Data Processing And Application Of3D Laser Scanning To Dam Subsidence Monitoring In Mining Area

Three-dimensional laser scanning technology as another new innovation ofscience and technology, in many engineering fields currently get a very goodapplication, reflects its technological advantages. But the technology also there is abig problem in dealing with point cloud data, especially in the face of massive pointcloud data, in order to obtain high-precision quality data, the need to seek a way toefficient, stable and high degree of automation of the point cloud data processingmethod, which is the research focus of the technology of today, so the study ofthree-dimensional laser scanning point cloud data processing technology, seeking arobust, efficient and a more systematic method of point cloud data processing isnecessary.Based on analysis of previous studies,this paper focuses on point cloud datafiltering and registration included in point cloud data processing steps. The main workand results are as follows:1. For ordered point cloud data containing noise, we did three experiments usingthe median filtering, mean filtering and Gaussian filtering algorithm, respectively.Theexperimental results show that the Gaussian filtering method is better than others andable to remove the noise while maintain the original appearance of the original pointcloud data.2. For scattered point cloud data containing noise, we did two experiments usingthe bilateral filtering and the Laplace filtering algorithms, respectively.Theexperimental results show that the Laplace algorithm can effectively smoothprocessing of point cloud data, but can not effective to maintain point cloudcharacteristics, bilateral filtering algorithm can remove noise and can maintain thecharacteristics of the point cloud data at the same time.3. The model and the solution method of point cloud registration based onfeature were discussed in detail. The real scan data was used to experiment with themethod mentioned above and a good registration result was obtained.4. On the basis of analysis of the traditional ICP registration method,a methodbased on Delaunay triangulation ICP registration method was used, the method canquickly search the nearest point required by ICP registration process through buildingDelaunay triangulation of the original point cloud. The method does not need to select the feature points as the initial registration, which simplifies the process of ICPregistration method, the feasibility of the method is verified by experiments.5. Take the dam in the mining area for example, three-dimensional laser scanningtechnology was applied to mining subsidence monitoring and the subsidence curvesof dam was obtained. We compared3D laser scanning monitoring values withleveling data, the results showed that both of them had a high level of consistency,thus proved the feasibility of the technology for mine subsidence monitoring.