Rejestracja chmur punktów: porównanie wariantów wzajemnej rejestracji

pol Article in Polish DOI: 10.14313/PAR_224/5

Marta Jolanta Łępicka *, send Tomasz Kornuta ** * Politechnika Warszawska, Instytut Automatyki i Informatyki Stosowanej ** IBM Research - Almaden, USA

Download Article

Streszczenie

W dwuczęściowym artykule skupiono uwagę na problemie rejestracji chmur punktów. W pierwszej części omówiono kluczowe komponenty systemu V-SLAM uzupełnione przykładowymi algorytmami i rozwiązaniami stosowanymi w tych komponentach. W poniższej, drugiej części omówiono różne rodzaje wariantów algorytmu ICP, atrybuty punktów oraz operujące na nich metryki. Następnie omówiono metodykę badań oraz przedstawiono wyniki porównania wybranych wariantów wzajemnej rejestracji.

Słowa kluczowe

chmura punktów, ICP, KAZE, obraz RGB-D, rejestracja, SIFT, wzajemne łączenie

Registration of RGB-D images: comparison of pairwise registration variants

Abstract

The two-part article focuses on the problem of registration of point clouds. The first part briefly discussed the main components of V-SLAM systems and presented the main steps of the ICP (Iterative Closes Point) algorithm. In the following, second part of the paper, we analyse and compare diverse variants of the ICP algorithm. In particular, we discuss different attributes of points along with operating on them metrics that the ICP can employ. Finally, we present the research methodology and discuss the results of comparison of selected variants of ICP.

Keywords

ICP, KAZE, pairwise registration, point cloud, RGB-D image, SIFT

Bibliography

  1. Alahi A., Ortiz R., Vandergheynst P., FREAK: Fast Retina Keypoint. 2012 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 510–517, IEEE, 2012.
  2. Alcantarilla P.F., Bartoli A., Davison A.J., KAZE features. European Conference on Computer Vision (ECCV), 2012, 214–227, DOI: 10.1007/978-3-642-33783-3_16.
  3. Belongie S., Malik J., Puzicha J., Shape matching and object recognition using shape contexts. IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 24, Iss. 4, 2002, 509–522, DOI: 10.1109/34.993558.
  4. Besl P., McKay N., A method for registration of 3-D shapes. “IEEE Transactions on Pattern Analysis and Machine Intelligence”, 14(2):239–256, 1992.
  5. Chen Y., Medioni G., Object modelling by registration of multiple range images, 1991 IEEE International Conference on Robotics and Automation, Proceedings Vol. 3, 2724–2729.
  6. Feldmar J., Ayache N., Rigid, affine and locally affine registration of free-form surfaces. “International Journal of Computer Vision”, Vol. 18, Iss. 2, 1996, 99–119, DOI: 10.1007/BF00054998.
  7. Figat J., Kornuta T., Kasprzak W., Performance evaluation of binary descriptors of local features. [in:] Proceedings of the International Conference on Computer Vision and Graphics, Vol. 8671, Lecture Notes in Computer Science, Springer Berlin /Heidelberg, 2014, 187–194. DOI: 10.1007/978-3-319-11331-9_23.
  8. Frome A., Huber D., Kolluri R., Bülow T., Malik J., Recognizing objects in range data using regional point descriptors. European Conference on Computer Vision (ECCV), 2004, 224–237, 10.1007/978-3-540-24672-5_18.
  9. Hänsch R., Weber T., Hellwich O., Comparison of 3D interest point detectors and descriptors for point cloud fusion. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol. II-3, 2014, 57–64, DOI: 10.5194/isprsannals-II-3-57-2014.
  10. Holz D., Ichim A.E., Tombari F., Rusu R.B., Behnke S., Registration with the point cloud library – A modular framework for aligning in 3-D. “IEEE Robotics & Automation Magazine”, 22(4):110–124, 2015, DOI: 10.1109/MRA.2015.2432331.
  11. Kornuta T., Łępicka M.J., Rejestracja chmur punktów: komponenty systemu. „Pomiary Automatyka Robotyka”, R. 21, Nr 1, 2017, 19–24, DOI: 10.14313/PAR_223/19.
  12. Łępicka M., Kornuta T., Stefańczyk M., Utilization of Colour in ICP-based Point Cloud Registration, [in:] Proceedings of the 9th International Conference on Computer Recognition Systems CORES 2015. Advances in Intelligent Systems and Computing, Vol. 403. Springer International Publishing, 2016, 821–830, DOI: 10.1007/978-3-319-26227-7_77.
  13. Low K.-L., Linear Least-Squares Optimization for Point-to-Plane OCP Surface Registration, Technical Report TR04-004, 2004, 1–3.
  14. Lowe D., Object recognition from local scale-invariant features. The Proceedings of the Seventh IEEE International Conference on Computer Vision, 1999, Vol. 2, 1150–1157.
  15. Men H., Gebre B., Pochiraju K., Color point cloud registration with 4D ICP algorithm. 2011 IEEE International Conference on Robotics and Automation (ICRA), 1511–1516, DOI: 10.1109/ICRA.2011.5980407.
  16. Pomerleau F., Colas F., Siegwart R., A review of point cloud registration algorithms for mobile robotics. Foundations and Trends in Robotics, 4(1):1–104, 2015, DOI: 10.1561/2300000035.
  17. Rusinkiewicz S., Levoy M., Efficient variants of the ICP algorithm. Third International Conference on 3-D Digital Imaging and Modeling, Proceedings, 145–152, IEEE, 2001, DOI: 10.1109/IM.2001.924423.
  18. Rusu R.B., Blodow N., Beetz M., Fast point feature histograms (FPFH) for 3D registration. IEEE International Conference on Robotics and Automation, ICRA ’09, 3212–3217. IEEE, 2009, DOI: 10.1109/ROBOT.2009.5152473.
  19. Stefanczyk M., Laszkowski M., Kornuta T.. WUT Visual Perception Dataset – a dataset for registration and recognition of objects. [in:] Szewczyk R., Zieliński C., Kaliczyńska M. (eds) Challenges in Automation, Robotics and Measurement Techniques. Advances in Intelligent Systems and Computing, Vol. 440. Springer 2016, 635–645, DOI: 10.1007/978-3-319-29357-8_55.
  20. Tombari F., Salti S., Di Stefano L., Unique signatures of histograms for local surface description. Proceedings of the 11th European Conference on Computer Vision Conference on Computer Vision: Part III, ECCV ’10, 356–369, Berlin, Heidelberg, 2010. Springer-Verlag.
  21. Tuytelaars T., Mikolajczyk K., Local invariant feature detectors: a survey. Foundations and trends in computer graphics and vision, Vol. 3, Iss. 3, 2008, 177–280, DOI: 10.1561/0600000017.