© 2021 JETIR May 2021, Volume 8, Issue 5 www.jetir.

org (ISSN-2349-5162)

QUALITY AND PESTICIDES DETECTION IN

FRUITS AND VEGETABLES
1
Aradhana B S, 1Aiswarya Raj, 1Praveena K G, 1Joshy M, 2Reshmi B S
1
UG Scholar, Department of Computer Science and Engineering,
2
Asst. Prof, Department of Computer Science and Engineering,
Dr. APJ Abdul Kalam Technological University, Kerala, India.

Abstract: The existence of pesticides in fruits and vegetables has been a growing worry in all over the world. This work tries to solve the
problem by customer side. The consumer should get to know whether the selected fruits and vegetables are safe to consume or not. The main
aim of this work is to detect the amounts of pesticides on fruits and vegetables by calculating its normalized difference vegetation index using
sensor. Here the work is done in two parts. The first part performs the identification of fruits and vegetables by using CNN. Here a set of image
features such as color, shape, texture is trained using CNN for the identification of fruits and vegetables. In module 2, detect pesticides in fruits
and vegetables by three ways calculating its NDVI, by using IR sensor and gas sensor then compare the result. The program in the Arduino
gives for the output display. The detection information will be displayed on the screen. And the output graph is plotted.

Keywords: Convolutional Neutral Network, Machine learning.

I. INTRODUCTION
A wide verity of pesticides is wont to increase the production, quality in agriculture sector. For human diet fruits and vegetables are
the important part since they dispense essential nutrients that are mandatory for many of the reaction occurrence within the body. The continual
uses of pesticides have results in many problems including the human health. so as to scale back the loss and maintain the standard of fruits and
vegetables harvest pesticides were used. Today, the utilization of pesticides has increased significantly due to the growing population and therefore
the demands this population. The high increase in community means more food must be produced and spreading of diseases is more like. Pesticide
use in commercial agriculture has led to a rise in farm productivity, so why the farmers use the pesticides. Several reports suggest that prime levels
of pesticides in food can cause the expansion of diseases like cancer, kidney and lung ailments. Children have developing organs, susceptible to
catching infections and diseases. Residual amount of those chemicals is present on the food when it gets to the grocery. The importance of food
quality has become a significant issue thanks to the widespread use of pesticides. once you consume a fruit, the pesticide residue goes into the body
as they grind to a halt and may cause you to sick. The consumers should get to understand whether the chosen fruits and vegetables are safe to
consume. to make sure the security of food for consumers and to guard consumer health, the monitoring of pesticide residues in food products must
be pursued.
Our system can sense the fruit and indicate the level of pesticides substance in the fruit. It is user-friendly where the level of accuracy
is very high. Fruits which seem fresh from outside can be decayed inside, so the inner quality of the fruits is assessed using IR sensor. The main
aim of this work is to detect the amounts of pesticides on fruits and vegetables by calculating its normalized difference vegetation index using
sensor. Here the work done in two modules, in module 1, identify the fruit. The system is trained by image dataset. When the fruit image is captured
by the camera the system identifies the fruit using convolutional neutral network (CNN) algorithm. Fruits-360 dataset which is publicly available
on GitHub and Kaggle for evolution purpose. During feature extraction, at this point the network will perform a number of convolutions and
pooling operation during which the features are detected. Here the features extracted are color, texture and shape of a fruits. Features are extracted
from the input image which is further converted into pixel values. Convolution is utilized for conversion of image into pixel value which preserves
the relationship by learning image features. In module 2, pesticides in fruits and vegetables are detected in three ways by calculating its NDVI, by
using IR sensor and gas sensor then compare the value. Firstly, calculate the NDVI for that capture the reflected light using LDR and send to
Arduino where the signals are analyzed. secondly, by using IR sensor reflected and transmitted light is sends to Arduino and signal is analyzed.
Finally, by using gas sensor pesticide is determined from threshold value, compare all the methods result and final output graph is plotted. Since
the device is portable, it will enhance the efficiency of the food and safety authorities and ensure quality.

II. RELATED WORK

Quality Detection of Fresh Fruits and Vegetables: This paper developed a prototype to detecting quality of the fruit and to provide organic
consumption levels. The system consists of hardware module to detect the quality of the fruit. This is done in three modes, they are, inner
quality of the fruits is determined using ethylene gas sensor and IR sensor which determine the ethylene present in the fruit. Artificially ripened
fruits are detected and finally detect the pesticide residue level in fruits is indicated. This whole paper revolves around sensor integration and
application. [1] Automatic Fruit and Vegetable Detection and Disease Identification system: This paper developed the automatic vision
technology. This system is to identify the fruit image and it will be labeled if the fruit is defected it will send as input for disease identification.
Neutral Network method is used for image classification and identification of fruits and vegetables. The proposed work is composed of mainly
three steps. In the first step the data is accumulated. In the next step, features are extracted and the machine is created. In the third step, the
trained machine will find the fruit and label it. In the final step it will identify the disease related with the identified fruit. The driving objective
JETIR2105829 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org g227
© 2021 JETIR May 2021, Volume 8, Issue 5 www.jetir.org (ISSN-2349-5162)

