Zihao Liu
ABSTRACT
With the rapid development of drones, drones are widely used in various fields and bring convenience to people's production and life. However, they also bring security problems to society and the country. Especially in airports or military areas, the flight of drones can cause some problems. In order to effectively supervise the drone, this paper proposes a real-time detection drone algorithm HR-YOLACT which is based on instance semantic segmentation, and designed a new drone data set. The proposed algorithm combines the real-time instance semantic segmentation algorithm YOLACT with the deep high-resolution representation classification network HRNet. Firstly, feature maps are extracted by HRNet's backbone network. Secondly, the feature pyramid network is used to further extract image features, so that the network has better classification ability. Finally, the improved prediction head is utilized to detect the boxes of drones. In addition, this paper uses cross entropy instead of focal loss as the loss function to obtain better network training speed and quality. The experimental results show that HR-YOLACT has faster detection speed and higher detection precision than existing popular real-time object detection and real-time instance semantic segmentation algorithms.
- Gokce, F., Ucoluk, G., Sahin, E., and Kalkan, S. 2015. Vision-Based Detection and Distance Estimation of Micro Unmanned Aerial Vehicles. Sensors 2015, 15, (9), 23805--23846. DOI= https://www.mdpi.com/1424-8220/15/9/23805Google Scholar
- Rozantsev, A., Lepetit, V., and Fua, P., 2017. Detecting Flying Objects Using a Single Moving Camera. IEEE Transactions on Pattern Analysis and Machine Intelligence 2017, 39, (5), 879--892. DOI= https://ieeexplore.ieee.org/document/7466125Google Scholar
- Aker, C., Kalkan, S. 2017. In Using deep networks for drone detection, 14th IEEE International Conference on Advanced Video and Signal Based Surveillance, AVSS(Lecce, ITALY, AUG 29-SEP 01, 2017)DOI= https://ieeexplore.ieee.org/document/8078539Google ScholarCross Ref
- Redmon, J., Divvala, S., Girshick, R., and Farhadi, A. 2016. You Only Look Once: Unified, Real-Time Object Detection. 2016 IEEE Conference on Computer Vision and Pattern Recognition CVPR( Seattle, WA, JUN 27-30, 2016), pp 779--788. DOI= https://ieeexplore.ieee.org/document/7780460Google Scholar
- He, K., Girshick, R., and Sun, J. 2017. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence 2017, 39, (6), 1137--1149. DOI= https://ieeexplore.ieee.org/document/7485869Google Scholar
- He, K., Gkioxari, G., Dollar, P., and Girshick, R. 2017. Mask R-CNN. 16th IEEE International Conference on Computer Vision ICCV (Venice, ITALY, OCT 22-29, 2017), pp 2980--2988. DOI= https://ieeexplore.ieee.org/document/8237584Google Scholar
- Bolya, D., Chong, Z., Fanyi, X., and Lee, Y. J. 2019. YOLACT: Real-time Instance Segmentation arXiv. arXiv 2019, 10 pp.-10 pp. DOI= https://arxiv.org/abs/1904.02689Google Scholar
- Sun, K., Xiao, B., Liu, D., and Wang, J. 2019. Deep High-Resolution Representation Learning for Human Pose Estimation. In arXiv e-prints, 2019.DOI= https://arxiv.org/abs/1908.07919Google Scholar
- Nguyen, P., Ravindranathan, M., Nguyen, A., Han, R., and Vu, T. 2016. Investigating Cost-effective RF-based Detection of Drones. 2nd Workshop on Micro Aerial Vehicle Networks, Systems, and Applications for Civilian Use (Singapore, SINGAPORE, JUN 26, 2016), p 17--22. DOI= https://dl.acm.org/citation.cfm?doid=2935620.2935632Google ScholarDigital Library
- Anwar, M. Z., Kaleem, Z., and Jamalipour, A. 2019. Machine Learning Inspired Sound-Based Amateur Drone Detection for Public Safety Applications. IEEE Transactions on Vehicular Technology 2019, 68, (3), 2526--2534. DOI= https://ieeexplore.ieee.org/document/8616877Google Scholar
- Thai, V. P., Zhong, W. X., Pham, T., Alam, S., and Duong, V. 2019. Detection, Tracking and Classification of Aircraft and Drones in Digital Towers using Machine Learning on Motion Patterns. 2019 Integrated Communications, Navigation and Surveillance Conference (Herndon, VA, APR, 09-11, 2019) DOI= https://ieeexplore.ieee.org/document/8735240Google Scholar
- Shinde, C., Lima, R., and Das, K., 2019. Multi-view Geometry and Deep Learning Based Drone Detection and Localization. 2019, p 289--294. DOI= https://ieeexplore.ieee.org/document/8715593Google Scholar
- Lin, T.-Y., Dollár, P., Girshick, R., He, K., Hariharan, B., and Belongie, S. 2016. Feature Pyramid Networks for Object Detection. 30th IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR (Honolulu, HI, JUL 21-26, 2017). DIO= https://ieeexplore.ieee.org/document/8099589Google Scholar
- Hosang, J., Benenson, R., and Schiele, B. 2017. Learning non-maximum suppression, 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR(Honolulu, HI, JUL 21-26, 2017).DIO=https://ieeexplore.ieee.org/document/8100168Google Scholar
- Kaiwen, D., Song, B., Lingxi, X., Honggang, Q., Qingming, H., and Qi, T. 2019. CenterNet: Keypoint Triplets for Object Detection arXiv. arXiv 2019, 10 pp.-10 pp. DOI= https://arxiv.org/abs/1904.08189Google Scholar
- Dai, J., Li, Y., He, K., and Sun, J., 2016. R-FCN: Object Detection via Region-based Fully Convolutional Networks. In arXiv e-prints, 2016. DOI= https://arxiv.org/abs/1605.06409Google Scholar
- Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Dollar, P., and Zitnick, C. L.,2014. Microsoft COCO: Common Objects in Context. In Computer Vision - Eccv 2014, Pt V, Fleet, D.; Pajdla, T.; Schiele, B.; Tuytelaars, T., Eds. 2014; Vol. 8693, pp 740--755.DOI = https://authors.library.caltech.edu/94215/2/1405.0312.pdfGoogle Scholar
Index Terms
- A Real-Time Detection Drone Algorithm Based on Instance Semantic Segmentation
Recommendations
Real-time high-resolution omnidirectional imaging platform for drone detection and tracking
AbstractDrones have become steadily affordable, which raises privacy and security concerns as well as interest in drone detection systems. On the other hand, drone detection is a challenging task due to small dimensions of drones, difficulty of long-...
Comments