非下采样轮廓波域混合统计模型红外图像降噪方法

针对红外图像易受噪声污染、成像质量差等问题,提出一种非下采样Contourlet域变换系数混合统计建模的降噪方法。将图像变换到非下采样Contourlet域,对噪声系数、信号系数分别按照高斯分布和广义拉普拉斯分布建模,利用贝叶斯框架下的最大后验估计理论,推导原始信号系数估计公式,并对包含噪声的图像系数进行处理,实现非下采样Contourlet域红外图像降噪。实验结果表明,该方法能够对红外成像过程中产生的高斯噪声实现有效抑制,较完整地保持图像的边缘等细节信息,在峰值信噪比提高与视觉效果上优于部分经典算法。

Infrared Image Denoising Using Mixed Statistical Model in Nonsubsampled Contourlet Domain

For the denoising problem of infrared image , an algorithm is proposed based on mixed sta-tistical model of transform coefficients in nonsubsampled Contourlet domain .The image is transformed into nonsubsampled Contourlet domain , and noise coefficients and signal coefficients are modeled respectively by Gaussian distribution and generalized Laplacian distribution .By using maximum a posteriori estimation theory under Bayesian framework , a new estimation formula for the original signal coefficients is deduced , and the noise image coefficients are processed to realize the infrared image denoising of nonsubsampled Contourlet domain .The experimental results show that the proposed method can suppress the Gaussian noise effectively and keep most of the image ’ s detail information , which has higher value of peak signal noise ratio and better visual effect than some standard algorithms .