冰川对气候变化极为敏感, 其变化对区域气候、 生态及水资源等有重要影响。 对于高原山地区域而言, 使用遥感数据开展冰川变化研究时, 影像经常会有较大面积的山体阴影。 阴影使地物目标反映的信息量有所损失或受到干扰, 在遥感影像数据上难以判读。 因此, 基于遥感影像的山体阴影区冰川识别成为一个技术难点。 选择青藏高原上的大型山地冰川群为实验区, 基于Landsat 8 OLI影像数据, 分析了山体阴影区冰川与非冰川的波段反射特征, 结果表明由于阴影区直射光被遮挡, 波长较短的蓝光波段因具有更高的散射强度, 是阴影区冰川识别的优势波段; 长波波段在阴影区无论是冰川还是非冰川区域反射率都很低, 难以区分。 在此基础上提出针对山体阴影区冰川信息提取的增强指数算法, 并与常规的冰川信息提取方法进行效果对比, 结果表明增强指数方法得到的直方图分割阈值更为明显。 从冰川信息提取结果来看, 无论是空间分布还是面积误差比例, 采用优势波段的增强指数法效果最好。 在高原山地区域进行大规模冰川提取时, 采用所提出的山体阴影区冰川信息增强指数算法, 有助于提高整体工作效率。

Glaciers are extremely sensitive to climate change. And glacier changes have great impacts on the regional climate, ecology, water resources and so on. Remote sensing images are often used to study glacial changes. For plateau mountainous areas, the images usually have a larger area of the mountain shadows. Shadows cause loss or distraction of the information reflected by the ground target, making remote sensing image difficult to understand. Therefore, the identification of glaciers in the mountain shadow area based on remote sensing images becomes a technical difficulty. In this study, a large mountain glacier on the Qinghai-Tibet Plateau was chosen as experimental subject. Based on Landsat 8 OLI data, this study first analyzed the reflection characteristics of different bands for glacier and non-glacier in shadow area. The results showed that due to the fact that direct light is blocked and target objects in a shadow area are mainly irradiated by the scattered light, the blue band which has shorter wavelength and higher intensity of the scattered light is preferred band for glacier identification in shaded area. For longer wavelength band, the reflectance of ground target in the entire shadow region is very low, and it is difficult to distinguish between glaciers or non-glacial regions. On this basis, a shaded glacier information enhanced index is proposed. Compared with the conventional glacier information extraction methods, the proposed method can give a result to identify the segmentation threshold more clearly in the histogram; and get the best result both in accuracy of the extracted boundary and the total area. For large-scale glacier extractionin the plateau mountainous area, it is recommended to use the proposed method which can be helpful in improving the overall work efficiency.