Development of Smart IoT-Based CNN Technique for Harmful Maize Insects Recognition in Precision Agriculture

Ezeofor Chukwunazo Joseph¹

Section:Research Paper, Product Type: Journal-Paper
Vol.9 , Issue.5 , pp.48-60, Oct-2021

Online published on Oct 31, 2021

Copyright © Ezeofor Chukwunazo Joseph . This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
 

View this paper at   Google Scholar | DPI Digital Library

XML View     PDF Download

How to Cite this Paper

Abstract :
this paper presents Development of Smart Internet of Things (IoT)-Based Convolutional Neural Network (CNN) technique for Harmful Insects Recognition in Precision Agriculture. There has been a reported case of aggressive influx of maize moths (Spodoptera species) in maize farms, which has caused tremendous damage to maize crops yield yearly in Nigeria. Nigerian Farmers are still worried because reliable solutions have not been provided that would eliminate the attack of these moths. The maize moths were captured in the farm using installed delta pheromone traps and laboratory moths’ breeding. The captured images stored in the online Google drive dataset folder for the model training. Android mobile app was designed using Figma Editor and developed in Android studio using Kotlin programming language. The IoT system comprises of Raspberry Pi with in-built sensors for sensing, detecting and capturing of the targeted maize moths in the farm. A custom Convolutional Neural Network, called MothNet model was developed and trained in Google Colab using various Python libraries. During training and validation tests, MothNet achieved more than 90% accuracy with low losses but could not generalize well when tested on site because of the small images in the dataset. In order to ensure robust and generalizable system, transfer learning with data augmentation techniques was used to develop a better-performing model. Residual Neural Network (ResNet 50) Architecture was selected, modified and fine-tuned for the image classifications. During on-site testing, the IoT system was set up in the maize farm and internet connectivity established. Maize moth’s image was captured by the IoT system and used cloud based model to detect the image. The image was sent to the Amazon cloud via secure file transfer protocol (FTP) for classification. The modified ResNet-50 model stored in the cloud was called for prediction which achieved 85% accuracy. The predictions were sent to the Amazon Web Services (AWS) S3 database for storage and maize app used to download the predicted moths from the database for the farmers’ attention.

Key-Words / Index Term :
Spodoptera species, IoT, Maize, CNN, Google Colab, MothNet, ResNet 50, Python

