改进的HSI空间形态学有噪彩色图像边缘检测

摘要

Since it was difficult to effectively distinguish the similarity between colors in RGB space, this paper implemented color image processing in HSI color space which reflected the features of human vision, proposed a method of noise color image edge detection based on improved morphology. By applying the iterative operation of open and close and the operation of double-structure and multi-scale elements to traditional morphological gradient operator, it calculated the edge information of H, S, 1, then obtained the edge information of color image by combining three components' edge information according to their specific weights. The experimental result shows that the edge which detected by the proposed method meets human visual characteristics , this algorithm is better than traditional ones and others in the terms of resistance to noise, and is able to keep more complete outline of the original color image, it has a relatively small amout of calculation and better practicality and versatility.%针对在RGB空间中很难有效区分颜色相似性问题,选择了更加符合颜色视觉特性的HSI颜色空间进行图像处理,提出了一种改进的形态学有噪彩色图像边缘检测方法,将开闭的迭代运算和双结构元多尺度运算应用到传统形态学梯度算子中,然后计算图像H、S、I三个分量的边缘信息,根据H、S、I所占比重对三分量进行加权融合得到彩色图像边缘.实验结果表明,该方法所检测的边缘符合人眼视觉特性,在抗噪声方面的效果比传统方法及其他多种方法更佳,能够更完整地保留原彩色图像的轮廓,计算量相对较小,有很好的实用性和通用性.