Scholarlink

<p>The necessity for advanced methodologies in fruit grading and disease detection is underscored by the increasing demand for food safety and quality in the agricultural sector. Traditional methods often fail to accurately identify and classify diseases due to their inability to handle high-dimensional data and complex feature interactions. This work addresses these limitations by proposing an innovative framework that integrates multiple optimization techniques for enhanced feature selection and disease detection in fruits. The proposed model combines Particle Swarm Optimization (PSO), Genetic Algorithms (GA), Simulated Annealing (SA), and Coot Optimization (CO) for superior feature selection. This amalgamation leverages the unique exploration capabilities of each algorithm, thus facilitating a more comprehensive evaluation of the feature space and enabling the model to overcome local optima and navigate complex feature landscapes more effectively. The integration improves feature selection accuracy by 5–10% and reduce computational time by 10–20% compared to utilizing CO alone. Furthermore, the framework enhances deep learning algorithms for disease detection by integrating Convolutional Neural Networks (CNNs) with attention mechanisms and Quad Long Short-Term Memory networks (QLSTMs) with a Graph-based Generative Adversarial Network (Graph GAN). This integration allows for the effective capture of spatial relationships and temporal dependencies within thermal image datasets, leading to significant improvements in disease detection accuracy and a reduction in false-negative rates. To enhance the robustness and generalizability of the model, an extensive dataset comprising diverse fruit types, disease severities, and environmental conditions has been curated, including thermal sample images that depict various disease stages. This comprehensive dataset ensures that the model is well-equipped to detect both external and internal symptoms of diseases, resulting in an anticipated increase of 10–15% in model accuracy and a reduction of 20–30% in false-positive rates. The impacts of this work are multifaceted, offering significant advancements in the accuracy, efficiency, and applicability of fruit grading and disease detection methodologies. By addressing the limitations of existing techniques and introducing a robust, integrated framework, this research paves the way for safer and more reliable agricultural practices, ultimately contributing to enhanced food safety and quality.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Design of an iterative method for enhanced disease detection and feature selection using particle swarm optimization, genetic algorithms, and convolutional neural networks

  • 期刊论文

  • 发表日期：

    2025年7月29日

查看全文
  • Archana Ganesh Said,
  • Deepali Jawale

摘要

The necessity for advanced methodologies in fruit grading and disease detection is underscored by the increasing demand for food safety and quality in the agricultural sector. Traditional methods often fail to accurately identify and classify diseases due to their inability to handle high-dimensional data and complex feature interactions. This work addresses these limitations by proposing an innovative framework that integrates multiple optimization techniques for enhanced feature selection and disease detection in fruits. The proposed model combines Particle Swarm Optimization (PSO), Genetic Algorithms (GA), Simulated Annealing (SA), and Coot Optimization (CO) for superior feature selection. This amalgamation leverages the unique exploration capabilities of each algorithm, thus facilitating a more comprehensive evaluation of the feature space and enabling the model to overcome local optima and navigate complex feature landscapes more effectively. The integration improves feature selection accuracy by 5–10% and reduce computational time by 10–20% compared to utilizing CO alone. Furthermore, the framework enhances deep learning algorithms for disease detection by integrating Convolutional Neural Networks (CNNs) with attention mechanisms and Quad Long Short-Term Memory networks (QLSTMs) with a Graph-based Generative Adversarial Network (Graph GAN). This integration allows for the effective capture of spatial relationships and temporal dependencies within thermal image datasets, leading to significant improvements in disease detection accuracy and a reduction in false-negative rates. To enhance the robustness and generalizability of the model, an extensive dataset comprising diverse fruit types, disease severities, and environmental conditions has been curated, including thermal sample images that depict various disease stages. This comprehensive dataset ensures that the model is well-equipped to detect both external and internal symptoms of diseases, resulting in an anticipated increase of 10–15% in model accuracy and a reduction of 20–30% in false-positive rates. The impacts of this work are multifaceted, offering significant advancements in the accuracy, efficiency, and applicability of fruit grading and disease detection methodologies. By addressing the limitations of existing techniques and introducing a robust, integrated framework, this research paves the way for safer and more reliable agricultural practices, ultimately contributing to enhanced food safety and quality.