摘要
基于HSV彩色空间的色调值融合RGB色彩模型中的蓝色分量信息,提出了鲁棒而有效的二维彩色图像阴影区域自动分割方法.根据阴影与非阴影区域间存在色调差异,利用HSV彩色模型,提取可能阴影区域.为消除提取出的阴影区域中偏蓝物体影响,采用RGB彩色空间中的蓝色分量为模板,计算该模板与提取出的阴影区域间的直方图.采用单阈值化分割方法,确定该直方图阈值.将蓝色分量值低于该阈值的阴影区域确定为有效阴影区域.通过对不同光照下的实际自然场景图像的阴影检测,实验结果表明文中所提方法是有效可行的.
Based on combining both hue value in HSV color space and blue component in RGB color space,a robust and effective algorithm was presented to automatically extract shaded regions from two-dimensional color images. According to hue difference between shaded and non-shaded regions, possible shaded regions were extracted by HSV color model at first. In order to eliminate the influence of objects with bluer color within the segmented shaded regions, the blue component in RGB color space was taken as a template. A single threshold segmentation method was used to determine the optimal threshold value of histogram obtained from the template multiplied by the extracted shaded regions pixel-wise. The segmented regions with the blue value lower than the threshold value were taken as an efficient shaded region. Experimental results show that the developed algorithm is feasible after shadow detection on actual natural scenes in different illumination conditions.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2006年第4期624-627,共4页
Acta Electronica Sinica
基金
上海市教委与教育发展基金委曙光项目(No.04CX72)
上海市教委发展基金(No.05AZ38)
关键词
分割
阴影检测
特征提取
闽值
segmentation
shadow detection: feature extraction: threshold value