Local Directional Patterns for Plant Leaf Disease Detection

Abstract

Plant leaf disease detection is an attractive research issue for artificial intelligence and computer vision community, who aims to develop intelligent systems that make an automatic detection of plant leaf diseases. In the recent past years, deep learning techniques were extensively used since they allow developing end-to-end systems. However, for several case studies detection scores still need improvement. Presently, we propose a new descriptor that is based on local Directional Patterns to perform feature generation from plant leaves. This descriptor is associated with SVM classifier to develop the full detection system. Experiments are conducted by considering three crop species that are Tomato, Potato, and Bell pepper diseases. The proposed LDP features are evaluated comparatively to convolutional neural networks features as well as to the histogram of oriented gradients. The results obtained highlight the effectiveness of the proposed system which outperforms the LeNet-5 convolutional neural network by 3% in the over-all accuracy.