53 / 2025-04-14 09:06:11

Multi-target dairy cow rumination behavior recognition method based on Yolov11-StrongSort and BiLSTM

Object detection,object tracking,behavior recognition

3. 智慧牧场与动物福利 > （4）智慧牧场应用的案例实践

In the context of expanding breeding operations and advancing information technology, intelligent modernization has become an imperative trend in dairy farming. Automated monitoring of bovine rumination behavior plays a critical role in precision feeding and health management. This study presents a multi-target dairy cow rumination behavior recognition method based on Yolov11-StrongSort and BiLSTM. Our approach employs the Yolov11-StrongSort model for continuous tracking of key rumination-related anatomical features, coupled with a bidirectional long short-term memory network (BiLSTM) for temporal pattern analysis, enabling accurate determination of rumination status and duration calculation. Comparative experiments demonstrate superior performance over conventional motion-based detection methods, achieving over 90% recognition accuracy with temporal measurements closely aligned to ground truth observations. This technological advancement facilitates the advancement of smart dairy farming through enhanced behavioral monitoring capabilities, providing a robust solution for large-scale herd management and precision livestock farming practices.