基于视觉的智能车辆横向偏差测量方法

针对横向偏差传统测量手段的不足,提出了一种智能车辆横向偏差的视觉测量方法.该方法是根据摄像机的成像特点和高等级路面的特征,通过重构路面的空间关系来实现的.对其中的关键环节车载摄像机的标定采用了非线性分步优化的过程,该过程不需要知道不同取向的靶标图像的变动情况,能适应室外道路简便、快速的标定要求.试验表明,所提出的横向偏差测量方法在近视野情况下对于坡度不大的一般道路可以取得厘米级的实时测量精度,能够满足多传感器组合导航对横向偏差测量的要求.

Measurement method of intelligent vehicle's lateral deviation error based on vision

To overcome the shortcomings of traditional measurement methods for lateral deviation error,a new measurement method based on machine vision is proposed.This method is realized by reconstructing the highway spatial relationship according to the camera's imaging characteristic and the highway feature.For the key step of the proposed method,the camera calibration adopts the nonlinear,multi-step and optimized process,which does not need to know the relative positions between different target images and can satisfy simple and feasible calibration demands on the outdoor road.Experimental results demonstrate that the proposed vision measurement method for the lateral deviation error can achieve the cm-level real-time accuracy on the general road with low gradient under the condition of near view,which can satisfy the measurement demands of the multi-sensor integrated navigation system of intelligent vehicles.