Open Positions

Currently available topics for a semester project or Master thesis:

Development of a tracking algorithm for Brown Swiss and Holstein dairy dairy cows in a free-stall barn environment

Monitoring dairy cows is important for evaluating their health and welfare status. This requires the individual identification of cow and their continuous, precise tracking to obtain relevant factors such as lying/standing behaviours on individual level. There's growing interest in developing non-contact, computer vision (CV)-based methods for these tasks. However, recent research on multi-object tracking for cattle has faced challenges, including: (1) Brown Swiss cows appear very similar, making it hard for computer vision algorithms to distinguish between individuals; (2) Cows frequently leave and re-enter the barn for various reasons, such as milking, requiring the re-identification of previously seen individuals when they come back into the camera's field of view.
The goal of this project is to design, develop, and implement a CV-based tracking system that can effectively track Brown Swiss and Holstein cows in a free-stall barn environment. This initiative is focused on tackling the challenges previously outlined. We have collected data on Brown Swiss and Holstein cows from the dairy barn at Agro-Vet Strickhof, providing a solid foundation for this work. While there's no expectation to create an entirely new algorithm from scratch, having prior experience with implementing classic CV algorithms, such as YOLO, would be highly advantageous.
 

Assessment of dairy cow heat stress by behaviour monitoring using computer vision approach

Recognizing and managing heat stress in dairy cows is essential for their welfare and the productivity of dairy farms. Heat stress leads to adverse effects such as decreased milk yield, fertility issues, and increased health complications. Current assessment methods, like the temperature-humidity index, cannot accurately capture the comprehensive effects of heat stress. This project is focused on developing a computer vision-based approach to monitor behaviour changes in dairy cow, such as lying and standing patterns, which are indicative of heat stress. By leveraging advanced computer vision techniques, we aim to offer a detailed understanding of how heat stress affects dairy cow behavior in different environmental conditions. Data for this project were collected across different temperature and humidity levels at the Agro-Vet Strickhof dairy barn. While extensive experience in computer vision isn't required, a foundational knowledge in animal science, coupled with an enthusiasm for applying state-of-the-art computer vision algorithms in animal science, would be valuable.