Three Dimensional Image Reconstruction System Based On Multiple Kinect Camera

Three-dimensional reconstruction and three-dimensional display technology has become a hot research topic. At present, most of the reconstruction systems are of high cost and complex operation. Aiming at these problems,the main research work includes:1. Three Kinect cameras are used to build a 3D reconstruction system with simple operation and low cost. Image acquisition and processing software are compiled, and integrated three-dimensional display software, which can be manually rotated by the choice of view and perspective.2.Based on the construction of laboratory prototype system and according to the noise characteristics of the color image, I propose an improved median filtering algorithm, which can ensure that the gray value is invariable.3. In this paper, for the camera angle of view overlap problem,based on Zhang’s calibration method, three cameras pairwise calibrated to convert each image coordinate system to the same world coordinate system, and obtain the corresponding parameters.. In order to get accurate rotation matrix and translation matrix, we can calibrate it by manual rotation and translation, to make the camera more accurate external reference.4. The system uses an iterative closest point(ICP) algorithm to match the three splice resulting point cloud camera module, to obtain the full three-dimensional model. For data redundancy cloud overlap between the points, Paper choose one based on K-d tree search algorithm to eliminate redundant data and complete data integration.Research and experimental results show that the Kinect camera can obtain the depth image and color image at the same time, and form a simple, cost-effective three-dimensional system. Through further improvement in terms of accuracy and speed of image reconstruction,the system can be applied to medical imaging, medical surgical procedure of reconstruction and virtual fitting room 3D reconstruction applications. Therefore, the project has a more important research significance.The project has been supported by Natural Science Foundation of Shanxi Province(2015011041)and Research Project Supported by Shanxi Scholarship Council of China(2014-053).