Scholarlink

<p>This research presents an energy-efficient precision agriculture system that integrates deep learning with sustainable farming practices for plant pest detection and management. This study proposes a hybrid neural architecture that combines EfficientNetB0 for feature extraction with a Spiking Neural Network (SNN) classifier, designed to balance high accuracy with reduced computational and energy costs. The system analyzes plant leaf images to identify 27 pest categories and provides tailored organic solution recommendations from a curated database of over 1000 verified management strategies. To ensure statistical robustness, experiments were conducted on a comprehensive dataset of 12,900 images, demonstrating that the proposed hybrid model achieves a mean classification accuracy of 84.57% ± 1.26% across multiple runs. More importantly, the hybrid architecture delivers substantial efficiency gains, reducing power consumption to 52W and lowering computational complexity to 2.5G FLOPs. These improvements make the system particularly suitable for deployment in resource-constrained agricultural environments, including edge devices and field monitoring systems. By coupling accurate pest detection with actionable organic solution recommendations, this work advances sustainable farming practices, offering a practical tool that addresses both environmental responsibility and technological feasibility.</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

Energy-Efficient Plant Pest Detection with Organic Solution Recommendations for Sustainable Farming

  • 期刊论文

  • 发表日期：

    2026年8月3日

查看全文
  • S. Kalai Vani,
  • K. N. Divyaprabha,
  • Sanjaya Sankar

摘要

This research presents an energy-efficient precision agriculture system that integrates deep learning with sustainable farming practices for plant pest detection and management. This study proposes a hybrid neural architecture that combines EfficientNetB0 for feature extraction with a Spiking Neural Network (SNN) classifier, designed to balance high accuracy with reduced computational and energy costs. The system analyzes plant leaf images to identify 27 pest categories and provides tailored organic solution recommendations from a curated database of over 1000 verified management strategies. To ensure statistical robustness, experiments were conducted on a comprehensive dataset of 12,900 images, demonstrating that the proposed hybrid model achieves a mean classification accuracy of 84.57% ± 1.26% across multiple runs. More importantly, the hybrid architecture delivers substantial efficiency gains, reducing power consumption to 52W and lowering computational complexity to 2.5G FLOPs. These improvements make the system particularly suitable for deployment in resource-constrained agricultural environments, including edge devices and field monitoring systems. By coupling accurate pest detection with actionable organic solution recommendations, this work advances sustainable farming practices, offering a practical tool that addresses both environmental responsibility and technological feasibility.