模糊边缘检测在机器视觉图像系统仿真研究

Simulation Research on Fuzzy Edge Detection Method in Machine Vision Imaging System

研究图像优化问题,针对精确定位边缘和抑制噪声,视觉图像中最重要的是进行边缘检测,用于轮廓抽取、特征检测和纹理分析。传统的Pal.K ing增强算法在速度和检测效果等方面存在缺陷,提出了一种新的检测算法,算法简化了Pal.K ing复杂的变换和逆变换。先通过阈值来定义一个新的隶属函数将原始图像映射到模糊特征平面,利用模糊增强处理来提高区域之间的层次,加强边缘两侧的对比度,最后根据一定的判别准则提取出图像的边缘。实验结果表明,改进算法提高了运算效率,而且提取的边缘比较精细。新算法中具有唯一的参数且可以自动确定,保证了算法的自适应性。

Edge detection is one of the most important parts of image analysis and computer vision which is widely used in outline collect,analyses of characteristics and textures.The main task of edge detection is to pinpoint the edge and noise restraint.Because of the inherent defect of Pal.King enhance algorithm in speed and effect,a new detection algorithm is given.The algorithm proposed in the paper simplifies the calculations of complex transformation and inverse transformation in Pal.King algorithm.Firstly,a new membership function is defined by automatically selecting the threshold of the image in order to map the original image into the fuzzy feature plane.Secondly,the level of different regions is improved after performing fuzzy image enhancement,so that the contrast between both sides of the edge is enhanced.Finally,the edges of the image are extracted according to a certain discriminative rule.The experiment indicates that the algorithm is efficient and the extracted edges are more precise.There exists just one parameter in the modified algorithm and it can be determined automatically,which guarantees the algorithm is adaptive.