引导滤波和三维块匹配结合的红外图像去噪

Infrared image denoising based on three-dimensional block matching

针对红外图像存在噪声较大的问题,提出了一种引导滤波和三维块匹配结合的红外图像去噪方法。该方法通过将引导滤波与三维块匹配方法相结合,采用二次联合滤波的策略减轻了三维块匹配去噪方法在平滑区产生明显的"抓痕"现象。通过BM3D(block method of 3-dimension)去噪方法得到初始滤波图像;对噪声图像在低尺度进行BM3D去噪得到比较平滑的引导图像,用引导图像指导初始滤波图像进行二次滤波。仿真结果表明,提出的算法既保持了BM3D去噪算法的优秀性能,且具有更好的视觉效果和更高的峰值信噪比及结构相似度。

Due to high noise exists in infrared image,this paper presents an infrared image denoising method that combines guide filtering and the three-dimensional block matching. By combining the three-dimensional block matching and guide filtering,this method reduces the phenomenon of that three-dimensional block matching method which produces a significant＂scratch＂in the smooth area by the strategy of second joint filtering. First,the initial filtered image is got by BM3D（ Block Method of 3-Dimension） denoising method; Second,smoother boot image is got by BM3 D denoising in low scales,and the initial filtered image is filtered again with its guidance based on the principles of guiding image filter. Simulation results show that the proposed algorithm can maintain the excellent denoising performance of BM3 D algorithms,and has better visual effects and peak signal to noise ratio and higher structural similarity.