摘要
基于GIS技术,介绍一种数学重采样的遥感土地利用调查精度验证方法。利用该方法对定西县2000年土地利用调查结果的一级分类进行了分层验证,结果表明,定西县土地利用调查总体精度较高。人工用地的分类精度估计值为0.95,农业用地的分类精度估计值为0.99,自然/半自然植被的分类估计值为0.97,水体的正确分类估计值为1,未利用土地的分类估计值为0.94,总的分类精度估计值为0.9791。这说明复杂地区土地利用调查中用TM图像作为主要信息源是可行的。
Based on remote sensing and GIS technology, a mathematical method to verify the precision of remote sensing land use classification was introduced It was used to check the first grade classification result of land use of Dingxi County in 2000 The precision is relatively high The variance for correctly classifying the areas of man\|use\|land is 0000000027, of agricultural land is 0000021871, of natural and semi\|natural vegetation is 0000005427, of water body is 0, and of unused land is 0000009201 The total classification precision is 09791 The verification result reflects the reliability of using TM image as a main resource for land use investigation in complicated region
出处
《遥感技术与应用》
CSCD
2003年第2期81-86,共6页
Remote Sensing Technology and Application
基金
国家重点科技项目(攻关)计划(2000-K-01-06-03-02)。