基于小波多尺度的图像增强新算法

Novel algorithm for image enhancement based on wavelet multiscale

传统图像增强算法在增强对比度的同时,也很大地提升图像噪声,需要对图像进行降噪处理。小波增强方法兼顾图像信号的空域和频域特性,但没有充分考虑到视觉的非线性特性。针对现有图像增强技术的这一缺陷,在分析小波变换对噪声影响规律的基础上,结合小波多尺度的特性,提出了一种基于小波多尺度的图像增强新算法,利用不同尺度上的小波系数间的相关性和小波分析的时频局部化特性来有效区分噪声和图像信息,有效改善了图像增强过程中的噪声放大问题。

Traditional image enhancement algorithm enhances the contrast grade and noise of image at the same time.So,noise reduction is needed.Wavelet enhancement considers the spatial domain and frequency domain of image signal,but not the nonlinearity of vision.Aim at the limitation of the image enhancement technology,a new algorithm for image enhancement based on wavelet multiscale is proposed by analyzing the effects of wavelet on signal and noise regularity.Making use of the correlation of the wavelet coefficient from different scales and the frequency localization of wavelet,the noise and image information is divided effectively.It can alleviate the noise amplification while enhancing the image feature.