Sensors: Design and Development of A Low Cost, Non-Contact Infrared Thermometer With Range Compensation
Sensors: Design and Development of A Low Cost, Non-Contact Infrared Thermometer With Range Compensation
Sensors: Design and Development of A Low Cost, Non-Contact Infrared Thermometer With Range Compensation
Article
Design and Development of a Low Cost, Non-Contact Infrared
Thermometer with Range Compensation
Nicholas Wei-Jie Goh 1 , Jun-Jie Poh 1 , Joshua Yi Yeo 1 , Benjamin Jun-Jie Aw 1 , Szu Cheng Lai 2 ,
Jayce Jian Wei Cheng 2 , Christina Yuan Ling Tan 2 and Samuel Ken-En Gan 1,3, *
1 Antibody & Product Development Lab, EDDC, Agency for Science, Technology and Research (A*STAR),
Singapore 138672, Singapore; nicholas_goh@alumni.sutd.edu.sg (N.W.-J.G.); anson_12@hotmail.com (J.-J.P.);
joshua_yeo@eddc.a-star.edu.sg (J.Y.Y.); benjaminaw021@gmail.com (B.J.-J.A.)
2 Institute of Materials Research and Engineering, A*STAR, Singapore 138634, Singapore;
sc-lai@imre.a-star.edu.sg (S.C.L.); jayce_cheng@imre.a-star.edu.sg (J.J.W.C.);
yl-tan@imre.a-star.edu.sg (C.Y.L.T.)
3 Department of Psychology, James Cook University, Singapore 387380, Singapore
* Correspondence: samuel_gan@eddc.a-star.edu.sg; Tel.: +65-6407-0584
Abstract: Fever is a common symptom of many infections, e.g., in the ongoing COVID-19 pandemic,
keeping monitoring devices such as thermometers in constant demand. Recent technological ad-
vancements have made infrared (IR) thermometers the choice for contactless screening of multiple
individuals. Yet, even so, the measurement accuracy of such thermometers is affected by many factors
including the distance from the volunteers’ forehead, impurities (such as sweat), and the location
measured on the volunteers’ forehead. To overcome these factors, we describe the assembly of an
Arduino-based digital IR thermometer with distance correction using the MLX90614 IR thermometer
and HC-SR04 ultrasonic sensors. Coupled with some analysis of these factors, we also found ways to
Citation: Goh, N.W.-J.; Poh, J.-J.; Yeo,
programme compensation methods for the final assembled digital IR thermometer to provide more
J.Y.; Aw, B.J.-J.; Lai, S.C.; Cheng, J.J.W.;
accurate readings and measurements.
Tan, C.Y.L.; Gan, S.K.-E. Design and
Development of a Low Cost,
Keywords: infrared; Arduino; nano; contactless thermometer; HC-SR04; MLX90614
Non-Contact Infrared Thermometer
with Range Compensation. Sensors
2021, 21, 3817. https://doi.org/
10.3390/s21113817
1. Introduction
Academic Editor: Stefano Sfarra Early methods of measuring body core temperature utilizing contact mercury ther-
mometers are replaced by the safer and more convenient electronic thermometers at the
Received: 6 May 2021 sublingual, armpits, ear canals, and in some rare occasions, the rectum and axillary for
Accepted: 28 May 2021 accuracy [1]. Many of these surface measurement sites, specifically the temporal and cen-
Published: 31 May 2021 tral forehead, reflect lower readings than internal sites such as the tympanic temperature
readings, the current gold standard to represent the body core temperature [2], especially
Publisher’s Note: MDPI stays neutral given the impracticality of rectal/anal temperature takings.
with regard to jurisdictional claims in While screening for disease in the ongoing pandemic, rapid temperature measure-
published maps and institutional affil-
ments of many individuals quickly and safely without allowing the thermometer to be a
iations.
vector of pathogen transfer are crucial, thus making contact infeasible, ruling out many of
the above measurement sites.
