Instrumentation & Measurement Magazine 25-9 - 51

evaluate and judge the quality status of pepper powder. The
test results showed that the recognition accuracy reached to
over 98% for all types of the pepper samples with RBF-SVM
model.
Moreover, another two machine leaning models including
decision tree and random forest tree were compared with
SVM model to recognize the adulterations in pepper samples.
The test results showed that the pure pepper could be distinguished
from the adulteration samples successfully with the
accuracy of near 100% using machine leaning models. The
type of adulterations in the pepper powders could also be recognized
with high accuracy although corn flour and rice bran
were identified with lower accuracy than rosin and wheat
bran.
The comparison results of SVM, DT and RFT indicated
that the recognition performance of random forest was better
than that of decision tree algorithm in the classification of
pepper powder adulteration. However, SVM has the highest
recognition accuracy both for specific adulteration and total
evaluation among three algorithms discussed in the research.
Finally, the results from other more machine learning models
using Machine Learning Tool of MATLAB proved that the
SVM performed the best among different machine learning
classification algorithms to identify the adulterations in pepper
samples.
Accordingly, it is feasible to recognize the adulterations in
pepper powders with high accuracy with AOS. Further, the
method discussed in this research can also be applied to identify
the quality status of condiments as well as safety detection
for other food materials.
However, the data source of pure pepper and adulterated
species are still not enough, and the quantitative detection for
the specific VOCs released from adulterated materials are not
available now with AOS. Hence, more sample data captured
with our AOS and an intelligent algorithm to identify the kind
and concentration of the VOCs in different quality status for
the pepper powders will be our future work.
Acknowledgment
Financial support was provided by the National Natural
Science Foundation of China (21808181, 61903164), China Postdoctoral
Science Foundation (2019M653651,2021T140544),
Basic research project of natural science in Shaanxi province
(2020JM-021).
References
[1] L. Zhang, Y. Liu and P. Deng, " Odor recognition in multiple
e-nose systems with cross-domain discriminative subspace
learning, " IEEE Trans. Instrum. Meas., vol. 66, no. 7, pp. 1679-1692,
Jul. 2017.
[2] W. Zhang, L. Wang, J. Chen, W. Xiao and X. Bi, " A novel gas
recognition and concentration detection algorithm for artificial
olfaction, " IEEE Trans. Instrum. Meas., vol. 70, pp. 1-14, 2021, Art
no. 2509514.
[3] C. Krutzler, A. Unger, H. Marhold, T. Fricke, T. Conrad and A.
Schütze, " Influence of MOS gas-sensor production tolerances on
December 2022
pattern recognition techniques in electronic noses, " IEEE Trans.
Instrum. Meas., vol. 61, no. 1, pp. 276-283, Jan. 2012.
[4] A. Ren et al., " Machine learning driven approach towards the
quality assessment of fresh fruits using non-invasive sensing, "
IEEE Sensors J., vol. 20, no. 4, pp. 2075-2083, Feb. 2020.
[5] M. Pardo and G. Sberveglieri, " Coffee analysis with an electronic
nose, " IEEE Trans. Instrum. Meas., vol. 51, no. 6, pp. 1334-1339,
Dec. 2002.
[6] P. K. Kundu, A. Chatterjee and P. C. Panchariya, " Electronic
tongue system for water sample authentication: a slantlettransform-based
approach, " IEEE Trans. Instrum. Meas., vol. 60,
no. 6, pp. 1959-1966, Jun. 2011.
[7] M. Anly Antony and R. S. Kumar, " A comparative study on
predicting food quality using machine learning techniques, " in
Proc. 2021 7th Int. Conf. on Advanced Computing and Communication
Syst. (ICACCS), pp. 1771-1776, 2021.
[8] S. S. Wang, " Study on volatile compound and aroma
characteristics of Chinese red pepper, " Southwest Jiaotong
University Master Degree Thesis, Chengdu, China, 2016.
[9] C. Gao, " Composition analysis of the aroma volatile compounds
in different stabilized rice bran during storage, " Jiangnan
University Master Degree Thesis, Jiangsu, China, 2021.
[10] L. J. Cui, " Research on volatiles in corn and the odor fingerprint
technology applied to detection, " Henan University of
Technology Master Degree Thesis, Zhengzhou, China, 2011.
Denglong Ma (denglong.ma@xjtu.edu.cn) is an Associate Professor
at the School of Mechanical Engineering, Xi'an Jiaotong
University in Xi'an, China. He received his Ph.D. degree in
power engineering and engineering thermophysics from Xi'an
Jiaotong University in 2014. His current research interests include
artificial olfactory system, process equipment safety and
food safety detection.
Chang Liu is pursuing a master degree in mechanical engineering
at Xi'an Jiaotong University in Xi'an, China. He
received the bachelor degree from Dalian University of Technology,
Dalian, China in 2019. His research interest is artificial
olfactory systems.
Fangjun Wu is an Engineer in Changsha, China. He received
the bachelor degree in mechanical engineering from Xi'an Jiaotong
University in Xi'an, China in 2020. His research interest
is detection instruments.
Zekang Li is pursuing a master degree in mechanical engineering
at Xi'an Jiaotong University in Xi'an, China. His research
interest is artificial olfactory systems. He received the bachelor
degree from Northwest Agriculture and Forestry University,
in Xi'an, China in 2020.
Xiuben Wu is pursuing a master degree in mechanical engineering
at Xi'an Jiaotong University in Xi'an, China. His
research interests include artificial olfactory systems. He
received the bachelor degree from Tongji University, in Shanghai,
China in 2020.
IEEE Instrumentation & Measurement Magazine
51
Instrumentation & Measurement Magazine 25-9

Table of Contents for the Digital Edition of Instrumentation & Measurement Magazine 25-9
