• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>International Conference on Electronics Technology >An Efficient Multi-threshold Selection Method for Lane Detection Based on LiDAR
【24h】

An Efficient Multi-threshold Selection Method for Lane Detection Based on LiDAR

机译：基于LIDAR的车道检测有效的多阈值选择方法

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

A premium transportation sign identification technique contributes a lot to the environmental perception, and real-time processing guarantees safer driving. This paper proposes an efficient lane marking extraction method by using a 16-beam light detection and ranging sensor in bad illumination conditions. A road surface point cloud is extracted, and background points are eliminated by applying the Random Sample Consensus segmentation method with a plane model. Then, a multi-threshold selection method is developed to recognize lane marking points with high echo intensities from a relatively large cluster instead of transforming the point cloud to a 2D grid, thus the candidate point cloud representing traffic lines can be determined. Additionally, a line model is carried out to further improve the detection precision. Performance of the proposed method is evaluated by Precision, Recall, F-score, and Accuracy through using samples from two datasets. Compared with other approaches, this method drastically reduces the processing time of lane identification and in the meantime, is capable of offering accurate results.
机译：高级运输标志识别技术对环境感知有很大贡献，实时处理保证更安全的驾驶。本文通过在不良照明条件下使用16梁光检测和测距传感器提出了一种高效的车道标记提取方法。提取道路表面点云，通过使用平面模型应用随机样本共识分割方法来消除背景点。然后，开发了一种多阈值选择方法以识别具有来自相对大的集群的高回波强度的车道标记点，而不是将点云转换为2D网格，因此可以确定表示交通线的候选点云。另外，执行线模型以进一步提高检测精度。通过使用来自两个数据集的样本来评估所提出的方法的性能，通过使用来自两个数据集的样本来评估。与其他方法相比，该方法大大降低了车道识别的处理时间，并且在此期间能够提供准确的结果。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号