论文部分内容阅读
Laser altimetry for forestry applications has been massively investigated recently, especially after the development of full waveform laser altimeter systems. They can record the entire emitted and backscattered signal of each laser pulse while the conventional airborne multi-echo laser scanner systems are only able to record a few echoes. Generally, the profiles of the forestry full waveform data can be considered as Gaussian mixture and solved by using the Gaussian decomposition, however, the number of Gaussian in the signal must be known in advance. In this study, we propose an adaptive Gaussian decomposition. Instead of estimating or guessing the number of Gaussian, it iteratively fits Gaussian locally and removes it from the signal and the stopping criterion is met when the magnitude of residual signal is less than the preset threshold. The performance analysis will be conducted with both simulated and real full waveform laser altimeter signal. The decomposition result can also provide information for land-cover classification including the high and low vegetation, urban, and bare land.