Résumé
Unlike airborne multi-echo laser scanner systems, full-waveform systems are able to digitize and record the entire backscattered signal of each laser pulse. It has been demonstrated that decomposing the return waveforms into a mixture of Gaussian components was suitable. In this paper, we focus on the improvement of peak detection and of raw signal modelling. Refined peak detection greatly increased the number of detected targets as well as their positional accuracy. Models more complex than the Gaussian model, such as the Lognormal or generalized Gaussian functions, were introduced and their contribution to waveform processing was studied. In this way, fitting of asymmetric, peaked or flattened echoes located both in urban and forested areas could be improved. Moreover, introduction of new echo parameters allowed the extraction of additional information on the target shape. This should make easier the decorrelation of geometric and radiometric influences on the signal and, as a consequence, the improvement of point cloud classification algorithms.