| From 1980s on, many countries have developed the research on mobile robottechnology. Visual aided navigation is one of the hot spot on mobile robot navigation. Androad obstacle detection and visual positioning navigation system is an importantcomponent in visual aided navigation system. In this thesis, the obstacle detection and roadpositioning algorithm based on image are researched.At first the components of AS-R mobile robot and visual system are introduced, thenthe relationship between target coordination in robot and computer mapping coordinationare investigated. The current algorithms of image preprocessing are analyzed. By thespecific requirements of the image processing in this thesis, the appropriate algorithm ischosen.In order to improve the real-time performance of roads and anti-jamming capability,edge extraction and integration of regional growth ways is used to identify the road edge.The edge line expressed according to marginal and regional information should occupyspace in the same or similar position, corresponding information on the location of discretepoints are calculated, and then the robot location in roads is determined. In the algorithm,including initial detection and tracking-detection algorithms, straight road model isconsidered as the constraints, and Hough transform is used for linear feature extraction. Byusing the features between image-sequence roads in the border tracking algorithm, theinteresting regions for restricting the seizure regions are established. And the real-timeperformance of algorithms can be improved. |
PDF Full Text Request