论文部分内容阅读
准确提取农村居民点用地规模及分布,对合理利用土地资源、改善农村生态环境及促进城市化发展具有重要意义。根据农村居民点用地的POLSAR散射特性及光谱特征,提出一种基于POLSAR极化散射特征与光学归一化差异指数的农村居民点用地提取方法,并结合实验分析了POLSAR极化相关系数在区分农村居民点用地与林地的不适用性。所述方法可有效解决单一数据源在农村居民点用地提取中裸地(光学数据)、林地(POLSAR数据)与农村居民点用地混分的问题,精确提取农村居民点用地(用户精度为91.7%,制图精度为95.2%,总体精度为95.9%)。相比基于POLSAR极化目标分解的H/α/Wishart迭代分类,该方法用户精度提高了34.9%,制图精度提高了14.4%,总体精度提高了16.2%;相比基于归一化植被指数和归一化建筑指数的监督分类,本文的用户精度提高了24.3%。
Accurately extracting the size and distribution of rural residential land use is of great significance to rational utilization of land resources, improvement of rural ecological environment and promotion of urbanization. According to the characteristics of POLSAR scattering and spectral characteristics of rural residential land use, a method of rural residential land extraction based on POLSAR polarization scattering characteristics and optical normalized difference index was proposed. Combined with experimental analysis, the POLSAR polarization correlation coefficient was used to distinguish rural Incompetence of residential land and woodland. The method can effectively solve the problem that a single data source can be used to extract the bare land (optical data) and the land (POLSAR data) in the rural residential land and the residential land in the rural residential area to accurately extract the land for the rural residential land (the user precision is 91.7% , Drawing accuracy of 95.2%, overall accuracy of 95.9%). Compared with the H / α / Wishart iterative classification based on POLSAR polarimetric target decomposition, the user accuracy is improved by 34.9%, the accuracy of mapping is improved by 14.4% and the overall accuracy is improved by 16.2%. Compared with the normalized vegetation index A building index of the supervision and classification, the accuracy of the user in this article increased by 24.3%.