Research Article
BibTex RIS Cite

DERİN ÖĞRENME YÖNTEMLERİ KULLANARAK GERÇEK ZAMANLI ARAÇ TESPİTİ

Year 2020, Volume: 13 Issue: 3, 1 - 14, 30.12.2020

Abstract

İnsansız hava araçları, sağlamış olduğu hareketlilik ve yüksek irtifa sayesinde günümüzde; alan tespiti, trafik izleme ve trafik kontrol gibi birçok alanda artan bir kullanıma sahiptir. İnsansız hava aracı kullanılarak yapılması hedeflenen önemli işlerden birisi de; alan resimleri yardımıyla gerçek zamanlı araç tespiti ve araç sayımı olarak görülmektedir. Bu amaç doğrultusunda derin öğrenme, makine öğrenmesi, gerçek zamanlı sınıflandırma ve tanımlama gibi birçok görüntü işleme tekniği ön plana çıkmaktadır. Fakat bu tekniklerin performansı, kullanılan veri ve işlenen alan doğrultusunda farklılık göstermektedir. Bu çalışma kapsamında derin öğrenme algoritmalarından YOLO algoritması referans alınarak, algoritmanın küçük obje tespitlerinde gösterdiği düşük performansı, tasarlanan ön tanımlı bir yapay sinir ağı yardımıyla iyileştirilmeye çalışılmıştır. Çalışma için uygun veri setleri toplanmış, algoritmaya uygun halde etiketlenmiş, sonrasında algoritma saf haliyle çalıştırılarak 50m, 75m, 100m ve 200m üzerinde araç tespit testleri uygulanmıştır. Paralelinde konvolüsyonel sinir ağları kullanılarak tasarlanan bir yapı yardımıyla, YOLO algoritmasının küçük obje tespitlerini iyileştirmek hedeflenmiştir. Tasarlanan ağ yardımıyla öğrenme sırasında algoritmanın objeler hakkında daha fazla bilgi sahibi olması sağlanmıştır. Çalışma sonucunda YOLO’ya yardımcı olarak sunulan yapının farklı veri setleri kullanılarak gerçekleştirilen testlerinde, YOLO’nun tespit oranını %4.3 arttırdığı ve 400x400 giriş değerlerinde 60fps değerine ulaşılabildiği görülmüştür. Çalışma kapsamında gerçek zamanlı uygulamalarda araç tespiti için kullanılabilecek bir yapı ortaya konmuştur.

References

  • Ammour, N., Alhichri, H., Bazi, Y., Benjdira, B., Alajlan, N., & Zuair, M. (2017). Deep learning approach for car detection in UAV imagery. Remote Sensing. https://doi.org/10.3390/rs9040312
  • Chen, X., Xiang, S., Liu, C. L., & Pan, C. H. (2014). Vehicle detection in satellite images by hybrid deep convolutional neural networks. IEEE Geoscience and Remote Sensing Letters. https://doi.org/10.1109/LGRS.2014.2309695
  • Deng, Z., Sun, H., Zhou, S., Zhao, J., & Zou, H. (2017). Toward Fast and Accurate Vehicle Detection in Aerial Images Using Coupled Region-Based Convolutional Neural Networks. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 10(8), 3652–3664. https://doi.org/10.1109/JSTARS.2017.2694890
  • Kyrkou, C., Timotheou, S., Kolios, P., Theocharides, T., & Panayiotou, C. G. (2018). Optimized vision-directed deployment of UAVs for rapid traffic monitoring. In 2018 IEEE International Conference on Consumer Electronics, ICCE 2018 (Vol. 2018-Janua, pp. 1–6). https://doi.org/10.1109/ICCE.2018.8326145
  • Liu, K.; Mattyus, G. DLR 3k Munich Vehicle Aerial Image Dataset. Available online: http://pba-freesoftware.eoc.dlr.de/3K_VehicleDetection_dataset.zip (accessed on 31 December 2015)
  • Liu, X., Yang, T., & Li, J. (2018). Real-time ground vehicle detection in aerial infrared imagery based on convolutional neural network. Electronics (Switzerland), 7(6). https://doi.org/10.3390/electronics7060078
  • Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. https://doi.org/10.1109/CVPR.2016.91
  • Ren, S., He, K., Girshick, R., & Sun, J. (2017). Faster R-CNN: Towards Real-Time Object Detection with. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(6), 1137–1149. https://doi.org/10.1109/TPAMI.2016.2577031
  • Sommer, L. W., Schuchert, T., & Beyerer, J. (2017). Fast deep vehicle detection in aerial images. In Proceedings - 2017 IEEE Winter Conference on Applications of Computer Vision, WACV 2017 (pp. 311–319). https://doi.org/10.1109/WACV.2017.41
  • Yang, M. Y., Liao, W., Li, X., Cao, Y., & Rosenhahn, B. (2019). Vehicle detection in aerial images. Photogrammetric Engineering and Remote Sensing, 85(4), 297–304. https://doi.org/10.14358/PERS.85.4.297