Infrared (IR) thermometers can fulfil this gap by measuring the surface temperature
without direct contact, which is through detecting the amount of thermal or black-body
Copyright: © 2021 by the authors. radiation emitted by the object. Additionally, these thermometers are now commonly used
Licensee MDPI, Basel, Switzerland. in clinical practices [3], as well as routinely during the pandemic for self-monitoring and
This article is an open access article screening at the entrances of public places.
distributed under the terms and
Typically, IR thermometer casings are manufactured by the expensive injection mold-
conditions of the Creative Commons
ing (due to mold production and tooling costs), producing significant waste material.
Attribution (CC BY) license (https://
The increased adoption of three-dimensional (3D) printing technology has revolutionized
creativecommons.org/licenses/by/
prototyping and reduced cost by shortening the lead time to manufacture with significantly
4.0/).
less waste. Coupling with 3D printing, the use of ‘off-the-shelf’ microcontroller kits such
as Arduino, Raspberry Pi, and Micro: bit can now allow novel electronic products to be
cost-effectively assembled, even by non-engineers without specialized equipment. It is
with such enabling technology that even home-made measurement devices can be made
easily, e.g., spectrophotometers [4,5], including those for psychology research use [6].
While thermometers can be easily assembled, IR thermometers are often thought to be
less reliable [7] than traditional contact thermometers. Non-contract infrared thermometers
were previously reported to have a sensitivity between 4.0% to 89.6%, specificity between
75.4% to 99.6%, and a positive predictive value between 0.9% to 76.0% [8]. In fact, there
are recommendations for its repeated measurements at hospital gantries [9], given that
IR thermometers are highly prone to external interferences by surrounding temperatures,
relative humidity [10], the site of measurement [11], and the presence of oil (sebum) and
sweat on the forehead [12], as well as other factors in the immediate environment [13].
Apart from these innate factors, intrinsic human physiological factors such as fever [14]
or exercise [15] can produce sweat to affect the measurements. With further confounding
influence from the surrounding temperature and humidity that affect blood supply to the
skin surface, which by generally lower than the expected body temperature [16,17], many
IR thermometers, especially self-assembled ones can be inaccurate as they lack ambient
temperature and distance sensors [7] for compensation.
At the point of writing, many IR sensors have in-built radiation emitter and receiver
devices [18] and can be used to provide reliable measurements at predetermined distances.
Yet, the onus is still on the user to operate them correctly for accurate measurements.
To help alleviate the above problems, we describe the design and assembly of a low
cost IR thermometer with distance and environmental temperature sensing capabilities to
provide more accurate measurements. Experiments were conducted to validate compensa-
tion adjustments made in the algorithm, as well as the effects of measurements on different
locations of the forehead.
Figure
Figure1.1. (a) Back
(a)Back
1.(a) view
viewofof
Backview prototype
ofprototype segmented
prototypesegmented into
segmentedinto the
intothe top
thetop (yellow)
top(yellow) and
(yellow)and bottom
andbottom (black)
bottom(black) half
(black)half with
halfwith color
withcolor differentiation
colordifferentiation for
differentiationfor
for
Figure
display purposes. (b) Front view of prototype displaying the IR sensor and ultrasound sensor in the top half. The bottom
displaypurposes.
display purposes.(b) (b)Front
Frontview
viewofofprototype
prototypedisplaying
displayingthe theIRIRsensor
sensorand
andultrasound
ultrasoundsensor
sensorininthe
thetop
tophalf.
half.The
Thebottom
bottom
half of the device includes the push button switch and battery holder.
half
halfofofthe
thedevice
deviceincludes
includesthethepush
pushbutton
buttonswitch
switchand
andbattery
batteryholder.
holder.
2.2.2. Circuitry Design and Implementation
2.2.2.
