Leaf-Rust and Nitrogen Deficient Wheat Plant Disease Classification using Combined Features and Optimized Ensemble Learning

Abstract

Automatic approaches for detecting wheat plant diseases at an early stage are critical for protecting the plants and improving productivity. In the traditional system, farmers use their naked eyes to identify the disease, which is time-consuming and requires domain knowledge. In addition, the domain experts in many remote areas are not available in time and are expensive. To address the above issues, this study proposed an automatic wheat plant disease classification using combined features and an optimized ensemble learning algorithm. The main objective of the proposed system is to detect and classify the normal vs leaf rust vs nitrogen-deficient in wheat plants. Further, we used 1459 wheat leaf images from a public dataset to evaluate the suggested method. From the experimental results (ACC=96.00% for normal vs nitrogen deficient, ACC=98.25% for normal vs leaf rust and ACC=97.39% for normal vs leaf rust vs nitrogen deficient), it is observed that the suggested ensemble method outperformed the other benchmark machine learning algorithms.