Precision Agriculture: Automated Potato Disease Identification Using Deep Learning
Rohit Kumar Singh, Swapnil Srivastava, Swati Sharma, Anshika Mittal, Priyanshu Kumar, Nikhil Kumar
Abstract
Due to their widespread production, potatoes are crucial for food security and livelihood. With quantity comes quality, which is often affected by diseases like late blight and early blight. If these diseases are not identified at an early stage, then they can cause huge losses for farmers, affecting the nation’s food basket. Traditional disease identification requires manual inspection, which is slow and subjective for farming carried out on a large scale. This research introduces deep learning models to automate potato disease detection, providing farmers a faster and more reliable solution. The four deep learning models used in our study are Custom CNN, VGG16, ResNet50, and InceptionV3. These models were selected based on their ability to classify potato leaf diseases more accurately. Their performance was measured through accuracy, F1 score, and recall to determine which one is the most effective model. The findings show that deep learning can help detect disease early and promote sustainable farming.