基于特征空间与坐标卷积的小目标检测算法

doi:10.3778/j.issn.1002-8331.2306-0407

摘要/Abstract

摘要： 为解决无人机高空拍摄面临的小目标聚集不易识别、可提取特征少的问题，提出一种特征空间与坐标卷积结合的小目标检测算法。在YOLOv5网络架构中加入特征空间模块（feature spatial model，FSM），利用卷积为不同特征的感受野分配自适应权重，增强主干网络特征提取能力；将坐标卷积模块（coordinate convolution model，CCM）嵌入模型颈部，精准定位目标所在位置，提高密集场景下小目标检测精度；删减原始层并添加小目标检测层，减少语义损失，充分提取浅层特征图中信息，强化高空图像中微小目标检测性能。实验结果表明，在VisDrone2019数据集上，改进后模型的精确率较YOLOv5提高4.1个百分点，mAP@0.5和mAP@0.5:0.95分别提高4.6个百分点和3.2个百分点，从而验证了提出模型在无人机检测小目标场景中具备有效性。

关键词: 目标检测, YOLOv5, 小目标, 特征增强, 坐标卷积

Abstract: In order to solve the problem that the small object detdction is difficult to identify and the few extractable features faced by UAV high-altitude shooting, a small target detection algorithm combining feature space and coordinate convolution is proposed. The feature space module (FSM) is added to the YOLOv5 network architecture, and the convolution is used to assign adaptive weights to the receptive fields of different features, so as to enhance the feature extraction ability of the backbone network. It embeds the coordinate convolution model (CCM) in the neck of the model, accurately locates the location of the target and improves the detection accuracy of small targets in dense scenes. It deletes the original layer and adds a small target detection layer to reduce semantic loss, fully extracts the information in the shallow feature maps, and strengthens tiny objects in the high-altitude images detection performance. The experimental results show that on the VisDrone2019 dataset, the accuracy of the improved model is 4.1 percentage points higher than that of YOLOv5, and mAP@0.5 and mAP@0.5:0.95 are 4.6 percentage points and 3.2 percentage points higher respectively, thus verifying that the model proposed in this paper is effective in drone detection effective in the target scenario.

Key words: object detection, YOLOv5, small target, feature enhancement, coordinate convolution

程梦洋, 葛海波, 何文昊, 马赛, 安玉. 基于特征空间与坐标卷积的小目标检测算法[J]. 计算机工程与应用, 2024, 60(19): 209-220.

CHENG Mengyang, GE Haibo, HE Wenhao, MA Sai, AN Yu. Small Object Detection Based on Feature Space and Coordinate Convolution[J]. Computer Engineering and Applications, 2024, 60(19): 209-220.

