本文为瑞典查尔姆斯理工大学（作者：PATRIKNYGREN and MICHAEL JASINSKI）的硕士论文，共54页。

主动安全已经成为当前汽车工业的一个重要组成部分，因为它被证明有潜力使驾驶更快乐，同时减少事故和伤亡。主动安全系统中使用不同的传感器来感知环境并实现驾驶员辅助和防撞系统。光探测和测距（LIDAR）传感器是这些系统中常用的传感器之一，激光雷达从周围环境中产生点云，可以用来探测和分类汽车、行人等物体。在本论文中，我们对几种方法进行了比较研究。对城市环境中的点云（支持向量机、前向神经网络、随机森林和k近邻）进行分类和评价。KITTI数据库中的数据用于验证使用PCL和Shark库实现的方法。我们评估了分类方法在两个不同的已开发特征集上的性能。实验表明，用支持向量机对7类不同对象的数据集进行处理，其最高精度可达96.3%。

Active Safety has become an important partof the current automotive industry due to its proven potential in makingdriving more joyful and reducing number of accidents and causalities. Differentsensors are used in the active safety systems to perceive the environment andimplement driver assistance and collision avoidance systems. Light detectionand ranging (LIDAR) sensors are among the commonly utilized sensors in thesesystems; a LIDAR produces a point cloud from the surrounding and can be used todetect and classify objects such as cars, pedestrians, etc. In this thesis, weperform a comparative study where several methods to both segment RegionGrowing and Euclidian Clustering) and classify (Support Vector Machines, FeedForward Neural Networks, Random Forests and K-Nearest Neighbors) point cloudsfrom an urban environment are evaluated. Data from the KITTI database is usedto validate the methods which are implemented using the PCL and Shark library.We evaluate the performance of the classification methods on two different setsof developed features. Our experiments show that the best accuracy can beobtained using SVMs, which is around 96.3% on the considered data set with 7different classes of objects.

1 引言

2 项目背景

3 相关理论

4 具体实现

5 结果

6 讨论

7 结论

更多精彩文章请关注公众号：