摘要
提出了改进的广义模糊算子(GFO)边缘检测新方法。该方法把 Otsu 理论应用到 GFO 参数确定中,从而实现了 GFO 中关键参数的自适应获取,使参数确定困难的问题得以解决。经该方法处理后的图像具有失真度小、细节分明等优点,优于目前现有的边缘检测方法。另外,为使得该方法在复杂图像边缘检测中取得良好效果,采用了先分割后处理的方案,即将复杂图像分割成若干子图像,然后根据各个子图像的灰度性质应用本文所提方法进行处理,再合成,从而为此方法在图像边缘检测中广泛应用提供了切实可行的解决方案。
A new method for edge detection with improved General Fuzzy Operator (GFO) is proposed. The method applies Otsu theory to the determination of GFO parameters, which makes the key parameters of GFO be obtained self adaptively. The problems of parameter selection is solved. Images processed by this method have less distortion and clear details. This shows that it is better than the present edge detection method. In addition, aiming at the complex image which contains a great amount of pixels and gray values, a scheme of segmenting the image before processing is presented, namely, dividing complex image into several sub-image and processing sub-images with the presented new method, composing the result of sub-images at last. Consequently, a practical and feasible solution is provided, which makes the new method applicable widely.
出处
《光电工程》
EI
CAS
CSCD
北大核心
2005年第2期80-83,共4页
Opto-Electronic Engineering
基金
国家自然科学基金资助项目(60274023)