The Method of Detecting Nearest Distance between Obstacles and Vehicle Tail Based on Binocular Vision System

摘要

In order to provide auxiliary information for car parking, a strategy of detecting the nearest distance between obstacle and vehicle tail was studied based on the binocular vision system and a parking assist system was designed. The captured images processed and analyzed by using binocular vision system, the image distortion was improved by using the histogram correction method, and the influence of noise in space was eliminated by using GAUSS FILTER. Harris algorithm was used to detect the characteristic points accurately, and the feature points were matched. A kind of strategy about detecting the recent distance between obstacles and the rear of vehicle was put forward. When fitting the curve, the curve of car tail was instead by polygon approximation. The detection of closest distance between obstacles and the rear of the vehicle was accomplished by using the binocular vision principle. The distance could be used to provide useful parking information to drivers.