of this work is to upgrade the esteem of fruit quality discovery.[2] Color Identification Technology: This paper uses color identification
technology to find pesticides in fruits and vegetables. The system consists of Acetyl pesticide testing, applying Raspberry pi 3B and network
cam for color identification method. The solution is prepared from acetylcholine, acetylcholinesterase and yellow color reagent. The solution
is bright yellow, lower blue value (225,225,0) in the non-appearance of pesticide. In the appearance of pesticide, solution becomes light yellow
with higher blue value (225,225,200). The constitutive area of blue value over time is used for converting pesticide residue concentration. The
proposed system is cost effective and requires less testing time.[3] Fruit Quality Detection Using OpenCV/Python: This work use a vision based
technology for fruit quality detection. This paper identify disease in tomato based on the color, the color of the fruit indicates its maturity and
the presence of defects.it is done in three steps, firstly acquiring the image of the tomato it is done by camera and collect the rishika 225 tomatoes
database and these images as input images in this system. Next, detect the disease by its color. This will enable the technology to be applied in
many products.[4]

III. PROPOSED SYSTEM

This work is used for both fruits or vegetables identification and pesticides detection in them. The system includes two modules.
Module 1 is for the detection of fruits and vegetables. Here a fruit recognition frame work utilizing CNN is proposed. The proposed
strategy utilizes profound learning methods for the grouping. The work uses the fruits size, color and texture to recognize each picture. For
training and testing, all the input images were selected from the 360 dataset which is publicly available on GitHub and Kaggle. The dataset
contains 90,380 different fruits and vegetables images of 131 categories. A white paper is placed back of the fruits as a background. Due to the
inconsistency in the lighting a flood fill type algorithm was evolve which extract the fruit from the background. After removing the background
all the fruits were scaled down to 100×100 pixels of standard RGB fruit images. Different varieties of the same fruits and vegetables are stored
as belonging to different classes. From each class, feature extraction of every images should be done. The feature extraction process was done
by three layers of CNN namely convolutional layer, Pooling layer and rectified linear unit layer (ReLU). The convolution layer (CONV) uses
filters that bring about convolution operations as it is scour the input with respect to its dimensions. ReLu layer will apply an elementwise as
shown in figure 1.[5]

Figure 1 process in CNN

After the involvement of these three layers, all the images in the dataset would get resized to same measurements and same filter. A CNN model
including all these images will be created and get saved. Then as an input the webcam would capture an image of the fruit or vegetable that was
used for the experiment. As explained earlier the feature extraction of the input takes place and the resultant image would get compared with
the CNN model that was already created and saved. Hence the fruit or vegetable used get identified. Output of CNN modules identify the fruits
apple, banana, orange and lemon as shown in the figure 2.[6]

JETIR2105829 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org g228
© 2021 JETIR May 2021, Volume 8, Issue 5 www.jetir.org (ISSN-2349-5162)

Figure 2 output of CNN module identify the fruits

Module 2 is for the detection of pesticides in fruits or vegetables. In this project three ways are used for the detection of pesticides. Firstly,
the NDVI method. The light from LED is made to incident on fruit and the reflected rays from the fruit is received by LDR. The output from
LDR is transmitted to Arduino. The ADC in Arduino converts the analog values to digital values. This process got repeated 30 times in order
to check accuracy (for eliminating errors). After completing this 30 times, an array containing the values would be generated and average of
the whole values get displayed on the screen. A graph corresponding to the values obtained by the repeated process also get generated.
Second way is using the IR sensors. An IR sensor has 2 parts, the transmitter and the receiver. The transmitter can transmit light rays of
wavelength up to 960 nm. The rays from transmitter are made to incident on fruit and the reflected rays from the fruit is received by receiver.
An IR receiver can receive rays of wavelength from 400-1000 nm. The output from IR receiver is transmitted to Arduino. This process got
repeated 30 times in order to check accuracy (for eliminating errors). Here signal analyzation takes place. After completing this 30 times, an
array containing the output values of Arduino would be generated and average of the whole values get displayed on the screen. A graph
corresponding to the values obtained by the repeated process also get generated.
Last method is by using the gas sensor. One of the legs of sensor should be grounded, other would be connected to the Arduino and next is
positive. The more chemicals in the fruit, the more there is in the air. This pesticide content gives a good indication for if a fruit or vegetable is
safe or not. The sensor would be already assigned by a threshold value. If the value obtained in the experiment become more than threshold
value then the fruit contains pesticides otherwise not. As that of the previous methods here also the process got repeated 30 times and a graph
would get generated. The repeated process also got generated. Compare all three sensors result and graph plotted as shown in the figure 3.

