基于Landsat长时间序列数据估算树高和生物量

摘要

以Landsat长时间序列数据为研究对象,旨在以光谱序列信息反演森林参数为视角,应用Landtrendr算法从时间序列数据中提取森林扰动变量,使用随机森林计算方法建立扰动变量、反射率和GLAS激光点森林参数之间的关系模型,获取树高和生物量的空间分布信息.为多源遥感数据反演森林参数提供参考,研究证明基于Landsat长时间序列数据获得的森林扰动变量能够增强反射率和森林参数之间的相关性,可提高预测精度.%This paper, taking the long time series data of Landsat as the research object, aims at retrieving forest parameters from spectral sequence information.The variables of forest disturbance are extracted from the time series data using the Landtrendr algorithm.With the random forest method to establish the model of the relationship among the disturbance variables, the reflectivity and the GLAS laser point, the spatial distribution information of the tree height and biomass is obtained, which provides a reference for retrieving forest parameters from multi source remote sensing data.The research proves that the forest disturbance variables obtained from the Landsat long time series data can enhance the correlation between the reflectance and forest parameters to improve the prediction accuracy.