Three-dimensional Shape Reconstruction For Dynamic Object Based On Color-coded Method

The world we observe is three dimensional (3D), but the image captured by the image acquisition device is two dimensional information, which misses the depth information. Three dimensional (3D) shape reconstruction techniques can recover the three dimensional information from a two dimensional image.3D measurement techniques draw a lot of attention in industrial production, reverse engineering, medical diagnosis and so on. With the rapid increase of the needs of practical application and computers, the measurement technique of three dimensional (3D) makes astounding advances. The color coded structure light method which is rapid and high precision is becoming a hot topic.In this paper, the technique for 3D measurement based on color-coded method is studied both theoretically and experimentally.The main content of the paper are as follow.Firstly, kinds of structure light methods are introduced in the paper, and a brief introduction of the development status and development trend of the measurement technique of 3D is given. Around color-coded method, this paper comparied the different methods, including timing sequence encoding method and spatial encoding method. Moreover, the spatial encoding method based on color-coded is studied on account of the 3D shape reconstruction for dynamic object.Secondly, the 3D shape reconstruction for dynamic object is discussed in this paper, and also an encoding method based on color-encoded digital fringe projection is presented. It is used for the 3D reconstruction of dynamic programming. Only one image is needed to be captured by the camera. The projected structured light is made up by red, green, blue and yellow stripes. Each group has two stripes, which are not the same, with a white stripe between groups. Deformed fringe images are grabbed by a CCD camera. Then, dates are processed with a suitable filtering. We get the information of height from the difference between the reference image and the obtained image, which has been encoded and decoded. Besides, the method uses three primary colors and their complementary color as the color stripes. It ensures that different colors can be discriminated exactly. This method discriminate varieties of colors dynamically, and can effectively avoid recognition error which is caused by the influence of the surface color of object and the outside environment.Finally, the 3D measurement experiments are designed and the results of two objects are presented.Error analysis of the experiment results is given, which proves the feasibility and validity of the proposed method.