Compressing and Classifying Lidar Waveform Data

Today’s advanced LiDAR systems are able to record the entire laser echo pulse, provided that sufficient data storage is available on the airborne platform. The recorded echo pulses, frequently called waveform data or full-waveform, can then be used to analyze the properties of the reflecting surface, such as classifying objects based on their material signatures; for example, land classification. However, both the efficient storage of waveform data and the waveform-based classification still present formidable challenges. In this paper, solutions based on state-of-the-art numerical methods, including the Discrete Wavelet Transform and Kohonen’s SelfOrganizing Map, are proposed to carry out these tasks. Using the Discrete Wavelet Transform has two advantages: first, it is an efficient tool to compress waveform data, and second, the wavelet coefficients describe the shape of the echo pulse, and, therefore, they can also be used for classification. The performance of the proposed method is evaluated using actual waveform data.