V律指数法提取城市建筑用地遥感影像的研究

Research a New Method for Urban Built-up Land Extraction from Remote-sensing Images

通过对Landsat ETM+不同波段遥感影像主要土地利用类型光谱特征曲线进行分析,提出一种新的提取城市建筑用地的方法:V律指数法(V law index,VLI).该方法采用3个波段的遥感影像,并结合约束条件,通过简单的逻辑判断和乘法运算实现建筑用地的提取.随机矩阵点实验测试表明:VLI法能够获得较好的提取效果,与其他建筑用地提取指数如NDBI单指数、NDBI/SAVI/MNDWI三指数方法相比,具有更低的误判率.

This paper makes an analyses of the spectral characteristic curves of the major land use categories based on Landsat ETM ＋ different bands remote-sensing images and proposes a new VLI（V law index ） method to extract images of urban built-up land. It adopts three-band remote-sensing images and combines with constraints to extract images of urban built-up land using simple logic determination and multiplication operations. Experimental random matrix point tests show that the VLI index, compared with other building land index, such as NDBI index and NDBI/SAVI/MNDWI index, can achieve better extraction results and has a lower misjudged rate.