Date of Graduation

5-2023

Document Type

Thesis

Degree Name

Bachelor of Science in Computer Science

Degree Level

Undergraduate

Department

Computer Science and Computer Engineering

Advisor/Mentor

Le, Thi

Committee Member/Reader

Gauch, John

Committee Member/Second Reader

Zhang, Lu

Abstract

Poultry is an important food source across the world. To facilitate the growth of the global population, we must also improve methods to oversee poultry with new and emerging technologies to improve the efficiency of poultry farms as well as the welfare of the birds. The technology we explore is Deep Learning methods and Computer Vision to help automate chicken monitoring using technologies such as Mask R-CNN to detect the posture of the chicken from an RGB camera. We use Meta Research's Detectron 2 to implement the Mask R-CNN model to train on our dataset created on videos of chickens in a controlled environment. We include the numeric results from different training sessions of varying datasets to showcase the improvement in the model over time. Our findings show that Deep Learning and Computer Vision technologies can effectively enhance poultry farming, and we believe that our study can serve as a foundation for future research in this field.

Keywords

Keypoint; Estimation; Computer; Vision

Share

COinS