Fast 3d Measurement System Based On Digital Fringe Projection

Three Dimensional morphology of objects can provide more plentiful data than 2D does , and it is needed more and more in industry .Currently, the acquirement of 3D data is not technical problems, the trend has shifted to higher speed and accurate measurement. In this paper, According to the purpose of get 3D morphology of the object in a fast way, a new measurement system is designed which is based on the grating patterns and Fourier transform profilometry(FTP).The system is convenient and simple in structure, and easy to be calibrated, it need not assistant by the reference plane when measurement and calibration is proceeding, high speed in capturing data and also with a good accuracy. This research will be widely used in rapid establishment of three-dimensional database, product online testing, animation, virtual reality and other fields.The main contents of this paper is as follows:1. Fourier transform profilometry is studied, including the theory of FTP, and the background noise reduction, frequency filtering and other means to improve the measurement accuracy of Fourier transform profilometry.2. Two-dimensional phase unwrapping algorithm is researched, according to the different measurement conditions, the adaptive method were supplied. The several different algorithm is compared, including their speed , the effort of the unwrap phase. And the unwrap method according to time direction is also discussed.3. A novel grating projection system based on assistant line and its calibration method is porposed .The equation of translation from relative phase to absolute phase is deduced according the assistant line. Space mapping method and its optimization algorithm is employed, by which the object's real 3d data can be calculated.4. A circle liner target is designed, with which the extraction of feature points and feature line can be accomplished by one image. In the system calibration, Zhang's camera calibration is adopted, the procedure of the calibration is simplified.5. The accuracy of the result is proved by measuring a standard hemispheroid with 220mm in diameter five times , the average of the fitting diameter is 219.95mm,and the mean square error is 0.48mm. and the mean square error of the distance between the 3D points and the spherical surface is arranged in 0.28-0.46 mm. The face with different expressions of a student is measured by the system , the 3D data is acquired by the system with only one fringe pattern and shows a vivid profile of the face.6. A binocular vision system which aims at the 3D texture mapping is designed. In this system, one gray camera captures the 3D coordinates, and one color camera acquires color information. The pixel matching model of binocular cameras was established, in which the intrinsic and structural parameters were defined using one simple calibration method.3D texture mapping model was created with the definition of the higher pixel matching model. The process of the proposed 3D texture mapping method was introduced in detail.