Research on Computer Vision Based Assistant Navigation and Location Technique for Vehicle

摘要

A lot of vehicles in troops are required to locate precisely and it is not convenient to operate due to the over-length and over-width of vehicle body. The underlying problem of the current methods for augmented vision is that the actual scene is not as big as that in image, so the precision of navigation and location for vehicle is low. It is necessary to develop a set of vision augment system with convenient and reliable operation so as to assist the driver to locate the vehicle quickly. A new navigation and location method based on computer vision is presented in the paper. By utilizing the characteristic of navigation ruler with higher gray value than outside environment, the approximate position of the ruler can be detected. At last, the exact position is got by Hough Transform in the vicinal region of the approximate position. The required information needed for vehicle navigation and location has been analyzed in the article. Based on the line of the detected navigation gauge and navigation symbols on the gauge, the information of navigation and location can be calculated such as the deflection angle of the vehicle, the distance between the vehicle and the target, transverse deviation of the vehicle. Experiments show that the method makes system configuration simple and easy to achieve, and has good real-time performance.