基于多线激光雷达的道路和障碍物检测
作者单位:北京工业大学
学位级别:硕士
导师姓名:段建民
授予年度:2016年
学科分类:080904[工学-电磁场与微波技术] 0810[工学-信息与通信工程] 0809[工学-电子科学与技术(可授工学、理学学位)] 08[工学] 081105[工学-导航、制导与控制] 081001[工学-通信与信息系统] 081002[工学-信号与信息处理] 0825[工学-航空宇航科学与技术] 0811[工学-控制科学与工程]
主 题:drivable road recognition Theil–Sen estimator multilayer LIDAR k-mean clustering Hough method Сanny edge detection
摘 要:Today there is a huge number of vehicles(cars, trains, planes) in the world. In one hand, it greatly simplifies our daily lives but in the other hand, it poses many problems that must be a solved. Research about vehicle recognition and adaptive cruise control is recently emerging. Many well-known vehicle manufacturers introduce new technologies in their production that gives the opportunity to increase the safety of our lives. Vehicle recognition and adaptive cruise control using various sensory resources. There are many sensors like cameras, GPS, accelerometers, radars etc., that are used for detecting edges of the road, traffic signs, cars or pedestrians. There are many different sensors to access the objects data *** paper presents obstacle road detection. To accomplish this task, from the entire set of available sensors the multilayer LIDAR(Light, Detection and Ranging) with four levels was chosen. This allows obtaining reliable information about the location of objects. The so-obtained data have parameters relative to the coordinate system, the distance of each point from the sensor, which opens up great opportunities for ***, this paper considers how to obtain data from the laser, with detailed consideration of its options. At this stage, the important step is data preprocessing, which includes consideration of only those points that are included in the scan range. The elimination of *** second step is the most important task is how to detect the road and objects on it. The solution to this problem requires the use of a variety of methods. It was viewed a large number of sources in this field and the following data processing methods. It is known that for the detection of the road can be considered a method of determining direct. Today, one of the most widely used methods is the Hough method. Its detailed implementation is proposed in this work. For the detection of objects is regarded, the least squares method. This metho