论文部分内容阅读
小波包变换在处理图像中的平滑区域时能够起到较好的效果,而Curvelet变换可以更好地逼近线性奇异高维函数,对图像的边缘区域有最稀疏的表示。在此基础上提出了基于二者联合的图像去噪算法,在对含噪图像进行分割后,分别对线性区域和平滑区域采用Curvelet阈值去噪处理和小波包阈值去噪处理。该方法充分发挥了二者各自的优势,实验表明,它对图像的去噪效果要优于单纯的Curvelet或小波包去噪方法。
Wavelet packet transform can work well in the smooth region of the image, while the Curvelet transform can better approximate the linear singular high dimension function, and has the sparse representation of the edge region of the image. On this basis, an image denoising algorithm based on the combination of the two is proposed. After the noisy image is segmented, Curvelet threshold denoising and wavelet packet threshold denoising are used respectively for the linear region and the smooth region. The method gives full play to their respective advantages, experiments show that it is better than the simple Curvelet or wavelet packet denoising method for image denoising.