Research On Calibration Of 3D Imaging System With Structured Illumination And Detection Of Sub-pixel

As non-contact, high speed and resolution measurement is approached by 3D sensor technique based on structured illumination, it has been widely applied into computer vision, industrial inspection, quality control, movie stunts and biomedicine etc. With the development of computer and image processing technology, there are enormous progresses in optical non-contact measurement. 3D imaging system based on structured illumination is considered to be the most practical 3D measurement method in future application. At present, it is widely applied into robot vision, automatically process and can be used to measure object with complex surface.The main work in this paper has three parts: Firstly, principle of 3D imaging system based on structured illumination is studied. A novel method is proposed to simulate the system. Then suitable parameters can be provided when designing or calibrating actual measurement system. Secondly, according to the reversibility of light path principle, the projector can be regarded as camera acting in reverse. Thus, the model of camera can also be used for projector. The intrinsic and extrinsic parameters of camera and projector are calibrated by placing a planar target in measurement volume at least three orientations. The planar target can be moved freely. The motion need not be known. Then parameters of measurement system are calibrated. Therefore, difficulties in making and carrying target are decreased. It is suitable for real-time calibration. Thirdly, the imaging character of optical system is analyzed. For the spatio-temporal encoding 3D imaging system with structured illumination, a novel method is proposed to detect the sub-pixel of projector coordinates. The experimental result shows that precision of the system is improved. The novel ideas of the paper are as follows: First of all a 2D planar target is introduced to calibrate the camera. Virtual planes are introduced to calibrate 3D imaging system based on structured illumination. Then, temporal sinusoidal fitting is proposed to detect sub-pixel of projector coordinate in 3D imaging system using binary spatio-temporal encoded illumination. Furthermore, according to the model of measurement system, a novel method with high precision is proposed to simulate the 3D imaging system based on structured illumination. At last, the relation between projector coordinate and height is derived from camera model. Five orders polynomial equation is used to express mapping equation between projector coordinate and height.