融合DBSCAN的改进YOLOv3目标检测算法

doi:10.3778/j.issn.1002-8331.2010-0251

摘要/Abstract

摘要： 针对YOLOv3（you only look once）检测算法对小目标、遮挡目标检测时存在识别率低和识别精度不高的问题，提出一种融合DBSCAN（density-based spatial clustering of applications with noise）的改进YOLOv3目标检测算法。首先在YOLOv3网络中增加DBSCAN聚类算法，其次对检测目标进行提取，实现数据集多尺度聚类，得到初代特征图，然后通过改进[K]-means聚类算法确定锚点位置达到更好的聚类，最后在VOC2007+2012数据集和MS-COCO数据集上对改进YOLOv3算法进行训练和测试。实验结果表明改进的YOLOv3算法使检测目标在VOC数据集和MS-COCO数据集上mAP（mean average precision）分别提高了14.9个百分点和12.5个百分点。与其他深度学习目标检测算法相比，改进YOLOv3检测算法具有更好的检测效果，同时具有良好移植性和更好的鲁棒性。

关键词: YOLOv3, 卷积神经网络, 目标检测, DBSCAN聚类算法

Abstract: Aiming at the problems of low recognition rate and low recognition accuracy when the YOLOv3（you only look once） detection algorithm detects small targets and occluded targets, an improved YOLOv3 algorithm is proposed in combination with DBSCAN（density-based spatial clustering of applications with noise） for target detection. Firstly, DBSCAN clustering algorithm is added to YOLOv3 network, and the detection target is extracted to achieve multi-scale clustering of the dataset to obtain the first generation feature map, and then the anchor point location is determined by improving [K]-means clustering algorithm to achieve better clustering. Finally, the improved YOLOv3 algorithm is trained and tested on VOC2007+2012 dataset and MS-COCO dataset. The experimental results show that the improved YOLOv3 algorithm increases the mAP of detection target by 14.9 percentage points and 12.5 percentage points on the VOC dataset and MS-COCO dataset, respectively. The improved YOLOv3 detection algorithm has better detection results in comparison with other deep learning target detection algorithms, as well as good portability and better robustness.

Key words: YOLOv3, convolutional neural network, target detection, DBSCAN clustering algorithm

李云红, 张轩, 李传真, 苏雪平, 聂梦瑄, 毕远东, 谢蓉蓉. 融合DBSCAN的改进YOLOv3目标检测算法[J]. 计算机工程与应用, 2022, 58(10): 208-215.

LI Yunhong, ZHANG Xuan, LI Chuanzhen, SU Xueping, NIE Mengxuan, BI Yuandong, XIE Rongrong. Improved YOLOv3 Target Detection Algorithm Combined with DBSCAN[J]. Computer Engineering and Applications, 2022, 58(10): 208-215.

参考文献

[1] 胡侯立，魏维，胡蒙娜.深度学习算法的原理及应用[J].信息技术，2015，4（2）：175-177.
HU H L，WEI W，HU M N.Principles and practices of deep learning[J].Information Technology，2015，4（2）：175-177.
[2] 刘博，王胜正，赵建森，等.基于Darknet网络和YOLOv3算法的船舶跟踪识别[J].计算机应用，2019，39（6）：1663-1668.
LIU B，WANG S Z，ZHAO J S，et al.Ship tracking and recognition based on Darknet network and YOLOv3 algorithm[J].Journal of Computer Applications，2019，39（6）：1663-1668.
[3] SZEGEDY C，TOSHEV A，ERHAN D.Deep neural networks for object detection[C]//Advances in Neural Information Processing Systems，Feb 15，2013：2553-2561.
[4] ERHAN D，SZEGEDY C，TOSHEV A，et al.Scalable object detection using deep neural networks[J].arXiv：1312.2249,2013.
[5] GIRSHICK R.Fast R-CNN[C]//2015 IEEE International Conference on Computer Vision，Nov，24，2015：1440-1448.
[6] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot multibox detector[C]//14th European Conference on Computer Vision.Cham：Springer，2016：21-37.
[7] REDMON J，DIVVALV S，GIRSHICK R，et al.You only look once：unified，real-time object detection[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition，Las Vegas，Jun 26-Jul 1，2016.Piscataway：IEEE，2016：779-788.
[8] REDMON J，FARHADI A.YOLO9000：better，faster，stronger[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition，Hawaii，Jul 21-26，2017.Piscataway：IEEE，2016：65-84.
[9] HE K M，GKIOXARI G，DOLLAR P，et al.Mask R-CNN[C]//2017 IEEE International Conference on Computer Vision，Venice，Oct 22，2017.Piscataway：IEEE，2017：386-397.
[10] 朱伯伟，庾农，肖松.红外极小目标检测算法研究[J].航空电子技术，2011，42（3）：5-11.
ZHU B W，YU N，XIAO S.Research on infrared minimal target detection algorithm[J].Avionics Technology，2011，42（3）：5-11.
[11] 何清，李宁，罗文娟，等.大数据下的机器学习算法综述[J].模式识别与人工智能，2014，27（4）：327-336.
HE Q，LI N，LUO W J，et al.Overview of machine learning algorithms under big data[J].Pattern Recognition and Artificial Intelligence，2014，27（4）：327-336.
[12] KUZNETSOVA A，MALEVA T，SOLOVIEV V.Using YOLOv3 algorithm with pre-and post-processing for apple detection in fruit-harvesting robot[J].Agronomy，2020，10（7）：645-679.
[13] YASWANTH V，KUMAR K K，HARSHITH N，et al.Outfit of exemplar-SVMs for object detection and beyond[J].Indian Journal of Science and Technology，2016，9（30）：854-879.
[14] STAHL T，PINTEA S L，VAN GEMERT J C.Divide and count：generic object counting by image divisions[J].IEEE Transactions on Image Processing，2019，28（2）：1035-1044.
[15] MOSTAFAA A，PEPER K E，BERRY C R.Use of cardiac sphericity index and manubrium heart scores to assess radiographic cardiac silhouettes in large-and small-breed dogs with and without cardiac disease[J].Journal of the American Veterinary Medical Association，2020，256（8）：888-896.