Abstract：In order to improve the accuracy of multi-target detection in complex scenes,a detection method for multiple moving objects based on ReInspect algorithm is proposed. The algorithm is based on the OverFeat algorithm and Faster R-CNN algorithm, adding LSTM (long short-term memory)recurrent network structure to record the feature sequence of multiple targets. By adjusting the LSTM network feature label information to preprocess the loss function, and the confidence segmentation method is used after tracking to match the detection results to solve the problem of repeated detection of the same target and target occlusion. The experimental results show that the algorithm has good anti-interference ability against traditional overlapping and occlusion problems, and the recognition accuracy is higher than 90% in different scenarios.
［1］Zhan B B, Monekosso D N, Remagnino P, et al. Crowd analysis: A survey［J］. Machine Vision and Applica-tions, 2008, 19 (5-6): 345-357.
［2］Junior J C S J, Musse S R, Jung C R. Crowd analysis using computer vision techniques［J］. IEEE Signal Processing Magazine, 2010, 27 (5): 66-77.
［3］Thida M, Yong Y L, Climent-Pérez P, et al. A literature review on video analytics of crowded scenes ［C］//In: Proc of the Intelligent Multimedia Surveillance, Berlin, Heidelberg: Springer-Verlag, 2013: 17-36.
［4］Li T, Chang H, Wang M, et al. Crowded scene analysis: A survey ［J］. IEEE Trans. on Circuits and Systems for Video Technology, 2014, 25 (3): 367-386.
［5］ 谭芳,穆平安,马忠雪. 基于YOLOv3检测和特征点匹配的多目标跟踪算法［J］. 计量学报, 2021, 42(2): 157-162.
Tan F, Mu P A, Ma Z X. Multi-target Tracking Algorithm Based on YOLOv3 Detection and Feature Point Matching［J］. Acta Metrologica Sinica, 2021, 42(2): 157-162.
［6］ 杨璐. 一种基于人体轮廓特征的教室人数统计算法［J］. 计量学报, 2021, 42(2): 178-183.
Yang L. A Classroom Population Statistics Algorithm Based on Human Contour Features［J］. Acta Metrologica Sinica, 2021, 42(2): 178-183.
［7］Kong D, Gray D, Tao H. A viewpoint invariant approach for crowd counting ［C］//In: Proc. of the 18th Int’l Conf. on Pattern Recognition, 2006, 1187-1190.
［8］ 张立国, 李晓松, 肖磊, 等. 基于单目视觉的四旋翼飞行器目标跟踪算法研究 ［J］. 计量学报, 2018, 39 (3): 342-347.
Zhang L G, Li X S, Xiao L, et al. Research on Target Tracking Algorithm of the Quadrotor Aircraft Based on Monocular Vision ［J］. Acta Metrologica Sinica, 2018, 39 (3): 342-347.
［9］ 陈卫东, 陈磊, 邓志巍, 等. 基于混合相关滤波信息融合再检测的目标跟踪算法 ［J］. 计量学报, 2019, 40 (6): 1006-1012.
Chen W D, Chen L, Deng Z W, et al. Target Tracking Algorithm Based on Hybird Correlation Filtering and Inf-ormation Fusion Redetection ［J］. Acta Metrologica Sinica, 2019, 40 (6): 1006-1012.
［10］ Benabbas Y, Ihaddadene N, Djeraba C. Motion pattern extraction and event detection for automatic visual survei-llance ［J］. Journal on Image and Video Processing, 2011, 2011 (1): 413-447.
［11］Sermanet P, Eigen D, Zhang X, et al. OverFeat: Integ-rated Recognition, Localization and Detection using Con-volutional Networks ［J］. arXiv preprint arXiv: 1312. 6229, 2013.
［12］Ren S Q, He K M, Girshick R, et al. Faster R-CNN: towards real-time object detection with region proposal networks ［C］//International Conference on Neural In-formation Processing Systems, MIT Press, 2015: 91-99.
［13］Graves A. Long Short-Term Memory［M］//Supervised Sequence Labelling with Recurrent Neural Networks. Berlin: Springer Berlin Heidelberg, 2012: 1735-1780.
［14］Zhao F, Feng J S, Zhao J, et al. Robust LSTM-Autoencoders for Face De-Occlusion in the Wild ［J］. IEEE Transactions on Image Processing, 2018, 27 (2): 778-790.
［15］Liu J, Shahroudy A, Xu D, et al. Spatio-Temporal LSTM with Trust Gates for 3D Human Action Recogni-tion ［J］. Computer Science, 2016, 9907: 816-833.
［16］Kalpana K, Amritha P P. Image manipulation detection using Deep Learning in tensor flow ［J］. International Journal of Control Theory & Applications, 2016, 9 (40): 221-225.
［17］Glenski M, Ayton E, Arendt D, et al. Fishing for Clickbaits in Social Images and Texts with Linguistically-Infused Neural Network Models ［J］. arXiv preprint arXiv: 1710. 06390, 2017.
［18］Chen X, Gupta A. An Implementation of Faster RCNN with Study for Region Sampling ［J］. arXiv preprint arXiv: 1702. 02138, 2017.
［19］Kang B, Tripathi S, Dane G, et al. Low-complexity object detection with deep convolutional neural network for embedded systems ［C］//Applications of Digital Image Processing XL. 2017: 60.
［20］Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions ［C］//Computer Vision and Pattern Recog-nition, IEEE, 2015: 1-9.
［21］Szegedy C, Ioffe S, Vanhoucke V, et al. Inception-v4, Inception-ResNet and the Impact of Residual Connec-tions on Learning ［J］. arXiv preprint arXiv: 1602. 07261, 2016.
［22］Wen Y D, Zhang K P, Li Z F, et al. A Discriminative Feature Learning Approach for Deep Face Recognition ［C］//European Conference on Computer Vision, Springer, Cham, 2016: 499-515.
［23］Kuhn H W. The Hungarian method for the assignment problem ［J］. Naval research logistics quarterly, 1955, 2 (1/2): 83-97.
2022, Vol. 43 
