Object Identification for Hyperspectral Image Based on Exhaustive Method

A new approach for object identification based on the exhaustive method is proposed.Each pixel spectrum in the hyperspectral imagery has an optimal matching spectrum selected from all possible reference spectra that are correlated with the study area.The selection is done by computing the maximal angle cosine or correlation coefficient.Thus,identification maps are created by the identification results of all pixels.Some reference spectra are changed against the information in the maps to obtain better identification results.This approach is applied to a hyperspectral image of the peripheral area of Pulang porphyry copper deposit,Zhongdian County,Yunnan Province,China.The results show that this method is effective and practical.