2.2.2.Circuitry
CircuitryDesign
Designand
andImplementation
Implementation
The Arduino Nano [23] microcontroller, in which the copper wires from the sensor
The
The Arduino Nano [23]
[23]microcontroller, in which the copper wires from the
thesensor
pins of theArduino Nano
ultrasonic sensor, IRmicrocontroller,
sensor, OLED, in
andwhich
pushthe copper
button wires
switch werefrom sensor
soldered, are
pins
pins ofofthe ultrasonic
the ultrasonicsensor,
sensor, IRIRsensor,
sensor,OLED,
OLED, and
and push
push button
button switch
switch were
were soldered,
soldered, are
are
shown in Figure 2, with a 9 V battery supplying 3.3 to 5 V to the circuit (Figure 2).
shown
shownininFigure
Figure2,2,with
witha a9 9VVbattery
batterysupplying
supplying3.3
3.3toto5 5VVtotothe
thecircuit
circuit(Figure
(Figure2).2).
Figure 2. A circuit schematic of the assembled digital IR thermometer consisting of: Ultrasonic sensor HC-SR04, push
Figure
Figure2.switch,
2.AAcircuit
circuitschematic
schematicofofthe
theassembled
assembleddigital
digitalIR
IRthermometer
thermometer consisting of:
of:Ultrasonic
consistingsensor
Ultrasonicsensor
sensorHC-SR04,
HC-SR04,push
push
button Arduino Nano board, OLED display SSD1306, and IR temperature MLX90614.
button switch, Arduino Nano board, OLED display SSD1306, and IR temperature sensor MLX90614.
button switch, Arduino Nano board, OLED display SSD1306, and IR temperature sensor MLX90614.
Figure3.3.OLED
Figure OLEDdisplay
displayillustrating
illustrating
thethe distance
distance to the
to the target
target surface,
surface, measured
measured surface
surface tempera-
temperature,
ture, and the measurement status.
and the measurement status.
2.4. The
Sensor Calculations adjustment equation was derived from the regression analysis
compensation
of the relationship
2.4.1. between the oral temperature values and the measured values. Com-
Infrared Sensor
pensatedThe equations the
values by andIR thermometer
calculations were
of the usedtemperature
target as the dependent value,
are based on awhile the
previous
measurement values by the IR thermometer and ambient temperature were each viewed
work by others [24]. For precise measurement of the absolute temperature of the target
as
(Tindependent variables.
k), the device temperature (Tdev) should be kept small with a stable ambient temperature
(Tamb). To compensate for the proximity effects, a distance-to-spot ratio (D/S ratio) is built
2.4. Sensor Calculations
in the algorithm. Specifically, the area measured increases as the distance increases. The
2.4.1. Infrared Sensor
selected IR temperature sensor has a field of view of 80 degrees. This translates to a D/S
ratioThe equations
of 1:1.68. andthe
Setting calculations of the of
average height target temperature
a human forehead areofbased on aas
58.3 mm previous work
the constraint
by
[25], the maximum horizontal distance that the IR temperature sensor can reliably ),meas-
others [24]. For precise measurement of the absolute temperature of the target (T k the
device temperature
ure the temperature of (T ) should be kept small with a stable ambient temperature
devthe target, is approximately 4 cm. Beyond this distance, flanking (T amb ).
To compensate for the proximity effects, a distance-to-spot
areas of the forehead would also be measured, affecting the accuracy. ratio (D/S ratio) is built in the
algorithm. Specifically, the area measured increases as the distance increases. The selected
IR temperature
2.4.2. Ultrasound sensor has a field of view of 80 degrees. This translates to a D/S ratio of
Sensor
1:1.68. Setting the average height of a human forehead of 58.3 mm as the constraint [25],
The ultrasound
the maximum sensor
horizontal consists
distance thatofthea IR
transmitter
temperaturesending
sensorancan
ultrasound wave and
reliably measure thea
receiver detecting the reflected wave by the targeted physical object. The
temperature of the target, is approximately 4 cm. Beyond this distance, flanking areas of time taken be-
tween the transmission and detected wave is registered
the forehead would also be measured, affecting the accuracy. for the calculation of the distance
from the speed of ultrasound waves at 330 m/s by the Arduino Nano.
