Theses and Dissertations

ORCID

https://orcid.org/0000-0002-0104-0317

Issuing Body

Mississippi State University

Advisor

Zhang, Xin

Committee Member

Lu, Yuzhen

Committee Member

Sukumaran, Anuraj T.

Committee Member

Wijewardane, Nuwan

Committee Member

Chen, Jingdao

Other Advisors or Committee Members

Lowe, John Wes

Date of Degree

8-8-2023

Document Type

Graduate Thesis - Open Access

Major

Biological Engineering

Degree Name

Master of Science (M.S.)

College

James Worth Bagley College of Engineering

Department

Department of Agricultural and Biological Engineering

Abstract

Manual inspection is a prevailing practice for quality assessment of poultry meat, but it is labor-intensive, tedious, and subjective. This thesis aims to assess the efficacy of an emerging structured illumination reflectance imaging (SIRI) technique with machine learning approaches for assessing WS and microbial spoilage in broiler breast meat. Broiler breast meat samples were imaged by an in house-assembled SIRI platform under sinusoidal illumination. In first experiment, handcrafted texture features were extracted from direct component (DC, corresponding to conventional uniform illumination) and amplitude component (AC, unique to the use of sinusoidal illumination) images retrieved from raw SIRI pattern images build linear discriminant analysis (LDA) models for classifying normal and defective samples. A further validation experiment was performed using deep learning as a feature extractor followed by LDA. The third experiment was on microbial spoilage assessment of broiler meat, deep learning models were used to extract features from DC and AC images builds on classifiers. Overall, this research has demonstrated consistent improvements of AC over DC images in assessing WS and spoilage of broiler meat and that SIRI is a promising tool for poultry meat quality detection.

Sponsorship

U.S. Department of Agriculture, National Institute of Food and Agriculture [Grant No. 2022-67018-36625] and the Mississippi Agricultural and Forestry Experimentation Special Research Initiative

Share

COinS