Summary: | 碩士 === 國立中興大學 === 電機工程學系 === 92 === In this paper, a computer vision system is developed for the guidance of the unmanned vehicle in corridor environments. The boundaries are first detected based on the camera imaging and pattern recognition. Then the guidance of the unmanned vehicle can be guaranteed through the calculations of the vanish point and inverse perspective transformation.
In order to overcome the discontinuous edges in an image, the Radon transformation and the weighted averages method are adopted to form the continuous edges. With the ratio of the inverse perspective transformation to the edge coordinate, and the known width of the corridors, we can get the position of the unmanned vehicle and the distance between the unmanned vehicle and the walls.
In our experiments, by the use of the characteristics of the parallels of the obtained edges, we can get sufficient image features for the guidance and operation mode of the unmanned vehicle.
|