都江堰震后土地利用/覆被变化信息提取方法研究

快速和精确评估"5·12"汶川地震后的土地利用/覆被变化对科学减灾、灾后重建及生态环境恢复具有重要意义。常规方法从遥感图像上提取土地利用/覆被变化信息时,多以研究区整体为处理对象,直接对全图像进行分类提取,容易忽略地形地貌和地质构造等背景因素对分类结果的影响。本文以都江堰为试验区,根据地质构造展布特征及地形地貌发育形态将其分成平原区(Ⅰ区)、低山区(Ⅱ区)、中高山区(Ⅲ区)和高山区(Ⅳ区)。对Ⅰ区采用ISODATA方法进行非监督分类;对Ⅱ区和Ⅲ区分别采用最大似然(ML)分类法进行监督分类;对Ⅳ区采用人机交互解译方法进行分类。试验结果表明,基于地理地质环境的分块分类方法是高效、精确的。

The Information Extraction Method for Land Use and Cover Change (LUCC) after the Earthquake in Dujiang Dam

It is very important to immediately and accurately acquire LUCC reconstruction and restoration of the ecological environment. Remote Sensing images are characterized by a wide range of information,timeliness and rapidness in LUCC information extraction. Traditional methods in LUCC information extraction from remote sensing images often take the overall area as a study target,classify the images and acquire the information directly. Nevertheless, these methods are liable to ignore the influence of topography, geological structure and other background factors on the classification results. The solution lies in dividing the image into sub-blocks,sub-block classification and comprehensive analysis based on geographic and geological environment. In this paper, Dujiang Dam as the study area was divided into four blocks: plain district (Ⅰ),middle-mountain district (Ⅱ),middle and high mountain district (Ⅲ) and high mountain district (Ⅳ) based on the distribution characteristics of geological structures and landform development patterns. ISODATA method was used for non-supervised classification in District I, maximum likelihood (ML) classification method for supervised classification was used in DistrictⅡ and Ⅲ, and man-machine interactive interpretation method for classification was used in District Ⅳ. The results show that the classification method based on the division of geographic and geological environment is efficient and precise.