A method for simplifying surface point cloud data of coal mine roadways based on secondary feature extraction
-
摘要:
采用三维激光扫描技术提取的煤矿巷道表面点云数据量大且存在较多的冗余数据，而现有点云数据精简方法存在大数量级点云处理过程中细节保留不足的问题。针对上述问题，提出了一种基于二次特征提取的煤矿巷道表面点云数据精简方法。首先对采集到的原始巷道点云数据进行去噪预处理；其次建立K−d树，并利用主成分分析法对去噪后点云数据估算来拟合邻域平面的法向量；然后通过较小的法向量夹角阈值对点云进行初步的特征区域与非特征区域划分，保留特征区域并随机下采样非特征区域，接着依据较大的法向量夹角阈值将特征区域点云划分为特征点和非特征点，并对非特征点进行体素随机采样；最后将2次点云精简结果与特征点合并得到最终的精简数据。仿真结果表明，该方法在百万数据量级点云和高精简率条件下，相较曲率精简方法、随机精简方法和栅格精简方法，在特征保留和重构精度方面都取得了更好的效果，三维重构后计算所得标准偏差平均可低于相同精简率下其他方法30%左右。
Abstract:The surface point cloud data of coal mine roadways extracted using 3D laser scanning technology has a large amount of redundant data. The existing point cloud data simplification methods have the problem of insufficient detail preservation in the processing of large-scale point clouds. In order to solve the above problems, a surface point cloud data reduction method for coal mine roadways based on secondary feature extraction is proposed. Firstly, the method performs denoising preprocessing on the collected original roadway point cloud data. Secondly, the method establishes a K-d tree and uses principal component analysis to estimate the denoised point cloud data to fit the normal vector of the neighborhood plane. Thirdly, the point cloud is preliminarily divided into feature and non-feature regions using a smaller normal vector angle threshold, retaining the feature regions and randomly downsampling the non-feature regions. Fourthly, based on the larger normal vector angle threshold, the feature region point cloud is divided into feature points and non-feature points. And voxel random sampling is conducted on the non-feature points. Finally, the method merges the two point cloud simplification results with the feature points to obtain the final simplified data. The simulation results show that under million data level point clouds and high precision conditions, this method achieves better results in feature preservation and reconstruction precision compared to curvature simplification methods, random simplification methods, and grid reduction methods. The average standard deviation calculated after 3D reconstruction can be about 30% lower than other methods under the same reduction rate.
-
图 1 煤矿巷道表面点云数据精简方法流程
Figure 1. Flowchart of point cloud data simplification method of coal mine roadway surface
图 2 基于法向量夹角的特征区域选取原理
Figure 2. Principle of feature area selection based on normal vector angle
图 4 精简率为50%的简化及三维重建结果
Figure 4. Simplification and 3D reconstruction results with reduction rate of 50%
图 6 精简率为10%的简化及三维重建结果
Figure 6. Simplification and 3D reconstruction results with reduction rate of 10%
图 5 精简率为30%的简化及三维重建结果
Figure 5. Simplification and 3D reconstruction results with reduction rate of 30%
表 1 不同特征提取次数下最大偏差与标准偏差
Table 1. The maximum deviation and standard deviation under different feature extraction times
特征提取方式 精简率/% 最大偏差（正向/负向）/m 标准偏差/m 一次特征提取 10 1.340 6/−2.807 2 0.038 93 二次特征提取 0.918 3/−0.748 4 0.036 44 一次特征提取 30 3.049 9/−2.398 7 0.035 16 二次特征提取 2.754 1/−2.375 9 0.033 27 一次特征提取 50 1.582 8/−1.835 1 0.021 50 二次特征提取 2.739 3/−2.226 8 0.020 66 
下载: 导出CSV
表 2 不同精简方法下最大偏差与标准偏差
