A Self-adaptive Dual Radius Filtering Algorithm Based on LiDAR Point Cloud

doi:10.12382/bgxb.2022.1093

Abstract

Abstract:

Point cloud denoising is a key step for intelligent driving vehicles to perceive the surrounding environment information. To solve the problem of low operation speed of high-precision denoising in the LiDAR point cloud denoising method, a self-adaptive dual radius filtering method is proposed for complex scenes and multi-scale noise. The 3D point cloud is first simplified by voxel filtering under the constraint of the minimum number of points, and the outliers are preliminarily filtered. Then KD-tree is used to build an index to calculate the average density of point clouds.The adaptive large- and small-radius models are constructed according to the point cloud density to filter drift noise voxels. To verify the effectiveness of the algorithm, in the simple and complex scenes with multiple noise types, the noise removal accuracy and operation speed are compared with other algorithms. In the case of slightly reduced noise removal accuracy, the operation time is less than 0.6 seconds in simple scenes and less than 2 seconds in complex scenes. The new algorithm has high noise removal accuracy and operation speed, as well as a wide range of applications.

Key words: LiDAR point cloud, point cloud denoising, adaptive filtering, dual radius filtering

CLC Number:

TP391.9

LIU Bin, LI Xuemei. A Self-adaptive Dual Radius Filtering Algorithm Based on LiDAR Point Cloud[J]. Acta Armamentarii, 2023, 44(9): 2768-2777.

Figures/Tables 10

Fig.1 Radius filtering calculation

Fig.2 Denoising scheme process

Fig.3 Self-adaptive dual radius filtering program flow

Fig.4 Visualization of each filtering algorithm for the Bunny model

Fig.5 Visualization of each filtering algorithm for the Road scene

Table 1 Parameter setting of each filtering algorithm

滤波算法	参数	Bunny模型	Road场景	滤波算法	参数	Bunny模型	Road场景
体素滤波	格子体积V_g/cm³	0.2×0.2×0.2	12×12×12		V_g/cm³	0.2×0.2×0.2	10×10×10
半径滤波	R/cm	0.15	30		M_P	2	1
	R_T	1	10	自适应双半径滤波	M_L	30	40
中值滤波	N_mf	5	5		N_L	1	5
统计滤波	N_sf	20	20		M_S	40	60
	k	0.5	0.5		N_S	1	4

Fig.6 Comparison of denoising accuracy (The horizontal scale has comitted duplicate noise mean values)

Fig.7 Comparison of noise recall rate (The horizontal scale has comitted duplicate noise mean values)

Fig.8 Comparison of original poretention rate (The horizontal scale has comitted duplicate noise mean values)

Fig.9 Comparison of runtime (The horizontal scale has comitted duplicate noise mean values)