参考文献

[1] SILVA A, BASSO M, MENDES P, et al. A map building and sharing framework for multiple UAV systems[C]//Proceedings of the 2022 International Conference on Unmanned Aircraft Systems (ICUAS), 2022: 1333-1342.
[2] DING J, ZHANG J, ZHAN Z, et al. A precision efficient method for collapsed building detection in post-earthquake UAV images based on the improved NMS algorithm and Faster R-CNN[J]. Remote Sensing, 2022, 14(3): 663.
[3] YANG C, HUANG Z, WANG N. QueryDet: cascaded sparse query for accelerating high-resolution small object detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022: 13668-13677.
[4] LI W, CHEN Y, HU K, et al. Oriented reppoints for aerial object detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022: 1829-1838.
[5] WEI J, LIU G, LIU S, et al. A novel algorithm for small object detection based on YOLOv4[J]. PeerJ Computer Science, 2023, 9: e1314.
[6] BENNETT K P, BREDENSTEINER E J. Duality and geometry in SVM classifiers[C]//Proceedings of the Seventeenth International Conference on Machine Learning, 2000: 57-64.
[7] TANG Y, HAN K, GUO J, et al. An image patch is a wave: phase-aware vision MLP[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022: 10935-10944.
[8] AZIZ L, SALAM M S B H, SHEIKH U U, et al. Exploring deep learning-based architecture, strategies, applications and current trends in generic object detection: a comprehensive review[J]. IEEE Access, 2020, 8: 170461-170495.
[9] ZHAI S, SHANG D, WANG S, et al. DF-SSD: an improved SSD object detection algorithm based on DenseNet and feature fusion[J]. IEEE Access, 2020, 8: 24344-24357.
[10] CAI Z, VASCONCELOS N. Cascade R-CNN: high quality object detection and instance segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019, 43(5): 1483-1498.
[11] REN Y, ZHU C, XIAO S. Object detection based on fast/faster RCNN employing fully convolutional architectures[J]. Mathematical Problems in Engineering, 2018, 2018: 1-7.
[12] LIANG Y, HAN Y, JIANG F. Deep learning-based small object detection: a survey[C]//Proceedings of the 8th International Conference on Computing and Artificial Intelligence, 2022: 432-438.
[13] XU D, WU Y. Improved YOLO-V3 with DenseNet for multi-scale remote sensing target detection[J]. Sensors, 2020, 20(15): 4276.
[14] CAO C, WU J, ZENG X, et al. Research on airplane and ship detection of aerial remote sensing images based on convolutional neural network[J]. Sensors, 2020, 20(17): 4696.
[15] ZHU X, LYU S, WANG X, et al. TPH-YOLOv5: improved YOLOv5 based on transformer prediction head for object detection on drone-captured scenarios[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021: 2778-2788.
[16] HUANG M, ZHANG Y, CHEN Y. Small target detection model in aerial images based on TCA-YOLOv5m[J]. IEEE Access, 2022, 11: 3352-3366.
[17] ZHANG X, FENG Y, ZHANG S, et al. Finding nonrigid tiny person with densely cropped and local attention object detector networks in low-altitude aerial images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2022, 15: 4371-4385.
[18] LIU Z, GAO Y, DU Q, et al. YOLO-extract: improved YOLOv5 for aircraft object detection in remote sensing images[J]. IEEE Access, 2023, 11: 1742-1751.
[19] VANHERLE B, MOONEN S, VAN REETH F, et al. Analysis of training object detection models with synthetic data[J]. arXiv:2211.16066, 2022.
[20] ZHAO W, SYAFRUDIN M, FITRIYANI NL. CRAS-YOLO: a novel multi-category vessel detection and classification model based on YOLOv5s algorithm[J]. IEEE Access, 2023, 11: 11463-11478.
[21] WANG Q, WU B, ZHU P, et al. ECA-Net: efficient channel attention for deep convolutional neural networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020: 11534-11542.
[22] ILIEV A, KYURKCHIEV N, MARKOV S. On the approximation of the step function by some sigmoid functions[J]. Mathematics and Computers in Simulation, 2017, 133: 223-234.
[23] LIANG D, GENG Q, WEI Z, et al. Anchor retouching via model interaction for robust object detection in aerial images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2021, 60: 1-13.
[24] SHEN R, ZHEN T, LI Z. YOLOv5-based model integrating separable convolutions for detection of wheat head images[J]. IEEE Access, 2023, 11: 12059-12074.
[25] XU B, CHEN M, GUAN W, et al. Efficient teacher: semi-supervised object detection for YOLOv5[J]. arXiv:2302. 07577, 2023.
[26] TONG K, WU Y, ZHOU F. Recent advances in small object detection based on deep learning: a review[J]. Image and Vision Computing, 2020, 97: 103910.
[27] TIAN G, LIU J, ZHAO H, et al. Small object detection via dual inspection mechanism for UAV visual images[J]. Applied Intelligence, 2022, 52(4): 4244-4257.
[28] SUN W, DAI L, ZHANG X, et al. RSOD: real-time small object detection algorithm in UAV-based traffic monitoring[J]. Applied Intelligence, 2022,52(8): 8448-8463.