REAL-TIME VEHICLE DETECTION BY USING DEEP LEARNING METHODS

Year 2020, Volume: 13 Issue: 3, 1 - 14, 30.12.2020

Abstract

Thanks to unmanned aerial vehicles’ mobility and high
altitude; it has an increasing use in many areas such as area detection,
traffic monitoring and traffic control in today. Real time vehicle detection
and vehicle count are the one of the important works to be done by using
unmanned aerial vehicles.
For this
purpose, many image processing techniques such as deep learning, machine
learning, real-time classification and identification come to the fore.
However, the performance of these techniques differs
according to the data used and the area processed.
Within the scope of this study, the YOLO algorithm,
which is one of the deep learning algorithms, has been tried to be improved
with the help of a predefined artificial neural network. Firstly, appropriate
data sets for the study were collected, then labeled in accordance with the
algorithm. After that the algorithm was run in its pure form, and vehicle
detection tests were applied over 50m, 75m, 100m and 200m.
With the help of a structure designed using
convolutional neural networks in parallel, it is aimed to improve the small
object detection of the YOLO algorithm.
With
the help of the designed network, more information about objects are provided.
As a result of the study, it was observed that the structure, which was offered
as an aid to YOLO, increased the detection rate of the YOLO by 4.3% and reached
60fps at 400x400 input values. Within the scope of the study, a structure that
can be used for vehicle detection in real-time applications has been revealed.

References

  • Ammour, N., Alhichri, H., Bazi, Y., Benjdira, B., Alajlan, N., & Zuair, M. (2017). Deep learning approach for car detection in UAV imagery. Remote Sensing. https://doi.org/10.3390/rs9040312
  • Chen, X., Xiang, S., Liu, C. L., & Pan, C. H. (2014). Vehicle detection in satellite images by hybrid deep convolutional neural networks. IEEE Geoscience and Remote Sensing Letters. https://doi.org/10.1109/LGRS.2014.2309695
  • Deng, Z., Sun, H., Zhou, S., Zhao, J., & Zou, H. (2017). Toward Fast and Accurate Vehicle Detection in Aerial Images Using Coupled Region-Based Convolutional Neural Networks. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 10(8), 3652–3664. https://doi.org/10.1109/JSTARS.2017.2694890
  • Kyrkou, C., Timotheou, S., Kolios, P., Theocharides, T., & Panayiotou, C. G. (2018). Optimized vision-directed deployment of UAVs for rapid traffic monitoring. In 2018 IEEE International Conference on Consumer Electronics, ICCE 2018 (Vol. 2018-Janua, pp. 1–6). https://doi.org/10.1109/ICCE.2018.8326145
  • Liu, K.; Mattyus, G. DLR 3k Munich Vehicle Aerial Image Dataset. Available online: http://pba-freesoftware.eoc.dlr.de/3K_VehicleDetection_dataset.zip (accessed on 31 December 2015)
  • Liu, X., Yang, T., & Li, J. (2018). Real-time ground vehicle detection in aerial infrared imagery based on convolutional neural network. Electronics (Switzerland), 7(6). https://doi.org/10.3390/electronics7060078
  • Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. https://doi.org/10.1109/CVPR.2016.91
  • Ren, S., He, K., Girshick, R., & Sun, J. (2017). Faster R-CNN: Towards Real-Time Object Detection with. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(6), 1137–1149. https://doi.org/10.1109/TPAMI.2016.2577031
  • Sommer, L. W., Schuchert, T., & Beyerer, J. (2017). Fast deep vehicle detection in aerial images. In Proceedings - 2017 IEEE Winter Conference on Applications of Computer Vision, WACV 2017 (pp. 311–319). https://doi.org/10.1109/WACV.2017.41
  • Yang, M. Y., Liao, W., Li, X., Cao, Y., & Rosenhahn, B. (2019). Vehicle detection in aerial images. Photogrammetric Engineering and Remote Sensing, 85(4), 297–304. https://doi.org/10.14358/PERS.85.4.297
There are 10 citations in total.

Details

Primary Language Turkish
Subjects Engineering
Journal Section Naturel Sciences
Authors

Hüseyin Seçkin Dıkbayır

Halil İbrahim Bülbül

Publication Date December 30, 2020
Acceptance Date November 10, 2020
Published in Issue Year 2020 Volume: 13 Issue: 3

Cite

APA Dıkbayır, H. S., & Bülbül, H. İ. (2020). DERİN ÖĞRENME YÖNTEMLERİ KULLANARAK GERÇEK ZAMANLI ARAÇ TESPİTİ. TÜBAV Bilim Dergisi, 13(3), 1-14.
ISSN: 1308 - 4941