摘要
基于彩色编码结构光中的RGB彩色条纹,提出一种基于彩色分量交点的条纹边缘亚像素检测方法。将彩色条纹图像由RGB模型转为l1l2l3归一化彩色模型,模型各分量不受环境光、拍摄角度和物体形状的影响,利用模型中各分量交点实现彩色条纹边缘检测。给出了彩色模型转换推导过程和条纹边缘检测算法,实验结果表明,本文算法的条纹边缘检测误差小于0.4pixel,条纹边缘检测平均误差小于0.2pixel,方差小于0.1pix-el2。
By using RGBcolor strips of color encoding structure-light,a method of color stripe edge detection based on color component intersection point is presented.The method changes the color strip images from RGBmodel to normalized color model l1 l2 l3which is not influenced by the ambient light,shooting angles of CCD and the shape of the 3D objects.Then the color stripe edge detection is implemented by the color component intersection point of l1 l2 l3model.And the deduce process of the color model transfer is given.The experiments prove that the maximum error of stripe edge detection is less than 0.4 pixel,and the average error of the stripe detection is less than 0.2 pixel,the variance is less than 0.1 pixel2.
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2010年第5期714-717,共4页
Journal of Optoelectronics·Laser
基金
国家自然科学基金资助项目(60572030)
哈尔滨市科技攻关计划资助项目(2005AA1CG152)
