基于邻域加权的多层次模糊边缘检测算法

Weighted neighbor- region based multi- level fuzzy edge detection method

针对目前边缘检测方法在低对比度图像、噪声图像中检测效果不理想的问题,本文结合微分算子和模糊边缘检测的优点,提出一种基于邻域加权的多层次模糊边缘检测方法。首先,利用微分算子计算图像梯度特征,依据图像梯度特征对图像进行自适应地分层;然后构造模糊函数,用模糊函数增强不同强度的图像梯度特征,取得了较好的边缘检测结果。仿真实验表明:基于邻域加权的多层次模糊边缘检测算法能较好地检测低对比度图像的边缘,同时能有效抑制椒盐噪声、高斯噪声对图像边缘检测的干扰。

Most existing edge detection methods can not perform well in low contrast image and noise images. Combing the advantages of differential operator and fuzzy edge detection, a novel multi-level fuzzy edge detection method based on weighted neighbor-region is proposed. First, this method computes the image gradient features, and utilizes the adaptive method to divide image into multiple tiers based on gradient. Then, a fuzzy function is constructed to strengthen different image gradient features in different levels. Experiment results show that the proposed approach can highlight the gradient features in low contrast region of an image with the help of strengthening. Also this approach outperforms the state-of-art methods in terms of both visual quality and noise （Salt and Pepper noise and Gaussian noise） suppression.