Automated Fruit Disease Detection using Convolutional Neural Networks

Abstract

This study proposes a convolutional neural network (CNN)-based method for the automated detection of fruit diseases. Using deep learning techniques, the model is trained on a large collection of fruit images to accurately recognize and classify different types of diseases affecting fruits. The system provides a rapid, reliable, and cost-effective approach for early disease identification, which can support better crop management and minimize the excessive use of harmful chemicals. Experimental results demonstrate high classification accuracy, highlighting the significant potential of artificial intelligence in enhancing modern agricultural practices. In the proposed approach, the application first captures an input image from the user. The image then undergoes segmentation to extract the relevant region of interest. The segmented image is subsequently provided as input to the CNN model, which extracts important feature vectors for accurate fruit disease detection and classification. The proposed model attains 95% detection accuracy. Keywords— Fruit disease detection, CNN, Image classification, Agricultural, automation, Disease identification.

Plant phenotyping relevance

果実画像から病害状態を直接推定するCNN手法の開発・評価が中心であり、植物の病害表現型を対象とする画像ベースのフェノタイピング研究。

Code and data availability

The paper uses a Kaggle fruit disease dataset but provides no dataset URL, no author code/model availability statement, and no public repository. No paper-specific public asset is actionable.

No evidence-backed public reproduction asset is currently recorded.