基于彩色形态学的图像边缘检测算法

Edge Detection of Color Morphology

基于对彩色图像形态学的研究,提出了一种面向HSI颜色空间的多结构元多尺度的彩色形态学图像边缘检测算法。首先借助彩色形态学的膨胀、腐蚀、开运算和闭运算给出彩色多结构交替顺序滤波算子的相关定义,同时利用交替顺序滤波器(ASF)的优势得到一种混合滤波器(HF)对图像进行滤波;其次引入一种全方位形态结构元再结合多尺度结构元的思想对图像不同方向的边缘进行提取;最后再使用信息熵对各方向的边缘进行融合得到图像整体边缘信息。实验表明,所提出的算法不仅具有良好的抗噪性,而且能够保留更多的图像细节信息。

In view of the research of color morphology, a color image edge detection method was proposed based on color morphology multi - structural and multi - scale elements in HSI color space. Firstly, the related definitions of the alternating se- quence filter operator of color structure were provided with color morphological dilation, erosion, opening operation and closing operation and a hybrid filter （HF） was presented with the advantage of Alternating Sequence Fiher （ASF） operator simultaneous- ly. Then an omnidirectional morphological structural element was introduced and combined with the thought of multi - scale ele- ments to extract edge information of images from different directions. Finally, the edges of each direction were fused to get the o- verall edge information using the knowledge of information entropy. Experiments show that the proposed algorithm has good noise resistance and retains more image characteristics.