吉林西部多时相遥感数据分类方案的构建及应用

为了深化遥感监测方法在生态环境调查中的应用,本文以吉林西部为试验区,设计了一种多时相遥感数据分类方案。该方案以物候信息为主,结合地物特征变量(植被、水体和土地信息)构建的多维特征空间数据集用于土地覆被分类。该遥感分类方案提取了9种地表覆被类型,结果表明:地表植被季节变化信息和土地利用信息的引入能明显改善土地覆被的分类精度;与基于原始波段的分类方案相比,多时相遥感数据分类方案的分类精度最好,总体分类精度为95.50%,Kappa系数为95.04%。

Land Cover Classification Method Based on Multi-Temporal Satellite Images: Taking Western Jilin Region as an Example

With the rapid development of 3S (remote sensing (RS), geographical information system (GIS), global positioning system (GPS)) technology, the satellite image data used for monitoring the surface vegetation cover is vast.Western Jilin region was selected as the experimental zone.Using various functions, the land cover classification scheme was proposed to quickly and accurately extract the land cover information in the test area based on multi-temporal satellite images, coupled with the main classified variables including the seasonal variation information of vegetation, the water information and the land use information.Furthermore, the extracted data were statistically analyzed to verify the feasibility and rationality of the method.Finally, the results are as follows: 1) The way combined these classification features for extracting land cover type effectively improved the overall classification accuracy.Especially, the introduction of the changed information, including the seasonal variation of vegetation cover and the land-use and land-cover information, could significantly improve the classification accuracy of land cover;2) The overall classification accuracy of the algorithm was 95.5%, the Kappa coefficient of classification was 95.04%.