基于L-K局域光流的空间视线坐标法面形测量技术

A L-K Local Optical-flow-based Coordinate Profilometry by Using the Capture Line in Space

提出了一种基于L-K局域光流的空间视线面形测量技术.测量系统由一个投影仪和一个CCD摄像机组成,采用小投影角度的投影方式.通过投影光线与视线(观测光线)的交点坐标,直接计算得出被测物体三维面形高度分布,其中条纹图中观察位置的变化由L-K光流算法计算得到.建立了在点光源投影条件下投影光线与视线交点坐标、光流与被测物体面形高度之间的关系.模拟与实验结果证明该方法能够准确恢复被测物体高度.与传统面形测量方法不同,视线光流面形测量技术不需要采集多幅条纹图像,也无需计算条纹频率和物体相位值,只需两幅条纹图即可恢复物体面形高度分布.

A technique based on L-K local optical flow algorithm for retrieving the height distribution of a measured surface is proposed by the calculation of the coordinates between the capture line and projection line in space.The measuring system is composed of a CCD camera and a projector which project a grating pattern with a small angle.The height distribution of a measured surface can be calculated directly by using the intersection coordinates of the projected line and the captured line,where the changes of the observation position between two fringe patterns caused by the measured surface can be calculated by LK optical flow algorithm.The theoretical relationship between the optical flow and the height of the measured object at the intersection of the projected light and the line of sight under the projection condition of the point source is established.Simulation and experimental results show that the proposed method can retrieve the height of the measured object surface accurately.Different from traditional shape measurement methods,the proposed surface shape measurement technique using captured line and optical flow does not need to collect multiple fringe images,nor need to calculate the fringe frequency and the phase distribution,but only needs two fringe images to retrieve the height distribution of the measured surface.