Real-time person detection using an embedded system for UAV Search and Rescue operations

Abstract

Unmanned Aerial Vehicles (UAVs) are becoming more of a viable solution for tasks that pose difficulties to emergency services during Search and Rescue (SAR) operations. The ability to quickly identify and locate a lost or injured victim in varying environments minimizes the time needed to accomplish a rescue. The aim of this paper is to develop a real-time victim detection algorithm to address this problem. More specifically, we apply colour segmentation and graph theory techniques in order to locate a static target in an image that differs in colour to its surroundings. The algorithm is then implemented using OpenCV, a fast open-source computer vision library, on a single-board computer (SBC) equipped with a high performance processor. We present empirical results demonstrating the performance of the algorithm for an outdoor environment using an embedded platform for tasks drawn from the SAR domain. Additionally, we demonstrate the number of false positives of the proposed algorithm and compare the computational speed of the algorithm for varying sizes of images.