Scholarlink

<p>Agricultural productivity and profitability heavily depend on effective crop disease management, which is fundamental to sustainable farming practices. The presence of crop diseases presents considerable challenges, necessitating precise and timely identification to minimize potential reductions in yield. Traditional methods are laborious and costly, prompting exploration into Deep Learning techniques such CNN, FasterR-CNN. However, these methods have often fallen short of achieving high generalization accuracy. To overcome these limitations, this paper introduces a novel approach combining image enhancement techniques with the YOLOv9 framework for potato disease recognition using UAV-captured datasets. The proposed method achieves significant improvements, achieving a mean average precision of 97% across all disease classes at a confidence interval from 0.5 to 0.95. Leveraging UAV imagery and advanced deep learning models, the proposed framework significantly improves predictive accuracy and substantiates its scalability for extensive deployment in agricultural domains.</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

Integrating YOLOv9 and image enhancement on UAV-derived data enhances precision in potato disease detection

  • 期刊论文

  • 发表日期：

    2025年8月19日

查看全文
  • Mahendra Kumar Gurve,
  • Sankar Behera,
  • Yamuna Prasad,
  • Nitin

摘要

Agricultural productivity and profitability heavily depend on effective crop disease management, which is fundamental to sustainable farming practices. The presence of crop diseases presents considerable challenges, necessitating precise and timely identification to minimize potential reductions in yield. Traditional methods are laborious and costly, prompting exploration into Deep Learning techniques such CNN, FasterR-CNN. However, these methods have often fallen short of achieving high generalization accuracy. To overcome these limitations, this paper introduces a novel approach combining image enhancement techniques with the YOLOv9 framework for potato disease recognition using UAV-captured datasets. The proposed method achieves significant improvements, achieving a mean average precision of 97% across all disease classes at a confidence interval from 0.5 to 0.95. Leveraging UAV imagery and advanced deep learning models, the proposed framework significantly improves predictive accuracy and substantiates its scalability for extensive deployment in agricultural domains.