Author(s):
1. JOHANNES FARRELL LANDUTAMA:
Computer Science Department, BINUS Graduate Program – Master of Computer Science, Bina Nusantara University, Jakarta, Indonesia
2. GALUH PUTRA WARMAN:
Computer Science Department, BINUS Graduate Program – Master of Computer Science, Bina Nusantara University, Jakarta, Indonesia
3. BENFANO SOEWITO:
Computer Science Department, BINUS Graduate Program – Master of Computer Science, Bina Nusantara University, Jakarta, Indonesia
Abstract:
The development of technology in the agriculture sector has shown promise in the last few years. Crop management, disease detection, and irrigation management are examples of activities that can be implemented with modern technological approaches using artificial intelligence and IoT. Modern disease detection in agriculture uses machine learning to classify the health condition of plants by processing input images of leaves or branches into pre-trained machine learning models. Studies for plant disease detection have been done by other researchers using high computing capability devices; in contrast, there is still little research on implementing such machine learning models for mobile or small IoT-based devices. This study explores and proposes a model of machine learning algorithm namely CNN MobileNet and SVM, to run on small IoT-based devices. After some experiments, it was found that MobileNet can be used for this particular purpose. Furthermore, this research also shows a new contribution regarding the implementation of machine learning for disease detection into an IoT microcontroller commonly used for irrigation and soil moisture observation. The proposed model in this study has been tested in real-world experiments for Tomato plant disease detection with an approximate accuracy of 91.45%.
Page(s):
6420-6428
DOI:
DOI not available
Published:
Journal: Journal of Theoretical and Applied Information Technology, Volume: 100, Issue: 21, Year: 2022
Keywords:
machine learning
,
disease detection
,
Arduino
,
MobileNet
References:
References are not available for this document.
Citations
Citations are not available for this document.