Large Field of View Array System Using Low Cost RGB-D Camerasin

Driven by the miniaturization, lightweight of positioning and remote sensing sensors as well as the urgent needs for fusing indoor and outdoor maps for next generation navigation, 3D indoor mapping from mobile scanning is a hot research and application topic. It has been applied in indoor modeling, indoor localization and other rising fields. In general, 3D indoor mobile mapping systems are equipped with sensors include laser scanner, panoramic camera, IMU (inertial measurement unit) system and odometry. The IMMS system can achieve indoor 3D indoor point cloud data acquisition, but the range sensor, laser scanner, is very costly and poor portability. RGB-D (RGB-depth) camera can offer an alternative way to capture data from indoor scene. However, the narrow field of view of RGB-D sensors cannot provide sufficient efficiency and wholeness of data acquisition, and may cause tracking failed or false match more frequently in SLAM system. This paper builts a low cost indoor 3D mobile mapping system prototype device, and proposes a calibration approach which integrates multiple consumer level RGB-D cameras to large field of view system to solve the deficiency mentioned above. The overall analysis shows the precision of this system meets the demanding of basic application of indoor data collection.