University of Illinois Urbana-Champaign

Human-assisted high throughput livestock tracking using computer vision

Senthil, Pradeep

Content Files
SENTHIL-THESIS-2023.pdf

Permalink

https://hdl.handle.net/2142/120131

Description

Title
Human-assisted high throughput livestock tracking using computer vision
Author(s)
Senthil, Pradeep
Issue Date
2023-05-01
Director of Research (if dissertation) or Advisor (if thesis)
  • Caesar, Matthew
  • Dilger, Ryan N
Department of Study
Computer Science
Discipline
Computer Science
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
M.S.
Degree Level
Thesis
Keyword(s)
  • Pig Monitoring
  • Livestock
  • Precision Livestock Farming
  • Object Detection
  • Computer Vision
  • Annotation System
  • Data Management
Abstract
As Precision Livestock Farming (PLF) gains prominence in the livestock industry, the need for efficient data synthesis has grown significantly, posing new challenges. This thesis focuses on the application of novel systems using computer vision to enhance the efficiency of livestock monitoring and accelerate research in the domain. Firstly, we introduce a robust annotation system designed to minimize the time required for annotators of any skill level to process large volumes of raw video data while catering to the needs of both Animal Science and Computer Science researchers. We implement these system designs in an open-source tool called the Animal Video Annotation Tool (AVAT). Secondly, we address a critical challenge in analyzing large amounts of pig video research data. While significant progress has been made in tracking pigs across pens, most existing methods primarily focus on commercial farming applications, neglecting essential research aspects. To bridge this gap, we propose a high-throughput video analysis pipeline that tracks pig movement over several weeks, offering insights into the movement patterns over time. These insights facilitate understanding behavioral changes resulting from varying stimuli and treatment plans, leading to the development of improved nutritional strategies and practices that promote pig health and welfare. To evaluate our pipeline, we analyzed 13TB of video data collected from a live animal study, allowing us to observe movement trends across time and empowering animal science researchers to draw conclusions about pig behavior in response to their treatment plans.
Graduation Semester
2023-05
Type of Resource
Thesis
Handle URL
https://hdl.handle.net/2142/120131
Copyright and License Information
Copyright 2023 Pradeep Senthil

Owning Collections

Graduate Dissertations and Theses at Illinois PRIMARY
Graduate Theses and Dissertations at Illinois