摘要
针对现有的阴影检测算法对较亮阴影和较暗地物中的阴影不能同时较好地检测等问题,提出一种结合多种特征的高分辨率遥感影像阴影检测方法.该算法首先结合主成分分析、颜色特征和直方图的分割构建多种阈值检测条件,然后综合多种特征来进行遥感影像阴影的初步检测,最后通过分析RGB模型在阴影与非阴影地物上的差别,利用颜色特性最终检测出阴影区域.实验结果表明,本文算法能有效检测较亮阴影和较暗地物中的阴影.与现有方法相比,较亮阴影的平均总错误率从水平集法的31.85%降至24.61%,较暗地物中阴影的平均总错误率从自动检测法的37.75%降至23.30%.
To aim at the problem that shadow detection algorithms cannot simultaneously well detect partial-bright shadows and shadows in dark object, a kind of high resolution remote sensing images shadow detection method that combine a multiple features is proposed. The algorithm firstly combines principal component analysis, color features and histogram segmentation to construct the detection conditions of various thresholds, then integra various features of remote sensing image for initial detection, finally by analyzing the difference of the RGB models in the shadow and non shadow,uses the color characteristics to detect the shadow region. Experimental results show that the algorithm proposed in this paper can detect partial uses bright shadows and shadows in dark object effectively. Compared with the existing methods,the average total error rate goes from the level set method 31.85 % down to 24.61 % for partial shadow, and the average total error rate is reduced from the automatic detection method 37.75 % to 23.30 % for shadows in dark object.
出处
《自动化学报》
EI
CSCD
北大核心
2016年第2期290-298,共9页
Acta Automatica Sinica
基金
国家自然科学基金(61373180
61461047)
西南交通大学2015年研究生创新实验实践项目(YC201504106)资助~~
关键词
高分辨率遥感影像
阴影检测
主成分分析
颜色特征
直方图的分割
High resolution remote sensing images
shadow detection
principal component analysis
color features
histogram segmentation
