3D Surface Reconstruction Based On Binocular Vision Using Structured Light

The methods of3D surface measurement include contact3D measurement andnon-contact3D measurement. Non-contact3D measurement includes active andpassive. The measurement methods based on optical are widely used in two kindsof method.This paper studies the current research work of3D surface measurement in market andresearch lab, compares the technologies they used and chooses an efficient3D visionmeasure method based on coded-structured-light, and realizes the measurement of each3Dsurface and correction work based on the method. Finally, realizing the3D reconstructionof the complete target by using the method of artificial marking feature pointsand splicing various3D surface.The accuracy of3D surface measurement mainly depends on three links.1. The calibration and correction of camera. In this paper the principle ofcamera calibration and correction are studied and expounded, and solve theproblem in tests.2. The accuracy of the decoding of Gray code views and Phase-sift views. Weimprove boundary accuracy by the methods which compares gray code and theirinverses.， and solve the problem of period dislocation between Gray code andphase-shift in course of decoding. Do not meet the accuracy requirements ofthe points are filtered in this process. Calculate the three-dimensional points at last,and filter again.3. This paper use the method of artificial marking and identification of featurepoints to improve the precision and efficiency of splicing.In this paper,3D surface reconstruction of face of gypsum model is realized by VC.