摘要
针对三维视觉检测技术应用中动目标多视点去模糊的难题,提出了一种动目标多视点基于标定的去模糊方法.首先估计部分视点模糊图像(以两个视点的模糊图像为例)点扩展函数.其次根据多视点几何关系,利用相机标定方法,得到各个相机的参数矩阵;结合估计出部分视点的点扩展函数和相关参数矩阵,以直线运动为例,建立多视点模糊图像的点扩展函数关系.最后在确保各视点模糊路径对应的关系前提下,对多视点图像进行整体去模糊.结果表明该方法对于对运动模糊图像恢复有很好的效果.
To solve the problem of multi-view motion blur in application of three-dimensional vision inspection,a new multi-view motion deblurring method was proposed.Firstly,the point spread function of some view motion-blurred images was estimated(We took two view motion-blurred images as examples).Then,the parameter matrix of each camera was obtained by a camera calibration method.The corresponding relationship between these point spread functions was built by using the parameter matrix of each camera.Finally,the multi-view motion blurred images were restored as a whole based on the premise that the corresponding relationship between these point spread functions was ensured.The results show that the proposed method improves the effect of image restoration for the multi-view motion blurred images.
出处
《武汉工程大学学报》
CAS
2014年第12期56-62,共7页
Journal of Wuhan Institute of 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项目(2012FFA046)
武汉市国际科技合作计划项目(2014030709020310)
关键词
点扩散函数
多视点
运动模糊
图像复原
point spread function
multi-view
motion blur
image restoration
