Real-Time 3D Path and Map Estimation Using a Multi-Camera System and a FastSLAM Algorithm

This work proposed a real-time 3D map building system based on the visual odometry information derived from a Multi-Camera hardware. The accurate odometry information and images derived from the proposed intuitive Multi-Camera hardware, so called Multi-Camera Unit (MCU), are used as an input for the 3D map building process which cooperates the FastSLAM algorithm as a map and path estimator. Experiment results show that the proposed system yield good estimation of 3D path and map in real-time. Using the estimated 3D path and map, a showcase application of 3D photorealistic map building can be implemented with ease where it can be run on top of all process in real-time manner using a standard PC. Copyright © 2013 IFSA.