基于偏微分方程的图像修复数值模型  

Numerical Model of Image Inpainting Based on Partial Differential Equation

作　　者：霍俊蓉 

机构地区：[1]沈阳师范大学数学与系统科学学院,辽宁 沈阳

出　　处：《计算机科学与应用》2021年第8期2035-2041,共7页Computer Science and Application

摘　　要：图像修复是图像处理的一个分支,是利用周围区域的信息填充图像的缺失或损坏区域。本文为今后能够更好地对破损图像进行修复与技术处理,进而针对图像修复中常见的边缘问题总结了几类基于偏微分方程进行图像修复的数值模型,这几种数值模型能够在保证修复效果较好且稳定的同时减少计算量,并提高修复效率,进一步完善修复效果,能够更好地为图像修复发展研究提供理论与数值模型基础。Image inpainting is a branch of image processing, which uses the information of the surrounding area to fill in the missing or damaged area of the image. In order to better repair and process the damaged image in the future, this paper summarizes several kinds of numerical models based on partial differential equation for image restoration, aiming at the common edge problems in image restoration. These numerical models can ensure the good and stable repair effect, reduce the amount of calculation, improve the repair efficiency, and further improve the repair effect, it can provide a theoretical and numerical model foundation for the development of image inpainting.

关 键 词：图像修复 图像处理 TV模型 偏微分方程 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]