考虑角度不定的车距视觉自动测量方法

在新一代的智能汽车中,对车距测量结果存在很大的误差,主要是因为没有考虑车辆之间摄像机高度与成角对结果的影响。计算区域过大,导致性能不佳。提出引入兴趣区移动传感网络的车距测量方法,对采集设备的设计进行了详细分析。根据设备确定车辆中的感兴趣区域,在规定的感兴趣区域中,从上到下,依据水平线逐行进行扫描,求出各行灰度的平均值,通过灰度平均值的变化情况对前方车辆进行检测。在此基础上,通过路面上近视场点与CCD摄像机之间的纵向距离,依据针孔模型下的摄像机成像关系求出前方车辆和本车之间的距离。实验结果表明,这种方法对车距进行测量,不仅测量精度高,而且不受摄像机高度和角度的影响,具有很高的鲁棒性。

Distance measurement method based on the mobile sensor network research and design

In the new generation of smart car, made the car distance measurement results on a lot of errors, mainly because of not considering vehicle camera height and Angle between the influence of the results. The calculation area is too large, resulting in poor performance. Proposed introduction of interest in the area of mobile sensor network distance measurement method, the design of acquisition device are analyzed in detail, according to the interested area of the equipment to determine vehicles, in the regulation of interested area, from top to bottom, according to the horizontal line by line scan, marching out all the average gray level, through the change of average gray level to test the vehicles ahead, on this basis, through the pavement myopia sites and the vertical distance between CCD camera, based on the pinhole model of camera imaging and the relationship between the front and the default of the distance between the. Simulation experiment results show that this method of distance measurement, high measuring accuracy, not only and is not affected by the camera height and Angle, has the very high robustness.