Research On Image Matching Technology In Slope Monitoring Based On Photogrammetry

With the increase of slope engineering scale,slope monitoring is more and more important to ensure the safety of life and property of its surrounding people.Considering that non-contact close-range photogrammetry has the advantages of low equipment cost,comprehensive information acquisition and convenient monitoring operation,this paper uses close-range photogrammetry to monitor and analyze the slope in three-dimensional position,and studies the image matching based on the characteristic information in the slope image.The main results of the study are as follows:(1)The principles of close-range photogrammetry and camera calibration are expounded.It expounds the relevant parameters,common coordinate system and its mutual conversion relationship,col linear equation and common plane equation in close field photo gram me try,introduces the common analytical processing method of close field photography,and the direct linear transformation method is used to analyze the image.By Python the script language,the calibration experiment of Zhang Zhengyou plane camera is completed,and the application possibility of smartphone camera in close-range photogrammetry is judged by comparing the pixel accuracy.(2)According to the condition of slope environment and considering the imaging effect of slope image,the artificial mark point of slope with low cost and easy imaging is designed.On the basis of image HSV color space theory,image mathematical morphology,image mask convex hull,contour detection and contour moment,and so on,the Python language is used to compile the software of slope mark point coordinate positioning.(3)The mapping software involves image preprocessing,morphological processing and contour processing in photogrammetry image processing,and puts forward a point matching location method based on image contour features.By adjusting the threshold of image brightness,image mask,Gauss blur,corrosion expansion and contour shrinkage in the localization program,the coordinate location in different environments can be realized,which has certain universality.The three-dimensional coordinates of mark points are solved by direct linear transformation method.(4)The applicability of three-dimensional coordinate calculation of slope mark is verified by close indoor experiment and remote outdoor experiment.By comparing the three-dimensional coordinates collected by the self-built coordinate system of the total station,the measurement error of the close-range photogrammetry method is analyzed,and a point arrangement method suitable for solving coordinates by direct linear transformation method is proposed.(5)Based on the three-dimensional reconstruction theory,the digital surface model reconstruction of slope image obtained by non-measured camera is carried out.The reconstruction steps include feature extraction and matching of slope image,monocular incremental sparse reconstruction,dense reconstruction,triangulation of 3D point cloud and texture mapping of triangular mesh model.The slope reconstruction model based on smart phone lays a foundation for visual visualization,comprehensiveness and analogy of slope monitoring data.