Date of Award

2011

Document Type

Thesis - ECU Access Only

Publisher

Edith Cowan University

Degree Name

Doctor of Philosophy

School

School of Engineering

Faculty

Faculty of Computing, Health and Science

First Supervisor

Dr Douglas Chai

Second Supervisor

Dr Alexander Rassau

Abstract

This thesis describes the development and evaluation of a data storage colour 2D barcode known as the MMCC barcode. 1D barcodes are widely used to provide inexpensive and accurate data capture capability. However, due to their low data capacity, they can only function as indexes to back-end databases. The desire for higher data capacity has led to the development of colour 2D barcodes. The recent increasingly ubiquitous availability of camera phones has created a new platform to facilitate this data capture. However, it is challenging to decode colour barcodes as the colour values vary greatly under different lighting conditions. In addition, images captured by mobile phone cameras can be blurry, which can cause the cells to be read at the wrong location. As such, it is even harder to accurately decode these barcodes when a mobile phone is used as the capture device. Hence, in order to work with phone cameras, some recent colour barcodes such as ColorCode and Microsoft Tag use a limited number of colours and cells which trades off data capacity for robustness. This type of barcode can still only function as an index, such as a URL pointing to content on the Internet. Index barcodes do not work all the time as network connectivity is not always available. Hence, the MMCC barcode was developed. To the best of the author’s knowledge, this barcode has a data capacity higher than any of the existing barcodes that can reliably work with camera phones. A number of innovative features including random interleaving, reference colours, appropriate selection of the colours used, alignment cells, adaptive use of multiple colour spaces to classify the cells, and RS coding as the FEC are proposed in this research to improve the robustness of the MMCC barcode. Together with RS coding, random interleaving protects the code against clustered errors by distributing codewords randomly across the barcode. Since the communication channels affect the reference colours and the cells in a similar way, classification of the cells based on the reference colours is significantly more accurate than classification based on the original encoded colour values. It is also observed that the colour pairs of Red-Magenta and Blue-Cyan are close to each other during capturing. Hence, to avoid using the colour pairs at the same time, the MMCC barcode uses a selection of colours that are located on the plane of Red, Black, White, Cyan in the RGB colour space. Border and alignment cells can be used to improve the alignment of the barcode and this thesis investigates their effectiveness in improving the capacity and robustness of the colour 2D barcode. The borders tested were found to be ineffective in the context of high data density barcodes. In the MMCC barcode, alignment cells were shown to improve the decoding of the barcode as each cell is more likely to be read at the correct location. The effects of JPEG artifacts are overcome as the centre of each alignment cell is determined using the number of black pixels that are enclosed within a white border. With the same physical area and a similar level of error correction, the MMCC barcode has been shown to achieve twice the data capacity of the QR code, the most common existing 2D barcode. With this increase, more applications, such as the use of an MMCC barcode as a mobile electronic readable business card, can be developed. With suitable further development, it is thought that the MMCC barcode can enable a range of novel mobile data capture applications.

Share

 
COinS