2.4.2. Ultrasound Sensor
The ultrasound sensor consists of a transmitter sending an ultrasound wave and
a receiver detecting the reflected wave by the targeted physical object. The time taken
between the transmission and detected wave is registered for the calculation of the distance
from the speed of ultrasound waves at 330 m/s by the Arduino Nano.
Three experiments were performed for this study. Each experiment was conducted in
an air-conditioned environment of 20 to 22 ◦ C for a relatively constant Tamb . Volunteers
were asked to avoid high intensity activities before the experiments, and temperature
measurements were taken 5 min after acclimatization to the test environment. A sterilized
Omron digital thermometer MC-343F was used to measure the oral temperatures, which
was also set as the target point for compensation adjustments. Temperature measurements
of the volunteers’ forehead were taken in triplicate experiments.
In the first experiment, the IR thermometer was used to measure the volunteers’
forehead temperatures at three different locations (left, right, and center of forehead) at
varying distances from the forehead (2–4 cm, at 0.5 cm intervals). The range of 2–4 cm is
based on the minimum and maximum allowable horizontal distance that the IR temperature
sensor can accurately measure (2–4 cm) as per the manufacturer’s recommendations. In
the second experiment, water was sprayed on the volunteers’ forehead using a spray bottle
to simulate wetness from sweat.
The IR and oral temperature readings from the first experiment set were used for
compensation adjustments and stored. Then, the calibrated IR thermometer was used to
measure the center and lateral forehead temperatures of volunteers and test its performance
and deviation from the oral temperature. Data from each experiment were plotted with the
distance against the temperature.
3. Results
3.1. Performance of IR Thermomter on the Forehead
Prior to the compensation adjustments of the IR thermometer, an inverse correlation
between the temperature readout and distance from the forehead was measured for all the
five volunteers at all three forehead locations (left, right, and center, see Figure 4).
The variation of the mean temperature measurements between the center and lateral
areas of the forehead were calculated to be between −0.69 to 0.55 ◦ C (Table 1). The center
is taken as the most accurate region for compensation, in view of the stronger correlative
relationship between the recorded temperatures and distance.
ors 2021, 21, 3817 sensor used in this experiment is shown to have high reproducibility for
6 ofmeasuring
13 the
temperature (based on the IR temperature sensor) and distance.
Figure 4. Measured
Figureforehead temperatures
4. Measured forehead against the distance
temperatures against between 2–4 cm
the distance for each2–4
between volunteer taken
cm for each on the (a) left of
volunteer
the forehead, (b) right
taken on of
thethe
(a)forehead,
left of theand (c) center
forehead, of theofforehead
(b) right prior and
the forehead, to the
(c)compensation adjustments.
center of the forehead prior Error
to bars
the compensation
represent the standard deviation adjustments.
from triplicateError bars represent the standard deviation from triplicate experi-
experiments.
ments.
Table 1. Variations in the forehead temperature measured from the center and lateral positions of the forehead of five
The
volunteers using our variation of prior
IR thermometer the mean
to thetemperature
compensationmeasurements
adjustments. between the center and lateral
areas of the forehead were calculated to be between −0.69 to 0.55 °C (Table 1). The center
Volunteer isTemperature
taken as theDifference between
most accurate the Center
region of the Temperature
for compensation, in view ofDifference between
the stronger the Center of the
correlative
Forehead and Left Forehead ( ◦ C) Forehead and Right Forehead (◦ C)
relationship between the recorded temperatures and distance.
