Font Size: a A A

Feature Extraction And Texture Analysis Of High-Resolution Images For Ground Fissures

Posted on:2021-08-28Degree:MasterType:Thesis
Country:ChinaCandidate:H J YangFull Text:PDF
GTID:2480306110496144Subject:Geological Engineering
Abstract/Summary:PDF Full Text Request
As a type of ground deformation disaster,ground fissures not only cause damage to buildings,roads and land resources,but may also cause a series of ecological and environmental problems.In recent years,the development of remote sensing technology has been rapid,the resolution of sensors has been continuously improved,and the ability to recognize small-scale features such as ground fissures has been continuously improved.Comprehensive and systematic analysis and research on the characteristics of high-resolution images of ground fissures can obtain more accurate information on the distribution and destruction of ground fissures,improve the efficiency of ground fissure disaster investigation,and are important for mine production safety and ecological environment governance Meaning and role.The spatial distribution characteristics and ground manifestations of coal mining fissures,as well as high-resolution remote sensing image feature extraction and texture analysis methods are summarized.The general characteristics of ground fissures are summarized,and the spatial distribution of coal ground fissures and the ground manifestation of ground fissures in different locations and different topography are analyzed.The goaf area of Jinggong No.1 Mine in Pingshuo mining area was selected as the research area to investigate the ground fissure disasters in the research area.Based on the investigation,the image characteristics of ground fissures are analyzed.The ground fissures in the study area are mostly collapsed ground fissures.The vertical dislocations on both sides of the cracks are obvious.The steep ridges and graben are clearly visible on the image;the area collapses and collapses The situation is serious.With the collapse and landslide,the tensiontype ground fissure develops obviously,the crack length and width are small,there is no obvious dislocation on both sides,and it appears as a thin dark tone line on the image.Analyze the spectral and geometric characteristics of ground fissures,calculate the brightness,asymmetry,boundary index,compactness,density,main direction,rectangularity,roundness,shape index and other characteristic parameters,determine the threshold range,establish the ground Set of rules for crack extraction.Based on the analysis of the brightness and shape characteristics of ground fissures,the texture characteristics of ground fissures are analyzed.The characteristics of ground fissure texture are mainly manifested in two aspects,one is the regularity of the brightness change of the ground fissure,and the second is the structure texture feature of the ground fissure morphology and combination on the image.The gray level co-occurrence matrix is used to analyze the texture characteristics of ground fissure images.The influences of the three parameters of pixel pair distance,gray level and sliding window size on the construction of gray level co-occurrence matrix and the description of ground fissure texture are analyzed.Calculate the change of each feature value when the distance of the pixel is 1-15.When the distance is less than 5,it will have a greater impact on the feature value.Therefore,calculate the distance 1-5,the description effect of each feature value on the ground fissure,when the distance is 3 The effect is better;secondly,the description effects of 128,64,32,16,8 gray levels on the characteristic parameters and the characteristics of ground fissure texture are analyzed.The experiment found that selecting the gray level of 32 to describe the characteristics of ground fissure texture is more effective.Good;Finally,the effect of different window sizes is analyzed.The window size is selected as 3,9,15,21,27.According to the feature image,it can be concluded that when the window size is 9,the effect is better.Construct a gray level co-occurrence matrix with a pixel pair distance of 3,a gray level of 32,and a sliding window size of 9,to extract ground fissure texture features in the study area.The study found that the correlation reaches the maximum in the direction of ground fissure extension,and the contrast can be used to describe The edge information of ground fissures,entropy,angular second moment and inverse moment can be used to express the spatial structure complexity of ground fissures of different sizes.
Keywords/Search Tags:Ground Fissures, High-Resolution Images, Feature Extraction, Texture Analysis, Gray level co-occurrence matrix
PDF Full Text Request
Related items