Multi-object Detection in Night Time

  • Pavan Sai Vemulapalli
  • Ajay Kumar Rachuri
  • Heena Patel
  • Kishor P. Upla
Keywords: Thermal imaging; Deep Learning; object detection; convolutional neural network; Yolo; Multi class detection.

Abstract

This paper discusses the work on detecting multi-objects such as person and car in thermal image captured during night time using deep learning architecture. Thermal images are superior to the visible images when it comes to the amount of useful information required to detect the objects during night time. Thermal imager uses radiation emitted by the objects to create an image and improve the visibility of objects in a dark environment. Contrast to that, visible image does not provide useful information in darkness. Hence, it is better to use thermal images to detect objects present in darkness. The state-of-the-art, Yolo-v3, deep learning convolutional neural network model is the latest version of the Yolo model in which the feature extraction layer contains a much deeper network. The results of detecting person and car in the thermal images obtained by the proposed model are compared with the results of Yolo- v3. Experimental results show that there is a significant improvement in detecting person and car in the thermal images in terms of mean average precision (mAP) using the proposed method.

References

[1]R. Abbott, J. Del Rincon, B. Connor, and N. Robertson. Deep object classification in low resolution lwir imagery via transfer learning. In Proceedings of 5th IMA Conference on Mathematics in Defence, volume 2, 2017.
[2]M. Bertozzi, A. Broggi, C. H. Gomez, R. Fedriga, G. Vez- zoni, and M. DelRose. Pedestrian detection in far infrared images based on the use of probabilistic templates. In 2007 IEEE Intelligent Vehicles Symposium, pages 327–332. IEEE, 2007.
[3]T. P. Breckon, A. Gaszczak, J. Han, M. L. Eichner, and S. E. Barnes. Multi-modal target detection for autonomous wide area search and surveillance. In Emerging Technologies in Security and Defence; and Quantum Security II; and Un- manned Sensor Systems X, volume 8899, page 889913. In- ternational Society for Optics and Photonics, 2013.
[4]N. Dalal and B. Triggs. Histograms of oriented gradients for human detection. 2005.
[5]A. Dutta and A. Zisserman. The vgg image annotator (via). arXiv preprint arXiv:1904.10699, 2019.
[6]M. Everingham, L. Van Gool, C. K. Williams, J. Winn, and A. Zisserman. The pascal visual object classes (voc) challenge. International journal of computer vision, 88(2):303– 338, 2010.
[7]R. Girshick. Fast r-cnn. In Proceedings of the IEEE international conference on computer vision, pages 1440–1448, 2015.
[8]R. Girshick, J. Donahue, T. Darrell, and J. Malik. Rich feature hierarchies for accurate object detection and semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 580–587, 2014.
[9]K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016.
[10]S. Hwang, J. Park, N. Kim, Y. Choi, and I. So Kweon. Multispectral pedestrian detection: Benchmark dataset and baseline. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1037–1045, 2015.
[11]S. Ioffe and C. Szegedy. Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv preprint arXiv:1502.03167, 2015.
[12]M. Ivasiˇ c-Kos, M. Kri ´ sto, and M. Pobar. Human detection ˇ in thermal imaging using yolo. In Proceedings of the 2019 5th International Conference on Computer and Technology Applications, pages 20–24. ACM, 2019.
[13]K. Jo, J. Im, J. Kim, and D.-S. Kim. A real-time multi-class multi-object tracker using yolov2. In 2017 IEEE International Conference on Signal and Image Processing Applications (ICSIPA), pages 507–511. IEEE, 2017.
[14]A. Kathuria. What’s new in yolo v3? https://towardsdatascience.com/yolo-v3-object-detection53fb7d3bfe6b. Accessed: 2019-07-05.
Published
2020-03-26
How to Cite
Vemulapalli, P. S., Rachuri, A. K., Patel, H., & Upla, K. P. (2020). Multi-object Detection in Night Time. Asian Journal For Convergence In Technology (AJCT), 5(3), 01-07. Retrieved from http://www.asianssr.org/index.php/ajct/article/view/903