期刊文献+

雨雾环境下基于激光雷达的复杂运动目标检测方法

Complex moving target detection method based on LiDAR in rainy and foggy environment
在线阅读 下载PDF
导出
摘要 为提升雨雾条件下复杂运动目标识别准确性,提出基于激光雷达的复杂运动目标检测方法。通过高频发射的激光束采集待检测区域点云数据,使用特征加权下模糊C均值聚类方法区分正常数据和噪声数据,运用双边滤波方法平滑点云数据噪声,利用K最邻近方法计算点云局部邻域信息,采用自注意力机制实现激光点云分割,通过点云分割获得不同运动目标的激光点云特征,通过输入域、非线性调制域和脉冲产生器完成脉冲耦合神经网络训练,通过平均运动速度计算候选目标范围,精准锁定复杂运动目标。实验结果证明,所提方法在雨雾等恶劣天气下,复杂运动目标检测F1值超过0.8,目标检测精度高,具备极强的鲁棒性。 To improve the accuracy of complex moving target recognition under rainy and foggy conditions,a com-plex moving target detection method based on LiDAR is proposed.By using high-frequency laser beams to collect point cloud data in the area to be detected,the fuzzy C-means clustering method under feature weighting is used to distin-guish between normal data and noisy data.The bilateral filtering method is used to smooth out the noise in the point cloud data.The K-nearest neighbor method is used to calculate the local neighborhood information of the point cloud.The self attention mechanism is used to achieve laser point cloud segmentation.The laser point cloud features of differ-ent moving targets are obtained through point cloud segmentation.The pulse coupled neural network is trained through the input domain,nonlinear modulation domain,and pulse generator.The candidate target range is calculated by the average motion speed to accurately lock in complex moving targets.The experimental results demonstrate that the pro-posed method has an F1 value of over 0.8 for complex motion target detection in adverse weather conditions such as rain and fog,with high target detection accuracy and strong robustness.
作者 颉琦 李瀚涛 董成伟 JIE Qi;LI Hantao;DONG Chengwei(School of Semiconductor and Physics,North University of China,Taiyuan 030051,China)
出处 《激光杂志》 北大核心 2025年第8期60-64,共5页 Laser Journal
基金 国家自然科学基金青年项目(No.12205257) 山西省基础研究计划面上项目(No.202203021221095) 山西省基础研究计划青年项目(No.20210302124025) 山西省教育科学“十四五”规划课题(No.GH-240489)。
关键词 雨雾环境 激光雷达 目标检测 点云数据分割 脉冲耦合神经网络 rainy and foggy environment lidar object detection point cloud data segmentation pulse coupled neural network
作者简介 颉琦(1986-),男,博士,讲师,研究方向:半导体光学、非线性光学;李瀚涛(1988-),男,博士,副教授,研究方向:原子核贝塔衰变;董成伟(1985-),男,博士,教授,研究方向:非线性物理,复杂系统。
  • 相关文献

参考文献18

二级参考文献124

共引文献138

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部