Vision-based detection technique for effective line-tracking
autonomus vehicle
Abstract:
An alternative means of implementing line detection technique
to navigate autonomous mobile vehicles was investigated, specifically using
vision sensor as the front end detection device. A vision sensor has the
advantages of much higher spatial and optical resolutions as compared to the
conventional discrete photo-reflective sensors used in most line-following
mobile vehicles. It could offer better detection accuracy and reliability in
varying lighting conditions. A selective video line technique was introduced
and implemented which resulted in reduced number of video lines needs to be
processed. Hence, avoiding the needs to use high-end processor to process and
analyze the vast amount of information from the video images. There was enough
space of information from each video line signal to be manipulated and could
easily and flexibly calibrated with several lighting conditions. Abrupt changes
in ambient lighting can be overcome without affecting the accuracy of the
mobile vehicle navigation. A small inexpensive analog video camera, with
standard video output signal had been chosen with a decent 8-bit microcontroller
to process the video line signals and navigate the mobile vehicle.
0 comments:
Post a Comment