Date of Award

Spring 2013

Project Type

Thesis

Program or Major

Electrical Engineering

Degree Name

Master of Science

First Advisor

Richard Messner

Abstract

The main goal of this study is to research image processing methods in attempts to develop a robust approach to image pre-preprocessing of Data Matrix barcode images that will improve barcode read rates in an open source fashion. This is demonstrated by element state classification to re-create the ideal binary matrix corresponding to the intended barcode layout through pattern recognition theory.

The research consisted of implementing and evaluating the effectiveness of many image processing algorithms types, as well as evaluating key features that clearly delineate different element states. The algorithms developed highlight the use of morphological erosion and region growing for object segmentation and edge analysis and Fisher's Linear Discriminant as a means for element classification.

The results demonstrate successful barcode binarization for ideal barcodes with improved read rates in most cases. The techniques developed here provide ground work for a test bed environment to continue improvements by analyzing non-ideal barcodes for additional robustness.

Share

COinS