Improving the automatic target recognition algorithm’s accuracy through an examination of the different time-frequency representations and data augmentation

Boban Sazdic-Jotic; Boban Bondzulic; Jovan Bajcetic; Milenko Andric; Ivan Pokrajac; Danilo Obradovic; Bojan Zrnic

doi:10.1515/freq-2022-0015

Article

Improving the automatic target recognition algorithm’s accuracy through an examination of the different time-frequency representations and data augmentation

Boban Sazdic-Jotic , Boban Bondzulic , Jovan Bajcetic , Milenko Andric , Ivan Pokrajac , Danilo Obradovic and Bojan Zrnic

Published/Copyright: November 3, 2022

Published by

Become an author with De Gruyter Brill

Author Information

From the journal Frequenz Volume 77 Issue 5-6

Abstract

This research focuses on an improved automatic target recognition algorithm for solving the classification challenge of ground-moving targets from pulsed-Doppler radar. First, it was studied how decision-making intervals affect the proposed algorithm. Second, the altering of the data augmentation process was investigated. Third, a consideration of the three time-frequency signal representations and finally the use of different deep learning models for the classification issues were examined. It is proven that the proposed algorithm can efficiently recognize all targets enclosed in the publicly available RadEch dataset, with 4 s of radar echoes. When the decision-making time is only 1 s, a classification probability of 99.9% was obtained, which is an improvement related to the other research studies in this area. Furthermore, when the decision-making time is reduced 16 times the classification accuracy is reduced by only 1.3%. Moreover, the proposed algorithm was successful on another dataset enclosing ground-moving targets from comparable pulsed-Doppler radar.

Keywords: automatic target recognition (ATR) algorithm; classification; convolutional neural network (CNN); convolutional recurrent neural network (CRNN); deep learning (DL); short-time Fourier transform (STFT)

Corresponding author: Boban Sazdic-Jotic, Military Academy, University of Defence in Belgrade, Veljka Lukica Kurjaka 33, Belgrade, Serbia, E-mail: boban.sazdic.jotic@vs.rs

Funding source: Military Technical Institute (MTI) in Belgrade

Award Identifier / Grant number: III-47029

Funding source: University of Defence in Belgrade

Award Identifier / Grant number: VA-TT/3/20-22

Author contributions: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission. The findings achieved herein are solely the responsibility of the authors.
Research funding: This research is conducted under the project funded by the University of Defence in Belgrade (Grant No. VA–TT/3/20–22), the project RABAMADRIDS funded by Military Technical Institute (MTI) in Belgrade, and was partially supported by the Ministry of Education, Science and Technological Development of the Republic of Serbia (Grant No. III-47029).
Conflict of interest statement: The authors declare no conflicts of interest regarding this article.

References

[1] B. Boashash, “Part I: a the time-frequency tutorial,” in Time-Frequency Signal Analysis and Processing, Oxford, Elsevier, 2016, pp. 3–137.Search in Google Scholar

[2] N. E. Huang, Z. Shen, S. R. Long, et al., “The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis,” Proc. R. Soc. London. Ser. A Math. Phys. Eng. Sci., vol. 454, no. 1971, pp. 903–995, Mar. 1998. https://doi.org/10.1098/rspa.1998.0193.Search in Google Scholar

[3] N. E. Huang, Z. Wu, S. R. Long, K. C. Arnold, X. Chen, and K. Blank, “On instantaneous frequency,” Adv. Adapt. Data Anal., vol. 1, no. 2, pp. 177–229, 2009, https://doi.org/10.1142/S1793536909000096.Search in Google Scholar

[4] B. Boashash, N. A. Khan, and T. Ben-Jabeur, “Time-frequency features for pattern recognition using high-resolution TFDs: a tutorial review,” Digit. Signal Process. A Rev. J., vol. 40, no. 1, pp. 1–30, 2015, https://doi.org/10.1016/j.dsp.2014.12.015.Search in Google Scholar

[5] M. Andrić, B. Bondžulić, and B. Zrnić, “The database of radar echoes from various targets with spectral analysis,” 10th Symp. Neural Netw. Appl. Electr. Eng. NEUREL-2010 - Proc., 2010, pp. 187–190.10.1109/NEUREL.2010.5644074Search in Google Scholar

[6] I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning, London, England, MIT Press books, 2016.Search in Google Scholar

[7] F. J. Moreno-Barea, J. M. Jerez, and L. Franco, “Improving classification accuracy using data augmentation on small data sets,” Expert Syst. Appl., vol. 161, p. 113696, Dec. 2020, https://doi.org/10.1016/j.eswa.2020.113696.Search in Google Scholar

