Research On Characteristics Of Mining Surface Cracks Based On UAV Images

Surface cracks caused by underground coal mining are an intuitive manifestation of mining subsidence destruction,causing serious damage to buildings and land.Accurately obtaining the development characteristics of mining surface cracks is of practical significance for mining subsidence research and mining area ground protection.Conventional mining surface crack monitoring methods have obvious limitations in terms of cost efficiency and temporal and spatial resolution.Low-altitude drone surveying and mapping technology is an important means to collect large-scale and high-resolution surface space information.By setting reasonable aerial photography parameters,high-resolution images of subsidence areas can be quickly obtained,but the current universal remote sensing image surface information acquisition The method is difficult to adapt to the extraction of surface cracks in mining areas.This paper takes the surface of a fully mechanized caving mining face in the western Yushen mining area as the experimental area,and uses low-altitude drone aerial photography to obtain high-resolution images of the subsidence area.Based on surface modeling,crack extraction,and vegetation removal,based on the deformation of the subsidence area The distribution feature constructs a connection algorithm for mining fractures,which provides technical support for the efficient and accurate acquisition of surface fracture information in mining subsidence areas.The main content and results of the thesis research are as follows:(1)The aerial photography parameters of low-altitude UAV photography,which are suitable for the fine extraction of surface cracks in the Yushen mining area,are determined.Combining the accuracy of field acquisition and actual conditions on site,the influence of various aerial parameters in the photogrammetry system on the spatial resolution of the image is analyzed.Field tests show that under the aerial photography parameters designed in this paper,the ground resolution of the image is better than 0.015m,which meets the accuracy requirements for automatic extraction of surface cracks in mining areas.(2)The actual effects of Canny algorithm,support vector machine and maximum likelihood method in extracting mining cracks under the conditions of Yushen mining area are compared and analyzed.The crack extraction results of various algorithms are superimposed with the orthophoto map,and the applicability of different algorithms in the natural environment of the study area is analyzed in combination with the three-dimensional model of the subsidence area,the algorithm principle and the efficiency of the algorithm.The results show that the crack information obtained by different crack extraction algorithms is significantly different,and the maximum likelihood method has the best extraction effect.(3)The fine extraction of mining cracks is realized by filtering the image background noise.Based on the systematic analysis of the source and characteristics of the background noise in the crack information,the random forest algorithm is first used to filter the background noise,and then the above-mentioned maximum likelihood method is used to extract the fine features of the surface cracks.The actual effect is better.(4)According to the movement and deformation characteristics of the mining subsidence basin,the connection algorithm of the surface mining cracks is constructed.Aiming at the discontinuity problems in the extracted mining fractures,combined with the theoretical model of mining subsidence deformation,an automatic connection algorithm for intermittent fractures is constructed,and the algorithm is improved and the effect is verified through experiments.The results show that the fracture connection algorithm proposed in this paper helps to obtain the actual distribution characteristics of surface fractures in mining subsidence areas.