Study On The Extraction Of Corresponding Feature Points From3D Laser Scanning Data

3D laser scanning technology is the development trend of the engineering surveytechnology in the future. Its emergence has been hailed as a technology revolution inthe field of surveying and mapping after the GPS technology is the once again. Thetechnology has a more and more important effect on a number of fields due to itsquickness, precision and measure without contact. After introducing the terrestriallaser scanner into the study of mining subsidence,it solved a lot of problems intraditional observation station,such as inadequacy of data volume and long period ofobservation etc. However,there are still some defects in the application of thistechnique in mining subsidence monitoring. Currently using this technique caneffectively obtain the surface subsidence value, but not direct access to the horizontaldisplacement. It needs to take advantage of feature points in the point cloud data toacquire the surface horizontal displacement.This thesis studies how to acquire the horizontal displacement combining withthe characteristics of3D laser scanning data, and proposes a horizontal displacementcalculating method based on feature extraction of objects: First of all, using thedistance weighted average method to eliminate the ground data, and save thenon-ground data; Then using the method based on grid handling and binarization tosegment the independent ground object, according to the aggregation characteristicsof point cloud data; Then discriminating the feature type of point cloud data bycalculating the normal vector and curvature, and choosing the correspondingalgorithm to extract the feature; At last, matching the corresponding feature pointsextracted from two periods of observation data to acquire the surface horizontaldisplacement.This thesis selects the monitoring data of two mining areas to validate theeffectiveness of the method presented in this paper. This thesis acquires the horizontaldisplacement of the two mining areas, using the method presented in this paper, andcompares the results with that acquired by conventional measuring means. The resultverifies the validity of the algorithm in this paper.