Research And Application Of Vegetation Filtering Algorithm For Three-dimensional Laser Point Cloud Of Rock Slope In Medium And Low Density Vegetation Are

In recent years,the application market of 3D point cloud has grown rapidly,and it has become one of the four major fields alongside GNSS positioning,GIS spatial analysis and remote sensing in the traditional geospatial industry,and it is the fastest growing market.With the rapid development of smart cities,intelligent transportation,global mapping and other industries,the demand and requirements for point cloud data are getting higher and higher,but the original point cloud data obtained through photogrammetry or lidar technology is affected by many factors and produces many noise points,which hinders the construction of accurate3 D digital models.In rock mass engineering,the formation of rock mass structural surface is very important for geotechnical engineering research,and the rapid and accurate acquisition of structural surface information through 3D point cloud is one of the key links in the application of 3D point cloud in rock mass engineering.The rock mass point cloud data is seriously affected by the vegetation point cloud,which will interfere with the extraction of the surface shape information of the measured rock mass,and even cannot be extracted,so the vegetation noise point must be processed.In this paper,the vegetation filtering algorithm of rock mass point cloud in medium and low density vegetation area is studied,breaking through the idea that the traditional vegetation filtering algorithm only uses a single spatial position information or RGB color information,and combines the two to form a new vegetation filtering algorithm,which verifies the effectiveness of the algorithm and applies it to the extraction of rock mass structural surfaces.It mainly includes the following:1.Based on the characteristics of point cloud data provided by commonly used instruments,a new vegetation index,red-green difference index,is proposed,and vegetation filtering is carried out by setting thresholds,and the filtering results obtained by setting different thresholds are analyzed and compared.When the vegetation filtering of the high and steep rock slope is carried out,a good vegetation filtering effect is obtained,but there is a problem of filtering out more ground points,and a grid evaluation of the surface features is proposed to preliminarily evaluate the retention of surface features.2.Aiming at the situation that the threshold filtering algorithm based on red-green difference index deletes more ground points,this paper proposes a filtering algorithm combining spatial position information and red-green difference index.The algorithm selects vegetation seed points by using the point cloud distribution characteristics,and flexibly filters out vegetation points based on the red-green difference index,so as to ensure that more ground points are retained while filtering out vegetation points.3.The K-means clustering algorithm and principal component analysis method are used to simplify the algorithm structure,and the characteristics of great similarity of the point cloud data in the cluster are used to improve the accuracy of the algorithm in filtering out the vegetation point cloud,which can not only remove more vegetation points,but also retain most of the ground points.4.The engineering rock mass structure intelligent interpretation system developed by our research group is used to show the interpretation results of rock mass structural surface before and after filtering,which not only proves that the vegetation filtering algorithm is the key step to accurately interpret rock mass structural surface,but also reflects the effectiveness of the vegetation filtering algorithm proposed in this paper.