Image Processing-based Cracks Monitoring Technology In The Trailing Edge Of Landslide

As a common type of geological hazards,landslides can be monitored scientifically and reasonably,which can effectively reduce the threat to people's lives and property.The crack curve of the back edge of a landslide is the crack curve between the back edge of the landslide and the immovable mountain body after the landslide movement occurs.Its changing trend represents the movement track of the landslide body.The common monitoring method of the crack curve of the back edge of a landslide is to install angle and displacement sensors on the landslide disaster body and obtain the deformation data of the landslide through mechanical measurement.However,this method has high cost,low precision degree and measurement data are not comprehensive and other shortcomings.Digital image processing was born in the 1950 s.It uses image data collected by image acquisition equipment such as camera as data source.After image denoising,image enhancement,image segmentation and other operations,the ultimate goal is to extract the information of the target area.Compared with traditional mechanical measurement methods,digital image processing has the advantages of fast measurement,comprehensive data and low cost.In view of the existing problems in landslide monitoring and the technical advantages of image processing,this paper carries out the research of landslide rear edge crack displacement monitoring technology based on image processing.The specific contents are as follows:(a)Study on the characteristics of landslide hazards.Emphasis is laid on the study of the causes of landslide formation and the influencing factors of landslide sliding,as well as the classification of the shape of the crack curve at the back edge of the landslide.(b)Research on identification method of crack curve at the back edge of landslide.In-depth study of the theoretical feasibility of image processing technology applied to landslide hazard body,the use of camera to collect landslide hazard body image,and image enhancement and noise reduction processing in turn,by virtue of the large color contrast between the landslide rear edge and the immobile mountain body,the crack curve of the landslide rear edge can be identified.(c)Research on extraction method of displacement information of crack curve at the back edge of landslide.In this paper,the trend of crack identification curves of various landslides is analyzed in depth,and the "interval median comparison" algorithm is designed to simplify the original bending folding curve as a single strike curve.The angle difference of the landslide trailing edge curve is taken as the azimuth displacement value of the crack curve,and the distance of the landslide trailing edge crack curve is taken as the sliding displacement of the crack curve.At the interface,the azimuth deviation direction and the maximum displacement change point are marked to visually show the displacement change trend of the crack curve at the back edge of the landslide.(d)Research on three-dimensional display technology of crack curve at the back edge of landslide.In-depth study of the implementation conditions of three-dimensional display of crack curves,acquisition of the same landslide hazard body face image and overhead image,in turn,registration of two image two-dimensional pixels to form three-dimensional pixels.Two main achievements have been achieved:(a)Recognition of landslide crack curve.The corresponding code compilation and system debugging are completed.The designed system can identify the crack curve of the landslide trailing edge.The recognition curve results are basically consistent with the trend of the crack curve of the landslide trailing edge in the image source.(b)Description of displacement change of landslide crack curve.Quantitative description of azimuth deviation and displacement change of fracture curve is completed.The method is applied to 3D simulation model,gravel model,soil model and landslide of Panzhihua Airport.The results show that the design method is effective.The innovation of this paper is embodied in:(a)Data comprehensiveness.In traditional landslide monitoring,owing to the fact that the crack curve of the back edge of the landslide is generally long and the number of monitoring sensors is limited,geologists can only judge the possible sliding position of the next stage of the landslide hazard body based on experience and install sensors.This method can easily omit the real and effective landslide movement data,that is,there is a "monitoring blind area".The design method of this system makes use of the powerful computing power of the computer to calculate the displacement values of all sliding positions by traversal calculation,which makes the monitoring data more comprehensive.(b)The expression of landslide crack movement.In view of the unclear(or unified)expression of the change of the movement of landslide cracks,a simple and quantifiable expression method of the movement of landslide cracks is proposed,which fully considers the relationship between the movement of landslide disasters and the shape of the crack curve at the back edge of landslide.