Degree Type

Dissertation

Date of Award

2003

Degree Name

Doctor of Philosophy

Department

Agricultural and Biosystems Engineering

First Advisor

Hongwei Xin

Second Advisor

Julie A. Dickerson

Abstract

A real-time bioimaging system was developed to detect movement of group-housed pigs and to classify thermal comfort state of the pigs based on their resting patterns. This dissertation describes the background, theory, image features selection and analysis, image classification techniques, system architecture, and system performance with artificial and live pigs.;The system is based on the bioimaging control method: using animal's thermal behavior as the ultimate biosensor that is used to reflect animal thermal comfortable levels. An automated real time image processing and control system is developed that monitors animal behavior and adjusts external devices according to the image classification results.;The system contains three major components: image capture component that grabs live animal images for analysis and display purposes; image processing component that detects animal movement and analyzes its thermal behaviors; and external device control component that adjusts environmental temperature to keep the temperature within the thermoneutral zone (TNZ). The system operates on a real-time basis with multiple input channels. It also displays live animal images and system control states for selected channels.;The system contains a rich set of image features, including image moment invariants, run-length frequency, animal body occupation ratio, and animal group shape compactness that are extracted as feature vectors. Different image classification methods were also investigated and a minimum distance was adopted as the measurement to distinguish the thermal comfort states of cold vs. comfortable. The system was designed to be body weight independent, and a sliding window was employed to update reference image feature sets so that classification is always based on the most recent information.;Rather than providing an academic prototype, the system is commercial application oriented. It provides real-time configurations so that users can modify system variables such as animal group size, image threshold value, channel selection and temperature set points on the run.

DOI

https://doi.org/10.31274/rtd-180813-13206

Publisher

Digital Repository @ Iowa State University, http://lib.dr.iastate.edu

Copyright Owner

Bin Shao

Language

en

Proquest ID

AAI3143532

File Format

application/pdf

File Size

73 pages

Share

COinS