Underground personnel localization is highly important in the operations of coal mines. Considering the special underground environment, this paper introduces a novel localization scheme based on step detection and image recognition technologies, which makes use of unique characteristics of the underground environment like the dark environment and the miner’s lamp. Since the underground topology is relatively simple, the miner can be located only by step information. However, the localization with step information always causes the problem of cumulative error. To solve this problem, we rebuild a special base station with a camera in a dark underground environment. A miner’s lamp, which every miner carries, can simply transform to irradiate unique shapes (such as triangles, rectangles, and circles) and every coal miner at the base station can identify these shapes based on image recognition technologies. Thus, we can obtain the miner’s precise position when he/she is passing by a base station. In that way, we can correct the localization results to solve cumulative error. We implemented our algorithm in indoor and underground environments. The experimental results show that 96% of spatial errors were 2.5 m or less.
Read full abstract