一种改进的非局部均值井下图像去噪

An improved non-local means algorithm for underground image denoising

针对井下图像噪声的特点,提出一种改进的非局部均值去噪算法,该算法通过小波变换对井下图像进行多尺度分解,根据分解后图像噪声的分布特点,对低频部分采用非局部均值去噪,而对高频部分采用中值去噪,对去噪后的图像进行小波重构。实验结果表明,该算法在兼顾算法运算速度的同时,能有较高的峰值信噪比。

Considering the characteristics of noise of underground image,an improved non-local means denoising algorithms is proposed. In the algorithm,wavelet transform was applied to decompose the underground image in multi-scale. According to the noise distribution characteristics of decomposed image,the non-local means method was used to denosie the low frequency components while the median filtering was used to denosie the high frequency components. The paper restructed the undergound image by denoised components. The experimental results show that the algorithm can have a higher PSNR and reduce the execution time compared to NLM.