Scholarlink

<p>The growing global demand for accurate plant disease classification underscores the need for early intervention to enhance disease management and boost fruit productivity. This study presents a&#xa0;novel deep learning (DL)-based approach for classifying Apple Leaf diseases, aiming to improve agricultural outcomes. The proposed method, DensoMobileForestNet, employs feature fusion techniques combining outputs from MobileNetV2 and DenseNet201, followed by classification using a&#xa0;Random Forest (RF) model to improve disease classification accuracy (<i>AC</i>). The Apple Leaf disease dataset, sourced from three diverse origins, undergoes thorough preprocessing to ensure high-quality input for model training. Feature fusion is initially applied by extracting and merging features from multiple CNN models, enhancing the robustness of the classification. To address the challenges of manual hyperparameter tuning, metaheuristic optimization algorithms—including Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), and Differential Evolution (DE)—are explored for optimal parameter selection. Among these, the PSO-optimized DensoMobileForestNet achieves the best performance, attaining a&#xa0;classification AC of 95.65%. Comparative analyses demonstrate the statistical and computational superiority of the PSO-enhanced model, confirming its effectiveness for accurate and efficient Apple Leaf disease diagnosis.</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

Metaheuristic-Integrated DensoMobileForestNet for Apple Leaf Disease Identification

  • 期刊论文

  • 发表日期：

    2025年5月23日

查看全文
  • Shantilata Palei,
  • Puspanjali Mohapatra

摘要

The growing global demand for accurate plant disease classification underscores the need for early intervention to enhance disease management and boost fruit productivity. This study presents a novel deep learning (DL)-based approach for classifying Apple Leaf diseases, aiming to improve agricultural outcomes. The proposed method, DensoMobileForestNet, employs feature fusion techniques combining outputs from MobileNetV2 and DenseNet201, followed by classification using a Random Forest (RF) model to improve disease classification accuracy (AC). The Apple Leaf disease dataset, sourced from three diverse origins, undergoes thorough preprocessing to ensure high-quality input for model training. Feature fusion is initially applied by extracting and merging features from multiple CNN models, enhancing the robustness of the classification. To address the challenges of manual hyperparameter tuning, metaheuristic optimization algorithms—including Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), and Differential Evolution (DE)—are explored for optimal parameter selection. Among these, the PSO-optimized DensoMobileForestNet achieves the best performance, attaining a classification AC of 95.65%. Comparative analyses demonstrate the statistical and computational superiority of the PSO-enhanced model, confirming its effectiveness for accurate and efficient Apple Leaf disease diagnosis.