摘要： 高分辨率多光谱遥感影像中的建筑物阴影具有重要作用。针对检测到的建筑物阴影经常受高大树木阴影影响的问题，在可分离亮度的Lab颜色空间建立树木颜色模型，使用查表法准确检测遥感影像中的树木区域。提出根据树木区域和光照方向去除建筑物阴影中树木阴影的方法，通过实验证明其有效性。

关键词: 建筑物阴影, 树木阴影, 颜色空间, 颜色模型, 树木概率图

Abstract: Building shadows in high-resolution and multi-spectrum remote sensing images are important. The detected building shadows are always influenced by big trees shadows. Aiming at this problem, this paper models the colors of trees in the intensity detachable Lab color space and detects the trees exactly by look-up table method in the remote sensing images. A tree shadow removal method for building shadows based on the detected tree regions and the orientation of the sun light is proposed and is proved effectively by experiment.

Key words: building shadow, tree shadow, color space, color model, Tree Probability Map(TPM)

中图分类号: 

• TN911.73