Méthodologie de Recherche

Design and development of a smart contactless thermometer using face recognition

Author: Hatem ZEHIR

Abstract – The aim of this paper is to develop a smart thermometer based on Raspberry Pi 4b. Living with the Novel Coronavirus is becoming the new way of life in countries all over the world. However, to stop the virus from spreading, people who have the Coronavirus must be separated from the rest of the population. This work is an attempt to make a smart contactless thermometer based on the MLX90614 sensor and the Raspberry Pi.

Keywords: Facial Recognition, Raspberry Pi, Python, Database

INTRODUCTION

Today, we are going through one of the biggest health crises in history due to the spread of the COVID-19 and the increase in the number of Coronavirus cases. The COVID-19 is caused by the SARS-COV2 virus, this novel respiratory disease has affected most people’s way of life around the globe.

Common symptoms of COVID-19 [1] include high body temperature (above 38°C), dry cough, tiredness, in addition to other symptoms such as aches and pains headache, loss of taste or smell, and difficulty breathing or shortness of breath. To reduce the coronavirus cases and stop the spread of the virus, many safety procedures were taken by governments, such as obligatory mask-wearing and regularly test citizens. Most countries are also making temperature checkups and mask mandatory for schools, offices, and other workspaces. Unfortunately, in some countries, there is not enough tests for all people, which makes it impossible to identify all of the positive cases. In this pandemic, companies must continue to produce and provide their services, but without neglecting the coronavirus health protection regulations. During virus outbreaks like COVID-19 companies around the globe must play a major role to stop the spread of the disease, especially in detecting employees or customers that may have been infected by the virus.

Today, temperature checkups are done manually using IR contactless thermometer, which is inefficient and impractical especially in places with large crowds. To solve all of these problems, we have designed a small and portable device that automates the process of temperature checkup by using a contactless IR temperature sensor and facial recognition. Our solution involves building a device capable of detecting a person in an office, company, or home when he has a high body temperature. This can serve as an early warning system we can make a plan of action ahead of time. This product will serve as a way of protecting the company, the employees, and the customer.

The device is designed to recognize the faces and to determine whether the person has a high body temperature or not. Using the collected data, we can decide whether the concerned person can be allowed inside a public place or not.

For this, we will use the Raspberry Pi 4b [2]–[4] because of its low cost, vast peripheral support, small size, and huge processing power.

The rest of this paper is organized as follow: the first part is dedicated to explaining the general system conception. In the second chapter, we will focus on the hardware and techniques used in the project such as the Raspberry Pi 4b, the MLX90614 sensor[3][16], face recognition algorithms, QR code[5][8][9],and the SQL database. Finally, in the third chapter, we will discuss the final results of the project.

CONCEPTION OF THE SYSTEM

In this part we will discuss the hardware used in our system: a definition of the Raspberry Pi 4b, The MLX90614 IR temperature sensor and its characteristics, and the Raspberry Pi camera module.

The Raspberry PI 4b

The Raspberry Pi 4b was released in June 2019. It is a 64-Bit, 1.5 GHz quad core, with 2 GB, 4 GB or 8 GB of LPDDR4 SDRAM. it has also full-throughput Gigabit Ethernet, dual-band 802.11ac wireless networking, Bluetooth 5.0, two USB 3.0 and two USB 2.0 ports, dual monitor support at resolutions up to 4K 60FPS, VideoCore VI graphics supporting OpenGL ES 3.x, 4Kp60 hardware decode of HEVC video and complete compatibility with earlier Raspberry Pi products.

Power consumption is 3.4 watts at idle, that number jumps to 7.6 watts under full stress. Figure 1 shows a metal cap on the CPU chip that helps to dissipate heat without requiring a heat sink.

Figure 1:Raspberry 4b

Temperature sensor (MLX90614)

As previously stated, the MLX90614 sensor can measure an object's temperature without making physical contact with it. This is made possible by the Stefan-Boltzmann Law, which states that all objects and living beings emit infrared energy, the intensity of which is directly proportional to the temperature of the object or living being. As a result, the MLX90614 sensor calculates an object's temperature by measuring the amount of IR energy it emits.

Figure 2: MLX90614 IR temperature sensor

Raspberry Pi Camera Module v1.3

The Raspberry Pi camera has a resolution of 5 megapixels and is compatible with the Raspberry Pi 2B, 3, 3B, 3B+, and 4 models. It's also compatible with the Zero and Zero W models.

RESULTS AND DISCUSSION

The algorithm starts by classifying the real-time images and recognizes faces and check the temperature. The image will be loaded preprocessed for classification after importing the necessary packages and checking the temperature.

After checking the temperature and recognizing the faces the results will be stored an SQL database as follow:

• If the temperature is lower than 38°C the entry will be stored in the “employees” table in the database.

• If the temperature is higher than 38°C the entry will be stored in both “RedZone” and “employees” tables.

If the checked person has a normal temperature the green LED will turn on, else a red LED will turn on.

If a person has a high temperature it can be allowed to enter the building if he provides a negative COVID-19 test. The test has a form of QR-code and includes the following details:

• Person name

• Test result (positive or negative)

CONCLUSION

In this article, we have proposed a smart thermometer system that uses the Digital Non-Contact Infrared Thermometer MLX90614 and facial recognition techniques with the Raspberry Pi 4b and an SQLite database for data gathering.

By using this device to measure the temperature of people and detect fever, we can help to fight the COVID-19 pandemic that is currently affecting the world.

The functionalities of our fever detector system have been successfully tested; the facial recognition, the temperature measurement using the MLX90614 IR sensor, save the measurements using an SQLite database, and the QR code.

The paper was divided into two main parts: conception of the system, and results and discussion. The first part focused on the hardware and the tools needed and used to create our system. In the second part, we have discussed the results of our project, as well as a walk through the execution of our face recognition system, temperature checking, the SQLite database, and the results obtained from reading the QR code and removing a person from the database.

A suggestion for future projects is to use the MLX90640 thermal camera (Instead of the non-contact thermometer MLX90614) and a custom 3D printed case, this makes large-scale deployments simple for businesses of all sizes.

REFERENCES

[1] CDC, ‘Coronavirus Disease 2019 (COVID-19) – Symptoms’, Centers for Disease Control and Prevention, Mar. 22, 2022. https://www.cdc.gov/coronavirus/2019-ncov/symptoms-testing/symptoms.html (accessed Jul. 25, 2022).

[2] R. P. Foundation, ‘Teach, learn, and make with the Raspberry Pi Foundation’, Raspberry Pi Foundation. https://www.raspberrypi.org/ (accessed Jul. 25, 2022).

[3] J. W. Jolles, ‘Broad-scale applications of the Raspberry Pi: A review and guide for biologists’, Methods Ecol. Evol., vol. 12, no. 9, pp. 1562–1579, 2021, doi: 10.1111/2041-210X.13652.

[4] O. Nath, ‘REVIEW ON RASPBERRY Pi 3b+ AND ITS SCOPE’, vol. 4, no. 9, p. 3, 2020.