[1]陈立潮,解丹,曹建芳,等.改进光流法和GMM融合的车辆实时检测算法研究[J].智能系统学报,2021,16(2):271-278.[doi:10.11992/tis.201907051]
CHEN Lichao,XIE Dan,CAO Jianfang,et al.Research on vehicle real-time detection algorithm based on improved optical flow method and GMM[J].CAAI Transactions on Intelligent Systems,2021,16(2):271-278.[doi:10.11992/tis.201907051]
点击复制
CHEN Lichao,XIE Dan,CAO Jianfang,et al.Research on vehicle real-time detection algorithm based on improved optical flow method and GMM[J].CAAI Transactions on Intelligent Systems,2021,16(2):271-278.[doi:10.11992/tis.201907051]
改进光流法和GMM融合的车辆实时检测算法研究
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
16
期数:
2021年第2期
页码:
271-278
栏目:
学术论文—机器感知与模式识别
出版日期:
2021-03-05
- Title:
- Research on vehicle real-time detection algorithm based on improved optical flow method and GMM
- Keywords:
- IOFGMM detection algorithm; optical flow method; gaussian mixture background model; Information fusion; real-time detection; gradient; illumination; constraint
- 分类号:
- TP391
- 摘要:
- 针对传统光流算法受光照影响较大和在不同场景中检测效果差别较大等问题，提出一种改进的光流法与混合高斯背景模型相融合的运动车辆实时检测算法（improved optical flow and gaussian mixture model，IOFGMM）。首先，在光流算法中加入限制条件使得不同梯度点处采用不同约束；其次，融合高斯混合背景模型（gaussian mixture model，GMM）；最后，采用提出的融合算法比较目标框的数量和目标框之间的重叠面积，从而在监控视频中显示出融合后的车辆检测信息。实验结果表明：该算法在3种不同场景视频上的检测效果达到了84.80%的平均准确率，84.79%的平均召回率以及84.63%的平均F1值。与经典的光流法和高斯混合背景模型及基于这两种理论的算法相比，IOFGMM算法的各项指标平均有37%的提高，具有良好的检测效果。
- Abstract:
- To solve the problem of the optical flow algorithm being greatly affected by illumination and the highly variable detection effect in different scenes, in this paper, we propose an improved optical flow and Gaussian mixture model (IOFGMM) algorithm for the real-time detection of moving vehicles. First, a restriction is added to the optical flow algorithm whereby different constraints are used at different points. Then, a Gaussian mixture model (GMM) is fused. Finally, the number of target boxes and the area in which the target boxes overlap are compared by the proposed fusion algorithm. The vehicle detection information after fusion is displayed in the surveillance video. Experimental results show that the detection performance of the IOFGMM algorithm achieved an average accuracy rate of 84.80%, an average recall rate of 84.79%, and an average F1 value of 84.63% for videos of three different scenes. Compared with the classical optical flow method, the GMM, and the algorithm based on these two theories, the IOFGMM algorithm shows an average improvement of 37% in each metric. Therefore, we can conclude that the IOFGMM algorithm has good detection performance.
- 参考文献/References:
-
[1] 徐子豪, 黄伟泉, 王胤. 基于深度学习的监控视频中多类别车辆检测[J]. 计算机应用, 2019, 39(3):700-705
XU Zihao, HUANG Weiquan, WANG Yin. Multi-class vehicle detection in surveillance video based on deep learning[J]. Journal of computer application, 2019, 39(3):700-705
[2] WULFF J, SEVILLA-LARA L, BLACK M J. Optical flow in mostly rigid scenes[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, USA, 2017:6911?6920.
[3] 于晓明, 李思颖, 史胜楠. 混合高斯融合三帧差的运动目标检测改进算法[J]. 红外技术, 2019, 41(3):256-261
YU Xiaoming, LI Siying, SHI Shengnan. An improved algorithm for moving target detection using a Gaussian mixture with three-frame difference[J]. Infrared technology, 2019, 41(3):256-261
[4] 范文超, 李晓宇, 魏凯, 等. 基于改进的高斯混合模型的运动目标检测[J]. 计算机科学, 2015, 42(5):286-288, 319
FAN Wenchao, LI Xiaoyu, WEI Kai, et al. Moving target detection based on improved Gaussian mixture model[J]. Computer science, 2015, 42(5):286-288, 319
[5] GIRSHICK R. Fast R-CNN[C]//Proceedings of 2015 IEEE International Conference on Computer Vision. Santiago, Chile, 2015:1440?1448.
