结合非下采样轮廓波变换( NSCT),提出了一种红外图像改进非局部均值滤波算法( Improved Non-local Means Filtering,INLMF).该算法首先对红外噪声图像进行多尺度 NSCT变换,其次分别从相似图像块自适应划分方法以及滤波权重计算方法 2个方面对经典非局部均值滤波算法进行适当改进,将改进后的非局部均值滤波算法( INLMF)应用于处理高频分解系数,然后将滤波后的高频分解系数与低频分解系数进行重构,得到去噪后的图像,最后对去噪后图像采用非负支撑域有限递归逆滤波(Non-negativity and Support Constraints Recursive Inverse Filtering,NAS-RIF)算法进行图像复原,以尽可能消除因滤波造成的图像失真.测试结果表明,本文算法滤波效果优于 NLMF及其已有的改进算法.

Combined with nonsubsampled contourlet transform,an improved non-local means infrared image filtering algorithm is proposed.Firstly,the infrared image noise is conducted by nonsubsampled contourlet transform.Then,the similarity image block adaptive partition method and weighting calculation method are proposed so as to improve the non-local means filtering algorithm(INLMF) to deal with the high-frequency NSCT coefficients.Thirdly,the high-frequency NSCT coefficients after filtering and low-frequency NSCT coefficients are reconstructed and the denoised image is obtained.Finally,the denoised image is processed by the non-negativity and support constraints recursive inverse filtering algorithm(NAS-RIF),so the infrared image of better visual effect is obtained.The experimental results indicated that,the performance of the algorithm in this paper is superior to the already existed improved NLMF algorithm.