148

POSTER SESSION
POSTER SESSION W1: Medical Instrumentations
Noncontact-Body-Temperature-Measurement

N. T Hang1, P. T. N Thao1, Tin T. Dang1, Anh Dinh2

1
Faculty of Electrical and Electronics Engineering, Ho Chi Minh City University of
Technology, Ho Chi Minh City, Vietnam
2
Department of Electrical and Computer Engineering, University of Saskatchewan, Saskatoon,
Canada

Abstract. Our study aims to design the prototype system, which includes the
non-contact temperature sensor, embedded system, and database. Using a
specialized mechanical fixture and a simple calibration technique, preliminary
results show desirable temperature measurement performance. To reduce the
inconvenience caused during direct measurement to the users such as having the
thermometer clamped to one part of the body (mouth, armpit, or rectum) and kept
fixed for a certain period, our device provides a new measuring method without
the need for contact and quickly displays results with high accuracy compare with
other devices. Moreover, our purpose is to use this device for making the timely
diagnosis of malaria through the human temperature in households, schools, and
hospitals.
Keywords: Non-contact, Infrared thermal, Temperature, Microcontroller,
Reducing noise.

1. Introduction

The temperature of the human body is one of four main vital signs which must be
observed daily to ensure safe, effective care and is being applied in all healthcare
environments [1]. Consequently, recent years have seen increased attention being given
to automatic temperature monitoring technology. In comparison with traditional
techniques including the Digital Axillary Thermometer (DAT), the automatic
thermometer can reduce the labor cost and capture continuous, reliable and rapid
recording.
This paper deals with non-contact measurement methods, using the infrared sensor.
Everything emits infrared light. Infrared sensor detects infrared light and
microcontroller analyze them to get the temperature. Non-contact thermometers use

membrane) or no (Non-contact Infrared thermometer NCIT) contact with the body [2].
Moreover, allowing for easier and faster use when monitoring huge numbers of
individuals like in school, in hospital, in airports or in border crossings (limit the
disinfection of the thermometer process), apply widespread following the SARS(Severe
Acute Respiratory Syndrome) infection, and particularly during the pandemic H1N1
outbreak [3]. In a similar situation when the outbreak of the disease in Hong Kong, the

BME7 in Vietnam, IFMBE Proceedings 2018

 149

student was required to measure their body temperature at the entrance of the school.
The design of an automatic and rapid measuring device is an urgent demand. However,
the accuracy of the device is a matter to be considered and it also causes a lot of
discomfort when users wear the device like a forehead mask [4]. In the researching
process, we found that infrared sensors give inaccurate results due to the effect of
external factors such as ambient temperature and distance from the sensor to the object.
Consequently, in our paper, we find an approximately function to reduce noise.
Besides the infrared sensor MLX90614 [5] for non-contact temperature measurement,
we use (SONAR) sensor [6] to measure distance and the thermal sensor LM35 [7] to
capture the environment temperature. The approximation functions based on the
distance, environment temperature and body temperature will be calculated by excel.
Then, we appl
each circumstance and display in the LCD [8]. By taking into account the influence of
environmental factors, the approximation gives a more realistic result, respond flexibly
to changes in the environment and contributes to simplifying the measurement process.
2 Hardware Design And Operating Principle
2.1 Hardware design
According to the method described above, the components on the board are divided into
four main units:
Receiver, transmitter, data processor: LM35 thermal sensor, ultrasonic sensor HC-
SR04, infrared sensor MLX90614.
Control Device: Arduino Nano [9], LCD Contrast Ratio Switch, On/Off Switch.
Display Device: LCD.
Independent DC power supply: source 9V DC.
The relationship between the components is shown in the diagram. We use the power
9V to power all the devices in our system. The environmental temperature is measured
by the LM35 sensor. To determine the human temperature, we use the Infrared sensor.
The other essential data is the distance between the human and the sensor which is
measured by the ultrasonic sensor. All data is transmitted to the Arduino Nano, which
will combine, process, output the result to the user. When all data is processed, we use
LCD to display the result.

