| 基于多传感融合的有轨电车在途障碍物检测方法研究 |
| |
| 引用本文： | 张勇,王磊,杨峥岭,徐梦.基于多传感融合的有轨电车在途障碍物检测方法研究[J].现代城市轨道交通,2021(2):22-25. |
| |
| 作者姓名： | 张勇 王磊 杨峥岭 徐梦 |
| |
| 作者单位： | 通号万全信号设备有限公司自动化研究院;通号万全信号设备有限公司项目中心 |
| |
| 摘 要： | 有轨电车是一种具有混合路权的轨道交通方式,其行车安全主要依赖于信号系统及司机,因此研发一种能够辅助司机检查轨行区域安全的检测方法显得尤为重要.文章以SSD卷积神经网络的视觉障碍物检测方法为基础,通过多传感融合的方式,在其检测结果上融合激光雷达传感器的障碍物检测结果,实现对轨行区内障碍物的有效识别.实验证明,该方法对于列...
|
| 关 键 词： | 有轨电车 多传感融合 卷积神经网络 SSD 障碍物检测 |
A multi-sensor fusion based tram obstacle detection method |
| |
| Abstract: | Tram is a kind of rail transit mode with mixed right of way,and its driving safety mainly depends on the signaling system and the driver.Therefore,a detection method assisting the driver to check the safety of the track area is particularly important.Based on the SSD convolution neural network based visual obstacle detection method,by using multi-sensor fusion,the obstacle detection results based on lidar sensor are fused in the detection results,so as to realize the effective identification and detect obstacles in the track area.The experimental results show that the method has a satisfactory ability to detect the obstacles on the line,and it is an effective auxiliary method for train operation safety. |
| |
| Keywords: | tram multi-sensor fusion convolutional neural network SSD obstacle detection |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
