Extract forest information using the mixed end-members from images and field survey data

摘要

Using linear unmixed model to obtain forest information is widely accepted at present, in which extraction of end-members is mainly acquired from remote sensing images directly. In this paper, the advantage of extracting the end-member images is analysed. The linear unmixed model was applied to the mixed end-members. The results of the research show that the computer simulation results are accurate and correct rate is higher than 85% with 60 random sample points from field investigation.