津渤地区陆地卫星影象数字处理的环境背景效果分析

ANALYSIS OF THE EFFECT OF ENVIRONMENTAL BACKGROUND BASED ON DIGITAL PROCESSING OF LANDSAT IMAGERY AT JING-BO REGION, CHINA

陆地卫星数字数据经计算机图象增强和分类处理,可获得多种增强影象和土地复盖分类图等。本文探讨了对上述各种计算机彩色图象的分析方法及其在环境背景研究中的作用和效果。

The main objective of the investigation is an evaluation of the method and effect of regional environmental synthetic analysis based on LANDSAT CCT data digital processing technique. Some distinct data handling methods for corresponding special natural environmental characteristics were undertaken in order to obtain expectant interpreting effects on environmental background. Here three types of data enhancement techniques, linear density slicing,, uniform distribution and local stretch were adopted. Colour composites generated from these enhancement processings provided clearer resolution, better definition of boundaries and more information. Besides, automated classifications using computer imagery processing program from STC/IIS 101 system were tested, and finally computer land cover classifying picture participated with visual interpretation was drawn. The three types of enhancement colour composites mentioned above, which was output respectively on the colour screen of monitor, displayed different effects on the interpretation.However, colour composites of uniform distribution and local stretch have special value in explaining environmental background. For example, the objectivees on the land surface, such as arid cultivated land, irrigated land, residential area of village or city, drainage system including pond, reservoir and river etc., and the sand spit in the mouth of the HAI RIVER, diffusion of industrial smoke and so on can be clearly recognized on the colour composite imagery of uniform distribution stretch. And the fine structure of the water surface is particularly clear on the colour composite imagery of the local enhancement. More information on the ocean surface along the coast of BO-HAI BAY, water surface of BEI-DA-GANG reservoir and circumstances of salt pool at the eastern coast of TANG-KU city are visible too. Therefore they are useful stretch techniques, which improves the brightness contrast of the original imagery and obtains more valuable information than the original imegery for the interpreter.Afterword, three types of cluster, minimum distance and maximum likelihood algorithm classifications were undertaken at this region. The best result that was verified by comparing with visual interpreting of the main types of land cover on streching imagery belongs to unsupervised cluster classifying technique.There are only eight classes displayed by different colours on the colour imagery of cluster classification. Finally, the result of land cover interpreting map was produced by combining visual explanation with the computer colour output of cluster classification. Land cover interpreting picture containing about sixteen types of three classes was drawn on colour imagery of computer automatic cluster classification