DOI:https://doi.org/10.15407/kvt199.01.005
Cybernetics and Computer Engineering, 2020, 1(199)
KYYKO V.M., PhD (Engineering),
Senior Researcher of Pattern Recognition Department
e-mail: vkiiko@gmail.com
International Research and Training Center for Information Technologies and Systems of the National Academy of Sciences of Ukraine and Ministry of Education and Science of Ukraine,
40, Acad. Glushkov av., Kyiv, 03187, Ukraine
MATCHING BASED MULTISTYLE LICENSE PLATE RECOGNITION
Introduction. A State-of-the-Art of license plate (LP) recognition from images is observed. Despite the fact that License Plate Recognition (LPR) is often regarded as a solved task, country-specific systems are mostly designed that limits their application. Pay attention to the increasing mobility, effective LPR systems must handle multistyle LP including multinational ones that have different fonts and syntax. Another bottleneck of LPR is that accuracy of recognition at varying environmental conditions as well as of low resolution or degraded LP usually is rather low.
The purpose of the paper is to develop algorithms for multistyle single line LP learning and recognition from images as well as for comparatively low resolution LP processing.
Methods. Randomized Hough transform is used for detecting horizontal frame lines and subsequent LP localization in image. Structural feature matching approach is used for
character recognition. Correction of recognition results is based on calculation of modified Levenstein distance (MGED) between LP description and templates.
Results. New algorithms for multinational license plate learning and recognition from images are proposed. Localization of LP in images is based on LP frame detection using a randomized Hough transform to detect horizontal contour frame line segments. Recognition of segmented characters inside LP is based on searching key points in skeletonized character images and matching these points with etalons. Correction of recognition LP output is carried out by matching and defining MGED between LP input description and templates. Online active learning for recognition of new LP symbols and templates is also proposed. Results of testing developed algorithms and software are described.
Conclusions. Algorithms for multistyle LP localization and recognition from images are proposed. Control and correction of recognition results is based on calculation of MGED between input LP description and templates which are more general in comparison conventional text lines. As future work, it is planned to increase accuracy by learning feature etalon weights, as well as to consider other LP types for recognition and to test developed means on more representative date samples.
Keywords: license plate localization and recognition, key points matching, Levenstein distance.
REFERENCES
1 Shan D., Mahmoud I., Shehata M., Badawy W.. Automatic License Plate Recognition (ALPR): A State-of-the-Art Review. IEEE Transactions on circuit and systems for video technology. 2013, Vol. 23, no. 2, pp. 311-325.
https://doi.org/10.1109/TCSVT.2012.2203741
2 Anagnostopoulos C.N.E., Anagnostopoulos I.E., Psoroulas I.D., Loumos V., Kayafas E. License plate recognition from still images and video sequences: A survey. IEEE Trans. Intell. Transp. Syst. 2008, Vol. 9, no. 3, pp. 377-391.
https://doi.org/10.1109/TITS.2008.922938
3 Anagnostopoulos C.N.E., Anagnostopoulos I.E., Loumos V., Kayafas E. A license plate-recognition algorithm for intelligent transportation system applications. IEEE Transactions on Intelligent Transportation Systems. 2006, 7 (3), pp. 377-392.
https://doi.org/10.1109/TITS.2006.880641
4 Oliveira-Neto F.M., Han L.D., Jeong M.K. An Online Self-Learning Algorithm for License Plate Matching. IEEE Transactions on intelligent transportation systems. 2013, Vol. 14, no. 4, pp. 1806-1816.
https://doi.org/10.1109/TITS.2013.2270107
5 Thome N., Robinault L. A cognitive and video-based approach for multinational license plate recognition. Mach. Vision Application. 2011, Vol. 22, no. 2, pp. 389-407.
https://doi.org/10.1007/s00138-010-0246-3
6 Laroca R., Severo E., Zanlorensi L.A., Oliveira L.S., Gonçalves G.R., Schwartz W.R., Menotti D. A Robust Real-Time Automatic License Plate Recognition Based on the YOLO Detector. 2018 International Joint Conference on Neural Networks (IJCNN), (Rio de Janeiro, 8-13th of Jul, 2018). Rio de Janeiro, Brazil, 2018, pp. 1-10.
https://doi.org/10.1109/IJCNN.2018.8489629
7 V.M. Kyyko. Localization and recognition of license plates from images. Upravlaûŝie sistemy i mašiny. 2017, no. 6, pp. 26-34, 40 (in Russian).
https://doi.org/10.15407/usim.2017.06.026
8 Kyyko V.M., Schlesinger M.I. Width – Independent Fast Skeletonization Algorithm for Binary Pictures. Intern. Journal of Imaging Systems and Technology. 1991, Vol. 3, no. 3, pp. 222-226.
https://doi.org/10.1002/ima.1850030307
9 Levenshtein V.I.Binary codes capable of correcting deletions, insertions, and reversals.Doklady Academy of sciences USSR. 1965, no. 163 (4), pp. 845-848 (in Russian).
10 Duda R.O., Hart P.E. Use of the Hough Transformation to Detect Lines and Curves. Pictures. Comm. ACM. 1972, Vol. 15, pp. 11-15.
https://doi.org/10.1145/361237.361242
11 Kiryati N., Eldar Y., Bruckstein A.M. A Probabilistic Hough transform. Pattern Recognition. 1991, no.24, (4), pp. 303-316.
https://doi.org/10.1016/0031-3203(91)90073-E
12 Schlesinger M.I. Fast implementation of one class of linear convolutions. Theoretical and applied issues of image recognition. 1991, Kyiv, pp. 61-69 (in Russian).
13 Ocuda T., Tanaka E., Kasai T. A method for correction of garbled words based on the Levenstein metric. IEEE Transactions on Computers. 1976, no. 25 (C-2), pp. 172-177.
https://doi.org/10.1109/TC.1976.5009232
14 G. Seni, Kripasundar V., Srihari R. Generalizing edit distance to incorporate domain information: handwritten text recognition as a case study. Pattern Recognition. 1996, no. 29 (3), pp. 405-414.
https://doi.org/10.1016/0031-3203(95)00102-6
15 Jonker R., Volgenant A. A shortest augmenting path algorithm for dense and sparse linear assignment problems. Computing. 1987, no. 38, pp. 325-340.
https://doi.org/10.1007/BF02278710
16 Wei J. Markov edit distance. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2004, no. 6(3), pp. 311-320.
https://doi.org/10.1109/TPAMI.2004.1262315
17 Schlesinger M.,Hlavac V. Ten Lectures on Statistical and Structural Pattern Recognition. Dordrecht / Boston / London, 2002, 520 p.
https://doi.org/10.1007/978-94-017-3217-8
18 URL: http://viatec.ua/product/nomerok-4 (Last accessed: 19. 09.2019)
19 URL: http://isscctv.com.ua/?page=SecurOSAuto (Last accessed: 10.07.2019)
Received 11.11.2019