本文为瑞典林雪平大学（作者：FredrikNilsson）的硕士论文，共82页。

高级驾驶员辅助系统（ADAS）是当前研究和开发的热门领域。由于90%的交通事故是由人为因素造成的，因此通过向车辆驾驶员提供必要的帮助，ADAS可以显著减少交通事故的数量。带有相机的ADAS提供了广阔的观察视野，并且由于当今先进的图像处理技术，可以从相机图像中提取出大量信息。

本文提出了一种利用单摄像机系统估计车辆航向的方法。该方法采用等速运动模型的扩展卡尔曼滤波器实现，通过机器学习算法的分类测量和角速率测量来预测车辆运动路径。本文所进行的蒙特卡洛模拟结果显示出良好的效果。对实际数据的研究表明，为了得到与仿真相同的结果，必须改进用于构造角速率测量的方法。为了进一步向滤波器提供有用的信息，引入了附加的测量，即车辆转角。研究表明，为了达到与立体相机系统相同的性能水平，单目相机系统需要进一步改进。

Advanced driver assistance systems (adas)is a popular and evolving area of research and development. By providingassistance to the vehicle drivers, adas could significantly reduce the numberof traffic accidents since 90 % of all accidents are caused by the humanfactor. adas with cameras provides a wide field of view and thanks to today’sadvanced image processing techniques, lots of information can be extracted fromthe camera image. This thesis proposes a method of estimating the heading ofvehicles using a mono camera system. The method consists of an extended Kalmanfilter with a constant velocity motion model to predict the vehicle’s path, fedby classification measurements from machine learning algorithms together withangular rate measurements. Monte Carlosimulations performed in this thesis show promising results. The results onreal-world data indicate that the method used to construct the angular ratemeasurements must be improved in order to reach the same results as obtainedfrom the simulations. An additional measurement, the vehicle’s corners, isintroduced in order to further provide the filter with information. The thesisshows that the mono camera system needs further improvements in order to reachthe same level of performance as a stereo camera system.

1 引言

2 理论背景

3 滤波器构建与方法论

4 性能评估

5 结论与未来工作展望

附录A 蒙特卡洛仿真

下载英文原文地址：

http://page5.dfpan.com/fs/5lcjb22152916630645/

更多精彩文章请关注微信号：