|LI Junjie,HE Longhua,DAI Jingfang,LI Jinlian.Extract enclosure culture in lakes based on remote sensing image texture information. J. Lake Sci.2006,18(4):337-342. DOI:10.18307/2006.0403
|关键词: 中巴资源卫星 纹理 湖泊围网养殖 决策树
|Extract enclosure culture in lakes based on remote sensing image texture information
LI Junjie1,2, HE Longhua1, DAI Jingfang1, LI Jinlian1,2
1.Nanjing Institute of Geography and Limnology, Chinese Academy of Sciences, Nanjing 210008, P. R. China;2.Graduate School, Chinese Academy of Sciences, Beijing 100039, P. R. China
|Most lakes in eastern China are exploited in the form of enclosure culture. The high density enclosure culture area is easy to cause lake eutrophication and deteriorate the water quality. So the lake enclosure culture area is focused by people and is usually extracted by using the multi-temporal remote sensing data and the multi-spectral classification or the manual digitization means. The multi-spectral classification is easy to confuse enclosure culture area with the natural water area and the manual digitization demands much time. This paper suggests a new method using the image texture information to extract lake enclosure area. The result is more accurate and the method demands less expense. Lake Baima is the study area and the image of the area is multi-spectral data of China Brazil Earth Resources Satellite 02. First the water area A the man-made area and the bare soil which has the similar special characteristics with the water are extracted by using the normalized difference water index. Then the principal component analysis was used to compress the multi-spectral data of the study area and enhance the texture informa-tion of the image. The third step Grey Level Co-occurrence Matrix (GLCM) was used to analyze the image, select a proper threshold of mean value which quantifies the GLCM. Finally, the lake enclosure culture area was extracted by using the decision tree.
|Key words: Chirm Brazil Earth Resources Satellite texture enclosure culture in lake decision tree