LAI Xin,CHEN Yongming,GUO Jun.Improved EAL-YOLOv8 for Small Target Detection in UVA Scenarios[J].Journal of Chengdu University of Information Technology,2026,41(02):160-166.[doi:10.16836/j.cnki.jcuit.2026.02.004]
改进的EAL-YOLOv8无人机场景小目标检测
- Title:
- Improved EAL-YOLOv8 for Small Target Detection in UVA Scenarios
- 文章编号:
- 2096-1618(2026)02-0160-07
- 分类号:
- TP183
- 文献标志码:
- A
- 摘要:
- 随着无人机技术的飞速发展以及计算机视觉技术的不断突破,小目标检测在实际生活中的应用更加广泛,使得无人机场景下的小目标检测成为研究热点。然而,在小目标检测任务中也面临着诸多技术挑战,尤其是小目标因像素占比低而导致的特征信息稀缺、容易漏检以及检测精度不高的问题,严重制约其在无人机场景等复杂环境中的应用效果。为此,提出一种针对小目标检测的EAL-YOLOv8算法。首先,在模型中引入ELSA注意力模块,在不降维的情况下实现精准定位; 其次,融入ASF-YOLO算法中的尺度序列特征融合模块和三重特征编码器,以实现精准分割并提高模型的特征融合能力; 最后,添加小目标检测层提升模型对小目标的感知能力,有助于减少漏检的情况。实验结果表明,在VisDrone2019数据集上EAL-YOLOv8算法比原始模型的平均检测精度YOLOv8提高了5.9个百分点且参数量减少11.4%。
- Abstract:
- With advancements in UVA tech and breakthroughs in computer vision and machine learning, the detection of small targets in real-world scenarios has gained broader application, particularly in drone scenarios. However, challenges such as feature scarcity, detection omissions, and low accuracy due to the small pixel ratio of targets hinder its effectiveness in complex environments. To tackle these, we propose the EAL-YOLOv8 algorithm. It incorporates an ELSA attention module for precise positioning without reducing dimensions, integrates SSFF and TFE from ASF-YOLO for better segmentation and feature fusion, and adds a small target detection layer to improve perception and reduce missed detections. Tests on the VisDrone2019 dataset show a 5.9 percentage points increase in average detection accuracy and an 11.4% reduction in parameters compared to the original YOLOv8 model.
参考文献/References:
[1] Yan Y,Deng C,Ma J,et al.A traffic sign recognition method under complex illumination conditions[J].IEEE Access,2023,11:39185-39196.
[2] Girshick R,Donahue J,Darrell T,et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2014:580-587.
[3] Girshick R.Fast R-CNN[C].Proceedings of the IEEE International Conference on Computer Vision.Santiago:IEEE,2015:1440-1448.
[4] Hechri A,Mtibaa A.Two-stage traffic sign detection and recognition based on SVM and convolutional neural networks[J].Iet Image Processing,2020,14(5):939-946.
[5] Redmon J,Farhadi A.YOLO9000:better,faster,stronger[C].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2017:7263-7271.
[6] Redmon J, Divvala S, Girshick R, et al.You only look once:unified,real-time object detection[C].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2016:779-788.
[7] Liu W,Anguelov D,Erha D,et al.SSD:Single Shot Multibox Detector[C].Proceedings of the European Conference on Computer Vision.Amsterdam:Springer,2016:21-37.
[8] 张润梅,肖钰霏,贾振楠,等.改进YOLOv7的无人机视角下复杂环境目标检测算法[J].光电工程,2024,51(5):92-103.
[9] 雷帮军,余翱,余快.基于YOLOv8s改进的小目标检测算法[J].无线电工程.2024,54(4):857-870.
[10] 聂源,赖惠成,高古学.改进YOLOv7+Bytetrack的小目标检测与追踪[J].计算机工程与应用,2024,60(12):189-202.
[11] 刘荣欣,卢胜男,刘晓天.基于改进YOLOv8的轻量型车辆目标检测算法[J].信息技术与信息化,2024(6):89-92.
[12] Siguana Q,Liang-Chieh C,Alan Y.DetectoRS:detec ting objects with recursive feature pyramid and switchable atrous convolution[C].IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).Nashville,TN,USA.IEEE,2021:10208-10219.
[13] 吴明杰,云利军,陈载清,等.改进YOLOv5s的无人机视角下小目标检测算法[J].计算机工程与应用,2024,60(2):191-199.
[14] Zhan W,Sun C F,Wang M C,et al.An improved Yolov5 realtime detection method for small objects captured by UAV[J].Soft Comput,2022,26(6):361-373.
[15] 李安达,吴瑞明,李旭东.改进 YOLOv7 的小目标检测算法研究[J].计算机工程与应用,2024,60(1):122-13.
[16] Xu W,Wan Y.ELA:Efficient Local Attention for Deep Convolutional Neural Networks[J].arxiv preprint arxiv:2403.01123,2024.
[17] 付孟丹,宣士斌,王婷,等.基于时间和空间注意力机制的视频异常检测[J].计算机技术与发展,2023,33(8):51-58.
[18] Kang M,Ting C M,Ting F F,et al.ASF-YOLO:A novel YOLO model with attentional scale sequence fusion for cell instance segmentation[J].Image and Vision Computing,2024,147:105057.
[19] 周飞,郭杜杜,王洋,等.基于改进YOLOv8的交通监控车辆检测算法[J].计算机工程与应用,2024,60(6):110-120.
[20] Du D W,Zhu P F,Wen L Y,et al.VisDrone-DET2019:the vision meets drone object detection in image challenge results[C].2019 IEEE/CVF International Conference on Computer Vision Workshop.Seoul:IEEE,2019:213-222.
备注/Memo
收稿日期:2024-08-22
基金项目:四川省科技厅资助项目(2022JDR0043); 四川省数值仿真重点实验室资助项目(KLNS-2023SZFZ002)
通信作者:郭俊.E-mail:junguo0407@cuit.edu.cn