摘要
为了提高图像划痕的修复效果,提出动态弹性云模型结合自适应方法。首先,建立弹性云模型,根据处理的像素数据使云节点的计算能力发生形变;然后,通过隶属度函数动态分配修复像素任务,把需要处理的像素数据分为1级、2级、3级,优先等级逐渐降低;接着在划痕位置定位的检测中加入对划痕颜色的判断,根据像素方差变化衡量修复模板间像素值的波动水平,自适应选择最优模板块,以4个像素点作为区域增大的步长;最后,给出了算法流程。实验仿真表明,动态弹性云模型修复效果没有断痕,处理时间较少,PSNR指标为37.009 9 d B,数据较优。
In order to improve the effect of image restoration,dynamic elastic cloud model algorithm and adaptive method were proposed. First,elastic cloud model was established,computing of cloud node was made deformation according to pixel data processing. Second,restoration pixel task was allocated dynamically using membership function,pixel data need to deal with three levels： 1,2,3,priority level gradually was reduced. Third,location of the scratches on the detection of the color scratch was detected,adaptive optimal plate was selected by pixel variance which was measured volatility level pixel value between repair template,with 4 pixel points as the region increasing the step size.Finally,the process was given. Simulation shows that result of repair image is no breaks,time is less,PSNR index is37. 0099 d B,data are better.
出处
《实验室研究与探索》
CAS
北大核心
2016年第3期46-50,共5页
Research and Exploration In Laboratory
基金
河南省科技发展计划项目(132102210522)
关键词
动态
弹性
云模型
自适应
划痕
dynamic
elasticity
cloud model
adaptive repair
scratch
