基于彩色空间多特征的高空间分辨率遥感影像阴影检测  

Shadow Detection Using Multi-features Based on Color Space in High Spatial Resolution Remote Sense Image

作　　者：李艳霞[1] 闫冬梅[2] 徐素妍[3] 崔玉红[1] 

机构地区：[1]装备指挥技术学院试验装备系,北京101416 [2]装备指挥技术学院光电装备系,北京101416 [3]装备指挥技术学院信息装备系,北京101416

出　　处：《装备指挥技术学院学报》2007年第3期94-98,共5页Journal of the Academy of Equipment Command & Technology

摘　　要：高空间分辨率遥感影像中的阴影检测对于影像的目标识别和信息恢复十分关键。针对单一特征阴影检测中普遍存在的强反射性地形、地物漏检和水体错检的情况，提出了一种基于彩色空间多特征的阴影检测方法：首先，利用改进的阈值法进行一般阴影的检测；然后利用色调值较高的特征检测包含强反射性地形地物的阴影区域；再利用水体在RGB不同波段的反射亮度差值来进行水体与阴影的区分。实验结果表明：该方法较好地克服了阈值反复试算、包含强反射性地形、地物阴影区域漏检和水体错检的问题，提高了阴影的检测精度。Shadow detection in high spatial resolution remote sense image is very critical for target identification and information recovering. In order to solve the problem of false dismissals in the presence of shadow including high reflective regions and false alarms in the presence of water, a shadow detection algorithm based on the color space multi-features of shadow detection is purposed. Firstly, general shadow is detected by improving threshold method, then shadow including high reflective regions is detected by taking advantage of high hue feature, lastly water is distinguished from shadow by intensity difference in RGB wave band. The experimental results show that the algorithm solves the follow problems： threshold calculating repeatedly, false dismissals in the presence of shadow including high reflective regions and false alarms in the presence of water. So the algorithm improves accuracy of shadow detection.

关 键 词：多特征 高空间分辨率 遥感影像 阴影检测 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]