Object tracking method based on joint global and local feature descriptor of 3D LIDAR point cloud

Qishu Qian; Yihua Hu; Nanxiang Zhao; Minle Li; Fucai Shao; Xinyuan Zhang

doi:doi:10.3788/COL202018.061001

Chinese Optics Letters, 2020, 18 (6): 061001, Published Online: May. 12, 2020

Object tracking method based on joint global and local feature descriptor of 3D LIDAR point cloud Download： 913次

Qishu Qian ^1,2Yihua Hu ^1,2,*Nanxiang Zhao ^1,2Minle Li ^1,2Fucai Shao ³Xinyuan Zhang ^1,2

Author Affiliations

¹ State Key Laboratory of Pulsed Power Laser Technology, National University of Defense Technology, Hefei 230037, China

² Anhui Provincial Key Laboratory of Electronic Restriction, National University of Defense Technology, Hefei 230037, China

³ The Military Representative Bureau of the Ministry of Equipment Development, Central Military Commission in Beijing, Beijing 100191, China

Figures & Tables

Fig. 1. Proposed object tracking method of point cloud based on JGLF.

下载图片查看原文

Fig. 2. Flow chart of the proposed object tracking method.

下载图片查看原文

Fig. 3. Comparison of the object recognition rate at different distances between LIDAR and the object.

下载图片查看原文

Fig. 4. Results of the PF point cloud tracking algorithm based on JGLF in frame $n$ : (a) $n = 1$ , (b) $n = 61$ , (c) $n = 121$ , (d) $n = 181$ , and (e) $n = 241$ . (f) Comparison between particles in frame 186.

下载图片查看原文

Fig. 5. Comparison of the object tracking effect between the two algorithms.

下载图片查看原文

Table1. Object Recognition Ability Comparison of Six Descriptors

	Object Recognition Rate (%)		Average Running Time (ms)
	Mean	Standard Deviation	Average Running Time (ms)
FPFH	62.37	13.17	5
VFH	64.34	11.25	3.6
CVFH	68.91	7.79	4.5
GRSD	39.85	16.28	31
ESF	87.59	16.71	39
JGLF	72.09	10.81	6

查看原文

Table2. Comparison of Tracking Results of Five Algorithms

	Tracking Accuracy (%)	R of Single Frame (%)		Average Running Time (ms)	CPU Utilized Percent (%)
	Tracking Accuracy (%)	Mean	Standard Deviation	Average Running Time (ms)	CPU Utilized Percent (%)
Basic algorithm	84.87	72.81	17.75	12.44	5
Algorithm based on FPFH	89.04	80.67	12.47	12.71	7
Algorithm based on VFH	90.76	82.21	14.59	12.68	7
Algorithm based on CVFH	91.25	80.33	9.07	12.82	8
Proposed algorithm	98.82	88.01	11.96	12.96	8

查看原文

Qishu Qian, Yihua Hu, Nanxiang Zhao, Minle Li, Fucai Shao, Xinyuan Zhang. Object tracking method based on joint global and local feature descriptor of 3D LIDAR point cloud[J]. Chinese Optics Letters, 2020, 18(6): 061001.

Object tracking method based on joint global and local feature descriptor of 3D LIDAR point cloud Download： 913次

Fig. 1. Proposed object tracking method of point cloud based on JGLF.

Fig. 2. Flow chart of the proposed object tracking method.

Fig. 3. Comparison of the object recognition rate at different distances between LIDAR and the object.

Fig. 4. Results of the PF point cloud tracking algorithm based on JGLF in frame $n$ : (a) $n = 1$ , (b) $n = 61$ , (c) $n = 121$ , (d) $n = 181$ , and (e) $n = 241$ . (f) Comparison between particles in frame 186.

Fig. 5. Comparison of the object tracking effect between the two algorithms.

Table1. Object Recognition Ability Comparison of Six Descriptors

Table2. Comparison of Tracking Results of Five Algorithms

关于本站 Cookie 的使用提示

全站搜索

Object tracking method based on joint global and local feature descriptor of 3D LIDAR point cloud Download： 913次

Fig. 1. Proposed object tracking method of point cloud based on JGLF.

Fig. 2. Flow chart of the proposed object tracking method.

Fig. 3. Comparison of the object recognition rate at different distances between LIDAR and the object.

Fig. 4. Results of the PF point cloud tracking algorithm based on JGLF in frame n: (a) n=1, (b) n=61, (c) n=121, (d) n=181, and (e) n=241. (f) Comparison between particles in frame 186.

Fig. 5. Comparison of the object tracking effect between the two algorithms.

Table1. Object Recognition Ability Comparison of Six Descriptors

Table2. Comparison of Tracking Results of Five Algorithms

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索

Fig. 4. Results of the PF point cloud tracking algorithm based on JGLF in frame $n$ : (a) $n = 1$ , (b) $n = 61$ , (c) $n = 121$ , (d) $n = 181$ , and (e) $n = 241$ . (f) Comparison between particles in frame 186.