Scholarlink

<p>Plant diseases significantly impact agriculture, causing significant financial losses and food insecurity. Early detection is essential for control and prevention. While farmers traditionally relied on visual inspection, this manual method is time- consuming and error-prone. Although bacteria, fungi, viruses, and other infec- tious organisms can infect many areas of a plant, our research primarily focused on identifying plant leaf disease. Numerous researchers have conducted various experiments to identify and prevent plant leaf diseases. This paper’s approach to identifying plant leaf diseases is based on XGBoost and the OPTUNA hyper- parameter. OPTUNA is a software framework for autonomous hyper-parameter optimization, while XGBoost is a machine learning approach that improves performance. The model is trained and evaluated using open-source datasets con- taining images of healthy and infected plant leaves. Experimental results show that the proposed approach performs comparatively well compared to other algo- rithms. On the PlantVillage dataset, the model had an accuracy of 82.32%, precision of 84.54%, recall of 86.42%, F1-score of 86.68%, and specificity of 92.22%. For the Tomato Leaf dataset, the model achieved 89.12% accuracy, 91.35% precision, 88.32% recall, 88.23% F1-score, and 91.61% specificity. The model performed notably well on the Rice Leaf dataset, with 98.44% accuracy, 98.58% precision, 98.44% recall, 98.45% F1-score, and 99.27% specificity, demonstrating its robustness in detecting plant leaf diseases across different crop types.</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

Plant Leaf Disease Detection Using XGBoost with OPTUNA Hyperparameter Optimization

  • 期刊论文

  • 发表日期：

    2025年10月13日

查看全文
  • Parijata Majumdar

摘要

Plant diseases significantly impact agriculture, causing significant financial losses and food insecurity. Early detection is essential for control and prevention. While farmers traditionally relied on visual inspection, this manual method is time- consuming and error-prone. Although bacteria, fungi, viruses, and other infec- tious organisms can infect many areas of a plant, our research primarily focused on identifying plant leaf disease. Numerous researchers have conducted various experiments to identify and prevent plant leaf diseases. This paper’s approach to identifying plant leaf diseases is based on XGBoost and the OPTUNA hyper- parameter. OPTUNA is a software framework for autonomous hyper-parameter optimization, while XGBoost is a machine learning approach that improves performance. The model is trained and evaluated using open-source datasets con- taining images of healthy and infected plant leaves. Experimental results show that the proposed approach performs comparatively well compared to other algo- rithms. On the PlantVillage dataset, the model had an accuracy of 82.32%, precision of 84.54%, recall of 86.42%, F1-score of 86.68%, and specificity of 92.22%. For the Tomato Leaf dataset, the model achieved 89.12% accuracy, 91.35% precision, 88.32% recall, 88.23% F1-score, and 91.61% specificity. The model performed notably well on the Rice Leaf dataset, with 98.44% accuracy, 98.58% precision, 98.44% recall, 98.45% F1-score, and 99.27% specificity, demonstrating its robustness in detecting plant leaf diseases across different crop types.