改进YOLOv5的无人机影像小目标物体识别算法
Improved Algorithm for Small Target Object Recognition in Drone Images Based on YOLOv5
DOI: 10.12677/MOS.2023.126490, PDF,  被引量   
作者: 窦同旭, 曾 勇, 杨 冲, 李 昂：盐城工学院机械学院，江苏 盐城
关键词: 无人机影像；YOLOv5；小目标；辅助训练头；Drone Images； YOLOv5； Small Targets； Auxiliary Head
摘要: 无人机图像目标检测是流行的图像识别任务之一。然而，由于无人机航拍影像具有目标尺度变化大、背景复杂等特点，现有的检测器难以准确检测小目标。为了解决这个问题，本文提出了改进的算法模型YOLO-ADOP，用于提高在无人机影像中小目标的检测效果。首先，通过调整检测分支以提高模型在小目标下的检测能力。其次，将分类任务与回归任务解耦，设计了解耦头结合辅助训练头的预测头。最后，使用EIOU优化损失函数，使用OTA改进标签分配策略。通过VisDrone数据集的实验结果表明，改进后的YOLO-ADOP模型较YOLOv5模型在640分辨率和1536分辨率上的AP50指标分别提升了7.18%和4.65%。对比其他主流模型对于小目标拥有更好的检测效果，能够有效完成无人机航拍影像的小目标检测任务。
Abstract: Unmanned aerial vehicle (UAV) image object detection is a popular task in image recognition. How-ever, existing detectors face difficulties in accurately detecting small objects in UAV aerial images due to their large scale variations and complex backgrounds. To address this issue, this paper pro-poses an improved algorithm model called YOLO-ADOP, aiming to enhance the detection perfor-mance of small objects in UAV images. Firstly, the detection branch is adjusted to enhance the mod-el’s detection capability for small targets. Secondly, the classification task and regression task are decoupled, and a decoupled head with an auxiliary head is designed for prediction. Finally, the EIOU optimization loss function and OTA improved label assignment strategy are employed. Experi-mental results on the VisDrone dataset demonstrate that the improved YOLO-ADOP model achieves a 7.18% and 4.65% improvement in AP50 metric at resolutions of 640 and 1536, respectively, compared to the YOLOv5 model. It outperforms other mainstream models in detecting small targets and can effectively accomplish the task of small target detection in drone aerial images.
文章引用：窦同旭, 曾勇, 杨冲, 李昂. 改进YOLOv5的无人机影像小目标物体识别算法[J]. 建模与仿真, 2023, 12(6): 5395-5407. https://doi.org/10.12677/MOS.2023.126490

