Recognition of Safety Helmet Wearing Of Operator Local Ground Floor Unit 2 Suralaya PGU Based On Improved YOLOv3

https://doi.org/10.22146/jmdt.85645

Firlan Maulana Ruaz(1*)

(1) PT. PLN Indonesia Power, Suralaya Power Generation Unit, Suralaya, Indonesia
(*) Corresponding Author

Abstract


PT. PLN Indonesia Power Suralaya Power Generation Unit requires its operators on the local ground floor to wear safety helmets to prevent head injuries while performing their tasks. Observation of workers wearing safety helmet must be ensured by the company to prevent head injury by operators. The Culture Transformation Program of K3 PT PLN Indonesia Power brings a new observation method to detect safety helmet wearing of operator local ground floor using real-time detection. YOLOv3 is an application which can be used to real-time detection helmet wearing of workers using the darknet53 algorithm based on YOLOv1 and YOLOv2 by using data images that have been collected. Based on YOLOv3 model, the improved version of YOLOv3 is proposed to improve accuracy and speed detection of safety helmet wearing by combining multi-scale detection training. Different YOLOv3 versions will be used to compare results of helmet recognition. The improved YOLOv3 show results 96.63% mAP50 and 725,711 milliseconds better than the other versions of YOLO to detect helmet safety. The experimental result show that the improved YOLOv3 have satisfying with the detection speed and accuracy of safety helmet wearing detection by operators at PT PLN Indonesia Power Suralaya Power Generation Unit.

Keywords


YOLO; safety helmet; deep learning; real-time detection; deep residual network.

Full Text:

PDF


References

Dollar, P., Appel, R., & Belongie, S. (2014). Fast feature pyramids for object detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 36(08), 1532–1545

Girshick, R. (2015) Fast R-CNN[C]//IEEE International Conference on Computer Vision. IEEE, 2015:1440-1448.

He K , Zhang X , Ren S , et al. (2016). Deep Residual Learning for Image Recognition[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE Computer Society: Page 770- 778.

Krizhevsky, A., Sutskever, I., Hinton, G.E., (2012). ImageNet Classification With Deep Convolutional Neural Networks[C]//International Conference on Neural Information Processing Systems. Curran Associates Inc, 1097-1105.

Liu, W., Anguelov, D., & Erhan, D. (2016). SSD: Single shot multibox detector. Computer Vision - ECCV 2016, PT I. Lecture Notes in Computer Science, 9905, 21–37.

Ma, Q.,Zhu, B,Zhang, H.W.,Zhang, Y., Jiang, Y.C. (2019) Low-altitude UAV Detection and Recognition Method Based on Optimized YOLOv3.J. Laser & Optoelectronics Progress. Vol 56, No 20.

Mazzia, V., Khaliq, A., & Salvetti, F. (2020). Real-time apple detection system using embedded systems with hardware accelerators: An edge AI application. IEEE ACCESS, 08, 9102–9114.

Redmon, J., Divvala, S., Girshick, R., et al. (2015) You Only Look Once: Unified, Real-Time Object Detection. J. Page 779-788

Redmon, J., Farhadi, A. (2018). YOLOv3: An Incremental Improvement. J. IEEE Conference on ComputerVision and Pattern Recognition. IEEE, 89-95.

Redmon, J., Farhadi. A. (2017). YOLO9000: Better, Faster,Stronger[C]//IEEE Conference on Computer Visionand Pattern Recognition. IEEE. Page 6517-6525.

Ren S , He K , Girshick R , et al. (2015). Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence. Vol 39, No 6:1137- 1149

Ren, S., He, K., Girshick, R., et al. (2015). Faster R-CNN: Towards Real-Time Object Detection With Region Proposal Networks[C]//International Conference on Neural Information Processing Systems. MIT Press, 91-99.

Zhang, Y., Shen, Y. L., & Zhang, J. (2019). An improved Tiny YOLOv3 pedestrian detection algorithm. OPTIK, 183, 17–23.



DOI: https://doi.org/10.22146/jmdt.85645

Article Metrics

Abstract views : 356 | views : 269

Refbacks

  • There are currently no refbacks.


Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.