Figure 3 graph plotted by three sensors

IV. CONCLUSION
The methodology consists of the quality and pesticides detection of fruits and vegetables. The proposed work is composed of mainly two
modules, module one it detect the fruit using image processing and the network used here is CNN. The images are used for train CNN, total of
360 fruits images is used for training. Input for the CNN is capture by camera and classification is done. In module 2, pesticides detection in
fruits and vegetables by three ways calculating its NDVI, by using IR sensor and gas sensor then compare the result. The detection information
will be displayed on the screen. And the output graph is plotted. Out of these method gas sensor give more accuracy to detect the pesticide on
the surface of fruits and vegetables as shown above. Since the device is portable, it will enhance the efficiency of the food and safety authorities
and ensure quality.
V. ACKNOWLEDGEMENT
We would wish to thank everyone who had contributed to the successful completion of this project. We would like to express our gratitude to
our project guide Ms. Resmi B S, Assistant Professor in Computer science and Engineering Department who gave valuable suggestion and guidance
for our project. We express our deepfelt gratitude to beloved HOD Dr. Ramani K, Head of the Department for providing necessar y information
regarding the project and also her support in completing it. We also thank our project co-ordinator, Mr. Jithin Jacob, assistant professor who gave
expert supervision, encouragement and constructive criticism amist his busy schedule throughout the project. We also grateful to all the authors of
books and papers which have been referred to publish this paper

REFERENCES
[1].Shalini Gnanavel, Saranya Manobar, Kanmani Erakuty Sridhar, Sumathi Sokkanarayanan, Mithleysh Sathiyanarayanan, “Quality
Detection of Fresh Fruits and Vegetables to Improve Horticulture and Agro-industries ” IEEE Xplore2017.
[2].R Chandramma, Ramya C, S Vintha, Shivani P R,Vidya MS, ”Automatic Fruit And Vegetable Detection And Disease Identification
system”, www.ijert.org NCAIT-2020 conference Proceedings.
[3].Chang-Chi Lee, Yu-Sung liu, Chun-Hsiang Tseng,Jeng-Yue Chen, Fuh-Shyang Juang, Yung-Ying Chang, Hung-Chih Hsu,”Pesticide
Residue Testing System for Fruits and Vegetables by Color Identification Technology” 2019 IEEE International Conference on Consumer
ElectronicsTaiwan (ICCE-TW).
[4].Miss. Supriya V Patil, Miss. Vaishnavi M. Jadhav, Miss. Komal K. Dalvi, Mr. B.P. Kulkarni, “Fruit Quality Detection Using
OpenCV/Python”, 2020 IRJET.
[5].Fruit Classification CNN:
https://docs.google.com/document/u/0/d/1itOnujaBcnz_S0YFM3T8gZJFzsMxGTbrXUTc0W3rVio/mobilebasic?usp=gmail_thread.
[6].R. Kanmani, R. Maheswar, A. Sureshkumar, A. Mugilan, Dr. V. Praveena,”Non-destructive Approach to Detect

JETIR2105829 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org g229
© 2021 JETIR May 2021, Volume 8, Issue 5 www.jetir.org (ISSN-2349-5162)

Pesticides in Fruits and Vegetables using IoT Technology,”2020 International Conference on Computer Communication and
Informatics(ICCCI2020),Jun 22-24,202.
[7].Yuefang Hua, Hongjian Zhang,”Qualitative and Quantitative Detection of Pesticides With Terahertz Time-Domain Spectroscopy”,
IEEE Transactions on microwave theory and techniques, VOL.58, NO.7, JULY 2010.
[8]. G. S. Gawande, P. S. Deshmukh,”ReviewDetermination of pesticides residue on fruits and vegetables”,Technical Research organisation
India.
[9]. Xia Sun, Xiangyou Wang,“Organophosphorous pesticides detection system based on amperometric enzyme biosensor”2010 IEEE.
[10]. Wang Changlong, Zhang Shengwei, Yuan Ting, Shen Maosong, “Rapid detection of pesticide residues in
vegetables with mini-colorimeter” The Ninth International Conference on Electronic Measurement & Instruments, ICEMI’2009.
[11] Archana B S, Athira S Kumar, Pooja M, Thaneema N, Indu V Nair, Different Methods to Determine Pesticides in Fruits and
Vegetables: A Review, International Research Journal of Engineering andTechnology (IRJET) Volume: 08 Issue: 04, Apr 2021
[12] Ms. Anisha M Nayak, Mr. Manjesh R, Ms. Dhanusha, Fruit Recognition using Image Processing International Journal of Engineering
Research & Technology (IJERT) Volume 7, Issue 08, 2019
[13] Aniket Harsh, Kishan Kumar Jha, Shashwat Srivastava, Abhinav Raj, Raghav S, Fruit Freshness Detection Using CNN Approach,
International Research Journal of Modernization in Engineering Technology and Science Volume:02/Issue:06, June-2020
[14] R Chandramma Ramya C Shivani P R S Vinitha Vidya M S Automatic Fruit and Vegetable Detection and Disease Identification
System International Journal of Engineering Research & Technology (IJERT) Volume 8, Issue 15, 2020
[15] Fouzia Risdin, Pronab Kumar Mondal, Kazi Mahmudul Hassan, Convolutional Neural Networks (CNN) for Detecting Fruit
Information Using Machine Learning Techniques IOSR Journal of Computer Engineering.

JETIR2105829 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org g230