Position localization is a necessary and important task for an intelli
gent robot, which moves and works in place of human workers. Various m
ethods for position localization tasks have been proposed and develope
d by many researchers; for example, can active method using ultrasonic
waves or laser and a passive method using images of bar codes or mark
s in the workspace of the robot. The passive method, using images of t
he object in the workspace of the robot, is particularly desirable, as
this method does not influence the workspace. Thus, an arrangement of
the workspace is not necessary and many robots can work simultaneousl
y. This paper describes the development of a position localization pro
gram for a mobile robot, using a colour image of an object in the nucl
ear plant workspace. The developed program finds edges and vertices in
nuclear plants as feature points for identifying the position of the
mobile robot. Colour information is utilized effectively for detecting
edges and extracting a necessary region.