The Machine Vision Unit researches the transformation of
raw signals into a symbolic representation, whether through initial
investigation, an intermediate stage or (mainly) a final interpretation of
the signal data into human concepts. Much of the research uses
distance data acquired from a range finder.
As an academic group we are privileged in being able to
research almost any interesting vision-related topics using a variety of
raw data such as intensity images, range data, sonar, or stereo
images. However, our major focus is the analysis of range data. We
have been involved in all stages of this analysis since 1979.
About half our current research involves high-level
analysis of range data, mainly reverse engineering of CAD models and
identification of articulated models.
Other current projects include
the automatic placement of cameras for generating virtual reality
environments, and iconic vision in 2D.
A visit to our past research
pages will give you a taste of some our our favourite past successes
which contributed to the high-quailty track record of this research