Development of a 3D mapping using 2D/3D sensors for mobile robot locomotion

This paper presents a simultaneous localization and mapping algorithm (SLAM) by using a new 3D sensor namely the photonic mixer devices (PMD). The PMD camera enables 3D image grabbing within a few milliseconds and gives an important impulse in visual 3D sensing. This camera is capable of capturing reliable depth images directly in real-time. The PMD is also compact and affordable, which makes it attractive for versatile applications including surveillance and computer vision. However, the PMD based device has still limited resolution and provides only gray scale information. To achieve the virtual geometric data, the rather new 3D PMD and 2D RGB color cameras are combined to generate visually realistic 3D maps. Visual input from the 2D camera not only delivers high resolution texture data but also enables the mobile robot to enhance the 3D data calculated from the range output of the PMD camera. Precision robot locomotion is implemented in order to register the 3D geometric data simultaneously.