Home > Published Issues > 2022 > Volume 10, No. 4, December 2022 >
JOIG 2022 Vol.10(4): 158-165
doi: 10.18178/joig.10.4.158-165

Camera Radial Distance-Based Accuracy of a Bacterial Blight, Brown Spot, and Rice Blast Plant Disease Identification System for Remote Communications

Christian Jeremy N. Canlas 1, Cayla Mari M. Cortez 1, Ravin Angelo M. Dela Cruz 1, Jasmine Rone G. Padua 1, Anna Grachiel G. Timbol 1, Jelly G. Yumul 1, Emmanuel T. Trinidad 1,2, and Lawrence Materum 2,3
1. Don Honorio Ventura State University, Bacolor, Philippines
2. De La Salle University, Manila, Philippines
3. Tokyo City University, Tokyo, Japan

Abstract—Rice plant diseases pose a high threat to rice production. However, the ability to produce better crops is required for any country's economic development. Thus, early detection of rice plant diseases is needed as it can also save the farmer's economic loss. This study presented an identification framework based on AlexNet architecture and transfer learning that can distinguish between healthy leaves, leaves infected with one of the three most common diseases, namely rice blast, brown spot, or bacterial blight, or leaves infected with a disease not covered by any of the three, and displays the results, nature, solutions, and interventions in an application. The Convolutional Neural Network (CNN) and the application were implemented using MATLAB. The datasets used to train the network were obtained from online repositories, and the trained network was tested on actual data taken from the farm. The training-testing division used for labeled images was 80%-20%, and thus the network obtained a validation accuracy of 99.84%. The images taken from the field were captured and proposed to be deployed for remote monitoring via Raspberry Pi connected through Wireless Local Area Network (WLAN) interfaced in a Graphical User Interface (GUI). The identification of the AlexNet achieved a classification accuracy of 94% in testing a 2-inch radial distance of the camera to the subject with images taken from the field. Furthermore, a computed average percentage rating of 80.89% based on the evaluation responses from crop experts and other evaluators proved that the framework was functional, reliable, and efficient.

Index Terms—AlexNet, transfer learning, convolutional neural network, rice plant disease, WLAN

Cite: Christian Jeremy N. Canlas, Cayla Mari M. Cortez, Ravin Angelo M. Dela Cruz, Jasmine Rone G. Padua, Anna Grachiel G. Timbol, Jelly G. Yumul, Emmanuel T. Trinidad, and Lawrence Materum, "Camera Radial Distance-Based Accuracy of a Bacterial Blight, Brown Spot, and Rice Blast Plant Disease Identification System for Remote Communications ," Journal of Image and Graphics, Vol. 10, No. 4, pp. 158-165, December 2022.

Copyright © 2022 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.