ItemArticle

METROLOGICAL IMPACT ON ORANGE FRUIT: A COMPARATIVE ANALYSIS OF DEEP LEARNING MODELS FOR PREDICTING FRUIT DISEASES

Title
METROLOGICAL IMPACT ON ORANGE FRUIT: A COMPARATIVE ANALYSIS OF DEEP LEARNING MODELS FOR PREDICTING FRUIT DISEASES
Creator
Kujur, Lawrence; Gupta, Varuna; Singhal, Abhinav
Description
This research offers a comprehensive comparative analysis of efficient deep learning models for predicting diseases in orange fruits, with a particular emphasis on the impact of meteorological factors on disease prevalence. Citrus diseases such as Blackspot, Canker, and Greening are significantly influenced by environmental conditions. Recognizing the crucial role of weather conditions in the development and spread of these diseases, we concentrate on enhancing prediction accuracy by integrating Convolutional Neural Networks (CNNs) with various classification algorithms to develop hybrid models that account for meteorological impacts. Specifically, we assess the performance of a CNN combined with Gradient Boosting (CNN-GB) and compare it against other hybrid models such as CNN integrated with Long Short-Term Memory networks (CNN-LSTM), Support Vector Machines (CNN-SVM), and Random Forest classifiers (CNN-Random Forest). These models are evaluated using different optimization algorithms to determine the most effective approach for disease prediction under varying meteorological conditions. A meticulously curated dataset comprising 1,600 training images and 300 testing images of orange fruits exhibiting a variety of disease symptoms was utilized for evaluation. The dataset reflects diverse environmental conditions to capture the meteorological impact on disease manifestation. All the models tested, the CNNGB hybrid model with NDAM optimizer exhibited superior performance (Accuracy 98.03) in comparison of other models like CNN (Accuracy 96.03), CNN+LSTM (Accuracy 96.16), CNN + SVM (Accuracy 97.13) and CNN + Random Forest (Accuracy 97.79). The exceptional performance of the CNN-GB model suggests that integrating CNNs with powerful classification algorithms like Gradient Boosting, along with considerations of meteorological data, can significantly enhance disease detection in crops. This advancement contributes to more proactive and effective disease management strategies, ultimately reducing economic losses and increasing productivity in the agricultural sector. 2025 Published by Faculty of Engineering.
Source
Proceedings on Engineering Sciences;Volume;7;Issue;3;pp.1781-1790
Date
01-01-2025
Publisher
Faculty of Engineering, University of Kragujevac
Subject
Convolutional Neural Networks (CNNs); Fruit Disease Prediction; Gradient Boosting (CNN-GB); Hybrid Models; Measurement Accuracy; Metrology
Coverage
Kujur L., Christ University, Bengaluru, India; Gupta V., Christ University, Bengaluru, India; Singhal A., Christ University, Bengaluru, India
Rights
All Open Access; Gold Open Access; Green Open Access
Relation
ISSN: 26202832;
Format
online
Language
English
Type
Article

Citation

Kujur, Lawrence; Gupta, Varuna; Singhal, Abhinav, “METROLOGICAL IMPACT ON ORANGE FRUIT: A COMPARATIVE ANALYSIS OF DEEP LEARNING MODELS FOR PREDICTING FRUIT DISEASES,” CHRIST (Deemed To Be University) Institutional Repository, accessed June 18, 2026, https://archives.christuniversity.in/items/show/23422.

Output Formats