基于噪声检测彩色图像滤波算法_李岳阳.pdfVIP

第 10卷第 2期 江南大学 学 报(自然科学 版) Vo.l 10 No. 2 2011年 4月 Journal of Jiangnan University(Natural Science Edition) Apr. 2011 基于噪声检测的彩色图像滤波算法 李岳阳, 孙 俊 (江南大学 物联网工程学院, 江苏无锡 214122) : 对于被脉冲噪声污染的彩色图像,提出了 于噪声检测的彩色图像滤波算法运用脉冲噪 声检测器,估计出图像中的噪声像素, 应用后续滤波算法, 只对检测出来的噪声像素进行滤波, 而 对非噪声像素(即信号像素) 保持其值不变并对 于相似度的自适应滤波算法加以改进,作为后 续的滤波算法实验结果表明,提出的滤波算法与传统滤波算法相比, 在有效消除噪声的同时, 更 能够保留图像中的边缘和细节特征 : 彩色图像滤波;脉冲噪声;噪声检测; 相 度 : TP 751 : A : 1671- 7147(2011)02- 0127- 08 Filtering A lgorithm for Color Images Based on Noise D etection LIYue-yang, SUN Jun ( School of Internet ofThings Engineering, JiangnanUniversity, Wuxi214122, China) A stract: As to the color mi ages corrupted by mi pulse noise, we propose filtering algorithms for color mi ages based on noise detection. Firstly, an mi pulsedetector ispresented to find outallpixels that are considered as noise before filtering. Then, if a pixel is considered as noise, it w ill be replaced by a pixel in the neighborhood which is noise-free during filtering procedure later. Otherw ise, it is unchanged. The sel-f adaptive algorithm based on similarities is mi proved as the filtering algorithms later. As demonstrated by the extensive expermi ental results, all the proposed filtering methods compare favorably w ith conventional techniques in the capabilities of noise attenuation and details preservation, in both quantitative and qualitativemeasures. Key words: color image filtering, impulse noise, noise detection, smi ilarities , R-, G-, B-, 3, , , , , , , , ,