基于小波和PCA的自适应颜色空间彩色图像去噪*

在彩色图像去噪任务中,传统的颜色空间通道之间存在很强的互相关性,使去噪后图像出现颜色突变,影响图像去噪效果。针对该问题,提出一种降低通道之间相关性的颜色空间构造方法,该方法以待去噪图像在小波变换域中系数的聚集特征为依据,利用主成分分析方法确定系数聚集的主方向、次主方向。由主方向和次主方向的基向量确定自适应的颜色空间,在该颜色空间中实现图像去噪。实验结果表明,相比传统的颜色空间,本文所构造的颜色空间去噪无论在视觉效果、峰值性噪比和稀疏特征保真度上,均取得了更好的去噪效果。

Color image denoising based on wavelet and PCA in adaptive color space

In color image denoising task, there is a strong correlation between color channels, and the correlation makes the denoised image appear color mutation, which affect the image denoising. In order to solve this problem, a method is put forward to reduce the mutual correlation between color channels. Based on the wavelet transform domain coefficient that has aggregation characeristics, the method use PCA to determine principal direction and secondary principal direction of coefficient aggregation ,an adaptive linear transform space. After that, their directions are used to determine adaptive color space where color image denoising can be achieved .The experimental results show that, our method is better in view of subjective parameters and objective parameters, which contain the peak signal to noise ratio and sparse feature fidelity .