|
| 基于多相机协同的大鼠行为分析系统 |
| Multi-camera collaborative tracking system for rat behavioral analysis |
| 投稿时间：2025-03-03 |
| DOI：10.3969/j.issn.1005-5630.202503030041 |
| 中文关键词: 定量分析 多相机协同 神经网络 物体遮挡 透视变换 |
| 英文关键词:quantitative analysis multi-camera collaborative neural network object occlusion perspective transformation |
| 基金项目: |
|
| 摘要点击次数: 0 |
| 全文下载次数: 0 |
| 中文摘要: |
| 动物行为的定量分析对生物医学和神经科学领域至关重要。然而，大鼠具有复杂的互动模式和行为，如追逐、探索、寻觅和交织运动，这使得在复杂环境中监测多只大鼠变得尤为困难。本文提出了一种基于多相机协同的追踪系统，以突破单摄像头的限制问题，如物体遮挡、外观特征模糊和视场限制。该系统采用级联神经网络，结合透视变换、监控重叠区域筛选和多摄像头数据融合技术，全面研究大鼠的行为。该系统可推广应用于动物行为学、药理毒理学和安全监控等领域。 |
| 英文摘要: |
| Quantitative analysis of animal behaviour is crucial for biomedical and neuroscience fields. However, rats have complex interaction patterns and behaviours such as chasing, exploring, seeking and intertwined motion which make monitoring multiple rats in a complex environment particularly challenging. A multi-camera collaborative tracking system is proposed in this paper to break through the limitation of single-camera tracking such as object occlusion, appearance feature ambiguities and field-of-view limitation of single camera. The system adopts cascaded neural network, combined with perspective transformation, surveillance overlapping area screening and multi-camera data fusion techniques to comprehensively studying the behaviours of the rats. We prospect that the system can be applied in the fields of animal behaviour research, pharmacological and toxicological research and security surveillance. |
| HTML 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
