Authors :
Mr. Ravindra P; Dr. Raghavender Raju L; Hemanth Reddy S; Harshavardhan Chary V
Volume/Issue :
Volume 8 - 2023, Issue 10 - October
Google Scholar :
https://tinyurl.com/25vp28he
Scribd :
https://tinyurl.com/y6357sc2
DOI :
https://doi.org/10.5281/zenodo.10032706
Abstract :
Using advanced machine learning techniques,
specifically Convolutional Neural Networks (CNNs), this
study focuses on detecting leaf diseases in three vital
crops: Bell Pepper, Potato, and Tomato, crucial for
global food production. A curated dataset contains
various healthy and diseased plant images, covering
diseases like Bacterial Spot, Early Blight, Late Blight,
and more. The methodology involves data preprocessing,
including augmentation techniques to improve the
model's robustness. CNN is superior to SVM, pretrained
models, Random Forest, MLP, and ensemble methods
because CNN provides high scalability, which is crucial
for our dataset consisting of 20,000 images. Additionally,
CNN outperforms other methods with an impressive
prediction accuracy of 98-99% on the training data. This
work offers a scalable and adaptable solution for early
disease detection, aiding farmers in implementing
targeted disease management strategies and reducing
crop losses. It represents a practical contribution to
agriculture, leveraging CNNs to combat plant diseases
effectively.
Keywords :
Convolutional Neural Networks (CNNs), Data Preprocessing, Data Augmentation, CNN Architecture, Machine Learning, Deep Learning, Image Classification, Food Security.
Using advanced machine learning techniques,
specifically Convolutional Neural Networks (CNNs), this
study focuses on detecting leaf diseases in three vital
crops: Bell Pepper, Potato, and Tomato, crucial for
global food production. A curated dataset contains
various healthy and diseased plant images, covering
diseases like Bacterial Spot, Early Blight, Late Blight,
and more. The methodology involves data preprocessing,
including augmentation techniques to improve the
model's robustness. CNN is superior to SVM, pretrained
models, Random Forest, MLP, and ensemble methods
because CNN provides high scalability, which is crucial
for our dataset consisting of 20,000 images. Additionally,
CNN outperforms other methods with an impressive
prediction accuracy of 98-99% on the training data. This
work offers a scalable and adaptable solution for early
disease detection, aiding farmers in implementing
targeted disease management strategies and reducing
crop losses. It represents a practical contribution to
agriculture, leveraging CNNs to combat plant diseases
effectively.
Keywords :
Convolutional Neural Networks (CNNs), Data Preprocessing, Data Augmentation, CNN Architecture, Machine Learning, Deep Learning, Image Classification, Food Security.
