基于模糊增强的图像区域分割算法研究

摘要

In order to effectively segment an image into three types of regions: smooth region, textured region and well-defined edges, a new image segmentation algorithm is presented. The algorithm adopts the Canny edge detection principle based on the fuzzy enhancement The pixel distribution characteristics of textured regions and well-defined edges are analyzed in detail. The algorithm enhances the pixel contrast of texture area and detects more texture details. Then it realizes image segmentation by using of inflation and region connection. The experimental results show that the algorithm can segment the image into smooth re gion, textured region and well-defined edges accurately. And it has the strong antinoise ability. The segmentation result can dis play more image details.%针对图像平坦区、纹理区和清晰边缘的分割问题,提出了一种基于模糊增强的图像分割算法.该算法依据基于模糊增强的Canny边缘检测原理,在充分分析图像纹理区和清晰边缘的像素分布特点的基础上,通过增强纹理区像素对比度,检测出更多的纹理区细节.并利用膨胀、区域连通等方法实现了图像的区域分割.实验结果表明,该算法能够准确地实现了图像平坦区、纹理区和清晰边缘的分割,并有较强的抗噪能力.图像分割结果可以反映更多的纹理细节信息.