基于改进的NDBI指数法的遥感影像城镇用地信息自动提取

Automatic extraction of urban land-use information from remote sensing images based on improved NDBI method

针对目前遥感影像城镇用地信息提取存在精度不高、不够客观等问题,在分析Landsat ETM影像上各主要地物光谱特征的基础上,提出了一种基于改进的归一化建筑指数(NDBI)的遥感影像城镇用地信息自动提取方法.利用南京市主城区的Landsat ETM影像对该方法进行了验证,结果表明:该方法能够有效提取研究区的城镇用地信息;相对于常规的NDBI指数法,改进后的方法消除了稀疏植被对城镇用地提取精度的影响,大大提高了城镇用地信息的精度,并具有快速、客观等优点.

With regard to low precision and objective deficiency in the extraction of urban land-use information from remote sensing images, the spectral characteristics of main land covers on Landsat ETM images were analyzed. An automatic extraction method of urban land-use information from remote sensing images based on the improved normalized difference built-up index （NDBI） was proposed. The proposed method was verified by use of the Landsat ETM images in Nanjing City. The results show that the improved NDBI method can effectively extract the urban land-use information. Compared with the conventional NDBI methods, the improved NDBI method can eliminate the effects of sparse vegetation on the extraction precision of urban land-use information so as to raise the precision of the urban land-use information. It is of advantages of accuracy and objectiveness.