【摘要】：作為圖像預處理手段之一,圖像去噪在圖像處理領域發(fā)揮著重要作用。基于非局部均值的圖像去噪算法是圖像去噪研究的重點。圖像質量評價是對圖像視覺質量進行評價的方法。本文研究了基于非局部均值和圖像質量評價的圖像去噪算法。在基于圖像自相似性進行圖像去噪的情況下,一般是利用歐式距離進行圖像塊的相似性度量。一些客觀的圖像質量評價方法不僅考慮圖像的像素,還考慮圖像的亮度、對比度、結構等信息。與只用歐式距離考慮圖像塊像素值的差異相比,本文研究了把客觀圖像質量評價方法與歐式距離結合起來尋找圖像相似塊的方法,尋找到的相似塊會在結構等圖像信息方面有更多相似。自然狀況下的圖像通常包含多種噪聲,并且不容易分辨。通常是用高斯分布模型對圖像分布模型進行模擬進而尋找相似塊,但是高斯分布模型不一定適用于多種噪聲混合下的圖像,因而需要考慮圖像分布模型的自適應情況。本文研究了自適應軟閾值為基礎的圖像去噪方法,經過實驗對比,本文提出的方法取得了較好的去噪效果。
[Abstract]:As one of the means of image preprocessing, image denoising plays an important role in the field of image processing. Image denoising algorithm based on non-local mean is the focus of image denoising. Image quality evaluation is a method to evaluate image visual quality. An image denoising algorithm based on non-local mean and image quality evaluation is studied in this paper. In the case of image denoising based on image self-similarity, Euclidean distance is generally used to measure the similarity of image blocks. Some objective image quality evaluation methods not only consider the pixels of the image, but also consider the brightness, contrast and structure of the image. Compared with only using Euclidean distance to consider the pixel value of image block, this paper studies the method of combining objective image quality evaluation method with Euclidean distance to find image similar block. The similar blocks will be more similar in structure and other image information. Natural images usually contain multiple noises and are not easily discernible. Gao Si distribution model is usually used to simulate the image distribution model and to find similar blocks. However, Gao Si distribution model is not necessarily suitable for images with multiple noises, so it is necessary to consider the adaptive image distribution model. The image denoising method based on adaptive soft threshold is studied in this paper.
【學位授予單位】：天津大學
【學位級別】：碩士
【學位授予年份】：2016
【分類號】：TP391.41

【參考文獻】

相關期刊論文 前1條

1 路文;高新波;王體勝;;一種基于WBCT的自然圖像質量評價方法[J];電子學報;2008年02期

，

本文編號：2302412