1 0.13 ± 0.44 0.31 ± 0.34
2 in the forehead temperature−measured
Table 1. Variations 0.16 ± 0.35from the center and lateral positions of the forehead
−0.51 ± 0.31
of five
volunteers using3our IR thermometer prior to the 0.04 ± 0.27
compensation adjustments. −0.69 ± 0.27
4 −0.51 ± 0.34 0.00 ± 0.56
Temperature
5 Difference between
−0.03 ± the Center of Temperature Difference between
0.23 the
0.55 ± Center of
0.40
Volunteer
the Forehead and Left Forehead (°C) the Forehead and Right Forehead (°C)
1 ± 0.44
0.133.4. ± 0.34
0.31 System
Test Performance of the Optimized IR Temperature
2 −0.16 ± 0.35 −0.51 ± 0.31
The performance of our implemented control system with distance-sensing capabilities
3 0.04 ± 0.27 −0.69 ± 0.27
was tested factoring the gradient into the algorithm. For compensation, a fixed arbitrary
4 ± 0.34 was added to the raw temperature outputs,
−0.51number 0.00 ± 0.56
given that this would differ across
5 ± 0.23sources of manufactured IR sensors. 0.55 ± 0.40
−0.03various
Table 2. Difference between the dry and wet forehead temperatures on the respective locations from
2–4 cm. The positive value implies that the dry forehead registers a higher temperature.
Figure
Figure5.
5.Combined
Combinedtrend
trendlines
linesfor
forall
allthe
thevolunteers
volunteers between
between 2–3
2–3 cm
cm (in
(in black)
black) and
and between
between 3–4
3–4 cm
cm (in
(in red).
red).
Following
3.4. Test compensation
Performance adjustments,
of the Optimized experiments
IR Temperature were repeated with the compensa-
System
tion applied to the measurements. We found only the left forehead of Volunteer 4 to have
The performance of our implemented control system with distance-sensing capabil-
significant differences (an average of 1.36 ◦ C, p < 0.001) from the oral temperature using
ities was tested factoring the gradient into the algorithm. For compensation, a fixed arbi-
trary number was added to the raw temperature outputs, given that this would differ
across various sources of manufactured IR sensors.
Following compensation adjustments, experiments were repeated with the compen-
sation applied to the measurements. We found only the left forehead of Volunteer 4 to
have significant differences (an average of 1.36 °C, p < 0.001) from the oral temperature
Sensors 2021, 21, 3817 8 of 12
Dunnett’s post-hoc test. Across the five volunteers, the mean temperature measured at
the center of the forehead (over the left and right) is closer to the mean oral temperature
(Figure S1). Thus, subsequent measurements were performed based on the center of the
forehead temperatures.
Based on the ANOVA results, there was no statistically significant difference in the
temperature measurement from the center or lateral area of the forehead against the
oral temperature (Figure S1). However, Dunnett’s post-hoc test showed that the p-value
differences between the measurements from the center of the forehead to the control was
Sensors 2021, 21, 3817 higher than that of the measurement of lateral areas to the control across all volunteers.
9 of 13
Between 2–4 cm, the measured temperatures were within a range of ±0.29◦ C of their
oral temperature (Figure 6). Despite the presence of outliers, the recorded temperature
displayed a standard deviation of ±0.3 ◦ C (Volunteer 3) from the oral temperature.
Figure
Figure 6.
6.Recorded
Recordedtemperatures
temperatures and
andthethe
oral temperature
oral temperaturefor for
each volunteer.
each For each
volunteer. volunteer,
For each the dots
volunteer, the represent indi-
dots represent
vidual datadata
individual points in the
points inrespective distance
the respective fromfrom
distance the foreheads.
the foreheads.
4.
4. Discussion
Discussion
We
We set
setout
outtoto
assemble
assemble andandcalibrate an Arduino-based
calibrate an Arduino-based thermometer
thermometer capable of com-
capable of
pensating for varying measurement distances from the forehead
compensating for varying measurement distances from the forehead for a more accuratefor a more accurate con-
tactless body
contactless temperature
body temperaturemeasurement.
measurement.
