基于非盲去模糊算法的图像复原  

Image restoration application based on non-blind de-convolution

作　　者：朱侃逸 霍欢[1] 龚如宾[1] 徐亚[1] 奚金金[1] 

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《信息技术》2014年第4期16-18,23,共4页Information Technology

基　　金：国家自然科学基金(61003031);上海市研究生创新基金项目(JWCXSL1202)

摘　　要：在图像复原中,对降质图像失真的点扩散函数求逆是研究难点之一。现根据先验概率将贝叶斯理论的函数关系化,提出遗传算法的种群繁殖模型,通过逐步演化来执行变换、调整和集成。同时,为了获取更高效的效果,构建Jenkins持续集成平台进行集成和测试模型参数。最后,从决策支持系统的角度阐述整个系统的框架。In image restoration, the inverse operation of point spread function（PSF） of degraded image is one of research difficulties. The paper demonstrated that the population propagation model of the genetic algorithm is gradually evolved to change, adjust and integrate based on prior probabilities in the function＇ s relation of bates theory. At the same time, Jenkins continuous integration is implemented .in integrating and testing the parameters of model to obtain more efficient result. Finally, the framework for whole system is stated by the view of decision support system.

关 键 词：非盲 贝叶斯理论 Jenkins持续集成 遗传算法 决策支持系统 

分 类 号：TN911.73[电子电信—通信与信息系统]