Deep Learning Techniques Toward Advancement of Plant Leaf Diseases Detection

Abstract

This paper presents a proposed system that is used to classify and detect plant leaf diseases using image processing and deep learning techniques. The proposed system consists of two methods for classification and makes a comparison between them. The first method is based on the support vector machine (SVM) algorithm and consisted of several stages leading to the classification stage. In our work, specific types of plants are selected, which are tomatoes, pepper, and potatoes, as they are the most common types of plants in the world in general and in Iraq in particular. The second method used the convolution Neural network (CNN) for classification. In these two methods, 15 classes were classified, including 12 classes for diseases of different plants that were detected, such as bacteria, fungi, etc., and three classes for healthy leaves. The result of the comparison shows the preference of the CNN algorithm over the SVM algorithm in terms of accuracy and time, which makes an effective and accurate system in the detection and classification of plant leaf diseases.

Plant phenotyping relevance

植物葉の画像から病害状態を検出・分類する画像処理および深層学習手法を提案し、SVMとCNNを比較評価しているため、植物フェノタイピング手法が中心です。

Code and data availability

公開本文の所在を確認できませんでした。非公開または購読が必要な可能性があります。

No evidence-backed public reproduction asset is currently recorded.