Fig. 1. Operational principle.

BME7 in Vietnam, IFMBE Proceedings 2018

 150

Arduino with LCD: The Arduino Nano is complete, small, board-friendly board
base on the ATmega328 (Arduino Nano 3.0) or ATmega168 (Arduino Nano 2.x), in
this product, we use Arduino 3.0. This board is easy to use and programming (using
Arduino IDE software based on C programming language) and connected to LCD
through 12, 11, 5, 4, 3, 2 pins. We use LCD 16x2, with 2 rows on the screen, each row
has a maximum of 16 characters. During measurement, the LCD shows the distance,
ambient temperature, and temperature from the IR sensor. After calculating the
influence of environment, the value of the MLX is replaced by the body temperature.

a. Arduino nano. b. LCD 16x2 .

Fig. 2. Arduino nano [9] and LCD 16x2 [8].

Thermal sensor LM35: LM35 [7] is an analog, linear temperature sensor. The voltage
output of the LM35 increases 10mV per degree Celsius rise in temperature. The
accuracy is high and depends on the temperature range, at a temperature of 25 degrees
Celsius, the error is ±1 degrees Celsius, the room temperature accuracy is 0.25 and the
error between -55 and 150 is ±1.5 degrees Celsius. We can calculate the temperature:

U=t*k (1)
t = U/k = 1.1/1024/0.01 = 0.10742188 oC (2)
Where U is the value read from the analog-Read function and U = 1.1 is the
maximum value of the reference voltage, connected to Arduino via A0 pin and k = 0.01.
This means that it will map input voltages between 0 and 1.1 volts into integer
values between 0 and 1023. This yields a resolution between readings of 1.1
volts/1024 units or 0.0010742188 volts (1.0742 mV) per unit. So, the temperature
is 1.0742/10 equal to 0.10742188 per unit.
Infrared sensor: Infrared sensor MLX90614 [5] uses I2C interface. It has a compact
design, high precision thermal sensor. This sensor calibrated in wide range: -40 to
125 for sensor temperature and -70 to 380 for the object temperature, high
accuracy of 0.5 within the range of 0-50 for both T ambient and T object. This
sensor can transmit continuously the measured object temperature for an object
temperature range of - -bit
PWM. Supported by the Adafruit-library Master [10] and is communicated with
Arduino by two pins A4 and A5. Follow is the image of MLX90614 sensor.

BME7 in Vietnam, IFMBE Proceedings 2018

 151

Fig. 3. MLX90614 sensor [5].

Ultrasonic (SONAR): Module Ultrasonic sensor HC-SR04 [6] is able to measure

the range from 2-400cm, precision 3mm and effectual angle is smaller than 15°. The
working principle of the ultrasonic sensor is as follows: the high-level signal is sent at
least 10us by using IO Trigger. The module sends eight 40 KHz signal automatically
and then detects whether a pulse is received or not. If the signal is received, through
high level, time of high duration is the time gap between sending and receiving the
signal.
Distance = (Time x Speed of Sound Air)/2 (3)

(A5) of Arduino. This sensor measures precisely the distance without actual contact
and used at several places like water level measurement, small distance measurement.

Fig. 4. Ultrasonic sensor [6].

