Design of Face Recognition Temperature Detector Based on YoloV2 Strategy

Because of the new champions league epidemic raged in the whole world, a variety of inconvenience to people’s production, life, at present, although the effective control of domestic outbreak, but the epidemic still cannot relax, requires a lot of manpower, material resources, financial resources, in particular schools, hospitals, shopping malls and other crowded places, as the main representation of the new champions league patients body fever, Temperature detection is needed for initial screening of COVID-19 patients. Based on the above reasons, a non-contact, long-distance body temperature detector was designed to detect the temperature of candidates to avoid problems such as temperature misdetection, false alarm, and omission. The method of manual verification by invigilators was changed, and the YoloV2 algorithm was used to obtain identity in real time and automatically. Validation results. This article first analyzes the existing problems of the current body temperature detector and the market demand for epidemic prevention and control, and then designs the non-contact, remote body temperature measurement and face recognition functions of the body temperature detector. Finally, it has been actually verified to meet the design requirements, especially for the success rate of face recognition of blurred images is over 74%. With the continuous improvement and upgrade of functions, it will have high engineering application value.