Table 2. The maximum deviation and standard deviation under different simplification methods
精简方法 精简率/% 最大偏差（正向/负向）/m 标准偏差/m 曲率精简方法 10 2.800 9/−1.894 9 0.057 88 随机精简方法 2.452 6/−2.663 3 0.060 37 栅格精简方法 1.310 7/−3.052 4 0.055 09 本文方法 0.918 3/−0.748 4 0.036 44 曲率精简方法 30 3.265 2/−2.712 2 0.043 39 随机精简方法 2.428 1/−2.465 9 0.037 65 栅格精简方法 3.268 8/−1.856 2 0.039 15 本文方法 2.754 1/−2.375 9 0.033 27 曲率精简方法 50 3.151 3/−1.562 2 0.025 77 随机精简方法 2.848 6/−1.216 0 0.020 82 栅格精简方法 2.857 8/−1.550 1 0.033 63 本文方法 2.739 3/−2.226 8 0.020 66
下载: 导出CSV
-
[1] 俞艳波,李小松,苏海华,等. 便携式三维激光扫描技术在矿山地下巷道可视化建模中的应用[J]. 北京测绘,2022,36(12):1703-1708.YU Yanbo,LI Xiaosong,SU Haihua,et al. Application of portable 3D laser scanning technology in visualization modeling of mine underground tunnel[J]. Beijing Surveying and Mapping,2022,36(12):1703-1708. [2] 朱杭琦,徐亮亮,方秀友,等. 基于三维激光扫描技术的地铁隧道健康检测与安全性评估应用[J]. 城市地质,2022,17(4):493-500. doi: 10.3969/j.issn.1007-1903.2022.04.014ZHU Hangqi,XU Liangliang,FANG Xiuyou,et al. Health detection and safety assessment of subway tunnel based on 3D laser scanning technology[J]. Urban Geology,2022,17(4):493-500. doi: 10.3969/j.issn.1007-1903.2022.04.014 [3] 高帅,马全明,符新新,等. 三维激光扫描技术在地铁隧道调线调坡测量中的应用研究[J]. 城市勘测,2022(5):146-148.GAO Shuai,MA Quanming,FU Xinxin,et al. Research on route and gradient adjustment measurement of 3D laser scanning in metro tunnel[J]. Urban Geotechnical Investigation & Surveying,2022(5):146-148. [4] 宋成航,李晋儒,刘冠杰. 利用特征点采样一致性改进ICP算法点云配准方法[J]. 北京测绘,2021,35(3):317-322.SONG Chenghang,LI Jinru,LIU Guanjie. Point cloud registration method using feature point sampling consistency initial alignment and improved ICP algorithm[J]. Beijing Surveying and Mapping,2021,35(3):317-322. [5] 彭海驹,严科文,林松,等. 融合kmeans聚类与Hausdorff距离的点云精简算法改进[J]. 地理空间信息,2022,20(8):59-63. doi: 10.3969/j.issn.1672-4623.2022.08.014PENG Haiju,YAN Kewen,LIN Song,et al. Point cloud simplification improved algorithm integrating K-means clustering and Hausdorff distance[J]. Geospatial Information,2022,20(8):59-63. doi: 10.3969/j.issn.1672-4623.2022.08.014 [6] 刘源,左小清,李勇发,等. 使用冯米塞斯分布提取特征的点云精简方法[J]. 激光与光电子学进展,2023,60(2):341-351.LIU Yuan,ZUO Xiaoqing,LI Yongfa,et al. Point cloud simplification method using Von Mises-Fisher distribution to extract features[J]. Laser & Optoelectronics Progress,2023,60(2):341-351. [7] 张亦芳,李立,刘光帅. 基于特征显著性的点云自适应精简[J]. 计算机工程与设计,2021,42(8):2211-2217.ZHANG Yifang,LI Li,LIU Guangshuai. Adaptive simplification of point cloud based on feature saliency[J]. Computer Engineering and Design,2021,42(8):2211-2217. [8] SHI Zhiyuan,XU Weiming,MENG Hao. A point cloud simplification algorithm based on weighted feature indexes for 3D scanning sensors[J]. Sensors,2022,22(19):1-21. doi: 10.1109/JSEN.2022.3206996 [9] 曹爽,赵显富,马文. 一种基于曲面变化的工业构件点云数据精简方法[J]. 测绘通报,2018(11):20-24.CAO Shuang,ZHAO Xianfu,MA Wen. A simplification method of point cloud data of industrial components based on surface variation[J]. Bulletin of Surveying and Mapping,2018(11):20-24. [10] 李琪琪,花向红,赵不钒,等. 一种基于曲率泊松碟采样的散乱点云精简方法[J]. 测绘通报,2020(增刊1):176-180.LI Qiqi,HUA Xianghong,ZHAO Bufan,et al. A method for scattered point cloud simplification based on curvature poisson dish sampling[J]. Bulletin of Surveying and Mapping,2020(S1):176-180. [11] 傅思勇,吴禄慎,陈华伟. 