With
With thethe availability
availability of
of off-the-shelf
off-the-shelf electronic
electronic sensors
sensors andand microcontroller
microcontroller kits, kits, itit is
is
possible
possible for
for non-engineers
non-engineers to to assemble
assemble their
their own
own devices
devices toto meet
meet times
times ofof shortages
shortages suchsuch
as
as that
that experienced
experienced by by the
the authors
authors when when self-monitoring
self-monitoring measures
measures were
were implemented
implemented
during the COVID-19 pandemic in 2020. While it will
during the COVID-19 pandemic in 2020. While it will take a significanttake a significant timetime
for such self-
for such
assembled thermometers to be approved by regulatory bodies and
self-assembled thermometers to be approved by regulatory bodies and pass the required pass the required cal-
ibrations
calibrationsas required forfor
as required commercial
commercial devices,
devices, atattimes
timesofofextended
extendedshortages,
shortages,suchsuchself-
self-
assembled
assembleddevices
devicesallow
allowan animmediate
immediatepatch patchsolution
solutionalso alsoknown
knownin inIndia
Indiaas as“Jugaad”.
“Jugaad”.
Yet,
Yet, the
the potential
potentialproblem
problemarising
arisingfrom frominaccurate
inaccurate measurements
measurements cancanhave dire
have effects
dire effects in
infection
in infectioncontrol measures
control measureswithwith
falsefalse
negatives, especially
negatives, givengiven
especially that the
thatforehead tem-
the forehead
perature
temperaturerarely goes goes
rarely aboveabove
35oC.35 In ◦recognition of suchofeffects,
C. In recognition we have
such effects, wedecided to release
have decided to
our results
release ourfor the community
results to tweaktotheir
for the community tweakown assembled
their thermometers.
own assembled thermometers.
4.1. Assembly
4.1. Assemblyofofthe
theArduino-Based
Arduino-BasedThermometer
Thermometer
We adopted
We adopted Arduino
Arduino for
for its
its ease
ease of
of use
use and
and variety
variety of
of sizes
sizes for
for assembly. While
assembly. While
Scratch programming with Micro:bit [27] may be simpler in terms of programming,
Scratch programming with Micro:bit [27] may be simpler in terms of programming, its its
larger size and pin connectivity pose a problem for small handheld thermometers.
larger size and pin connectivity pose a problem for small handheld thermometers. Size is Size
also a concern for the more powerful alternative Raspberry Pi [28]. To balance these con-
siderations, we adapted an online DIY thermometer assembly guide, using Arduino Nano
to connect the available devices.
To reduce the measurement error caused by improper targeting at the forehead, we
programmed the thermometer to take five readings and display the average. Similarly,
Sensors 2021, 21, 3817 9 of 12
is also a concern for the more powerful alternative Raspberry Pi [28]. To balance these
considerations, we adapted an online DIY thermometer assembly guide, using Arduino
Nano to connect the available devices.
To reduce the measurement error caused by improper targeting at the forehead, we
programmed the thermometer to take five readings and display the average. Similarly,
the sensors allow for ambient temperature sensing, that together with our added ultra-
sound module for distance measurements, allow for compensations to be applied once the
effects of some common parameters such as forehead location, wetness, and distance of
measurement, are established.
4.2. Differences in Temperature between the Center and Lateral Areas of the Forehead
We first sought to investigate if there were differences on the surface temperature
between the center and the lateral areas of the forehead. Differences between the center
and lateral areas of the forehead were observed and persisted even after compensation
adjustments (Figure S1). This difference in mean temperature is also attributed to the
physiological phenomenon explained by the individual’s unique forehead temperature
distribution as determined through the simulation software [29]. It should be noted that
the IR sensor in our assembled thermometer is commonly used to measure the surface
temperature of the skin and is unable to determine the muscle temperature near the
superficial temporal artery, a major artery located beneath the vascular bed of the head’s
skin [30]. While the differences between the forehead locations are small and perhaps
inconsequential in differentiating high fevers above 38 ◦ C, our studies show that using
the IR thermometer for the center of the forehead is the most accurate and consistent.
Given that this is a specific forehead locale variable, compensation mechanisms are not
easily programmed into the device so the onus of accuracy falls upon the correct usage by
the user.
