Autonomous MAV Navigation in Underground Mines Using Darkness Contours Detection

摘要: This article considers a low-cost and light weight platform for the task of autonomous flying inspection in underground mine tunnels. The main contribution this paper is integrating simple, efficient well-established methods computer vision community state art vision-based system Micro Aerial Vehicle (MAV) navigation dark These include Otsu’s threshold Moore-Neighborhood object tracing. can detect position low-illuminated tunnels image frame by exploiting inherent darkness longitudinal direction. In sequel, it converted from pixel coordinates to heading rate command MAV adjusting towards center tunnel. efficacy proposed framework has been evaluated multiple experimental field trials an Sweden, thus demonstrating capability resource-constrained aerial vehicles fly autonomously through tunnel confined spaces.