中文摘要：

为了响应土地整理开发项目的政策，吉林省西部地区开发了大面积的水田。为掌握新开发水田的动态变化，有必要对吉林省地区进行土地利用分类。以吉林省前郭县为例，利用2010年环境卫星数据进行不同土地利用分类方法的比较，进而对另外3景图像进行信息提取，分析4个年份新开发水田的分布及变化情况。结果表明，支持向量机法比最大似然法的耕地分类精度高约5％，其产品和用户精度分别为95％和84％。加入纹理信息没有显著提高分类精度。2009～2012年水田面积分别增加-67．7km2，111．7km2和265．01km2。

英文摘要：

In response to the policy of the land development and consolidation project in western Jilin Province, a large area of paddy field has been developed. To understand the dynamic change of newly reclaimed paddy fields in Jilin Province, high-precision land use classifications was required. Based on the remote sensing image of Qianguo County in 2010, different land using classification methods were evaluated and compared, so as to select a better method, land use information was extracted from other three re- mote sensing images in different years. The temporal and spatial changes and area of the newly reclaimed paddy field were analyzed. Results showed that： the user accuracy or production accuracy was 95% or 84% by using support vector machine method which 5% higher than that using maximum likelihood clas- sification ; involving texture information did not significantly improve the extraction accuracy. The areas of the newly reclaimed paddy field from 2009 to 2012 were increased by - 67.7 km2 , 111.7km2 and 265.0 km2 respectively.