参考文献

[1] Redmon, J., Divvala, S., Girshick, R., et al. (2016) You Only Look Once: Unified, Real-Time Object Detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, 27-30 June 2016, 779-788. [Google Scholar] [CrossRef] 
[2] Redmon, J. and Farhadi, A. (2017) YOLO9000: Better, Faster, Stronger. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, 21-26 July 2017, 7263-7271. [Google Scholar] [CrossRef] 
[3] Redmon, J. and Farhadi, A. (2018) Yolov3: An Incremental Improve-ment.
[4] Bochkovskiy, A., Wang, C.-Y. and Liao, H.-Y.M. (2020) Yolov4: Optimal Speed and Accuracy of Object Detec-tion.
[5] Li, C., Li, L., Geng, Y., et al. (2023) YOLOv6 v3.0: A Full-Scale Reloading.
[6] Wang, C.-Y., Bochkovskiy, A. and Liao, H.-Y.M. (2022) YOLOv7: Trainable Bag-of-Freebies Sets New State-of-the-Art for Real-Time Object Detectors. 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Vancouver, 17-24 June 2023, 7464-7475. [Google Scholar] [CrossRef] 
[7] Jocher, G., Chaurasia, A., Stoken, A., et al. (2022) Ultralyt-ics/yolov5: v7.0-YOLOv5 SOTA Realtime Instance Segmentation. Zenodo.
[8] Liu, W., Anguelov, D., Erhan, D., et al. (2016) Ssd: Single Shot Multibox Detector. Computer Vision-ECCV 2016: 14th European Conference, Amsterdam, 11-14 Oc-tober 2016, 21-37. [Google Scholar] [CrossRef] 
[9] Girshick, R. (2015) Fast R-CNN. Proceedings of the IEEE International Conference on Computer Vision, Santiago, 7-13 December 2015, 1440-1448. [Google Scholar] [CrossRef] 
[10] Ren, S., He, K., Girshick, R., et al. (2015) Faster R-CNN: Towards Re-al-Time Object Detection with Region Proposal Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39, 1137-1149.
[11] Law, H. and Deng, J. (2018) Cornernet: Detecting Objects as Paired Keypoints. Proceedings of the Euro-pean Conference on Computer Vision (ECCV), Munich, 8-14 September 2018, 734-750. [Google Scholar] [CrossRef] 
[12] Duan, K., Bai, S., Xie, L., et al. (2019) Centernet: Keypoint Triplets for Object Detection. Proceedings of the IEEE/CVF International Conference on Computer Vision, Seoul, 27 October-2 No-vember 2019, 6569-6578. [Google Scholar] [CrossRef] 
[13] Ge, Z., Liu, S., Wang, F., et al. (2021) Yolox: Exceeding Yolo Series in 2021.
[14] Vaswani, A., Shazeer, N., Parmar, N., et al. (2017) Attention Is All You Need. NIPS’17: Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, 4-9 December 2017, 6000-6010.
[15] Liu, Z., Lin, Y., Cao, Y., et al. (2021) Swin Transformer: Hierarchical Vision Transformer Using Shifted Windows. Proceedings of the IEEE/CVF International Conference on Computer Vision, Montreal, 11-17 October 2021, 10012- 10022. [Google Scholar] [CrossRef] 
[16] Du, D., Zhu, P., Wen, L., et al. (2019) VisDrone-DET2019: The Vision Meets Drone Object Detection in Image Challenge Results. Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops, Seoul, 27-28 October 2019, 213-226.
[17] Du, D., Qi, Y., Yu, H., et al. (2018) The Unmanned Aerial Vehicle Benchmark: Object Detection and Tracking. Proceedings of the European Conference on Computer Vision (ECCV), Munich, 8-14 September 2018, 370-386. [Google Scholar] [CrossRef] 
[18] Zhang, R., Newsam, S., Shao, Z., et al. (2021) Multi-Scale Adver-sarial Network for Vehicle Detection in UAV Imagery. ISPRS Journal of Photogrammetry and Remote Sensing, 180, 283-295. [Google Scholar] [CrossRef] 
[19] Deng, S., Li, S., Xie, K., et al. (2021) A Global-Local Self-Adaptive Network for Drone-View Object Detection. IEEE Trans Image Process, 30, 1556-1569. [Google Scholar] [CrossRef] 
[20] Gallo, I., Rehman, A.U., Dehkordi, R.H., et al. (2023) Deep Object De-tection of Crop Weeds: Performance of YOLOv7 on a Real Case Dataset from UAV Images. Remote Sensing, 15, Article No. 539. [Google Scholar] [CrossRef] 
[21] Ringwald, T., Sommer, L., Schumann, A., et al. (2019) UAV-Net: A Fast Aerial Vehicle Detector for Mobile Platforms. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, Long Beach, 16-17 June 2019, 544-552. [Google Scholar] [CrossRef] 
[22] Zhu, X., Lyu, S., Wang, X., et al. (2021) TPH-YOLOv5: Improved YOLOv5 Based on Transformer Prediction Head for Object Detection on Drone-Captured Scenarios. Proceedings of the IEEE/CVF International Conference on Computer Vision, Montreal, 11-17 Oc-tober 2021, 2778-2788. [Google Scholar] [CrossRef] 
[23] Zhao, Q., Liu, B., Lyu, S., et al. (2023) TPH-YOLOv5++: Boosting Object Detection on Drone-Captured Scenarios with Cross-Layer Asymmetric Transformer. Re-mote Sensing, 15, Article No. 1687. [Google Scholar] [CrossRef] 
[24] Zhang, Y.-F., Ren, W., Zhang, Z., et al. (2022) Focal and Efficient IOU Loss for Accurate Bounding Box Regression. Neurocomputing, 506, 146-157. [Google Scholar] [CrossRef] 
[25] Ge, Z., Liu, S., Li, Z., et al. (2021) Ota: Optimal Transport Assignment for Object Detection. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Nashville, 20-25 June 2021, 303-312. [Google Scholar] [CrossRef] 

为你推荐