Logo image
A novel deep learning model for detection of severity level of the disease in citrus fruits
Journal article   Open access   Peer reviewed

A novel deep learning model for detection of severity level of the disease in citrus fruits

Pankaj Dhiman, Vinay Kukreja, Poongodi Manoharan, Amandeep Kaur, M. M Kamruzzaman, Imed Ben Dhaou and Celestine Iwendi
Electronics, Vol.11(3), 495
08/02/2022

Abstract

deep learning graph based segmentation object detection transfer learning severity Disease Citrus Fruits
Citrus fruit diseases have an egregious impact on both the quality and quantity of the citrus fruit production and market. Automatic detection of severity is essential for quality productions of fruits. In current work, citrus fruits dataset is preprocessed by rescaling and establishing bounding boxes with labeled image software. Then Selective search, which combines the capabilities of both an extensive search and graph based segmentation, is applied. The proposed DNN (deep neural network) model is trained to detect targeted area of the disease with its severity level using citrus fruits that have been labeled by taking help of a domain expert with four severity level(high, medium ,low and healthy) as ground truth. Transfer learning using VGGNet is applied to implement multi- classification framework for each class of severity. The model predicts the low severity level with 99% accuracy, and the high severity level with 98% accuracy. Model produces 96% accuracy in detecting 1 healthy conditions and 97% accuracy in detecting medium severity levels. The result of the work 1 shows that the proposed approach is valid, and it is efficient for detecting citrus fruit disease at four 1 levels of severity.
pdf
A Novel Deep Learning Model for Detection of Severity Level of the Disease in Citrus Fruits1.40 MBDownloadView
Published (Version of record)CC BY V4.0 Open Access
url
Link to Published VersionView
Published (Version of record)Open AccessCC BY V4.0 Open

Metrics

5 File views/ downloads
50 Record Views
265 Times Cited - Scopus

Details

Logo image

Usage Policy