An Effective Method for Plant Leaf Disease Detection Using MDLDPTS and Fuzzy Based KNN Model

This study aims to develop an effective and capable approach for the prediction of various plant leaf disease image through the Multi-Direction and Location Distribution of Pixels in Trend Structure (MDLDPTS) descriptor and the Fuzzy k-N earest Neighbor (FkNN) approach. The effectiveness of the proposed image classification process of the leaf disease has been analyzed on a benchmark dataset based on the accuracy, the area under the curves of the receiver operating characteristic and the curve. Investigational outputs have illustrated that MDLDPTS and FkNN-based method highly outperforms the contemporary techniques and also ensure a reliable disease prediction model for analyzing the plant leaf diseases. The proposed MDLDPTS and FkNN-based method has outperformed the other plant leaf disease detection models with an accuracy 97.7% over other existing techniques.