Optimizing Cattle Behavior Analysis in Precision Livestock Farming: Integrating YOLOv7-E6E with AutoAugment and GridMask to Enhance Detection Accuracy

被引:1
作者
Sim, Hyeon-seok [1 ]
Kim, Tae-kyeong [1 ]
Lee, Chang-woo [2 ]
Choi, Chang-sik [2 ]
Kim, Jin Soo [3 ]
Cho, Hyun-chong [1 ,4 ]
机构
[1] Kangwon Natl Univ, Dept Grad Program BIT Med Convergence, Chunchon 24341, South Korea
[2] Gangwon State Livestock Res Inst, Hoengseong 25266, South Korea
[3] Kangwon Natl Univ, Coll Anim Life Sci, Chunchon 24341, South Korea
[4] Kangwon Natl Univ, Dept Elect Engn, Chunchon 24341, South Korea
来源
APPLIED SCIENCES-BASEL | 2024年 / 14卷 / 09期
基金
新加坡国家研究基金会;
关键词
AutoAugment; cattle behavior; deep learning; GridMask; precision livestock farming; object detection; YOLOv7-E6E;
D O I
10.3390/app14093667
中图分类号
O6 [化学];
学科分类号
0703 ;
摘要
Recently, the growing demand for meat has increased interest in precision livestock farming (PLF), wherein monitoring livestock behavior is crucial for assessing animal health. We introduce a novel cattle behavior detection model that leverages data from 2D RGB cameras. It primarily employs you only look once (YOLO)v7-E6E, which is a real-time object detection framework renowned for its efficiency across various applications. Notably, the proposed model enhances network performance without incurring additional inference costs. We primarily focused on performance enhancement and evaluation of the model by integrating AutoAugment and GridMask to augment the original dataset. AutoAugment, a reinforcement learning algorithm, was employed to determine the most effective data augmentation policy. Concurrently, we applied GridMask, a novel data augmentation technique that systematically eliminates square regions in a grid pattern to improve model robustness. Our results revealed that when trained on the original dataset, the model achieved a mean average precision (mAP) of 88.2%, which increased by 2.9% after applying AutoAugment. The performance was further improved by combining AutoAugment and GridMask, resulting in a notable 4.8% increase in the mAP, thereby achieving a final mAP of 93.0%. This demonstrates the efficacy of these augmentation strategies in improving cattle behavior detection for PLF.
引用
收藏
页数:13
相关论文
共 32 条
  • [1] Image analysis for individual identification and feeding behaviour monitoring of dairy cows based on Convolutional Neural Networks (CNN)
    Achour, Brahim
    Belkadi, Malika
    Filali, Idir
    Laghrouche, Mourad
    Lahdir, Mourad
    [J]. BIOSYSTEMS ENGINEERING, 2020, 198 : 31 - 49
  • [2] [Anonymous], 2016, CORDIS Final Report Summary-EU-PLF (Bright Farm by Precision Livestock Farming)
  • [3] Review: Precision Livestock Farming technologies in pasture-based livestock systems
    Aquilani, C.
    Confessore, A.
    Bozzi, R.
    Sirtori, F.
    Pugliese, C.
    [J]. ANIMAL, 2022, 16 (01)
  • [4] General introduction to precision livestock farming
    Berckmans, D.
    [J]. ANIMAL FRONTIERS, 2017, 7 (01) : 6 - 11
  • [5] Detection of aggressive behaviours in pigs using a RealSence depth sensor
    Chen, Chen
    Zhu, Weixing
    Liu, Dong
    Steibel, Juan
    Siegford, Janice
    Wurtz, Kaitlin
    Han, Junjie
    Norton, Tomas
    [J]. COMPUTERS AND ELECTRONICS IN AGRICULTURE, 2019, 166
  • [6] Chen PG, 2024, Arxiv, DOI [arXiv:2001.04086, DOI 10.48550/ARXIV.2001.04086]
  • [7] Socioeconomic and demographic drivers of red and processed meat consumption: implications for health and environmental sustainability
    Clonan, Angie
    Roberts, Katharine E.
    Holdsworth, Michelle
    [J]. PROCEEDINGS OF THE NUTRITION SOCIETY, 2016, 75 (03) : 367 - 373
  • [8] The Ethics of Touch and the Importance of Nonhuman Relationships in Animal Agriculture
    Cooke, Steve
    [J]. JOURNAL OF AGRICULTURAL & ENVIRONMENTAL ETHICS, 2021, 34 (02)
  • [9] AutoAugment: Learning Augmentation Strategies from Data
    Cubuk, Ekin D.
    Zoph, Barret
    Mane, Dandelion
    Vasudevan, Vijay
    Le, Quoc V.
    [J]. 2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019), 2019, : 113 - 123
  • [10] Deng J, 2009, PROC CVPR IEEE, P248, DOI 10.1109/CVPRW.2009.5206848