[8] S. C. Wong, A. Gatt, V. Stamatescu, and M. D. McDonnell, “Understanding data Augmentation for classification: when to warp?” in 2016 International Conference on Digital Image Computing: Techniques and Applications (DICTA), 2016, pp. 1–6.10.1109/DICTA.2016.7797091Search in Google Scholar

[9] P. Molchanov, J. Astola, K. Egiazarian, and A. Totsky, “Classification of ground moving radar targets by using joint time-frequency analysis,” in 2012 IEEE Radar Conference, vol. 2, 2012, pp. 0366–0371.10.1109/RADAR.2012.6212166Search in Google Scholar

[10] I. Bilik, J. Tabrikian, and A. Cohen, “GMM-based target classification for ground surveillance Doppler radar,” IEEE Trans. Aero. Electron. Syst., vol. 42, no. 1, pp. 267–278, Jan. 2006, https://doi.org/10.1109/TAES.2006.1603422.Search in Google Scholar

[11] A. Serir and Y. Bouhafsi, “Micro-Doppler radar signature classification by time-frequency and time-scale analysis,” in 2012 11th International Conference on Information Science, Signal Processing and Their Applications, ISSPA 2012, Montreal, QC, Canada, pp. 995–1000.10.1109/ISSPA.2012.6310701Search in Google Scholar

[12] L. Pallotta, C. Clemente, A. De Maio, J. J. Soraghan, and A. Farina, “Pseudo-Zernike moments based radar micro-Doppler classification,” in 2014 IEEE Radar Conference, 2014, pp. 0850–0854.10.1109/RADAR.2014.6875709Search in Google Scholar

[13] C. Clemente, L. Pallotta, A. De Maio, J. J. Soraghan, and A. Farina, “A novel algorithm for radar classification based on Doppler characteristics exploiting orthogonal Pseudo-Zernike polynomials,” IEEE Trans. Aero. Electron. Syst., vol. 51, no. 1, pp. 417–430, Jan. 2015, https://doi.org/10.1109/TAES.2014.130762.Search in Google Scholar

[14] J. Zabalza, C. Clemente, G. Di Caterina, J. Ren, J. J. Soraghan, and S. Marshall, “Robust PCA micro-Doppler classification using SVM on embedded systems,” IEEE Trans. Aero. Electron. Syst., vol. 50, no. 3, pp. 2304–2310, Jul. 2014, https://doi.org/10.1109/TAES.2014.130082.Search in Google Scholar

[15] C. Foued, M. Ammar, and Y. Arezki, “Detection and classification of ground targets using a Doppler radar,” in 2017 Seminar on Detection Systems Architectures and Technologies (DAT), 2017, pp. 1–8.10.1109/DAT.2017.7889168Search in Google Scholar

[16] A. Mesloub, K. Abed-Meraim, and A. Belouchrani, “Ground moving target classification based on micro-Doppler signature using novel spectral information features,” in 2017 European Radar Conference (EURAD), 2017, pp. 255–258.10.23919/EURAD.2017.8249195Search in Google Scholar

[17] H. Göksu, “Ground moving target recognition using log energy entropy of wavelet packets,” Electron. Lett., vol. 54, no. 4, pp. 233–235, Feb. 2018, https://doi.org/10.1049/el.2017.4267.Search in Google Scholar

[18] E. Al Hadhrami, M. Al Mufti, B. Taha, and N. Werghi, “Transfer learning with convolutional neural networks for moving target classification with micro-Doppler radar spectrograms,” In Int. Conf. Artif. Intell. Big Data, ICAIBD, vol. 2018, Chengdu, China, pp. 148–154.10.1109/ICAIBD.2018.8396184Search in Google Scholar

[19] E. Alhadhrami, M. Al-Mufti, B. Taha, and N. Werghi, “Learned micro-Doppler representations for targets classification based on spectrogram images,” IEEE Access, vol. 7, pp. 139377–139387, 2019, https://doi.org/10.1109/ACCESS.2019.2943567.Search in Google Scholar

[20] D. Lee, C. Cheung, and D. Pritsker, “Radar-based object classification using an Artificial neural network,” in Proceedings of the IEEE National Aerospace Electronics Conference, Dayton, OH, USA, NAECON, 2019, pp. 305–310.10.1109/NAECON46414.2019.9058319Search in Google Scholar

[21] A. Krizhevsky, I. Sutskever, and G. E. Hinton, “ImageNet classification with deep convolutional neural networks,” Commun. ACM, vol. 60, no. 6, pp. 84–90, May 2017, https://doi.org/10.1145/3065386.Search in Google Scholar

[22] T. A. F. R. L. (AFRL), The Air Force Moving and Stationary Target Recognition Database, 1996. https://www.sdms.afrl.af.mil/index.php?collection=mstar.Search in Google Scholar

