摘要
图像修补技术已经广泛的应用于各个领域。为了寻找在编程中容易实现且能够对常规的局部受损图像进行快速修复的算法,研究了大量的算法,通过实验对比选择了一种基于样本块的图像修补算法,并做了改进,最后利用MATLAB语言开发了基于样本块图像修复系统。对该系统进行了测试,实验结果表明该系统对于局部受损的纹理图像和结构图像有很好的修复效果且修复速度比现有的图像修复系统更快、信噪比更高,特别对于移除障碍物后所造成的图像受损区域其视觉修补效果更佳,值得推广。
Image inpainting is greatly applied in invarous areas.In order to find the algorithm which is easy to implement and can repair the conventional damage of image rapidly.A large number of algorithms are researched in this paper.Finally an image inpainting algorithm based on sample block is selected through the experimental comparison.Still it has improved here.At last an image inpainting system on the improved algorithm has been developed in MATLAB .The system is tested.And it shows that the system has very good effect in repairing the texture image and structural image damage through the experimental results.And the image restoration speed is also faster than the existing system.Especially for removing obstructions caused by the effect of repairing damaged areas.So the system is worth popularizing.
出处
《电子测量技术》
2015年第8期73-75,共3页
Electronic Measurement Technology
基金
陕西省自然科学基金(2014JM8357
2014JQ2-6031)
延安科学技术局科技规划(2013-kg15
2012kg-07)
延安大学青年基金(YG2013-15
YG2011-16)项目
关键词
图像修补
优先权
最佳样本块
置信度
image inpainting
priority
optimal sample-block
confidence value
