WAVE-FORM CLASSIFICATION AND INFORMATION EXTRACTION FROM LIDAR DATA BY NEURAL NETWORKS

Two different neural network schemes for the classification of light d etection and ranging (LIDAR) waveforms for the LARSEN 500 airborne sys tem and for extraction of ocean information are proposed. The first me thod employs a single layer of linear neurons for classification of wa veforms into various clusters. Both unsupervised and supervised learni ng algorithms have been employed to demonstrate the spatial distributi on of milt in near-shore waters, In the second method, a new multistag e multilayer feedforward architecture is used for the classification o f the waveforms and for the extraction of various types of ocean infor mation. The stage I networks work in a parallel fashion and map the in put waveforms to a set of characteristics. The networks in stage II us e these characteristics to assign a signature number to the waveform o r extract other information, Both the schemes are used with real-world data collected by the LARSEN 500 system, The paper concludes with exp erimental results and comparisons.