[23] S. Chen, H. Wang, F. Xu, and Y.-Q. Jin, “Target classification using the deep convolutional networks for SAR images,” IEEE Trans. Geosci. Rem. Sens., vol. 54, no. 8, pp. 4806–4817, Aug. 2016, https://doi.org/10.1109/TGRS.2016.2551720.Search in Google Scholar

[24] X. Bai, R. Xue, L. Wang, and F. Zhou, “Sequence SAR image classification based on bidirectional convolution-recurrent network,” IEEE Trans. Geosci. Rem. Sens., vol. 57, no. 11, pp. 9223–9235, Nov. 2019, https://doi.org/10.1109/TGRS.2019.2925636.Search in Google Scholar

[25] R. Xue, X. Bai, and F. Zhou, “Spatial–temporal ensemble convolution for sequence SAR target classification,” IEEE Trans. Geosci. Rem. Sens., vol. 59, no. 2, pp. 1250–1262, Feb. 2021, https://doi.org/10.1109/TGRS.2020.2997288.Search in Google Scholar

[26] L. Wang, X. Bai, C. Gong, and F. Zhou, “Hybrid inference network for few-shot SAR automatic target recognition,” IEEE Trans. Geosci. Rem. Sens., vol. 59, no. 11, pp. 9257–9269, Nov. 2021, https://doi.org/10.1109/TGRS.2021.3051024.Search in Google Scholar

[27] M. Yang, X. Bai, L. Wang, and F. Zhou, “Mixed loss graph attention network for few-shot SAR target classification,” IEEE Trans. Geosci. Rem. Sens., vol. 60, pp. 1–13, 2022, https://doi.org/10.1109/TGRS.2021.3124336.Search in Google Scholar

[28] MathWorks, “Get Started with Image Preprocessing and Augmentation for Deep Learning - MATLAB & Simulink,” https://www.mathworks.com/help/images/get-started-with-image-preprocessing-and-augmentation-for-deep-learning.html?searchHighlight=jitterColorHSV&s_tid=srchtitle_jitterColorHSV_4 [accessed: Jan 08, 2022].Search in Google Scholar

[29] L. Stankovic, M. Dakovic, and T. Thayaparan, Time-frequency Signal Analysis with Applications, Artech House, 2013.Search in Google Scholar

[30] D. Bujaković, M. Andrić, B. Bondžulić, S. Mitrović, and S. Simić, “Time–frequency distribution analyses of Ku-band radar Doppler echo signals,” Frequenz, vol. 69, no. 3–4, pp. 119–128, Jan. 2015, https://doi.org/10.1515/freq-2014-0093.Search in Google Scholar

[31] M. Golgowski, “Classical versus deep learning methods for anomaly detection in ECG using wavelet transformation,” Prz. Elektrotechniczny, vol. 1, no. 6, pp. 74–78, Jun. 2021, https://doi.org/10.15199/48.2021.06.13.Search in Google Scholar

[32] A. Datta, B. Kolwadkar, A. Rauta, S. Handal, and V. V. Ingale, “ECG heartbeat classification using Wavelet transform and different Neural network Architectures,” in 20216th International Conference for Convergence in Technology (I2CT), 2021, pp. 1–7.10.1109/I2CT51068.2021.9418101Search in Google Scholar

[33] S. Zaman, S. Md, and R. Islam, “Classification of FNIRS using Wigner-ville distribution and CNN,” Int. J. Image Graph. Signal Process., vol. 13, no. 5, pp. 1–13, Oct. 2021, https://doi.org/10.5815/ijigsp.2021.05.01.Search in Google Scholar

[34] MathWorks, Datastore for Image Data. https://www.mathworks.com/help/matlab/ref/matlab.io.datastore.imagedatastore.html [accessed: Jan 16, 2022].Search in Google Scholar

[35] I. Chen, F. D. Johansson, and D. Sontag, “Why is my classifier discriminatory?” in Advances in Neural Information Processing Systems, 2018, pp. 3539–3550.Search in Google Scholar

[36] R. Lee and I.-Y. Chen, “The time complexity analysis of neural network model configurations,” in 2020 International Conference on Mathematics and Computers in Science and Engineering (MACISE), 2020, pp. 178–183.10.1109/MACISE49704.2020.00039Search in Google Scholar

Supplementary Material

The online version of this article offers supplementary material (https://doi.org/10.1515/freq-2022-0015).

Received: 2022-01-18

Accepted: 2022-10-11

Published Online: 2022-11-03

Published in Print: 2023-06-27

You are currently not able to access this content.

Supplementary Material

Articles in the same Issue

https://doi.org/10.1515/freq-2022-0015

Keywords for this article

automatic target recognition (ATR) algorithm; classification; convolutional neural network (CNN); convolutional recurrent neural network (CRNN); deep learning (DL); short-time Fourier transform (STFT)