双目视图运动图像去模糊方法

Deblurring of double view motion image

现代工业生产线上采集到的产品图像存在运动模糊,对后期探测产品表面信息以及重建产品轮廓产生影响.为了解决这一问题,提出一种基于双目视图的运动图像去模糊方法.通过双目相机之间的对极几何关系以及相机模型,推导出两视点图像点扩散函数路径之间的内在联系,并运用像机标定模块求取标定矩阵中的参数.通过此内在约束条件求取精确的点扩散函数,并对双目视点图像进行统一去模糊.结果表明,通过双视点图像几何关系求取的点扩散函数路径来进行去模糊取得了较好的效果.

The images acquired from the actual production line in modern industrial production are motion-blurred, which affects the detection of the surface information and the reconstruction of the profile. To solve this problem, we proposed a deblurring method based on double view motion image. We deduced the relationship of the point spread function path between the double view images through the epipolar geometry and the camera model,and solved the parameters in the calibration matrix with the calibration system based on MATLAB, then we got the association point spread function(PSF) path based on the constraint condition,and deblurred the motion images. The experimental results show that the motion image deblurring using the PSF path solved by the geometric relationship has better effect.