Detecting and Locating People Using Low-Cost Thermal Imaging Camera

Aswe are approaching times of Industry 4.0 and smart homes, there is a need forautomatic reliable mechanism for detecting living beings and there are manyuse casesfrom aiding elderly people with their everyday life to increasing safetyin dangerous workplaces by guarding hazardous areas. We propose a method forcounting and locating people using single low-cost thermal camera module andRaspberry Pi-like computer. This paper describes the process of necessaryhardware configuration of the whole system and a software solution to theproblem of detecting and locating people, which is based on applying humantemperature filter, image processing with object detection using OpenCV and 3Dscene reconstruction with known environment parameters. This approach even in its simplestform provides accuracy over 90 % on our data set with various possibilitiesfor improvement.