A new filtering algorithm for lidar data fused with image segmentation information

The traditional LiDAR Point Cloud filtering and classification algorithms are mostly based on the elevation information, which are difficult to distinguish bare-ground and non-ground point with this single data source. In this article, fused image information such is used in the filtering process. Through the image segmentation, each segmentation object are extracted with the shape and spectral information, then these vector space objects and point cloud are matched and the decision tree is set up based on topology relationship for filtering. In order to test the performance of the proposed algorithm, test area with thick buildings and roads were applied, which could achieve a better filtering effect.