References :
[1] P. K. Natikar, R. A. Balikai and Ch. Anusha, “Pest Management Strategies in Precision Farming,” J. Exp. Zool. India Vol. 19, No. 1, pp. 1-8, 2016.
[2] P. F. Konstantinos,“Deep learning models for plant disease detection and diagnosis,” Computers and Electronics in Agriculture, Vol.145, no.1, pp. 311–318. 2008.
[3] G. L. Konstantinos, B. Patrizia, M. Dimitrios, P. Simon and B. Dionysis, “Machine Learning in Agriculture: A Review,” Published in: Sensors MDPI, Vol 18 issue 8. pp. 2674, 2018.
[4] N. N. Esther,“Africa’s most notorious insects – the bugs that hit agriculture the hardest.,” Sept.2017 [Online] available:https://theconversation.com/africas-most-notorious-insects-the-bugs-that-hit-agriculture-the-hardest-83107
[5] E.S. Biswajit, A. Kowsar, B. Premankur, and C. Amit, “Development of m-sahayak-the innovative Android based application for real time assistance in Indian Agriculture and Health Sectors,” UBICOMM 2012:The sixth International conference on mobile ubiquitious computing, system,services and Technology.
[6] T. Ankur, L.C. De, K. P. Ravi, and K.S. Rakesh, “Mobile Android Application on Orchid farming based on North Eastern States of India,” Short Communication-National Research Centre for Orchids, Pakyong 737106, Sikkim. Vol.76 No.4 pp.752-756 ,2019,
[7] P.S. Arockia, B.A Thanga, M. Komal, A.H. Khule, “Fungal Detection in Agricultural Crops Using IoT and SVM,” International Research Journal of Engineering and Technology (IRJET), vol.5 No.2, pp. 872-875, 2018.
[8] R. T. Ms.Nilam and N. Swati, “Early Disease Detection and Monitoring Large Field of Crop by Using IoT,” International Journal of Computer Science and Information Security (IJCSIS), vol.15 no.10, pp.236-248, 2017.
[9] G. Demin, S. Quan, H. Bin, and Z.Shuo, “A Framework for Agricultural Pest and Disease Monitoring Based on Internet-of-Things and Unmanned Aerial Vehicles,” published in MDPI sensor, vol.20 issue 5, pp.1-18 , 2020.
[10] M.V. Manoj, “IoT based Unmanned Aerial Vehicle System for Agriculture Applications,” published in 2018 International conference on Smart Systems and Inventive Technology (ICSSIT), Tirunelveli, India, 2018,
[11] S.C. Madhukar, “Automatic Arial Vehicle Based Pesticides Spraying System for Crops”, International Journal of Innovative Technology and Exploring Engineering (IJITEE), Vol.8 Issue 11, pp. 41-44, 2019.
[12] H. C. Simon, N. Mayumbo, P. Jackson, M. M. Kabemba, O. Y. N. Philip, H.S. Philemon, “An Application of Machine Learning Algorithms in Automated Identification and Capturing of Fall Armyworm (FAW) Moths in the Field,” Proceedings of the Information and Communication Technology Society of Zambia (ICTSZ) International Conference in ICTS, Lusaka, Zambia, pp.119-124. 2018.
[13] H. C. Simon, N. Mayumbo, P. Jackson, M. K. Monde, O. Y. N. Philip, H. S. Philemon, “Machine Learning Algorithms for Automated Image Capture and Identification of Fall Armyworm (FAW) Moths,” Zambia Information Communication Technology (ICT) Journal. Vol. 3 No.1, pp.1-4, 2019.
[14] A. M. A. Belal, and S. A. Samy, “Image-Based Tomato Leaves Diseases Detection Using Deep Learning,” International Journal of Academic Engineering Research (IJAER), vol.2 No.12, pp. 10-16, 2018.
[15] A. M. A. Belal, B, S. A Sammy., and S. A. Samy, “Plant Seedlings Classification Using Deep Learning,” International Journal of Academic Information Systems Research (IJAISR), vol.3 No.1, pp.7-14, 2019.
[16] K. S. Arnab , S. Jayeeta, R. Radhika, S. Sachet, D. Subhojit , P. C. Soummyo and N. S. Himadri, “IOT-Based Drone for Improvement of Crop Quality in Agricultural Field,” Published in: 2018 IEEE 8th Annual Computing and Communication Workshop and Conference (CCWC). Date of Conference: 8-10 Jan. 2018. Las Vegas, NV, USA.
[17] F. M. Rodrigo, N. Ibrahim, H. T. Anh Lan, “Precision Agriculture Using Remote Monitoring Systems in Brazil,” Published in: 2017 IEEE Global Humanitarian Technology Conference (GHTC), Date of Conference: 19-22 Oct. 2017. Location: San Jose, CA, USA.
[18] K. Thenmozhi, & U.R. Srinivasulu, “Crop pest classification based on deep convolutional neural network and transfer learning,” published in Elsevier-Computers and Electronics in Agriculture.vol 164 No 1. 2019.
[19] K. Thenmozhi and R. U. Srinivasulu, “Machine learning ensemble with image processing for pest identification and classification in field crops”, published in Neural Computing and Applications (springer), 02 January 2021
[20] M. K Nour eldeen , l. mohamed , N. T. mohamed hamed, “Insect Pests Recognition Based On Deep Transfer Learning Models,” Journal of Theoretical and Applied Information Technology, vol. 98 no 1, pp.60-68, 2020.
[21] C.J. Ezeofor, , K. Akpado, and A. Ulasi, “Predictive Model For Maize Stem Borers’ Classification In Precision Farming”, International Journal of Artificial Intelligence & Applications (IJAIA) vol.12 no 4, pp.33-49, 2021.
[22] C.J Ezeofor, E.C. Okafor, K Akpado, and U Zakka. “IOT Architecture for Real Time Maize Stem Borers’ Detection and Capturing in Precision Farming”, Journal of Engineering and Applied Sciences (JEAS), Vol.18 no1, pp.381-392, 2021