Motion Detection with Background Modelling and Optical Flow

Yıl 2021, Cilt: 14 Sayı: 3, 223 - 228, 31.07.2021

İbrahim Delibaşoğlu

https://doi.org/10.17671/gazibtd.846961

Öz

Motion detection is a challenging task and may be used as a pre-processing step in different computer vision tasks. Methods proposed for motion detection are mostly based on background modelling and subtraction. In this study, a method is proposed with background modelling and optical flow vectors. Farneback method is applied to estimate optical flow vectors. Optical flow is used to determine the threshold value of each pixel applied in background subtraction step of proposed method. The experimental results show that proposed approach using optical flow in background subtraction improves the performance according to a static threshold. The proposed method is evaluated on different subset images of CDNET-2014 dataset and has a reasonable performance against methods in the literature. Proposed method has similar performance in the F1 metric compared to the methods in the literature, but it is observed that it has the best average performance in the PWC metric, which gives the ratio of wrongly detected or missed moving pixels.

Anahtar Kelimeler

motion detection, moving object detection, optical flow

Arkaplan Modellemesi ve Optik Akış ile Hareket Tespiti

Öz

Hareket tespiti çeşitli bilgisayarlı görme problemlerinde ön-işlem aşamasında kullanılmaktadır ve üzerinde çalışılan önemli konulardan birisidir. Hareket tanıma için önerilen yöntemler çoğunlukla arka plan modellemesi ve çıkarımına dayanmaktadır. Bu çalışmada, arka plan modelleme ve optik akış vektörlerinin kullanıldığı bir yöntem önerilmiştir. Optik akış tahmin edilmesinde Farneback yöntemi kullanılmıştır. Önerilen yöntemde optik akış, arka plan çıkarımı aşamasında her bir piksel için uygulanacak eşik değeri belirlemek için kullanılmıştır. Deneysel sonuçlar arka plan çıkarımı yaparken optik akış bilgisini kullanmanın sabit eşik değer uygulamaya nazaran performansı arttırdığını göstermiştir. Önerilen yöntem CDNET-2014 veri kümesinden farklı altküme görüntüleri üzerinde değerlendirilmiş ve literatürdeki yöntemler karşısında iyi sonuçlar elde edilmiştir. F1 performans kriterinde literatürdeki yöntemlere çok yakın sonuçlar elde edilirken, hatalı tespit edilen veya ıskalanan hareketli piksel oranını veren PWC metriğinde ise en iyi ortalama performansa ulaşıldığı gözlemlenmiştir.

Anahtar Kelimeler

hareket tespiti, hareketli nesne tespiti, optik akış

Kaynakça

• C. Marie-Neige, T. Bouwmans. "Moving Objects Detection with a Moving Camera: A Comprehensive Review", arXiv preprint arXiv:2001.05238, 2020.
• H. Çakır, H. K. Babacan, "Hareketi Algılayan Kamera Destekli Güvenlik", Bilişim Teknolojileri Dergisi, 4(2), 2011.
• R. Collins, A. J. Lipton, T. Kanade, H. Fujiyoshi, D. Duggins, Y. Tsin, D. Tolliver, N. Enomoto, O. Hasegawa, P. Burt, et al., A system for video surveillance and monitoring, VSAM final report, 1–68, 2000.
• L. Zhao, Q. Tong, H. Wang., “Study on moving-object-detection arithmetic based on w4 theory”, 2011 2nd International Conference on Artificial Intelligence, Management Science and Electronic Commerce (AIMSEC), IEEE, 4387–4390, 2011.
• H. Junjie, et al., "Optical flow based real-time moving object detection in unconstrained scenes", arXiv preprint arXiv:1807.04890, 2018.
• I. Eddy, et al., "Flownet 2.0: Evolution of optical flow estimation with deep networks.", IEEE conference on computer vision and pattern recognition, 2462-2470, 2017.
• T. Bouwmans, "Traditional Approaches in Background Modeling for Video Surveillance.", Handbook Background Modeling and Foreground Detection for Video Surveillance, Taylor and Francis Group, T. Bouwmans, B. Hoferlin, F. Porikli, A. Vacavant, 2014.
• A. Gianni, et al., "EFIC: edge based foreground background segmentation and interior classification for dynamic camera viewpoints", International Conference on Advanced Concepts for Intelligent Vision Systems. Springer, Cham, 2015.
• M. Yi, Kwang, et al., "Detection of moving objects with non-stationary cameras in 5.8 ms: Bringing motion detection to your mobile device", IEEE Conference on Computer Vision and Pattern Recognition Workshops. 2013.
• Z. Zivkovic, "Improved adaptive Gaussian mixture model for background subtraction.", 17th International Conference on Pattern Recognition, IEEE, 2004.
• Z. Zivkovic, F. Van Der Heijden, "Efficient adaptive density estimation per image pixel for the task of background subtraction", Pattern recognition letters, 27(7), 773-780, 2006.
• M. De Gregorio, M. Giordano, "WiSARDrp for Change Detection in Video Sequences", European Symposium on Artificial Neural Network (ESANN), 453-458, 2017.
• B. Olivier, M. Van Droogenbroeck, "ViBe: A universal background subtraction algorithm for video sequences", IEEE Transactions on Image processing, 20(6), 1709-1724, 2010.
• Y. Kimin, J. Lim, J. Y. Choi, “Scene conditional background update for moving object detection in a moving camera”, Pattern Recognition Letters, 88, 57-63.
• Y. Yang, L. Kurnianggoro, K. Jo, “Moving object detection for a moving camera based on global motion compensation and adaptive background model”, International Journal of Control, Automation and Systems, 17(7), 1866-1874, 2019.
• G. Farnebäck, "Two-frame motion estimation based on polynomial expansion", Scandinavian conference on Image analysis, Springer, Berlin, Heidelberg, 363-370, 2003.
• Internet: http://changedetection.net/, 25.12.2020.
• A. Sanin, C. Sanderson, B. C. Lovell, "Shadow detection: A survey and comparative evaluation of recent methods", Pattern recognition, 45(4), 1684-1695, 2012.