5. Conclusions
Forehead temperature measurements using an IR thermometer play an important role
of rapidly screening for fever to identify the infected individual. The performance and
precision of an IR thermometer for forehead temperature screening were studied together
with the design and implementation of an improved infrared temperature sensor-based
system with distance sensing capabilities.
While minimal, temperature differences between the center and lateral areas of the
forehead highlight the importance of the user in targeting the right area, which is the center
of the forehead.
Additionally, we showed that perspiration and water on the forehead can cause a
significant decrease in the detected temperature, but were unable to make programmed
compensations. Therefore, it is necessary for the user to take precaution in ensuring that the
forehead is dry and the skin surface temperature is restored before accurate measurements
can be taken.
After implementation of a range sensor, the preliminary results show that our IR
thermometer can achieve a more accurate forehead temperature measurement over a short
distance range. This is achieved by obtaining an optimized algorithm from experimentally
plotted linear regression lines. Moreover, it is observed that the measured temperatures
were well within the ±0.29 ◦ C variation of their oral temperature over the distance of
2–4 cm, achieving a similar performance to commercial thermometers. Therefore, it could
be concluded that the designed control system has a high validity to measure human
forehead temperatures within the compensated range.
Acknowledgments: The authors thank Anthony Chong for help in coordination, and SKEG thanks
Ebenezer for behind the scenes help in this work.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Fulbrook, P. Core temperature measurement: A comparison of rectal, axillary and pulmonary artery blood temperature. Intensive
Crit. Care Nurs. 1993, 9, 217–225. [CrossRef]
2. Mogensen, C.B.; Wittenhoff, L.; Fruerhøj, G.; Hansen, S. Forehead or ear temperature measurement cannot replace rectal
measurements, except for screening purposes. BMC Pediatrics 2018, 18, 15. [CrossRef]
3. Quast, S.; Kimberger, O. The Significance of Core Temperature—Pathophysiology and Measurement Methods; Dräger Medical GmbH:
Stuttgart, Germany, 2014.
4. Poh, J.-J.; Wu, W.-L.; Goh, N.W.-J.; Tan, S.M.-X.; Gan, S.K.-E. Spectrophotometer on-the-go: The development of a 2-in-1 UV–Vis
portable Arduino-based spectrophotometer. Sens. Actuators A Phys. 2021, 325, 112698. [CrossRef]
5. Ng, K.M.; Wong, C.-F.; Liang, A.C.; Liew, Y.-H.; Liang, A.C.; Yeo, J.Y.; Lua, W.-H.; Qian, X.-J.; Gan, S.K.-E. Republication—APD
SpectBT: Arduino-based mobile vis-Spectrophotometer. Sci. Phone Apps Mob. Devices 2019, 5. [CrossRef]
6. Gan, S.K.-E.; Yeo, J.Y. Editorial: The promises of Microcontroller kits and Smartphone apps for Psychological research. Sci. Phone
Apps Mob. Devices 2020, 6. [CrossRef]
7. Yaffe-Bellany, D. Thermometer Guns’ on Coronavirus Front Lines are “Notoriously not Accurate”; The New York Times: New York, NY,
USA, 2020.
