立体视觉双目图像MAP的优化复原方法  

Method for MAP optimization image restoration in binocular stereo system

作　　者：洪汉玉[1] 范艳[1] 时愈[1] 华夏[1] 张天序[2] 

机构地区：[1]武汉工程大学电气信息学院,湖北武汉430205 [2]华中科技大学多谱信息处理技术国家级重点实验室,湖北武汉430074

出　　处：《武汉科技大学学报》2016年第1期62-68,共7页Journal of Wuhan University of Science and Technology

基　　金：国家自然科学基金面上项目(61175013);湖北省自然科学基金创新群体项目(2012FFA046)

摘　　要：为了有效地对立体视觉双目模糊图像进行复原,提出一种动目标双目图像MAP的优化复原方法,将双目运动模糊图像的复原问题转化为在模糊核路径关系约束下的双重循环MAP优化估计问题,在保证模糊核路径对应关系前提下,对双目图像进行去模糊。建立了双目图像模糊核路径的对应关系模型,将关系模型作为约束条件式,嵌入到基于模糊核和双目原图像的MAP优化估计过程中,通过双重循环迭代获取双目图像的模糊核和清晰图像。实验结果表明,该方法能够有效地去除立体视觉双目图像的模糊。In order to remove the motion image blurring in binocular stereo vision system effectivery, this paper presents a novel motion image restoration method based on MAP optimization. The problem of motion blurry binocular image restoration can be transformed into MAP optimization and estimation double loop problem within the constraint of the relationship of blur kernel paths. The binocular image deblurring is performed under the condition that the corresponding relationship of blur kernel paths is guaranteed. The corresponding relationship model of motion blur kernel paths for the binocular image is built, and embedded into MAP optimization estimation as the constraints. The clear image and blur kernel of the binocular image is obtained by double circulation iteration. Experimental results show that the proposed approach is effective to deblur the binocular image in the stereo vision system.

关 键 词：立体视觉 动目标 双目图像 MAP 模糊核 图像复原 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]