基于局部像素特征的图像去噪算法

An image denoising algorithm based on partial pixel characteristics

针对线阵相机拍摄的动车图像上的噪声对图像比对识别结果产生的不良影响,提出了一种基于局部像素特征的图像去噪算法。该算法首先对待处理图像的像素点进行逐行扫描,并计算每一行的像素均值,同时记录与该行对应且与其相邻的、等距的上下行的像素均值;然后将该行像素均值与等距下行的像素均值进行比较,若比较结果大于设定的阈值,则将等距上、下行的每个像素点之和的均值赋给与该行对应的每个像素点;反之,则保持该行的像素值不变。将该算法与经典去噪算法进行比较,比较结果充分证明了该算法的有效性。将该算法应用到动车图像比对识别的预处理中,大量实际应用表明,该算法的去噪功能可以大大降低由该噪声引起的比对识别误报率。

Considering the negative effect of the noise of an electric multiple unit （EMU） image taken by the lineararray cameras on the image identification results, this paper proposes an image denoising algorithm based on partialpixel characteristics. Firstly, this algorithm will scan the pixels of the image to be processed line by line, then cal-culate the mean value of pixels of each line, and at the same time, record the mean value of the pixels of the adja-cent equidistant lines up and down. Furthermore, the m^an value of pixels of the line is compared with the meanvalue of the equidistant lines down. If the value is larger than the set threshold, then assign the mean value of thepixels of the equidistant lines up and down to every pixel point corresponding to the line ; on the contrary, the pixelvalue of the line should be kept invariant. By comparing with the classic denoising algorithms, this new algorithm isproven to be effective. This algorithm was applied to the pre-treatment of the EMU image identification system,showing that this algorithm could greatly lower the rate of false identification caused by the noise.