基于RDB-YOLOv4的煤矿井下有遮挡行人检测

doi:10.3778/j.issn.1002-8331.2009-0449

摘要/Abstract

摘要： 针对煤矿井下数字化人员检测过程中行人易被大型设备遮挡而导致的误检、漏检等问题，提出了一种基于RDB-YOLOv4网络的煤矿井下有遮挡行人检测方法。该算法以YOLOv4为基础网络架构，在CSPDarknet-53特征提取网络中加入了残差密集块（residual dense block，RDB），对不同层次的特征实现跨层传递和融合，连续的连接保证了低级和高级特征信息的存储和记忆，使得完整有效的局部特征能准确预测被遮挡行人的信息。对比当前主流目标检测算法和遮挡处理检测算法，该算法在PASCAL VOC 2007公开数据集和煤矿井下行人数据集下有效提升了测试的平均精度（average precision，AP），相比YOLOv4在两组不同数据集测试的平均精度分别提升了2.74个百分点和3.5个百分点（IoU=0.5）。

关键词: 遮挡行人检测, YOLOv4网络, 残差密集块, 煤矿井下

Abstract: Aiming at the false detection and missed detection caused by pedestrians easily blocked by large equipments during the detection process in coal mines, this paper proposes the detection method of blocked pedestrians in underground coal mines based on RDB-YOLOv4 network. The algorithm is based on YOLOv4 network architecture, and joins the residuals dense block（RDB） in CSPDarknet-53 network, to achieve cross-layer transfer and integration features at different levels. The continuous connections ensure the storage and memory of low level and high-level feature information, so that complete and effective local features can accurately predict the information of blocked pedestrians. Compared with the current mainstream target detection algorithms and occlusion processing detection algorithms, the algorithm in this paper has effectively improved the average precision（AP） of the test under the PASCAL VOC 2007 open data set and underground coal mine pedestrian data set. Compared with YOLOv4, the average precision of the test in two different data sets has been improved by 2.74 percentage points and 3.5 percentage points respectively（IoU=0.5）.

Key words: blocked pedestrian detection, YOLOv4 network, residual dense block, coal mine

谢斌红, 袁帅, 龚大立. 基于RDB-YOLOv4的煤矿井下有遮挡行人检测[J]. 计算机工程与应用, 2022, 58(5): 200-207.

XIE Binhong, YUAN Shuai, GONG Dali. Detection of Blocked Pedestrians Based on RDB-YOLOv4 in Coal Mine[J]. Computer Engineering and Applications, 2022, 58(5): 200-207.

参考文献

[1] WOJEK C，DOLLAR P，SCHIELE B，et al.Pedestrian detection：an evaluation of the state of the art[J].IEEE Transactions on Pattern Analysis & Machine Intelligence，2012，34（4）：743.
[2] DALAL N.Histograms of oriented gradients for human detection[C]//Proc CVPR，2005.
[3] MITA T，KANEKO T，HORI O.Joint haar-like features for face detection[C]//Tenth IEEE International Conference on Computer Vision（ICCV’05），2005：1619-1626.
[4] AHONEN T，HADID A，PIETIK?INEN M.Face recognition with local binary patterns[C]//European Conference on Computer Vision，2004：469-481.
[5] WU B，NEVATIA R.Detection of multiple，partially occluded humans in a single image by bayesian combination of edgelet part detectors[C]//Tenth IEEE International Conference on Computer Vision（ICCV’05），2005：90-97.
[6] MATHIAS M，BENENSON R，TIMOFTE R，et al.Handling occlusions with franken-classifiers[C]//Proceedings of the IEEE International Conference on Computer Vision，2013：1505-1512.
[7] FELZENSZWALB P，GIRSHICK R，MCALLESTER D，et al.Cascade object detection with deformable part models[J].Communications of the ACM，2013，56（9）：97-105.
[8] OUYANG W，WANG X.A discriminative deep model for pedestrian detection with occlusion handling[C]//2012 IEEE Conference on Computer Vision and Pattern Recognition，2012：3258-3265.
[9] OUYANG W，WANG X.Joint deep learning for pedestrian detection[C]//Proceedings of the IEEE International Conference on Computer Vision，2013：2056-2063.
[10] TIAN Y，LUO P，WANG X，et al.Deep learning strong parts for pedestrian detection[C]//Proceedings of the IEEE International Conference on Computer Vision，2015：1904-1912.
[11] ZHANG S，YANG J，SCHIELE B.Occluded pedestrian detection through guided attention in CNNs[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：6995-7003.
[12] BELL S，LAWRENCE ZITNICK C，BALA K，et al.Inside-outside net：detecting objects in context with skip pooling and recurrent neural networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2016：2874-2883.
[13] ZHOU C，YUAN J.Multi-label learning of part detectors for occluded pedestrian detection[J].Pattern Recognition，2019，86：99-111.
[14] WANG X，XIAO T，JIANG Y，et al.Repulsion loss：detecting pedestrians in a crowd[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：7774-7783.
[15] 陈宁，李梦璐，袁皓，等.遮挡情形下的行人检测方法综述[J].计算机工程与应用，2020，56（16）：13-20.
CHEN Ning，LI Menglu，YUAN Hao，et al.Review of pedestrian detection with occlusion[J].Computer Engineering and Applications，2020，56（16）：13-20.
[16] ZHANG Y，TIAN Y，KONG Y，et al.Residual dense network for image super-resolution[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：2472-2481.
[17] BOCHKOVSKIY A，WANG C Y，LIAO H Y M.YOLOv4：optimal speed and accuracy of object detection[J].arXiv：2004.10934，2020.
[18] LIU S，QI L，QIN H，et al.Path aggregation network for instance segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：8759-8768.
[19] BODLA N，SINGH B，CHELLAPPA R，et al.Soft-NMS—improving object detection with one line of code[C]//Proceedings of the IEEE International Conference on Computer Vision，2017：5561-5569.
[20] RUSSAKOVSKY O，DENG J，SU H，et al.ImageNet large scale visual recognition challenge[J].International Journal of Computer Vision，2015，115（3）：211-252.
[21] ZHOU X，WANG D，KR?HENBüHL P.Objects as points[J].arXiv：1904.07850，2019.
[22] DUAN K，BAI S，XIE L，et al.Centernet：keypoint triplets for object detection[C]//Proceedings of the IEEE International Conference on Computer Vision，2019：6569-6578.
[23] YIN T，ZHOU X，KR?HENBüHL P.Center-based 3D object detection and tracking[J].arXiv：2006.11275，2020.
[24] 王丹峰，陈超波，马天力，等.基于深度可分离卷积的YOLOv3行人检测算法[J].计算机应用与软件，2020，37（6）：218-223.
WANG Danfeng，CHEN Chaobo，MA Tianli，et al.YOLOv3 pedestrian detection algorithm based on depth-wise separable convolution[J].Computer Applications and Software，2020，37（6）：218-223.