High-Speed and High-Precision Algorithm to Determine Angular Coordinates of a Three-Dimensional Textureless Object Based on the Contour Analysis

Abstract

The paper presents a high-speed algorithm to determine three angular coordinates of a textureless 3D object based on contour analysis of the monocular images. The algorithm could be applied to identify the aircraft, spacecraft, and other objects angular position. Angular coordinates' computation is based on comparing the observed contour descriptor with the reference contour descriptor. At the stage of preparing the object reference descriptor, its 3D surface is defined by a polygonal mesh and is transformed into a three-dimensional descriptor based on the three-dimensional discrete Fourier transform. Angular coordinates are determined by solving the problem of minimizing difference between the Fourier descriptor of the observed contour and the reference contour descriptor obtained by two-dimensional interpolation from the reference three-dimensional descriptor. To achieve high performance and high accuracy, the algorithm implements the principle of iteratively increasing accuracy in the reference descriptor interpolation. In order to raise the algorithm speed, optimization minima search process is parallelized. The paper gives description of the algorithm and demonstrates results of simulating the implemented algorithm operation with synthesized images of an aircraft at the landing stage. According to the simulation results, the algorithm speed without using a graphics processor was 0.25 ms per one object in the detection mode and less than 0.1 ms in the tracking mode. The standard error deviation in determining the angular coordinates was about 0.23--0.3°

Please cite this article in English as:

Artemyev A.E. Introduction of the high-speed and high-precision algorithm to determine angular coordinates of a three-dimensional texturless object based on the contour analysis. Herald of the Bauman Moscow State Technical University, Series Instrument Engineering, 2024, no. 3 (148), pp. 115--135 (in Russ.). EDN: XLRKVJ

References

[1] Wallace T.P., Wintz P.A. An efficient three-dimensional aircraft recognition algorithm using normalized Fourier descriptors. Comput. Graph. Image Process., 1980, vol. 13, no. 2, pp. 99--126. DOI: https://doi.org/10.1016/S0146-664X(80)80035-9

[2] Breuers M., de Reus N. Image-based aircraft pose estimation: а comparison of simulations and real-world data. Proc SPIE., 2001, vol. 4379, pp. 472--479. DOI: https://doi.org/10.1117/12.445395

[3] Malyshev O.V. Opredelenie prostranstvennogo polozheniya i raspoznavanie tipov letatelnykh apparatov na osnove konturnogo analiza. Dis. kand. tekh. nauk [Determination of spatial position and recognition of aircraft types based on contour analysis. Cand. Sc. (Eng.). Diss.] Rybinsk, RGATU, 2011 (in Russ.).

[4] Alpatov B., Babayan P. Contour-based object orientation estimation. Proc. SPIE, 2016, vol. 9897, art. 98970H. DOI: https://doi.org/10.1117/12.2224249

[5] Zhang X., Jiang Z., Zhang H., et al. Vision-based pose estimation for textureless space objects by contour points matching. IEEE Trans. Aerosp. Electron. Syst., 2018, vol. 54, no. 5, pp. 2342--2355. DOI: https://doi.org/10.1109/TAES.2018.2815879

[6] Fu T., Sun X. The relative pose estimation of aircraft based on contour model. Proc. SPIE, 2017, vol. 10250, art. 102502T. DOI: https://doi.org/10.1117/12.2267118

[7] Alpatov B., Babayan P., Ershov M., et al. The implementation of contour-based object orientation estimation algorithm in FPGA-based on-board vision system. Proc. SPIE, 2016, vol. 10007, art. 100070A. DOI: https://doi.org/10.1117/12.2241091

[8] Mirzoyan A.S. Razrabotka sistem raspoznavaniya i pozitsionirovaniya letatelnykh apparatov i nazemnykh obektov na osnove metodov vychislitelnoy geometrii. Dis. kand. tekh. nauk [Development of recognition and positioning systems for aircraft and ground objects based on computational geometry methods. Cand. Sc. (Eng.). Diss.] Rybinsk, RGATU, 2017 (in Russ.).

[9] Mirzoyan A.S., Malyshev O.V., Khmarov I.M., et al. Recognition of aircrafts by optical system in real time. Vestnik MAI [Aerospace MAI Journal], 2014, vol. 21, no. 5, pp. 145--156 (in Russ.). EDN: TGWHHX

[10] Fu D., Li W., Han S., et al. The pose estimation of the aircraft on the airport surface based on the contour features. IEEE Trans. Aerosp. Electron. Syst., 2023, vol. 59, no. 2, pp. 817--826. DOI: https://doi.org/10.1109/TAES.2022.3192220

[11] On S., Kim S., Yang K., et al. Monocular 3D pose estimation of very small airplane in the air. MMAsia’23, 2023, art. 82. DOI: https://doi.org/10.1145/3595916.3626456

[12] Li B., Raja S., Li J., et al. A Fourier descriptor and PSCS-RBF fusion method for pumping machine fault diagnosis. Durham, Research Suare, 2023. DOI: https://doi.org/10.21203/rs.3.rs-2878163/v1

[13] Andrews M., Kakarala R. Shape simplification through polygonal approximation in the Fourier domain. Proc. SPIE, 2015, vol. 9406, art. 94060D. DOI: https://doi.org/10.1117/12.2078148

[14] Liao N., Guo B., Li Z., et al. An advanced Fourier descriptor based on centroid contour distances. J. Phys.: Conf. Ser., 2021, vol. 1735, art. 012002. DOI: https://doi.org/10.1088/1742-6596/1735/1/012002

[15] Park K.S., Lee N.S. A three-dimensional Fourier descriptor for human body representation/reconstruction from serial cross sections. Comput. Biomed. Res., 1987, vol. 20, no. 2, pp. 125--140. DOI: https://doi.org/10.1016/0010-4809(87)90040-1

[16] Artemyev A. Formation of a 3D descriptor for reconstruction of an object contour for contour analysis algorithms. Durham, Research Suare, 2023. DOI: https://doi.org/10.13140/RG.2.2.30881.79208

[17] Artemyev A. Estimating 6D position of a three-dimensional object from 2D contour on image in real time. Durham, Research Suare, 2023. DOI: https://doi.org/10.13140/RG.2.2.29204.07048

[18] Open Source Computer Vision Library: opencv.org: website. Available at: https://opencv.org (accessed: 16.01.2024).

[19] Garcia-Molla V., Alonso P. Parallel border tracking in binary images for multicore computers. J. Supercomput., 2023, vol. 79, no. 7, pp. 9915--9931. DOI: https://doi.org/79.10.1007/s11227-023-05052-2

[20] Artemyev A. High-speed and high-precision algorithm for determining the spatial position of a helicopter based on its image for automatic landing systems. TsAGI Int. Aeronautics Congress, 2023.

[21] Leng D., Sun W. Contour-based iterative pose estimation of 3D rigid object. IET Comput. Vis., 2011, vol. 5, no. 5, pp. 291--300. DOI: https://doi.org/10.1049/iet-cvi.2010.0098

[22] Zhang X., Zhang H., Wei Q., et al. Pose estimation of space objects based on hybrid feature matching of contour points. IGTA. Springer, 2016, pp. 184--191. DOI: https://doi.org/10.1007/978-981-10-2260-9_21