摘要: |
【目的】为了在生猪养殖过程中生猪更加快速精准地监测其异常行为规律,避免由人工观察所产生的耗时、费力、主观性、随意性等缺点。【方法】本文通过查阅大量的相关文献,并使用文献综述法和对比分析法围绕对生猪的异常行为监测展开梳理。【结果】通过对诸多文献的概括分析可以看出,研究主要集中在猪的分割与行为识别这几个方面,找出了当前研究中存在的问题,并提出了解决这些问题的未来研究工作设想,提出了几点针对性建议,作为后续学者研究的参考方向。【结论】本文总结基于计算机视频监控、音频技术和传感器等技术与图像处理和深度学习等方法相结合来监测猪的行为的发展过程,深度学习逐渐被应用在动物行为识别方面,虽然关于生猪行为监测的技术已经取得了良好成果,但是在提高监测技术的实时准确性方面还有待提高。 |
关键词: 生猪 行为 视频技术 传感器 监测 |
DOI: |
分类号:S274.2 |
基金项目:黑龙江农垦总局项目(HKKY190201-02);大庆指导性科技计划项目(zd-2021-82);校青年创新人才项目(CXRC2017014);校启动计划(XDB201813)。 |
|
Research on pig behavior monitoring based on fuzzy reasoning |
Li Qi, Li Aichuan
|
College of Information and Electrical Engineering,Heilongjiang Bayi Agricultural University
|
Abstract: |
[Purpose] The purpose is to monitor the abnormal behavior of pigs more quickly and accurately in the process of pig breeding, so as to avoid the shortcomings of time-consuming, laborious, subjectivity and randomness caused by manual observation. [Method] By consulting a large number of relevant literature and using literature review and comparative analysis, this paper combs the monitoring of abnormal behavior of pigs. [Result] Through the summary and analysis of many literature, it can be seen that the research mainly focuses on the segmentation and behavior recognition of pigs, finds out the problems existing in the current research, puts forward the ideas of future research to solve these problems, and puts forward some targeted suggestions as the reference direction of follow-up scholars' research. [Conclusion] This paper summarizes the development process of monitoring pig behavior based on the combination of computer video monitoring, audio technology and sensor technology with image processing and deep learning. Deep learning is gradually applied to animal behavior recognition. Although the technology of pig behavior monitoring has achieved good results, it needs to be improved in improving the real-time accuracy of monitoring technology. |
Key words: pigs behavior video technology sensor monitoring |