Scholarlink

To tackle challenges like small distant targets and occlusion in factory security detection, this paper proposes a YOLOX-based algorithm combining a gradient flow module and coordinate attention. The method enhances fine-grained feature extraction and global context modeling. A BiFPN structure is used for effective multi-scale feature fusion. Experiments in real factory scenarios show that the proposed method achieves 94.78% accuracy and outperforms mainstream detection algorithms in performance and robustness.

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

An Intelligent Detection Method for Factory Security Risks by Fusing Gradient Flow Features

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Fuan Dong,
  • Xiaoning Wu,
  • Dong Gao,
  • Ruixin Li,
  • Li Sun,
  • Pengyu Liu,
  • Lele Yuan

摘要

To tackle challenges like small distant targets and occlusion in factory security detection, this paper proposes a YOLOX-based algorithm combining a gradient flow module and coordinate attention. The method enhances fine-grained feature extraction and global context modeling. A BiFPN structure is used for effective multi-scale feature fusion. Experiments in real factory scenarios show that the proposed method achieves 94.78% accuracy and outperforms mainstream detection algorithms in performance and robustness.