Potato Plant Leaf Disease Detection Using Deep Learning Method

Abstract

In agriculture, plant disease detection and cures for those diseases are crucial for high crop production and yield sustainably. Improvements in the automated disease detection and analysis areas may provide important benefits for early action that would allow intervention at earlier stages for cure and preventing spread of the disease. As a result, damages on crop yield could be minimized. This study proposes a new deep-learning model that correctly classifies plant leaf diseases for the agriculture and food sectors. It focuses on the detection of plant diseases for potato leaves from images by designing a new convolutional neural network (CNN) architecture. The CNN methodology applies filters to input images, extracts key features, reduces dimensions while preserving important characteristics in images, and finally, performs classification. The experimental results conducted on a real-world dataset showed that a significant improvement (8.6%) in accuracy was achieved on average by the proposed model (98.28%) compared to the state-of-the-art models (89.67%) in the literature. The weighted averages of recall, precision, and f-score metrics were obtained around 0.978, meaning that the method was highly successful in disease diagnosis.

Keywords

• agriculture
• disease diagnosis
• plantvillage
• smart farming
• image classification
• deep learning
• convolutional neural networks