[6] REN Shaoqing, HE Kaiming, GIRSHICK R, et al. Faster R-CNN:towards real-time object detection with region proposal networks[J]. IEEE transactions on pattern analysis and machine intelligence, 2017, 39(6):1137-1149.
[7] REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once:unified, real-time object detection[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, USA, 2016:779?788.
[8] REDMON J, FARHADI A. YOLO9000:better, faster, stronger[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, USA, 2017:6517?6525.
[9] SENGAR S S, MUKHOPADHYAY S. Motion detection using block based bi-directional optical flow method[J]. Journal of visual communication and image representation, 2017, 49:89-103.
[10] PAN Chengyi, ZHU Zhou, JIANG Liangwei, et al. Adaptive ViBe background model for vehicle detection[C]//Proceedings of the 2nd Advanced Information Technology, Electronic and Automation Control Conference. Chongqing, China, 2017:1301?1305.
[11] HORN B K P, SCHUNCK B G. Determining optical flow[J]. Artificial intelligence, 1980, 17(1/2/3):185-203.
[12] LIU Xiong, PAN Li, SUN Xiaoliang. Real-time traffic status classification based on Gaussian mixture model[C]//Proceedings of IEEE 1st International Conference on Data Science in Cyberspace. Changsha, China, 2016:573?578.
[13] 魏超, 贺光辉. 基于直方图的白平衡算法的研究[J]. 微电子学与计算机, 2018, 35(6):75-78
WEI Chao, HE Guanghui. Automatic white balance algorithm based on histogram[J]. Microelectronics & computer, 2018, 35(6):75-78
[14] WANG Yi, JODOIN P M, PORIKLI F, et al. CDnet 2014:an expanded change detection benchmark dataset[C]//Proceedings of 2014 IEEE Conference on Computer Vision and Pattern Recognition. Columbus, USA, 2014:393?400.
[15] YANG Honghong, QU Shiru. Real-time vehicle detection and counting in complex traffic scenes using background subtraction model with low-rank decomposition[J]. IET intelligent transport systems, 2018, 12(1):75-85.
[16] STAUFFER C, GRIMSON W E L. Adaptive background mixture models for real-time tracking[C]//Proceedings of 1999 IEEE Conference on Computer Vision and Pattern Recognition. Fort Collins, USA, 1999:246?252.
[17] RODRIGUEZ P, WOHLBERG B. Incremental principal component pursuit for video background modeling[J]. Journal of mathematical imaging and vision, 2016, 55(1):1-18.
[18] SHAKERI M, ZHANG Hong. COROLA:a sequential solution to moving object detection using low-rank approximation[J]. Computer vision and image understanding, 2016, 146:27-39.
备注/Memo
收稿日期:2019-07-29。
基金项目:山西省自然科学基金项目（201801D221179，201701D121059）；太原科技大学校博士科研启动基金项目（20162036）；山西省高等学校人文社会科学重点研究基地项目（20190130）；忻州市平台和人才专项（20180601）
作者简介:陈立潮，教授，博士，中国计算机学会高级会员，主要研究方向为智能信息处理。主持省部级科技项目20余项、获山西省科学技术奖二等奖2项。发表学术论文120余篇;解丹，硕士研究生，中国计算机学会会员，主要研究方向为图像处理与模式识别;曹建芳，教授，博士，中国计算机学会高级会员，主要研究方向为数字图像理解、大数据技术。近5年来，主持省部级项目11项，获山西省高等学校科学研究优秀成果（科学技术）自然科学奖二等奖1项、山西省优秀学术论文二等奖2项、忻州市科学技术奖（自然科学类）二等奖2项、三等奖1项。发表学术论文30余篇，出版学术专著2部
通讯作者:曹建芳.E-mail:kcxdj122@126.com
更新日期/Last Update:
2021-04-25