利用SPACE LAB彩色红外影像及数据复合技术对藏北牧草资源进行计算机分类的研究

A Study on Computer Aided Classification for Grass Resource Investigation in Northern Tibet by Using Spacelab Colour Infrared images and Data Composite Techniques

本文就在藏北草资源调查中,应用空间实验室量测摄影机(MC)摄得的彩色红外影像,及根据遥感影像目视判读所利用的信息空间特点和草场分类原理进行计算机自动分类的理论与方法,对利用辅助数据和影像数据复合技术的几种尝试进行了讨论,并依实地调查资料和专家判读结果对各种分类结果进行了评价。说明利用这种航天遥感资料,辅以Landsat影像生物量指标、DTM和坡度、坡向数据以及计算机自动分类技术,进行资源调查是可行的,可以达到最佳效果。

This paper discusses the theories and method of computer aided classification for the grassland investigation by using the Spacelab color infrared images in accordance with the grassland, classification principles and the characteristics of information space used in the visual interpretations of remote sensing images. It takes the several kind of experiments as the instances of utilizing auxiliary data and image data composite techniques. The results have been evaluated by comparing those with the results of experts' interpretation according tot the on-the-spot surveying records. It shows us that it is flexible to use this kind of remote sensing dato for resources investigation, and one can get the optimum effects by employing Landsat MSS biomass index. DTM image, stope and its orientation data.