Exploiting full-waveform lidar data and multiresolution wavelet analysis for vertical object detection and recognition

A current challenge in performing airport obstruction surveys using airborne lidar is lack of reliable, automated methods for extracting and attributing vertical objects from the lidar data. This paper presents a new approach to solving this problem, taking advantage of the additional data provided by full-waveform systems. The procedure entails first deconvolving and georeferencing the lidar waveform data to create dense, detailed point clouds in which the vertical structure of objects, such as trees, towers, and buildings, is well characterized. The point clouds are then voxelized to produce high-resolution volumes of lidar intensity values, and a 3D wavelet decomposition is computed. Vertical object detection and recognition is performed in the wavelet domain using a multiresolution template matching approach. The method was tested using lidar waveform data and ground truth collected for project areas in Madison, Wisconsin. Preliminary results demonstrate the potential of the approach.