【摘要】 传统的彩色图像边缘检测方法有以下一些弊端,边缘易丢失,边缘定位容易出现偏差以及边缘不连续等,结合传统的边缘检测方法和四元数超复数空间利用四元数对矢量旋转表示的便利性,在四元数的空间里处理彩色图像。然后通过对彩色图像旋转变换,进而分离HSV模型图像中的S分量实现彩色图像的边缘检测。通过对比试验结果可以看出该方法可以有效地降低色彩边缘的误检率,一定程度上提高边缘检测器的性能。更多还原

【Abstract】 The traditional image edge detection methods have many shortcomings such as edge lost, with edge position deviation and with discontinuous edges etc. In order to solve such problems a novel method was provided in this article which combined with the traditional edge detection methods and quaternion space. When a color image is projected to the quaternion space, quaternion’s vector rotating would be used. By separating the S component of HSV image model, the edge of color image would be spreaded. Compared with the traditional image edge detection methods, it reduces the color edge detection error rate and improves the edge detection performance.更多还原