A new approach for terrain description in mobile robots for humanitarian demining missions

The humanitarian demining missions require the use of robust systems such as efficient mobile robots and improved sensors. This application domain involves performing tasks in non structured scenarios and dynamically changing environments. This paper is focused on a new approach based on omnidirectional vision systems for terrain description. Computer vision systems are widely used in similar applications. However, conventional video cameras have limited fields of view which make them restricted for several applications in robotics. For example, mobile robots often require a full 360o view of their environment, in order to perform navigational tasks such localizing within the environment, identifying landmarks and determining free paths in which to move. The omnidirectional sensors allow the capture of much wider field of view; they can provide panoramic images of 360o around the robot. Certain techniques have been developed for acquiring panoramic images. This work outlines the fundamental principles of three of these techniques, and shows a direct application of a low-cost catadioptric omnidirectional vision sensor on-board a six-legged robot intended for antipersonnel landmine localization.