本文论述了ＴＭ卫星像片的图像处理技术，以日本北海道苫小牧地区森林为例，进行不同树种的计算机自动分类，并给出多光谱图像上不同年龄树种的灰度与年龄之间的相关图。研究结果表明：云杉属、冷杉、落叶松的反射率随年龄增大而降低，用最大似然法监督分类，可以很好地区分不同树种，得出苫小牧地区森林分布情况的遥感图像。 This paper discusses the image processing technology of TM satellite imagery. Taking the Tomo Komaki forest in Hokkaido, Japan as an example, the computer automatically classifies the different tree species and gives the correlation chart between the grayscale and age of tree species with different ages on multispectral images. The results showed that the reflectivity of Picea, Fir and Larix decreased with age, and using maximum likelihood method to supervise the classification could well distinguish different tree species and obtain the remote sensing image of forest distribution in Tiemao area.