References 21

[1]	CHUNG S H, LEE S W, LEE S K, et al. LIDAR system with electromagnetic two-axis scanning micromirror based on indirect time-of-flight method[J]. Micro and Nano Systems Letters, 2019, 7(1):1-5. doi: 10.1186/s40486-019-0080-y
[2]	FAYYAD J, JARADAT M A, GRUYER D, et al. Deep learning sensor fusion for autonomous vehicle perception and localization: a review[J]. Sensors, 2020, 20(15):4220. doi: 10.3390/s20154220 URL
[3]	SHI C H, WANG C Y, LIU X L, et al. Three-dimensional point cloud denoising via a gravitational feature function[J]. Applied Optics, 2022, 61(6): 1331-1343. doi: 10.1364/AO.446913 URL
[4]	RUSU R B, MARTON Z C, BLODOW N, et al. Towards 3D point cloud based object maps for household environments[J]. Robotics and Autonomous Systems, 2008, 56(11): 927-941. doi: 10.1016/j.robot.2008.08.005 URL
[5]	李朋超, 王金涛, 宋吉来. 基于PCL的3D点云视觉数据预处理[J]. 计算机应用, 2019, 39(增刊2): 227-230.
	LI P C, WANG J T, SONG J L. PCL-based preprocessing for point cloud data of 3D vision[J]. Journal of Computer Applications, 2019, 39(S2): 227-230. (in Chinese)
[6]	JULIE D, CARLO D F. The bilateral filter for point clouds[J]. Image Processing on Line, 2017, 7:278-287. doi: 10.5201/ipol URL
[7]	TIZIANA C, ALESSANDRO D, GAETANO S, et al. VIPDA: a visually driven point cloud denoising algorithm based on anisotropic point cloud filtering[J]. Frontiers in Signal Processing, 2022, 2:842570. doi: 10.3389/frsip.2022.842570 URL
[8]	霍健, 陈慧敏, 马云飞, 等. 基于MEMS激光雷达的车辆目标识别算法[J]. 兵工学报, 2023, 44(4):940-948.
	HUO J, CHEN H M, MA Y F, et al. Vehicle targets recognition algorithm based on MEMS lidar[J]. Acta Armamentarii, 2023, 44(4):940-948. (in Chinese)
[9]	LIU X Q, CHEN Y M, CHENG L, et al. Airborne laser scanning point clouds filtering method based on the construction of virtual ground seed points[J]. Journal of Applied Remote Sensing, 2017, 11(1): 016032. doi: 10.1117/1.JRS.11.016032 URL
[10]	焦亚男, 马杰, 钟斌斌. 一种基于尺度变化的点云并行去噪方法[J]. 武汉大学学报(工学版), 2021, 54(3): 277-282.
	JIAO Y N, MA J, ZHONG B B. Point cloud parallel de-noising algorithms based on scale change[J]. Engineering Journal of Wuhan University, 2021, 54(3): 277-282. (in Chinese)
[11]	赵鹏, 柴会轩. 基于不同尺度算法的激光雷达点云去噪研究[J]. 应用激光, 2023, 43(1):66-75.
	ZHAO P, CHAI H X. Research on lidar point cloud denoising based on different scale algorithm[J]. Applied Laser, 2023, 43(1):66-75. (in Chinese)
[12]	韩东升, 徐茂林, 金远航. 多源异构点云配准数据的滤波及精度分析[J]. 测绘科学技术学报, 2020, 37(5): 503-509.
	HAN D S, XU M L, JIN Y H. Filtering and accuracy analysis of multi-source heterogeneous point cloud registration data[J]. Journal of Geomatics Science and Technology, 2020, 37(5): 503-509. (in Chinese)
[13]	ZHENG Z, ZHA B T, ZHOU Y, et al. Single-stage adaptive multi-scale point cloud noise filtering algorithm based on feature information[J]. Remote Sensing, 2022, 14(2): 367. doi: 10.3390/rs14020367 URL
[14]	刘大峰, 廖文和, 戴宁, 等. 散乱点云去噪算法的研究与实现[J]. 东南大学学报(自然科学版), 2007, 37(6): 1108-1112.
	LIU D F, LIAO W H, DAI N, et al. Research and implementation for denoising noisy scattered point data[J]. Journal of Southeast University(Natural Science Edition), 2007, 37(6): 1108-1113. (in Chinese)
[15]	周圣涛. 基于多核的DSP的点云数据快速降噪技术研究[D]. 长春: 长春理工大学, 2020.
	ZHOU S T. Research on fast noise reduction technology of point cloud data based on multi-core DSP[D]. Changchun: Changchun University of Science and Technology, 2020. (in Chinese)
[16]	张彬, 熊传兵. 基于体素下采样和关键点提取的点云自动配准[J]. 激光与光电子学进展, 2020, 57(4): 101-109.
	ZHANG B, XIONG C B. Automatic point cloud registration based on voxel downsampling and key point extraction[J]. Laser & Optoelectronics Progress, 2020, 57(4): 101-109. (in Chinese)
[17]	李仁忠, 杨曼, 冉媛, 等. 基于方法库的点云去噪与精简算法[J]. 激光与光电子学进展, 2018, 55(1): 243-249.
	LI R Z, YANG M, RAN Y, et al. Point cloud denoising and simplification algorithm based on method library[J]. Laser & Optoelectronics Progress, 2018, 55(1): 243-249. (in Chinese)
[18]	朱志华, 刘晋, 胡晰远. 线形痕迹三维点云去噪技术研究[J]. 刑事技术, 2020, 45(6):556-561. doi: 10.16467/j.1008-3650.2020.06.002
	ZHU Z H, LIU J, HU X Y. Attempt to denoise into 3D-point cloud of linear trace[J]. Forensic Science and Technology, 2020, 45(6): 556-561. (in Chinese) doi: 10.16467/j.1008-3650.2020.06.002
[19]	LIU D, LI D J, WANG M Z, et al. 3D change detection using adaptive thresholds based on local point cloud density[J]. ISPRS International Journal of Geo-Information, 2021, 10(3): 127. doi: 10.3390/ijgi10030127 URL
[20]	MUNARO M, RUDU R B, EMANUELE M. 3D robot perception with Point cloud library[J]. Robotics and Autonomous Systems, 2016, 78: 97-99. doi: 10.1016/j.robot.2015.12.008 URL
[21]	GEIGER A, LENZ P, STILLER C, et al. Vision meets robotics: the KITTI dataset[J]. The International Journal of Robotics Research, 2013, 32(11):1231-1237. doi: 10.1177/0278364913491297 URL