Research Of Forest Tree Height Extraction For The Laser Altimeter Satellite Full Waveform Data

Spaceborne lidar which with high precision, speed, penetration and other advantages, was widely used in measuring the polar ice caps, aerosols thickness and vertical structure of the forest. The introduction of full waveform data provides a new avenue for spaceborne laser used to extract the forest tree height. In this paper,the foreign ICESat / GLAS(Ice, Cloud and Elevation Satellite / Geosciences Laser Altimeter System) data was used for example to research the full waveform data processing and tree height extracting for the spaceborn lidar.First, in the basis of the inflection point Gauss decomposition method the findpeaks characteristics was take into account, put out a new full waveform Gauss Decomposition Algorithm which based on the crest of automatic identification. Then we contrasted the three methods and selected any study area to extract the tree height in Changbai Mountain Woodland as experimental area to prove the effectiveness of the new method.In this article, some work was done which as follows:(1)In this paper, we sorted the development path of the spaceborne laser system, summed up the classification system and the laser altimeter measurement principle, then introduced the common types of spaceborne laser altimeter system;Described the full waveform data processing for spaceborne lidar, which including preprocessing(voltage conversion, waveform normalized background noise estimation and the Gaussian smoothing, etc.),waveform decomposition(Gauss decomposition, wavelet decomposition, waveform deconvolution and waveform simulation) and curve fitting(nonlinear least squares fitting);Compared and analysised several common forest tree height estimation models which use the spaceborne laser altimetry satellite ICESat / GLAS data, and described two common indexs to estimate the accuracy evaluation;(2) To desolve the intensive computation for waveform Gaussian decomposition of the parity inflection method, we take the findpeaks charcateristics into account, put out a new full waveform Gauss Decomposition Algorithm which based on the crest of automatic identification. We have decomposed some waveforms and contrasted the precision between the three methods.(3) Finally, in this paper,we selected 304,173 hectares forest area of Wangqing business in Changbai Mountain Woodland as experimental area, picked out part of GLAS sample data in the experimental area to process the waveform and extract the forest tree height,then we contrasted the results with the measurement data. We have drawed the conclusions: the tree height’s extraction precision of the new full waveform Gauss Decomposition Algorithm which based on the crest of automatic identification is lower than the parity inflection Gauss decomposition,but higher than the findpeaks function Gaussian decomposition. Although lower then parity inflection Gauss decomposition on the precision, but the new method has a significant reducing on the computation which can be used to quickly extracting the forest tree hight over a wide range area.