一种基于小波尺度相关性的自适应阈值去噪算法

针对图像去噪时造成的边缘模糊和细节损失问题,结合小波尺度相关性理论和经典的阈值去噪方法,提出了一种新的图像去噪算法,该算法利用不同尺度上小波系数间的相关性,对原图像和噪声对应的小波系数进行最大程度的区分,对前者予以保留,对后者则结合闽值去噪方法和该小波系数是原图像的置信程度进行处理。实验结果表明,该算法在达到很好的图像去噪的同时,能很好地保留图像的边缘和细节信息。

Adaptive Threshold De-noise Algorithm Based on Wavelet Scale Correlation

The image de-noise method brings about edge blur and the loss of details.According to wavelet scale correlation theory and classic threshold de-noise algorithm,the paper puts forward a new image de-noise algorithm.The method uses the correlation of different scales of wavelet coefficient to distinguish the corresponding wavelet coefficients between original images and noise.The former is reserved and the latter is treated with threshold de—noise algorithm on the basis of taking wavelet coefficients as the confidence degree of original image.The experiment results show that the method can achieve good image de-noise and reserve the edge and detail information of the original image well.