A Study Of 3D Point Cloud Capture System

3D-technology is paid with much attention nowadays, with the development of IT. And it meets the development tendency of science and technology as well as the requirement of market.3D reconstruction for outdoor buildings is one piece of the hot study fields.The core of outdoor scene 3D reconstruction is capture of high precision 3D point clouds,which is getting the cube data of the outdoor scene via accurate mea-sure, then through extracting, projecting and integrating to get appearance and size of the object with real scale factor.High precision 3D point clouds can be of great benefit to various computer vision applications,like 3D digital map and automatic driver.It’s important to build a capture system of high precision 3D point cloud.There exists kinds of capture systems in reality, including lidar and stereo system-s.Due to the complexity of outdoor environment, stereo system accuracy is not high. Common lidar system is made of laser radar,camera,GPS and IMU with high cost. Therefore, we build a capture system of high precision 3D point cloud with low cost, which is only consist of lidar, camera and synchronization module.We gain both image data and lidar data synchronously because of the synchronization module.To get the correspondence of image and lidar data,we propose a calibration algorithm between lidar and camera.Through processing the image data,structure from motion(sfm) al-gorithms is presented to get a camera trajectory and a 3D point-cloud without scale factor.We propose a point cloud fusion algorithm, which is to joint the lidar based poind-cloud using the camera trajectory and a 3D point-cloud of outdoor scene with real scale factor.