2.2 Operating Principle

Fig 5 is the flow chart which talks about the method that we can define and calculate
the approximating function. At first, we have to measure the body temperature in twelve
times by the IR sensor, so we create a counter begins at 1 and finish at 12, the initial
distance is 10cm (measured by the Ultrasonic sensor). Then, the LM35 sensor is used
to measure the ambient temperature, each time the value of Ta and T is saved, we
calculate with increase the distance to about 20cm, and increase the
counter to 1, while the counter is smaller than 12, we continue to measure and collect
the data from the device. If the variable counts to the last value, we perform the function
approximation by excel. Typically, this function is a quadratic or a fifth polynomial of

BME7 in Vietnam, IFMBE Proceedings 2018

 152

the variable that is the distance between the device and the meter. The accuracy of this
function will be checked by comparing the error with 3%. We have to do this process
until found the logical result. On the contrary, saving the function. We divide the
ambient temperature into a variety of gap, from 23 to 34, increase one degree for one
step, so we have an approximating function corresponding with each Ta. So, the final
result quite exactly with the error is smaller than 3%, this rate is the optimal value that
we chose from the experimental process. Our final product is shown in Fig 6.

Fig. 5. Operating system Algorithm.

BME7 in Vietnam, IFMBE Proceedings 2018

 153

a. The front side of package.

b. The back side of package with small tube

around the IR sensor.
Fig. 6. The final package.

Body temperature measurement by the infrared sensor (IR sensor) depends on factors
such as angle of measurement, vibration, ambient light, ambient temperature. Thus, the
result we get is not accurate. In order to have good results, we have to reduce the noise,
we made a small tube at the head of infrared sensor to allow transfer directly to the
infrared sensor. This tube is presented in Fig 3b. Next, we remove noise caused by
environmental distance and t
ambient temperature ranges.
r T

(4)

Tr (o C) is the body temperature measured with a mercury thermometer.

T (o C) is the temperature measured by the LMX90614 sensor.
L (cm) is the distance from the device to the object defined by the Ultrasonic.
The ambient temperature is determined by LM35.
Table 1 shows the temperature of the hand measured by 2 devices at an ambient
temperature of 28.36°C, where L is the distance from the device to the object, T is the
temperature obtained from IR sensor MLX90614, Tr is the temperature measured from
the mercury thermometer.

BME7 in Vietnam, IFMBE Proceedings 2018

 154

Table 5. The temperature of the hand is measured by IR sensor at 28.36oC (with noise).
1 L (cm) 2 T 3 Tr 4 Error

10.13 29.85 35.8 16.62011173

20 29.19 35.8 18.46368715
30.03 28.91 35.8 19.24581006
39.97 28.63 35.8 20.02793296
50.03 28.63 35.8 20.02793296
60.1 28.51 35.8 20.36312849
70.19 28.51 35.8 20.36312849
80 28.57 35.8 20.19553073
89.53 28.59 35.8 20.1396648
99.78 28.53 35.8 20.30726257
110.61 28.53 35.8 20.30726257
120.37 28.59 35.8 20.1396648

Average error 19.68342644

Before reducing noise, the result is far from with the reality, which the average is
approximately 19.7%. So, we calculate the function by the distance and environmental

ambient temperature.

Table 2 Ta).

5 Ta (o C) 6

23 24 -6E-05L5 + 2E-07L4 2E-05L3 + 0.0005 L2 -0.0127L+5.824

24 25 -3E-10L5 + 9E-08L4 7E-06L3 -0.0002L2 + 0.0359L+ 5.6334

25 26 2E-06L3 0.0006L2 + 0.0478L + 3.8664

26 27 -2E-08L4 + 7E-06L3 -0.0009 L2 +0.0512L+3.8116

27 28 -3E-09L5 + 9E-07L4 9E-05L3 + 0.0038 L2 -0.0279L+6.1046

28 29 -3E-08 L 4 + 1E-05 L 3 - 0.0012 L 2 + 0.0796 L + 4.9914

29 - 30.5 -4E-08 L 4 + 1E-05 L 3 - 0.0014 L 2 + 0.081 L + 4.8359

30.5 32 -3E-08 L 4 + 9E-06 L 3 - 0.001 L 2 + 0.0531 L + 3.7695

32 34 1E-07L4 + 3E-06L3 - 0.002L2 + 0.1423L + 1.2785

BME7 in Vietnam, IFMBE Proceedings 2018

 155

Fig. 7. Operation of device

Fig 7 that is the flow chart, represents the operation of the device. At first, we measure
the ambient temperature , if 23 < Ta < 34, we continue to measure the distance L and
receive the temperature from the IR sensor which named T. Then, calculating T r =
fTa
presenting to the LCD. If not, we print into the LCD screen and finish
this process.

