对图像去噪的Perona-Malik模型的改良及数值算法的研究.pdfVIP

基于扩散张量的自适应正则化变分模型。第三章在Perona—Malik模 型和Catte正则化模型的基础上多考虑了二阶导数给出了新的图像 去噪模型及其算法，并给出了模型的适定性证明，还做了数值仿真模 拟，实验结果表明该模型不但能有效地去除噪声，而且能增强边缘并 保持边缘位置。第四章加入了强度因子进一步改进了去噪模型，并做 了数值仿真模拟。由仿真的结果可以明显的看出处理后的图像清晰度 和对比度都有大大增加，看上去也比较真实自然。 关键词图像去噪，偏微分方程，正则化，扩散率函数，差分法 舢Y洲2删—● 删3删7删7舢8哪4 AB STRACT Image isaclassical in the denoising problemimage processing．With of of aretransmitted developmenttechnologyimage processing，images thechannelor the areconfinedtothe by preservedby medium．Images extemal noisewillaffectthe invision．In conditions，the physical image needclearand areas，we manyapplication highqualityimages．Therefore， isnot an imagedenoisingonlyimportantimageprocessingproblem，but alsoan for other importantpretreatmentprocess image on of bringgreatimpactsubsequentprocessing．Theprincipleimage istousethedifferenceofnoise andthe in denoising signal imagesignal domainonthedistribution．The is frequency imagesignalmainly distributedinthelow thenoise isin frequencyregion，and signal high duetoitsweakcorrelationwiththearound frequencyregion pixels． Traditionalmethodsof are basedonalow