【作者】 陈红顺； 贺辉； 肖红玉； 黄静；

【Author】 CHEN Hong-shun;HE Hui;XIAO Hong-yu;HUANG Jing;College of Information Technology,Beijing Normal University;Centre for Remote Sensing and GIS,Beijing Normal University;

【机构】 北京师范大学珠海分校信息技术学院； 北京师范大学珠海分校遥感与GIS中心；

【摘要】 为了准确检测高空间分辨率遥感影像中的阴影,提高影像目标识别和信息提取的精度,提出了基于光谱特征的空间高分辨率遥感影像阴影检测技术。根据阴影区蓝、红波段辐射亮度下降幅度的差异,构建了归一化蓝红波段指数,并利用归一化植被指数区分植被与阴影;最后构建了单一的阴影强度指数,通过对阴影强度指数进行直方图阈值分割,提取遥感影像中的阴影。利用该方法在分辨率为2m的Worldview-2多光谱影像上进行了阴影提取实验,结果表明:该方法的阴影检测精度均在85%以上。研究表明该方法阴影检测效果较好,且计算简单,易于实现。更多还原

【Abstract】 To accurately detect shadows in high resolution remote sensing images and to improve the accuracy of target identification and information extraction from remote sensing images,this paper studies how to detect shadows in high-resolution remote sensing images based on the spectral characteristics of the shadows.The Normalized Blue and Red band Indexes(NBRI)was proposed according to the different decrease amplitudes of radiance in blue band and red band.Then the Normalized Difference Vegetation Indexes(NDVI)were calculated to distinguish the vegetation from the shade.Finally,a single index-shadow intensity(SI)was constructed by subtracting the NDVI from the NBRI and the shadow was easily segmented from the SI image by an appropriate threshold.The proposed method was tested on Woldview-2multispectral images with a spatial resolution of 2m,experimental results indicate that the overall accuracy of shadow detection is above 85%.It concludes that the method proposed in this paper has excellet performance in detecting shadow,needs less computation and is easy to be implemented.更多还原