【摘要】 本文将彩色图像像素的R、G、B分量作为彩色空间矢量的三个分量,利用矢量场通量以及四元数关于矢量旋转的相关知识,得到两种彩色图像边缘检测方法。基于矢量场通量最大的检测算法,较传统的利用Sobel算子的方法能更全面地提取图像的边缘细节,并且针对不同的图像具有很强的自适应效果;基于四元数的检测算法,利用四元数对矢量旋转的方便表示,提出一种全新的彩色边缘检测方法。实验结果表明,本文提出的两种方法针对彩色图像的边缘检测具有较好的效果,特别是四元数矢量旋转方法,能够提取很多传统的灰度化方法所不能提取的图像纹理细节。更多还原

【Abstract】 Regarding R,G, B of color iamge pixel as the three weights of color space vector, two color iamge edge detec- tion algorithms have been proposed using corresponding knowledge of vector field flux and quaternion. Maximizing vec- tor flux based detection algorithm can detect more edge details comparing conventional algorithm, simultaneously, this algorithm has quite self-adaptive ability to different nature iamges. A novel edge detection algorithm also be proposed based on quaternion using the easy quaternion expressing of vector rotating. Experiment result illuminate that the two algorithms have good detection effects to color iamge, especially, quaternion vector rotating based algorithm can detect much more texture details which the conventional algorithm can’t realize it.更多还原