Date of Award

Fall 2014

Degree Type

Thesis

Degree Name

Master of Science (MS)

Department

Mechanical Engineering

First Advisor

Xinyan Deng

Committee Member 1

Justin Seipel

Committee Member 2

Reuben Goforth

Abstract

The lionfish is an invasive species that out-competes and overcrowds native sh species along the eastern seaboard of the United States and down into the Caribbean. Lionfish populations are growing rapidly. Current methods of monitoring lionfish populations are costly and time intensive. A bio-inspired robotic fish was built to use as an autonomous lionfish tracking platform. Lionfish are tracked visually using an onboard processor. Five different computer vision methods for identification and tracking are proposed and discussed. These include: background subtraction, color tracking, mixture of Gaussian background subtraction, speeded up robust feature (SURF), and CamShift based tracking. Each of these methods were compared and their accuracy analyzed. CamShift based tracking is determined to be the most accurate for this application. Preliminary experiments for system identification and control design are discussed.

Download

Share

COinS