Vision-based detection technique for effective line-tracking autonomus vehicle

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.