TLS-based Landslide Mass Data Preprocessing And Surface Feature Change Detection

In the southwest mountainous area of China,geological disasters such as landslides,collapses and mudslides occur frequently.Once disasters occur,they often cause serious economic losses and casualties.It is an important work to detect the change of landslides in the field of geological disasters.Efficient detection of landslides can effectively predict the disaster in advance and take countermeasures to prevent and reduce the hazards caused by landslides.Terrestrial laser scanning technology is a new surveying and mapping technology developed rapidly in recent ten years.It has been widely used in the field of landslide change detection because it can obtain the spatial 3D coordinates of the target object quickly,non-contact,high-precision and high-density.In this paper,an associated landslide body located in the debris flow ditch of the Dabaini River in Dong chuan District of Kunming City is used as the test object.The Terrestrial laser scanning technology is used to collect point cloud data from the landslide body for two periods.After the data preprocessing of point cloud registration and filtering,the DEM oriented to the landslide is generated by interpolation,and the ridge point,valley point and ridge line of the landslide are extracted based on the DEM of Arc GIS platform.Finally,based on the change detection and analysis of the point,line and surface features of the landslide,a set of change detection process oriented to the landslide is formed,which achieves the purpose of multi-directional and comprehensive change detection.In this paper,the following research results are finally obtained:(1)The point cloud filtering method for debris flow associated with landslide mass is studied:The first phase of landslide TLS point cloud data was used as the data source.Four filtering methods based on mathematical morphology filtering,slope filtering,progressive encryption irregular TIN filtering,and CSF algorithm filtering were used for filtering.Two aspects were evaluated for accuracy,and four methods were compared and summarized.The results show that:The total error of the filteringbased on the progressive encryption irregular TIN filtering method is 5.51%,and the effect is the best in the experimental area of this paper.It can well adapt to complex terrain and effectively filter dense low vegetation;filtering based on mathematical morphology method The total error is 7.61%,the process is tedious,and the practicability is poor;the slope-based filtering error is 14.61%,the filtering effect is the worst,and it is not suitable for test areas with complex terrain conditions;the total error of CSF filtering is 13.80%,which is relatively flat The area can achieve good filtering effect,but it is not suitable for the landslide body associated with debris flow ditch that contains both flat and steep areas.(2)The filtered point cloud is used as the data source,and the inverse distance weighting,natural neighborhood method,kriging method and ANUDEM algorithm are used to interpolate and generate DEM for landslide body.The checkpoint method is used to obtain the DEM interpolation elevation accuracy index(Medium error,average absolute error and fitting superiority),Matlab was used to draw the checkpoint point error distribution map,and the DEM interpolation accuracy was compared and analyzed.The results show:the kriging interpolation method is better for generating DEM elevation accuracy indicators,and it is more suitable for point cloud data types with uneven distribution of original data and high-altitude landforms.Based on this DEM,the Arc GIS platform is used for slope analysis to extract the ridge points,valley points,ridge lines and valley lines and other surface features.(3)The practical application of 3D laser scanning on the landslide associated with the mudstone ditch is analyzed and analyzed.The two phases of point cloud data obtained from the scanning are used as the basis for the characteristic points of the landslide surface.The comprehensive analysis of the change detection in the X,Y,and Z directions of characteristic features,characteristic lines,and "surface" features provides a new idea for landslide change detection.