立体视觉双目图像MAP的优化复原方法

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。

Method for MAP optimization image restoration in binocular stereo system

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.