【摘要】 本文叙述应用计算机自动识别分类技术在航天遥感图像中对森林资源进行分类的一些探索研究。本文介绍了采用自编程序“地理坐标系和图像坐标系的精密仿射变换”、“识别型增强技术”、“多光谱信息的特征提取”、“多光谱模糊算法聚类分析”等一系列算法后所取得的分类图像结果,与实地及目视解译相比较,森林分类相对精度达90％。更多还原

【Abstract】 An approach to the computerized recognation and classification of different types of forest and it’s distribution from satellite image data is described. Some algorithms, such as "THE RECOGNIZING ENHANCEMENT TECHNIQUE", "THE SUPERVISION OF CLASSIFICATION BY A MULTISPECTRUMS FUZZY ALGORITHM", will be discussed in detail. The result of the classification will be shown in the form of photographs for the region under test, ist area is about 165,000 ha. Compared with the results of other methods, such as a visual interpretation of images or on sight feild classification of the tree types, our classification achieved a relative statistical accuracy of approximately about 86-90%.更多还原