摘要
在实现大脚印激光雷达GLAS森林冠顶高度反演算法基础上,建立了复杂地形条件下森林地上生物量神经网络反演模型,制作了研究区森林地上生物量分布图。总体上,激光雷达GLAS森林冠顶高度和地上生物量估算精度较高。森林冠顶高度针叶林精度最好(R2=0.692);阔叶林次之(R2=0.5062);地上生物量反演结果与实测结果十分接近,在空间分布上与土地覆盖分布特征非常一致。
Based on the algorithm of forest canopy height for GLAS data,the neural net model of above ground biomass in complex terrain conditions was established.The map of forest aboveground biomass from BP neural net model was produced.Overall,forest canopy height and aboveground biomass have higher accuracy.The result of forest canopy height of needle-leaf forest has highest accuracy（R2=0.692）.The result of broadleaf forest has higher accuracy（R2=0.5062）.The results of forest aboveground biomass are very close to the fields measured results,and are consistent with land cover data in the spatial distribution.
出处
《北京大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2011年第4期703-710,共8页
Acta Scientiarum Naturalium Universitatis Pekinensis
基金
国务院三峡工程建设委员会项目(SX2002-004)
863计划(2009AA12Z150)资助
