Abstract: The remote sensing technique is applied extensively in mineral prospecting， but it is strongly restricted in highvegetated area. Multispectral data， i. e. TM5， TM4 and TM2/TM3 were enhanced by principal component analysis and linear stretch method， and it is found that the vegetations between mining area and background area showed obvious difference in the composite image. The vegetation in mining area is dark red， while fresh pink in background area， based on which， the composite image data is classified. The results indicate that the botanical anomalies caused by mineralized alteration are well extracted by the methods mentioned above， which is helpful to the remote sensing exploration in highlyvegetated area of Northeast China.

Key words: botanical anomaly, information extraction, remote sensing, mineral resource prospecting, Hongtoushan area