Design and Research of Non-contact Temperature Measurement and Identity Recognition System

Abstract

With the rapid spread of the new crown epidemic, prevention of the epidemic has become an important task in daily life. As of now, the cumulative number of confirmed cases in the world has exceeded 150 million, and the epidemic prevention situation is very serious. In the face of the spread of the epidemic, effective temperature measurement, mask recognition, and face recognition are necessary for mobile personnel. Traditional methods such as the deployment of Raspberry Pi, windows platform, etc. can better meet the above requirements, but there are disadvantages such as low number of recognition frames, low accuracy, and bulky equipment. Due to the large market demand, the cost of the chip has also risen. If the above-mentioned functions are deployed according to the traditional method, the cost is relatively high, which increases the economic pressure on small and medium-sized enterprises. The main work of this research is to solve related functions through a processor Kendryte K210 based on the open source instruction set RISC-V. So as to realize a portable and low-cost non-contact temperature measuring device. The paper introduces a low-cost solution based on K210 on the basis of comprehensive discussion of multi-party implementation methods. Based on this processor, the face recognition and mask recognition functions based on YOLOv2 training are deployed through its machine vision and machine hearing capabilities. The temperature is measured by an external MLX90614 sensor; the serial port screen and the voice module JQ8400 are used for user interaction, which meets the guidance of users to the greatest extent. In the design of the hardware circuit, a lithium battery charging and boosting mobile power board based on MH-CD42 is designed. The user can supply power through an external lithium battery to realize the portability of the system.