基于粒子群—最小二乘法的GLAS波形分解及树高反演方法

An Approach to Decompose ICESat/GLAS Data Waveform and Estimate Canopy Height Based on PSO-LSM Method

树高是陆地碳循环研究中的一个重要的生物物理参数。ICESat/GLAS系统是第一个实现全球连续观测的星载激光雷达系统,由于其具有光斑覆盖面积大、穿透力强的特征,在树高等森林结构参数反演方面具有显著的优势。该文设计了一种基于粒子群一最小二乘法的GLAS波形分解及树高反演方法,采用该方法对贡嘎山地区和茂县山区坡度小于10°的GLAS激光点波形进行了高斯分解和树高反演工作,拟合波形与原始波形吻合良好;树高估算结果与采用统计模型估算的树高非常接近(贡嘎山:R^2=0.80,RMSE=5.30 m;茂县:R^2=0.96,RMSE=3.11m)。研究结果表明该方法能很好应用于GLAS波形分解,并获得较好的树高反演结果。未来可以针对该方法在不同坡度地区的适用性和树高反演精度问题开展进一步的研究。

Canopy height is an important biophysical parameter for terrestrial car bon cycle study.The Geo science Laser Altimeter System(GLAS)onbo ard NASAcs Ice,Cloud and land Elevation Satellite(ICESat)provides a new potential for estimatingforest canopy height worldwide.Analysis of the retur ned waveform over area of low topog raphy allows for the direct retrieval ofcanopy heig hts.This study presented a new waveform decomposition method based on particle swarm optimizatio nleast squaremethod(PSO-LSM).This method decompo sed the GLAS waveform into a series of Gaussian com ponents.The initial parameters of Gaussian decomposition were obtained using particle swarm optimization algorithm,and then they were optimized usingthe LevenburgMar quardt method.The first peak from ground direction,the intensity o f which exceeded the tenth of max imumwaveform intensity,was identified as ground peak.Eventually,the canopy height wasest imated from vertical difference betweenthe signalcs start and the ground peak.The newly proposed method was applied to Gongga and Maox ian mo untain fo rest.Theresults indicated that the sum of Gaussian components could be an accurate r epr esentation of the original w avefo rm.And the est imated canopy height matched well with the height calculated by the empirical model established by Hayashi et al.in2013(Gongga mountain:R2=0.80,RMSE=5.30m,Maox ian:R2=0.96,RMSE=3.11m).However,the applicability and accuracyof the appr oach in ar eas wit h differ ent slopes need to be furt her st udied.