Downloads: 4 | Views: 204 | Weekly Hits: ⮙1 | Monthly Hits: ⮙1
Prototypes and Models | Computer Science and Information Technology | Zimbabwe | Volume 11 Issue 12, December 2022 | Rating: 5.2 / 10
A Model for Smart Farming through Cloud-Based Plant Leaf Disease Detection using Supervised Machine Learning
Levis Makurumure | F. D Mukoko
Abstract: Plant diseases reduce both quality and quantity, preventing a high yield. Leaf diseases must be identified at all plant stages so as to preserve the expected yield from the plants. However, agriculture specialists are not often available to identify plant diseases thereby delaying the plant disease management process slow and costly. With smart agriculture, digital image processing and deep learning have made significant breakthroughs in plant disease management. A real-time disease prediction model will cover the time gap caused by human involvement, offering early diagnosis and treatments to plant leaf diseases and greater yields. This study implemented CNN to detect and classify diseases in tomato plants. The results of the study showed that the model loss reduced from 3.0 at the beginning to 0.5 at epoch 1 and remained falling until the final epoch, when training and validation had the least loss. On the other hand, the accuracy of the model also increased as the training progressed registering a higher accuracy of 90% for both the training and validation scores. The classification result shows that the model accurately detected Tomato Mosaic Virus on infected leaves and healthy plants. Predicting late blight was 92% accurate. Infected eaves scored 100%, while late blight leaves got 91%. The model detected health leaves with a 99% accuracy. Late blight was predicted with 91% and tomato mosaic virus with 100%. The model predicted all 1000 sample leaves classes with 96% accuracy.
Keywords: Cloud computing, Tomato plant leaf disease, Machine learning
Edition: Volume 11 Issue 12, December 2022,
Pages: 402 - 407