Design And Implementation Of 3D Reconstruction System Based On RGBD

3D reconstruction of real scene is a very important research direction in computer vision and image processing,and it has been widely applied in many fields such as virtual reality,game development and so on.In recent years,for the reason of the development of deep camera technology and the improvement of 3D reconstruction technology requirement,3D reconstruction technology based on RGBD data is becoming more and more important.In this paper,we thoroughly study the process of 3D reconstruction based on RGBD data,then put forward some corresponding solutions in view of the multiple problems in it,and design new 3D reconstruction system by using the existing frame.The main work and innovation of this paper are as follows:1.Proposing a method of point cloud registration based on ORB feature points to solve the problem of too long time consuming in the traditional registration method.In this paper,we obtain the point cloud by using the matching ORB feature points,which reduces the number of point cloud and reduces the amount of computation.Then we use matching feature points to solve the initial position and pose of the camera,thus reducing the number of iterative computation.The experimental results show that this method can significantly reduce the time consuming of registration process under the premise of taking into account the registration accuracy.2.Proposing a method of local point cloud registration based on feature points to solve the problem of low success rate of traditional matching method in fast moving scene.In the fast moving scene,traditional registration method has a small proportion of matching data,and the point data is too small in method of innovation point one,which makes the success rate of point cloud registration is not high.Therefore,we use the local pixels of the matching feature points to get the local point cloud,thus increasing the proportion of matching point cloud to total point cloud,and the experimental results show that the success rate of point cloud registration improves without affecting the accuracy and speed.3.Deeply analyze the whole process of 3D reconstruction,and design a new 3D construction system for the problem in it.We divide the system into four modules:RGBD data acquisition,point cloud conversion,point cloud registration and point cloud fusion,and the point cloud conversion module and the point cloud registration module are used to implement and verify the above two methods.Finally,we analyze and realize part of the algorithm and engineering problems.