Abstract: A mixed adaptive denoising method is presented for noise removal of an infrared image, which contaminated by Guass noise and impulse noise. First, Guass noise and impulse noise is distinguished by the proposed method , then , an adaptive stack median filtering method is adopted to remove the impulse noise; an adaptive local filtering method based on the estimation of local squared error is used to suppress the Gauss noise. Experiments indicate the denoising ability of this method is better than the linear mean filtering method and the ordinary median filtering method by its higher SNR (signal- to- noise rate) and smaller MSE (Minimizes the mean squared error). The eminent performance of this method is also proved by its details preservation capacity and its fast filtering speed.

Key words: Infrared image, mixed denoising, adaptive stack median filtering, adaptive local filtering

摘要： 针对带有高斯噪声和脉冲噪声的红外图像，提出一种自适应混合去噪方法。首先将受脉冲噪声和高斯噪声污染的像素区分开来，然后对被脉冲噪声污染的像素采用自适应层叠中值滤波法，对被高斯噪声污染的像素采用基于局部噪声方差估计的自适应局部滤波法。实验表明，该方法在信噪比（SNR）和最小均方误差（MSE）的改善上明显优于线性平滑滤波法和普通中值滤波算法，并具有较好的图像细节保护能力和较快的滤波速度。

关键词: 红外图像, 混合消噪, 自适应层叠中值滤波, 自适应局部滤波