8. Bitar, D.; Goubar, A.; Desenclos, J.C. International travels and fever screening during epidemics: A literature review on the
effectiveness and potential use of non-contact infrared thermometers. Eurosurveillance 2009, 14, 19115. [CrossRef]
9. Hsiao, S.H.; Chen, T.C.; Chien, H.C.; Yang, C.J.; Chen, Y.H. Measurement of body temperature to prevent pandemic COVID-19 in
hospitals in Taiwan: Repeated measurement is necessary. J. Hosp. Infect. 2020, 105, 360–361. [CrossRef] [PubMed]
10. Hudoklin, D.; Drnovšek, J. The New LMK Primary Standard for Dew-Point Sensor Calibration: Evaluation of the High-Range
Saturator Efficiency. Int. J. Thermophys. 2008, 29, 1652–1659. [CrossRef]
11. Patel, N.; Smith, C.E.; Pinchak, A.C.; Hagen, J.F. Comparison of esophageal, tympanic, and forehead skin temperatures in adult
patients. J. Clin. Anesth. 1996, 8, 462–468. [CrossRef]
12. Teran, C.G.; Torrez-Llanos, J.; Teran-Miranda, T.E.; Balderrama, C.; Shah, N.S.; Villarroel, P. Clinical accuracy of a non-contact
infrared skin thermometer in paediatric practice. Child. Care Health Dev. 2012, 38, 471–476. [CrossRef] [PubMed]
13. Pušnik, I.; Miklavec, A. Dilemmas in Measurement of Human Body Temperature. Instrum. Sci. Technol. 2009, 37, 516–530.
[CrossRef]
14. Wang, K.; Gill, P.; Wolstenholme, J.; Price, C.P.; Heneghan, C.; Thompson, M.; Plüddemann, A. Non-contact infrared thermometers
for measuring temperature in children: Primary care diagnostic technology update. Br. J. Gen. Pract. 2014, 64, e681–e683.
[CrossRef]
15. Dean, W. Effect of Sweating. JAMA 1981, 246, 623. [CrossRef]
16. Ng, D.K.-K.; Chan, C.-H.; Chan, E.Y.-T.; Kwok, K.-L.; Chow, P.-Y.; Lau, W.-F.; Ho, J.C.-S. A brief report on the normal range of
forehead temperature as determined by noncontact, handheld, infrared thermometer. Am. J. Infect. Control. 2005, 33, 227–229.
[CrossRef] [PubMed]
17. Ariyaratnam, S.; Rood, J.P. Measurement of facial skin temperature. J. Dent. 1990, 18, 250–253. [CrossRef]
18. Chen, K.S.; Lin, Y.C.; Yang, H.O. Non-Contact Temperature-Measuring Device and the Method Thereof ; Google Patents: Mountain
Veiw, CA, USA, 2010.
19. ElecFreaks. Ultrasonic Ranging Module HC—SR04. Available online: https://cdn.sparkfun.com/datasheets/Sensors/Proximity/
HCSR04.pdf (accessed on 23 October 2020).
20. Melexis. Datasheet for MLX90614. Available online: https://components101.com/sites/default/files/component_datasheet/
MLX90614-Datasheet.pdf (accessed on 30 October 2020).
21. Systech, S. Advance Information 128 × 64 Dot Matrix OLED/PLED Segment/Common Driver with Controller. Available online:
https://cdn-shop.adafruit.com/datasheets/SSD1306.pdf (accessed on 30 October 2020).
22. Raj, A. Make a Non-Contact Infrared Thermometer with MLX90614 IR Temperature Sensor. Available online: https://circuitdigest.
com/microcontroller-projects/ir-thermometer-using-arduino-and-ir-temperature-sensor (accessed on 22 May 2021).
23. Arduino. Arduino Nano Datasheet. Available online: http://www.farnell.com/datasheets/1682238.pdf (accessed on
23 October 2020).
24. Chen, H.-Y.; Chen, A.; Chen, C. Investigation of the Impact of Infrared Sensors on Core Body Temperature Monitoring by
Comparing Measurement Sites. Sensors 2020, 20, 2885. [CrossRef]
25. Sirinturk, S.; Govsa, F.; Pinar, Y.; Ozer, M.A. Study of frontal hairline patterns for natural design and restoration. Surg. Radiol.
Anat. 2017, 39, 679–684. [CrossRef]
26. R Core Team. R: A Language and Environment for Statistical Computing; Version 3.6.2; R Foundation for Statistical Computing:
Vienna, Austria, 2019.
27. Micro: Bit Educational Foundation. BBC Micro:bit v2. Available online: https://tech.microbit.org/hardware/ (accessed on
23 October 2020).
Sensors 2021, 21, 3817 12 of 12