Abstract: Owing to the characteristics of the gray relation analysis and the advantage of the alpha-trimmed mean filter, an efficient algorithm for noisy images removal based on the gray relational analysis and the alpha-trimmed mean filter is proposed. This algorithm uses the gray relation analysis to adjust the filter window’s coefficients adaptively, and it can improve the validity of the algorithms. Experimental results show that the proposed algorithm not only has better filtering effect for noisy image which corrupted by Gaussian noise or mixed noise, but also can preserve the integrity of edge and keep the details of the original image.

Key words: improved gray relational analysis, adaptive filter algorithm, Alpha-trimmed mean filter

摘要： 利用灰色关联度的特性和阿尔法均值滤波算法的优点，提出一种基于改进灰色关联度和阿尔法Alpha均值滤波的噪声图像的自适应滤波算法。该算法采用灰色关联度自适应地确定滤波窗口的加权系数值，改善算法的滤波性能。实验结果表明算法对受到高斯噪声和混合噪声干扰的图像进行去噪能取得较好的滤波效果，同时还保护了原始图像的细节信息。

关键词: 改进灰色关联度, 自适应滤波, 阿尔法均值滤波算法