基于ETM^+的遥感影像信息提取研究  

Information extraction from remote sensing images based on ETM^+

作　　者：武文波[1] 陈静[1] 

机构地区：[1]辽宁工程技术大学测绘与地理科学学院,辽宁阜新123000

出　　处：《甘肃农业大学学报》2008年第5期142-146,共5页Journal of Gansu Agricultural University

基　　金：教育部博士点基金"基于3S的矿区资源环境监测与辅助决策系统研究"(编号:20050147002);辽宁省高等学校重点实验室项目"基于采矿对环境影响特征知识库的遥感监测研究"(编号:20060370)

摘　　要：以沈阳市苏家屯区为试验区,对ETM+图像的光谱信息和纹理信息进行综合分析,以达到提高影像分类精度的目的.利用光谱信息提取水体、植被;采用基于灰度共生矩阵的纹理量的分类法,通过TM5波段提取灰度共生矩阵和灰度联合矩阵,计算并提取最能反映类别差异的纹理量值将光谱信息混淆的水田、旱田、居民地用分离,得到最终的分类结果.结果表明:将纹理特征应用于图像分类中可区分光谱混淆的地类,光谱与纹理特征结合得到的分类精度要高于单纯光谱的分类精度.Taking the Sujiatun District of Shenyang City as an example,a comprehensive analysis of the spectrum information with texture information of ETM^＋ images had been analyzed in this paper n order to improve the classification accuracy of remote sensing images. The spectrum information was used for extraction of water bodies and vegetations. The classification method of texture quantity based on the gray level co-occurrence matrix was applied to classify ETM^＋ images. Among which, the gray level co-occurrence matrix and gray symbiotic matrix were extracted from TM5 band,and the optimal value of texture quantity able to reflect difference in ground objects was calculated and extracted, so as to distinguish the ground objects that spectrum information easy to be confused such as paddy field, glebe and residential area, and finally, the image classification result was obtained. Furthermore, the final result was compared with the classification based on the spectrum information only using ERDAS. The results indicated that application of texture feature in images classification was able to distinguish the ground objects whose spectrum information often had been confused,and the information combination of spectrum with texture feature improved the classification accuracy compared with only using spectrum.

关 键 词：遥感影像 光谱特征 纹理特征 灰度共生矩阵 分层提取 土地利用 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置] TP753[自动化与计算机技术—控制科学与工程]