基于點云密度的結(jié)構(gòu)化道路邊界增強檢測方法
發(fā)布時間:2018-11-10 19:39
【摘要】:為快速魯棒地檢測結(jié)構(gòu)化道路邊界,提出一種基于HDL-64E激光雷達點云密度的道路邊界增強檢測方法。通過建立虛擬雷達模型,利用點云密度特征,實現(xiàn)前景與背景分離,并利用隨機采樣一致性算法得到20m內(nèi)的道路邊界。為解決20~100m內(nèi)道路邊界點云稀疏、檢測準確性下降的問題,提出利用光線切割模型對道路邊界進行增強檢測。在校園道路和城市快速路進行實驗,道路邊界檢測率達到95%以上,有效檢測距離可達70m以上,檢測周期小于32ms。
[Abstract]:In order to quickly and robustly detect structured road boundary, a method of road boundary enhancement detection based on point cloud density of HDL-64E lidar is proposed. Based on the virtual radar model and the feature of point cloud density, the foreground is separated from the background, and the road boundary within 20m is obtained by random sampling consistency algorithm. In order to solve the problem that road boundary cloud is sparse and the accuracy of detection is decreased within 20 ~ 100m, an enhanced detection method based on optical wire cutting model is proposed. Experiments on campus road and urban expressway show that the detection rate of road boundary is more than 95%, the effective detection distance is more than 70 m, and the detection period is less than 32 Ms.
【作者單位】: 軍事交通學院研究生管理大隊;軍事交通學院軍用車輛系;
【基金】:國家自然科學基金(91220301) 國家重點基礎研發(fā)計劃項目(2016YFB0100903)資助
【分類號】:TN958.98
[Abstract]:In order to quickly and robustly detect structured road boundary, a method of road boundary enhancement detection based on point cloud density of HDL-64E lidar is proposed. Based on the virtual radar model and the feature of point cloud density, the foreground is separated from the background, and the road boundary within 20m is obtained by random sampling consistency algorithm. In order to solve the problem that road boundary cloud is sparse and the accuracy of detection is decreased within 20 ~ 100m, an enhanced detection method based on optical wire cutting model is proposed. Experiments on campus road and urban expressway show that the detection rate of road boundary is more than 95%, the effective detection distance is more than 70 m, and the detection period is less than 32 Ms.
【作者單位】: 軍事交通學院研究生管理大隊;軍事交通學院軍用車輛系;
【基金】:國家自然科學基金(91220301) 國家重點基礎研發(fā)計劃項目(2016YFB0100903)資助
【分類號】:TN958.98
【參考文獻】
相關期刊論文 前4條
1 朱株;劉濟林;;基于馬爾科夫隨機場的三維激光雷達路面實時分割[J];浙江大學學報(工學版);2015年03期
2 辛煜;梁華為;梅濤;黃如林;杜明博;王智靈;陳佳佳;趙盼;;基于激光傳感器的無人駕駛汽車動態(tài)障礙物檢測及表示方法[J];機器人;2014年06期
3 王肖;王建強;李克強;徐成;李曉飛;;智能車輛3-D點云快速分割方法[J];清華大學學報(自然科學版);2014年11期
4 俞先國;劉大學;戴斌;;基于非閉合Snakes和單線激光雷達的路邊檢測與濾波[J];機器人;2013年04期
【共引文獻】
相關期刊論文 前10條
1 蘇致遠;徐友春;彭永勝;王任棟;;基于點云密度的結(jié)構(gòu)化道路邊界增強檢測方法[J];汽車工程;2017年07期
2 常虹;桂偉;;激光傳感器采集系統(tǒng)的智能模型車設計[J];激光雜志;2017年06期
3 許f,
本文編號:2323417
本文鏈接:http://sikaile.net/kejilunwen/xinxigongchenglunwen/2323417.html
最近更新
教材專著