THE RESEARCH ON QUALITY INSPECTION OF LAND COVER CLASSIFICATION BASED ON LIDAR

Abstract

Aiming at the quality inspection of land cover classification in natural resources investigation and monitoring, a method and process for quickly inspecting land cover misclassification and omission by using airborne Lidar data are proposed. Firstly, the ground objects are separated from laser feet points by post-processing such as gross error elimination, multi echo analysis and filtering of airborne Lidar data. Secondly, the elevation information is used to distinguish the easily confused ground objects with the similar spectral characteristics, such as forest land and shrub grass, the building area and road, high-rise houses and low-rise houses. Finally, overlay and compare the classified laser feet points with land cover achieving the misclassification. Experiments shows the effectiveness of this method, improve the quality inspection efficiency of land cover classification results, reduce the workload of field data collection, and improve the accuracy of quality inspection.