Airborne LiDAR Filtering By Fusing Multi-features

Airborne LiDAR (Light Detection and Ranging), integrated GPS technology, inertial navigation technology and laser range finding technology, is a newly emerged instrument of terrain mapping. As an active remote sensing system, with the ability of fast capturing 3D coordinates of targets, LiDAR has become one of important data collecting method in the field of Remote Sensing.The paper presents airborne LiDAR system, LiDAR data's characteristics and LiDAR system errors. We discuss some key techniques in data processing of LiDAR, including outliers elimination method of Airborne LiDAR and filtering of Airborne LiDAR. The primary innovations and works are included as:With the continuously improvement of airborne LiDAR and frequency of the system increasing, there are lots of outliers in the data which received from airborne LiDAR. The existing methods of outliers removed cannot completely detect outliers for existing in"meteor shower". The paper improves neighboring points searching method for removing outliers. First, using analysis of histograms frequency remove outliers which higher or lower than the ground points and the object points. Then, the neighboring points searching detect the outliers which are near to ground points and the object points. The method can effectively remove outliers which exist in"meteor shower", reduce the influence for filtering and retain the original surface features.There are some differences in dealing with complex area using current filtering method of airborne LiDAR. So it requires a large amount of manual interaction, it will take 60%—80% processing time. The paper use multi pulses, height, slope to introduce the filtering method for airborne LiDAR of mixed multi-feature. The method puts only pulse and last pulse into a 3D boxing, divides the 3D boxing into n 3D boxing. Comparison elevation and slope in 3D boxing filter the object points. In the filtering, constantly changing the bottom area of 3D boxing meet the needs of filtering. The method using for complex object and dense vegetation area can get better filtering results.