基于多刚体三维图像检测的车辆短距离拐弯碰撞预判

Vehicle Collision Anticipation Short Turn Based on Multi-rigid-body Threedimensional Image Detection

对车辆短距离拐弯碰撞进行智能视觉预判,可有效的减少车辆发生车祸的次数,具有一定的实用性。提出基于多刚体三维图像检测的车辆短距离拐弯碰撞预判方法。依据车辆短距离拐弯碰撞的特点,将车辆短距离拐弯碰撞问题,转化成车辆与对面物体是否相交的问题,构建相应的碰撞模型,引入非刚体三维图像检测算法,获取模型最优解,实现车辆短距离拐弯碰撞的预判。实验结果表明,改进的预判算法有效的增强了车辆短距离拐弯碰撞预判的准确率,降低了车辆短距离拐弯发生碰撞的概率,具有一定的实用性。

Intelligent visual prognosis of short corner vehicle collision, which can effectively reduce thenumber of vehicles in a car accident, has a certain practicality. Based on multi- rigid- body threedimensionalimage detection of vehicle collision forecasting methods short turn. The characteristics ofthe short distance curve based on vehicle collision, the vehicle short corner collision problem, into avehicle and the question of whether the object intersection opposite, build corresponding collision model,the introduction of three- dimensional image non- rigid detection algorithm, to obtain model of theoptimal solution, realize vehicle collision the anticipation of short turn. Experimental results show thatthe improved forecasting algorithm effectively enhances the short corner vehicle collision forecastingaccuracy, reduce the probability of collision vehicle short turn, has a certain practicality.