The extraction of linear features from airborne full-waveform LiDAR

Full-waveform (FWF) lidar is an advanced technology inlidar system development. Multi-echo (ME) lidar only provides 3-D point clouds and intensity, whereas FWF lidar provides entirereturned signals. The waveform attributesarethe useful information in feature extraction. However, discrete point clouds without waveform attributes are usually utilized for geometrical feature extraction. In this research, we use the waveform information to detect edges in urban area and to compare the result from discrete point clouds. The proposed method includes two major parts. First, we compute the separation distance of 3-D point clouds according to the height jump of edge points. When theseparation distances exceed the predefined threshold, we set them as the candidate edge points. Second, the different edges show different characteristics and waveform attributes. So we utilize the echo width and amplitude of the candidate edge points to remove the vegetation area. Moreover, the waveform attributes are used to discriminate the objects such as "buildings" and "power lines". The experimental data acquired by Riegl LMS-Q680i is located in Tainan city, the southern part of Taiwan. The experimental result indicates that the proposed method may separate the edge points from different objects when considering the waveform attributes.