parameters.This thesis proposes a method for extracting surface movement information without manual targeting,identifying typical vegetation on the surface and estimating its key growth parameters to analyze the influence of surface movement on vegetation,and programming the process of estimating surface mining subsidence and typical vegetation key growth parameters in mining areas,which simplifies the process of monitoring surface subsidence and vegetation parameters in the past and improves the efficiency.The main research contents and conclusions of this thesis are as follows:(1)Based on UAV images and dense matching point cloud data,this thesis proposes a method for extracting surface movement information in mining areas without manual targeting,the core of which is the determination of eponymous features.The work efficiency is improved.Firstly,the semantic segmentation model is used to extract the contour information of specific features in the mine area,and the study area of this thesis takes stone as an example to determine the relative offset pixel number of the same target in two adjacent UAV images by contour matching method,and obtain the horizontal surface movement amount of the mine area by combining the spatial resolution of UAV images;then,considering the situation that it is difficult to find such features on the surface of some mine areas,a target-free The method is based on the dense matching point cloud of UAV,taking the voxel grid of the first phase as the monitoring target,searching its position in the next phase of the point cloud,and using the average offset of the 3D coordinates of the points in the two adjacent phases as the horizontal and vertical movements;finally,the effectiveness of the algorithm is verified with the ground observation station data acquired by GPS.The accuracy of the targetfree method is lower than that of the feature-based contour feature method,and the errors of the former are 1.3 cm and 5.13 cm larger than those of the latter in the d E and d N directions,respectively,but it makes up for the shortcomings of the feature-based contour feature method.(2)In this thesis,the typical vegetation of the study area,lemon bar,was used as an example,and its chlorophyll content was inverted using visible light images,and the inversion results were used to analyze the influence law of surface movement and deformation on vegetation in the mine area,and it was found that there was a significant negative correlation between vegetation chlorophyll content and surface horizontal movement,which indicated that when the topography of the mine area was undulating,the vegetation in the upper and lower slope areas were under the most severe stress,answering the The question of which areas of vegetation should be focused on during coal mining was answered.Firstly,the classification of lemon bars is identified,and for the problem of low separability between lemon bars and pine and poplar trees,the geometric features extracted from dense matching point clouds are projected onto a two-dimensional plane,and the classification effect is improved by multi-feature fusion.Then,based on the existing research,a regression model between the visible light index Ikaw and the chlorophyll content of lemon sticks was established for two of the UAV images,with coefficients of 0.72 and 0.68,respectively;finally,based on the results of chlorophyll content estimation in the whole study area,the spatial scale of the mine area was used to analyze the vegetation in the mine area under stress.Finally,based on the results of chlorophyll content estimation in the whole study area,we analyzed the pattern of vegetation affected by stress in the mine area from the spatial scale of the mine area,and found that there was a significant negative correlation between the increment of chlorophyll content and the horizontal movement of the ground surface with a correlation coefficient of 0.615.(3)In this thesis,an integrated system for subsidence prediction and vegetation parameter estimation based on RTK-UAV products is designed,which includes the functions of surface movement extraction,sinkhole estimation and vegetation parameter estimation,and solves the problem that most of the input data for mining sinkhole estimation are mostly ground observation station data or manually extracted from UAV data.First,the surface movement extracted from the products generated by UAV photogrammetry technology is used as the input quantity of the subsidence estimation module;then,the estimation of the subsidence estimation parameters and the display module of the prediction results are implemented by using the Python programming language,based on the principle of the probability integration method,and the curve fitting method and the mode vector method based on the subsidence or horizontal movement model of any point on the ground;finally,the vegetation classification Finally,the vegetation classification and the estimation of vegetation parameters are designed as one of the system modules,and a process-oriented operating system is realized.The thesis has 58 figures,25 tables,and 107 references.