基于多源遥感数据的森林植被类型分类方法研究详解.docVIP

基于多源遥感数据的森林植被类型分类方法研究 摘 要：森林是地球上最大的陆地生态系统，是人类赖以生存和发展的必要基础。它不仅给人类提供丰富的木材和林副产品，而且在调节气候、涵养水源、保护环境等方面均起到重要作用。因此，开展森林资源调查，掌握森林资源现状及其变化，对于提高林业发展决策水平，促进林业和社会经济乃至全球环境的可持续发展等具有极为重要的意义。 本文利用云南省迪庆藏族自治州五境乡、小中甸、上江乡的SPOT 2.5米分辨率全色影像与10米分辨率多光谱影像融合后形成的多遥感数据对森林植被类型分类方法进行研究，利用ENVI图像处理软件分别进行了传统的监督分类中的最小距离法、最大似然法和基于专家知识的决策树分类法的实验，而且还尝试了面向对象的新的分类方法，对不同的分类方法进行了分类精度的对比，从而提出一个基于多遥感数据的森林植被类型分类方法的可行性建议。 关键词：遥感；图像分类；SPOT5；最小距离法；最大似然法；面向对象 Classification of Forest Types Based on Multi-source Remote Sensing Data Abstract: Forest is the largest terrestrial ecosystem on the earth, it is a necessary basis for human existence and development. It is not only to provide people with rich timber and forest by-products, but also play an important role in climate regulation, water conservation and protect environment. Therefore, developing forest resources investigation, getting the status and changes of forest resources, for improving the level of decision-making to promote the forestry development, forestry and social economy and even the global environment sustainable development has the extremely vital significance Based on remote sensing data acquired by the Diqing Tibetan Autonomous Prefecture in Yunnan Province five Jing Xiang, Zhongdian, Jiang Xiang which is researched on the types of forest vegetation classification, including the minimum distance supervised classification in the traditional method, maximum likelihood method and decision tree based on expert knowledge classification experiments by ENVI and object-oriented classification by ENVI EX，on different classification methods were compared by the classification accuracy, and put forward a feasible suggestion based on the types of forest vegetation classification method . Keywords: Remote sensing; Image classification; SPOT5; minimum distance classifier; maximum likelihood classifier; object-oriented classification. 目 录 第一章 绪论 5 1.1 研究背景 5 1.2 研究内容及意义 6 1.3 研究方法与技术路线 6 第二章 图像分类原理 8 2.1遥感图像分类传统方法 8 2.2.1 监督分类 8 2.3 遥感图像分类新方法