Livestock Monitoring and Analysis: A Deep Learning Approach on Edge Devices

Livestock monitoring is a topic of great interest for livestock farming since it would help identify the most appropriate feed cycle, improve and predict animal health, thus reducing investment cost, and providing other important knowledge to improve the breeding process. This research project proposes the development of a non-invasive monitoring system for livestock using video sequences of groups of pigs as experimental subjects. The system will be able to identify the different pigs, classify the different activities that they perform, and analyze different wounds on the skin to locate the most probable cause of them. The proposed system will be based on artificial vision through the use of a camera with deep learning architectures and machine learning techniques, which will be located in strategic places to monitor the behavior of the animals and their environment.

Faculty Supervisor:

Seok-bum Ko

Student:

Partner:

SASKAI Technology Limited

Discipline:

Engineering

Sector:

Professional, scientific and technical services

University:

University of Saskatchewan

Program: