Research On Point Cloud Splicing And Color Texture Of Optical 3-D Measurement

Optical three-dimensional shape measurement technology is integrated with optoelectronics, image and graph processing, computer control, mechanical and so on, which developed rapidly from the 20 century with the modern science and technology. It has extensive applications and wide prospect in many areas such as quality control in manufacturing, machine vision, remote sensing, biomedicine, and social security as well as others. And grating projection three-dimensional shape measurement technology is a kind of rapidly developing active optical three-dimensional shape measurement technology in recent years, with its higher measurement speed, higher measurement accuracy and practical value as one of the most promising active vision technology. The main work of the thesis is about theory, method and key technology of three-dimensional measurement based on fringe projection, as follows:(1) Analysis the shortcoming of the method of calculation and unwrapping phase which is combined traditional phase shift technology and gray encode, and adoption a complementary gray encode phase unwrapping method to solve the problem of errors and make the encoded image binarization process become relatively simple.(2) Do targeted research on the existing three-dimensional point cloud splicing technology in theory, and conduct in-depth study on the stitching algorithm based on the marking points pasted recognition combined with experiments. The experimental analysis shows that the stitching algorithm based on the marking points pasted recognition has high accuracy stitching, which is applicable for multi-angle measuring three-dimensional point cloud stitching of simple device.(3) Inspired by the color space defined, use three colors images to synthesize the color texture of the object and give black and white camera the ability to obtain color texture, and adoption a simple color texture optimization algorithm to improve color texture obtained. Finally, the algorithm which can accurately obtain color texture of the measured object is verified by experiment.