基于傅里叶变换去隔行图像的动态3维面形测量

Dynamic 3-D shape measurement based on de-interlaced images by Fourier transform

为了解决动态面形测量中隔行扫描CCD相机记录动态物体表面变形条纹图像存在缺陷的问题,提出了傅里叶变换去隔行算法,即把隔行扫描CCD获取动态物体的错位模糊帧图像分成两个单场图像,分别对每一单场图像进行傅里叶变换去隔行处理,再利用条纹分析法重建对应时刻的3维面形。理论分析得出单场傅里叶变换去隔行图像与对应的准确满帧图像相同的结论。结果表明,该方法可以很好地恢复条纹和重建物体,且简单实用,可用于基于空间相位检测、相位测量轮廓术、傅里叶变换轮廓术等条纹分析方法的动态物体3维测量中。

In order to eliminate imperfect fringes captured by an interlaced CCD camera when measuring a dynamic object, the Fourier transform de-interlacing algorithm was proposed. The blurring fringe pattern grabbed by the interlaced CCD is divided into two single-field images, each of which is processed into the restored frame image with this algorithm, then their corresponding 3-D shape were reconstructed by means of fringe analysis method. Theoretical analysis shows that the singe-field image disposed by the Fourier transform de-interlacing algorithm is identical with the corresponding exact full frame image. The results of simulation and experiment demonstrate that the imperfect fringe patterns can be restored and real objects can be reconstructed with this method. The method is simple and practical, so it can be applied in 3-D shape measurement for dynamic objects based on spatial phase detection, phase measurement profilometry and Fourier transform profilometry.