3 Experimental Results

Fig 8 shows that the hand temperature measured by this device is 35oC, the result is not
much different from the actual temperature. Ta is the environmental temperature, T is
the body temperature from our product and L is the distance from the user to the
equipment .

Fig. 8. Temperature measuring device.

BME7 in Vietnam, IFMBE Proceedings 2018

 156

Table 3 shows the results after eliminating the noise. The correct operating range of
device is at an ambient temperature from 23oC to 28oC, an error rate from 1%-3%. IN
this table, the error 1 is the error of REFCO LP-79 which is one of the IR thermometer
in the market and the error 2 is the error of our product. So, in this case, the result is
precise with the small error-approximately 0.9%.

Table 3. Experimental result after noise cancellation.

Electronic 9 RE
7 T 8 L( 11 De
thermometer FCO 10 Error 1 12 Error 2
a cm) vice
LP-79
23.85 80.46 32 31 3.125 31.27 2.335
24.17 68.59 32 29 9.375 31.92 0.251
25.89 60.34 32 31 3.125 31.68 1.01
26.43 56.83 32 32 0 31.8 0.629
27.61 64.07 36 32 11.111 36.1 0.277
28.36 68.68 36.1 32 11.357 35.88 0.613
28.47 97.83 36.1 32 11.357 36.48 1.042
28.57 36.89 36.1 32 11.357 35.63 1.319
Average 0.934
Average error 7.601
error

4 Discussion

4.1 Advantage
Non-contact temperature measuring methods are used to overcome the limitations of
mercury and electronic thermometers. They could avoid cross-contamination because
it is non-contact with the object and also have shorter measuring time. In this product,
we concentrate on the noise reduce and bring the convenience to the user. Our product
obtains certain enhancements such as, it can measure in longer distance with limiting
the effect of some environmental factors (ambient temperature and distance) to the
result from the sensors, and its cost is lower than the Infrared thermometer on the
market.
4.2 Disadvantage

used because of its shortage in Vietnam market, the range of temperature is still small.

BME7 in Vietnam, IFMBE Proceedings 2018

 157

5 Conclusion And Future Work

Nowadays, there are some better IR sensors such as the medical version and new
thermometer sensors which the fresh technique, thus, the results will be more accurate.
Moreover, the precision can be improved by increasing the number of sampling times.
With industrial scale, we can reduce the size as well as the price of this product.
Conflict of Interest
The authors declare that they have no conflict of interest.

References
1. John Hopkins Medicine, https://www.hopkinsmedicine.org/healthylibrary/vital_signs,
last access 15/5/2018.
2. -Contact Thermometers for Detecting Fever: A
Ottawa, Canadian Agency for Drugs and Technologies in
Health, 2014 Nov.
3.
J Med Eng Technol (May 2010).
4. Ho- -Contact Body Temperature Measurement
-China (2016)
5. MLX90614 Datasheet, p.1, p.2, p.26, http://www.datasheetlib.com/datasheet, last
access 15/5/2018.
6. - http://centmesh.csc.ncsu.edu, Nov. 16
2014, last access: 15/5/2018.
7.
Instrument, SNIS 159H-August 1999, revised December 2017.
8. LCD 16x2 datashee -NSW-
FBS/3.3V, Xiamen Amotec Display co, LTD, 29/10/2008.
9. http://avrchip.com/arduino-
nano-datasheet-and-tutorial/, last access: 15/5/2018.
10.
System, last update:28/2/2014, last access 15/5/2018.

BME7 in Vietnam, IFMBE Proceedings 2018