City Research Online

Abstract

This report is a study of the role of colour and its use in machine vision. Intuitively, for low level image processing, colour provides greater discrimination than grey level for separating different homogenous regions in an image. The first part describes the use of colour in edge detection. In current vision systems, extracting object features such as lines and arcs relies on edge detection. The use of colour images (RGB) can offer additional confidence in the existence of an edge element in one plane when it is corroborated by pixels at the location on one or more of the other planes. The second part investigates the problems and techniques associated with colour image segmentation. A spectral segmentation algorithm based on locating the boundaries of each colour cluster in the spectral space is proposed. The third part investigates the use of colour features for object recognition. Colour information also provides a useful cue for object localisation and identification. The major issues that have to be addressed are colour constancy and representation, and also, their connections to segmentation. Finally, a system for locating object surfaces based on a simplified colour constancy and its colour representation is proposed.

Publication Type:	Thesis (Doctoral)
Subjects:	Q Science > QA Mathematics T Technology > TA Engineering (General). Civil engineering (General)
Departments:	School of Science & Technology > Engineering > Electrical & Electronic Engineering School of Science & Technology > School of Science & Technology Doctoral Theses Doctoral Theses

Export

Downloads