空间栅格动态划分的点云精简方法[J]. 光学学报,2017,37(11):253-261.FU Siyong,WU Lushen,CHEN Huawei. Point cloud simplification method based on space grid dynamic partitioning[J]. Acta Optica Sinica,2017,37(11):253-261. [12] 张匡宇,田庆. 一种基于曲率判别的点云去噪与精简算法[J]. 南方农机,2022,53(3):121-123. doi: 10.3969/j.issn.1672-3872.2022.03.036ZHANG Kuangyu,TIAN Qing. A point cloud denoising and simplification algorithm based on curvature discrimination[J]. China Southern Agricultural Machinery,2022,53(3):121-123. doi: 10.3969/j.issn.1672-3872.2022.03.036 [13] 韩一菲,杨紫骞,郑福,等. 基于FPFH和法向量的改进点云配准算法[J]. 半导体光电,2021,42(4):579-584.HAN Yifei,YANG Ziqian,ZHENG Fu,et al. Improved point cloud registration algorithm based on FPFH and normal vector[J]. Semiconductor Optoelectronics,2021,42(4):579-584. [14] 李海鹏,徐丹,付宇婷,等. 基于FPFH特征提取的散乱点云精简算法[J]. 图学学报,2022,43(4):599-607.LI Haipeng,XU Dan,FU Yuting,et al. A scattered point cloud simplification algorithm based on FPFH feature extraction[J]. Journal of Graphics,2022,43(4):599-607. [15] 何宽,孙瑞,官云兰,等. 基于逐点前进法的点云数据精简[J]. 测绘通报,2022(9):167-169. doi: 10.3969/j.issn.0494-0911.2022.9.chtb202209031HE Kuan,SUN Rui,GUAN Yunlan,et al. Point cloud data simplification based on point-by-point advancing method[J]. Bulletin of Surveying and Mapping,2022(9):167-169. doi: 10.3969/j.issn.0494-0911.2022.9.chtb202209031 [16] 李佩佩,崔凤英. 基于二分K−means聚类的曲率分级点云数据精简优化算法研究[J]. 电子测量技术,2022,45(4):66-71.LI Peipei,CUI Fengying. The optimization algorithm for curvature graded point cloud data based on dichotomous K-means clustering[J]. Electronic Measurement Technology,2022,45(4):66-71. [17] 章紫辉,官云兰. 基于邻域点位置特征的点云数据精简[J]. 激光与光电子学进展,2023,60(16):393-402.ZHANG Zihui,GUAN Yunlan. Point-cloud data reduction based on neighborhood-point position feature[J]. Laser & Optoelectronics Progress,2023,60(16):393-402. [18] 梁栋,蒲洁,李岩峰. 一种保留特征点的大数据量点云分类精简算法[J]. 测绘科学,2022,47(5):99-106,133.LIANG Dong,PU Jie,LI Yanfeng. A streamlined algorithm for large data point cloud classification with preserving feature points[J]. Science of Surveying and Mapping,2022,47(5):99-106,133. [19] 丁吉祥,董寰宇,秦训鹏. 面向结构件几何特征保留的点云精简方法[J]. 电子测量与仪器学报,2022,36(6):196-204.DING Jixiang,DONG Huanyu,QIN Xunpeng. Point cloud simplification method for geometric feature preservation of structural parts[J]. Journal of Electronic Measurement and Instrumentation,2022,36(6):196-204. [20] 刘江,张旭,朱继文. 一种基于K−D树优化的ICP三维点云配准方法[J]. 测绘工程,2016,25(6):15-18. doi: 10.3969/j.issn.1006-7949.2016.06.004LIU Jiang,ZHANG Xu,ZHU Jiwen. ICP three-dimensional point cloud registration method based on K-D tree optimization[J]. Engineering of Surveying and Mapping,2016,25(6):15-18. doi: 10.3969/j.issn.1006-7949.2016.06.004 [21] 杨必胜,韩旭,董震. 点云深度学习基准数据集[J]. 遥感学报,2021,25(1):231-240. doi: 10.11834/jrs.20210542YANG Bisheng,HAN Xu,DONG Zhen. Point cloud benchmark dataset WHU-TLS and WHU-MLS for deep learning[J]. National Remote Sensing Bulletin,2021,25(1):231-240. doi: 10.11834/jrs.20210542 [22] DONG Zhen,LIANG Fuxun,YANG Bisheng,et al. Registration of large-scale terrestrial laser scanner point clouds:a review and benchmark[J]. ISPRS Journal of Photogrammetry and Remote Sensing,2020,163:327-342. doi: 10.1016/j.isprsjprs.2020.03.013 -
点击查看大图
计量
- 文章访问数: 83
- HTML全文浏览量: 33
- PDF下载量: 6
- 被引次数: 0
