Estimation of Coniferous Forest Height using Stereo Images of GF-7 Satellite and Airborne Lidar Data

Forest height is a key parameter of forest spatial structure, which is useful for forest management and ecosystem modelling. High resolution stereo images of satellite can be used to reconstruct the surface of ground objects. The objective of this study is to estimate the coniferous forest height combining stereo images of Gaofen-7 (GF-7) satellite and digital terrain model (DTM) extracted from airborne light detection and ranging (LiDAR). The study site is characterized by boreal coniferous forest located in the northern China. The dense point cloud from stereo images was normalized by subtracting the LiDAR DTM. The forest height was estimated by linear regression between the extracted features of normalized point cloud and the field measurements of stand height. The results shown that the values of estimated height were high correlated with the height percentile of 50% (R2 = 0.79, RMSE = 2.34m). The stereo images of GF-7 satellite have the great potential for regional forest monitoring.