自适应线性预测图像椒盐噪声去除方法

针对灰度图像中的椒盐噪声，提出了一种基于自适应权值与线性预测方法相结合的噪声去除方法。用椒盐噪声模型确定图像中的噪声点，以及噪声点所在滤波窗口内的噪声密度，在密度较小时利用设计的权值函数进行加权平均计算，以达到去噪声的目的；在噪声密度较大时，利用线性预测方法预测噪声点所在位置的灰度值，以实现去噪。对于非噪声点像素则不做处理，较好地保持了图像的细节。实验结果表明，与同类方法相比，此方法有良好的去噪性能。

Image Salt-and-Pepper noise removal algorithm based on adaptive linear predication

An image Salt-and-Pepper noise removal algorithm based on adaptive weight and linear predication is presented.This algorithm uses the theory of noise removal after noise detection.Aiming at the noise pixels,adaptive weight method is used to remove the noise when the noise density is low,and linear predication algorithm is used when the noise density is high.The signal pixels are preserved without any disposal in proposed algorithm.So the noise is removed and more details of image are preserved.Experimental results prove that this algorithm is faster than some other noise filters and can preserve more details of image during the noise removal.