中文摘要：

随着遥感影像空间分辨率不断提高,阴影对影像解译产生的影响愈加突出,因此对阴影的识别更为重要。文中针对阴影在RGB空间中与非阴影区域内地物容易混淆的缺点,利用其在HIS空间中低亮度、高色调和高饱和度的特性,提出了一种基于新型阴影检测指标的阴影检测方法。该阴影指标突显阴影区域,且抑制非阴影区域的响应,根据最大类间方差法做直方图阈值分割,可得到检测结果。实验证明,该方法可有效地将阴影与其他地物区分开。

英文摘要：

With the improvement of the remote sensing images＇ resolution, the presence of shadow in High-resolution imagery is much more evident, which will be a serious obstacle for image interpretation. Therefore, the shadow identification is important in image analysis. Shadow areas are easy to be confused with other classes in RGB space, while having lower brightness, higher saturation, and greater hue value. In view of this, this paper proposes a shadow detection method based on a novel shadow index. The shadow index highlights the shadow areas, while suppressing the nonshadow areas. The histogram thresholding segmentation is performed by Ostu algorithm to acquire the shadow detection result. Experimental results demonstrate that the proposed approach can separate shadow from nonshadow effectively.