Predicting Agriculture Leaf Diseases (Potato): An Automated Approach using Hyper-parameter Tuning and Deep Learning

• Published in: 2023 Third International Conference on Secure Cyber Computing and Communication (ICSCCC) pp. 490 - 493
• Main Authors: Sharma, Ochin, Rajgaurang, Mohapatra, Srikanta, Mohanty, Jayashree, Dhiman, Pummy, Bonkra, Anupam
• Format: Conference Proceeding
• Language: English
• Published: IEEE 26.05.2023
• Subjects: Agriculture; classification; Crops; Deep learning; Diseases; employability; industry; machine learning; Production; Research and development; Task analysis; Training; Transfer learning; Tuning
• DOI: 10.1109/ICSCCC58608.2023.10176819

The classification of potato illnesses is a crucial undertaking in agriculture because it enables farmers to recognise and control the numerous diseases that may impact their crops of potatoes. Many different potato diseases, including early blight, late blight, black scurf and powdery scab can strike. To avoid crop loss and guarantee a healthy output, it is essential to correctly detect and classify these illnesses. When a model is trained on a dataset including labelled photos of diseased potatoes, deep learning techniques can be very useful for classifying potato diseases. The ability to distinguish between healthy and sick potatoes and can pinpoint the precise disease that is afflicting a given potato. A deep learning model can be trained on the dataset once it has been prepared or acquired using well-known frameworks like Tensorflow. To enhance the model's performance, methods like data augmentation, transfer learning, and hyperparameter tuning are used and 99.42% accuracy is attained.