基于块分类的广义加权鲁棒主成分分析图像去噪模型
A Generalized Weighted Robust Principal Component Analysis Image Denoising Model Based on Block Classification

为改善广义加权鲁棒主成分分析模型(GWSLRPCA)模型的去噪性能，解决GWSLRPCA模型去噪后图像平滑区域噪声残留以及图像边缘模糊等问题，本文对GWSLRPCA模型进行了改进。通过Canny边缘检测算子对经过高斯滤波初步去噪后的图像进行边缘提取，对边缘提取后得到的图像进行图像块分解进而对图像块分类，从而对相对位置相同的观测图像的图像块进行分类处理，提升GWSLRPCA模型的去噪性能，使去噪后的图像保留更多的细节纹理信息。实验结果表明，该算法相比于GWSLRPCA模型、WSRPCA模型以及GRPCA模型，有着更高的峰值信噪比和更低的差错率，在视觉效果上也更好。
In order to improve the denoising performance of the generalized weighted robust principal component analysis (GWSLRPCA) model, and to solve the problems of residual noise in the smooth region of the image and blurring of the image edges after the denoising of the GWSLRPCA model, this paper improves the GWSLRPCA model. The Canny edge detection operator is used to extract the edges of the preliminary denoised image after Gaussian filtering, and the image obtained after edge extraction is decomposed into image blocks and then classified into image blocks, so as to categorize the image blocks of the observed images with the same relative position, improve the denoising performance of the GWSLRPCA model, and make the denoised image retain more detailed texture information. The experimental results show that the algorithm has a higher peak signal-to-noise ratio and lower error rate than the GWSLRPCA model, the WSRPCA model, and the GRPCA model, and is also better in terms of visual effect.