Cheng Kai Wang

Airborne Dual-Wavelength LiDAR Data for Classifying Land Cover

This study demonstrated the potential of using dual-wavelength airborne light detection and ranging (LiDAR) data to classify land cover. Dual-wavelength LiDAR data were acquired from two airborne LiDAR systems that emitted pulses of light in near-infrared (NIR) and middle-infrared (MIR) lasers. The major features of the LiDAR data, such as surface height, echo width, and dual-wavelength amplitude, were used to represent the characteristics of land cover. Based on the major features of land cover, a support vector machine was used to classify six types of suburban land cover: road and gravel, bare soil, low vegetation, high vegetation, roofs, and water bodies. Results show that using dual-wavelength LiDAR-derived information (e.g., amplitudes at NIR and MIR wavelengths) could compensate for the limitations of using single-wavelength LiDAR information (i.e., poor discrimination of low vegetation) when classifying land cover.

Waveform-based point cloud classification in land-cover identification

Abstract Full-waveform topographic LiDAR data provide more detailed information about objects along the path of a laser pulse than discrete-return (echo) topographic LiDAR data. Full-waveform topographic LiDAR data consist of a succession of cross-section profiles of landscapes and each waveform can be decomposed into a sum of echoes. The echo number reveals critical information in classifying land cover types. Most land covers contain one echo, whereas topographic LiDAR data in trees and roof edges contained multi-echo waveform features. To identify land-cover types, waveform-based classifier was integrated single-echo and multi-echo classifiers for point cloud classification. The experimental area was the Namasha district of Southern Taiwan, and the land-cover objects were categorized as roads, trees (canopy), grass (grass and crop), bare (bare ground), and buildings (buildings and roof edges). Waveform features were analyzed with respect to the single- and multi-echo laser-path samples, and the critical waveform features were selected according to the Bhattacharyya distance. Next, waveform-based classifiers were performed using support vector machine (SVM) with the local, spatial features of waveform topographic LiDAR information, and optical image information. Results showed that by using fused waveform and optical information, the waveform-based classifiers achieved the highest overall accuracy in identifying land-cover point clouds among the models, especially when compared to an echo-based classifier.

A Wavelet-Based Echo Detector for Waveform LiDAR Data

This paper presents a wavelet-based (WB) echo detector that can recover the echoes missed by a light detection and ranging (LiDAR) system via on-the-fly detection. An on-the-fly detection method normally utilizes a simple threshold (TH) to register a target point. Points that belong to weak and/or overlapping echoes are much complicated and are easily missed by TH approaches. The proposed detector based on wavelet transformation is robust to noise and is capable of resolving overlapping echoes. It is thus expected to be good at handling missing echoes. A simulated waveform data set and a real waveform data set of a forest area were both used in this paper. The simulated waveform data were utilized to compare the proposed detector with zero crossing (ZC) and Gaussian decomposition (GD) detectors in terms of their ability to deal with weak or overlapping echoes. The real waveform data set acquired from Leica ALS60 was used to demonstrate a WB algorithm for exploring the missing echoes. Experiments using the simulated data showed that the WB and GD detectors are superior to the ZC detector in finding overlapping echoes. The WB algorithm performs well when dealing with overlapping echoes with a low signal-to-noise ratio. Experiments using the real waveform data show that 31.5% additional weak or overlapping echoes can be detected by the WB detector compared with the point cloud provided by the system. With such additional points, the mean and root-mean-square errors of the digital elevation model differences can be improved from 0.72 and 0.79 m to 0.16 and 0.59 m, respectively.

Dual-directional profile filter for digital terrain model generation from airborne laser scanning data

Abstract The most important aspect of digital terrain model generation from airborne laser scanning (ALS) data is that of filtering a point cloud to obtain ground points. Numerous automatic filters have been proposed since ALS data became available. However, to filter out nonground points, a slope threshold is usually introduced to classify points into ground and nonground points; this leads to frequent over-filtering problems in cliff-like terrains. A solution to this problem is proposed, using a dual-directional slope-based filter originating from a conventional slope-based filter is proposed. This filter is designed as a directional filter in one dimension and is applied to every profile of light detection and ranging (LiDAR) points. In this process, a directional filter is first applied to the profile, and another directional filter is then applied at an angle of 180 deg from the first one. Each directional slope-based filter is complementary to the others, thus avoiding over-filtering. We utilize ISPRS LiDAR data for the test. A comparison of this filter approach with existing methods is presented. The comparison result shows that the proposed method obtains a classification accuracy that is as good as most of the compared methods, but is superior to them with regard to handling data from abrupt surfaces.