基于神经网络与数据融合的行人检测方法*

doi:10.19562/j.chinasae.qcgc.2020.11.005

摘要/Abstract

摘要： 针对毫米波雷达对行人识别不够理想、摄像头对距离检测不准确的缺点,本文中提出一种基于神经网络与数据融合的行人检测方法。利用神经网络算法从毫米波雷达检测到的目标中筛选出疑似行人目标,并与Mobileye检测到的行人目标进行匹配;然后通过卡尔曼滤波算法对毫米波雷达和Mobileye匹配成功的行人目标进行数据融合。实车试验结果表明:基于神经网络与数据融合的行人检测方法能对两个传感器检测到的行人目标进行准确匹配,并提高了行人位置检测的精度。

关键词: 行人检测, 神经网络, 数据融合, 卡尔曼滤波

Abstract: In view of the poor results of pedestrian recognition with millimeter wave radar and the inaccuracy of camera in distance detection, a novel pedestrian detection method based on neural network and data fusionis proposed in this paper. The neural network algorithm is used to screen in the suspected pedestrian targets from the targets detected by millimeter wave radar and match it with the pedestrian targets detected by Mobileye. Then the data fusion of matched pedestrian targets between millimeter wave radar and Mobileye is carried out by using Kalman filter algorithm. The results of real vehicle test show that the pedestrian detection method based on neural network and data fusion can accurately match the pedestrian targets detected by the two sensors and enhance the accuracy of pedestrian location detection

Key words: pedestrian detection, neural network, data fusion, Kalman filtering

朱波, 黄茂飞, 谈东奎, 胡旭东, 顾家鑫. 基于神经网络与数据融合的行人检测方法^*[J]. 汽车工程, 2020, 42(11): 1482-1489.

Zhu Bo, Huang Maofei, Tan Dongkui, Hu Xudong, Gu Jiaxin. Pedestrian Detection Method Based on Neural Network and Data Fusion[J]. Automotive Engineering, 2020, 42(11): 1482-1489.

参考文献

[1] World Health 0rganization. Global status report on road safety 2018[R]. Geneva:World Health Organization ,2018.
[2] 陈文强,熊辉,李克强,等.基于深度神经网络的行人及骑车人联合检测[J].汽车工程,2018,40(6):726-732,725.
CHEN W Q, XIONG H, LI K Q, et al. Concurrent pedestrian and cyclist detection based on deep neural networks[J]. Automotive Engineering, 2018,40(6):726-732,725.
[3] ZHANG S,BAUCKHAGE C, KLEIN D A, et al. Fast moving pedestrian detection based on motion segmentation and new motion features[J]. Multimedia Tools and Applications,2016,75(11):6263-6282.
[4] 疏坤,蒋建国,齐美彬,等.基于改进的HOG与Sobel-LBP融合的快速行人检测[J].合肥工业大学学报(自然科学版),2017,40(7):898-903.
SHU K, JIANG J G, QI M B, et al. Fast pedestrian detection based on combinatory feature of improved HOG and Sobel-LBP[J]. Journal of Hefei University of Technology(Natural Science), 2017,40(7):898-903.
[5] 徐伟,周培义,张芬,等.视觉和毫米波雷达信息融合行人识别算法[J].同济大学学报(自然科学版),2017,45(S1):37-42,91.
XU W, ZHOU P Y, ZHANG F, et al. Pedestrian recognition algorithm based on information fusion of visual and millimeter wave radar[J]. Journal of Tongji University(Natural Science), 2017,45(S1):37-42,91.
[6] 余志生.汽车理论[M]. 北京:机械工业出版社, 2017:99.
YU Z S. Automobile theory[M]. Beijing:China Machine Press, 2017:99.
[7] 中华人民共和国工业和信息化部.汽车用真空助力器性能要求及台架试验方法:QC/T 307—2016[S].2016.
Ministry of Industry and Information Technology of the People’s Republic of China. Performance requirements and bench test methods of vacuum booster for automobile:QC/T 307—2016[S]. 2016.
[8] 沈晨,秦训鹏,刘昌业,等.基于滑移率变化模型的ABS汽车制动距离计算[J].武汉理工大学学报(信息与管理工程版),2015,37(6):865-869.
SHEN C, QIN X P, LIU C Y, et al. Braking distance calculation of ABS automobiles based on variation model of wheel slip rate[J]. Journal of Wuhan University of Technology (Information and Management Engineering), 2015,37(6):865-869.
[9] 中华人民共和国国家质量监督检验检疫总局.机动车运行安全技术条件:GB7258—2017[S].2017.
General Administration of Quality Supervision, Inspection and Quarantine of the People’s Republic of China. Technical specifications for safety of power-driven vehicles operating on roads:GB7258— 2017[S]. 2017.
[10] ZHANG Yizhen,ANTONSSON E K, GROTE K. A new threat assessment measure for collision avoidance systems[C]. 2006 IEEE Intelligent Transportation Systems Conference, IEEE, 2006:968-975.
[11] 张庆.基于多传感器数据融合的乘用车AEB控制策略研究[D].长春:吉林大学,2019.
ZHANG Q. Research on AEB control strategy of passenger car based on multi-sensor data fusion[D]. Changchun:Jilin University, 2019.
[12] 刘志峰,王建强,李克强.具有鲁棒特性的车载雷达有效目标确定方法[J].清华大学学报(自然科学版),2008(5):875-878.
LIU Z F, WANG J Q, LI K Q. Robust vehicular radar target determination[J]. Journal of Tsinghua University(Science and Technology), 2008(5):875-878.
[13] 张国祥.基于深度神经网络的人车分类算法[D].西安:西安电子科技大学,2016.
ZHANG G X. Vehicle-pedestrian classification based on deep neural networks[D]. Xian:Xidian University, 2016.
[14] 耿超.基于机器学习的大跨桥梁斜拉索损伤识别[D].南京:东南大学,2018.
GENG C. Identification of damaged stay cable of long-span bridge based on machine learning[D]. Nanjing:Southeast University, 2018.
[15] 张志.多传感器信息融合及其应用研究[D].西安:西安电子科技大学,2017.
ZHANG Z. Research on multi-sensor information fusion and its application[D]. Xian:Xidian University, 2017.

基于神经网络与数据融合的行人检测方法^*

Pedestrian Detection Method Based on Neural Network and Data Fusion

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

Metrics

本文评价

推荐阅读 10

[1]	张念忠,宋强,王冠峰,王明生. 车用永磁同步电机无电流传感器控制研究[J]. 汽车工程, 2024, 46(2): 281-289.
[2]	王彦鑫,李海岩,崔世海,贺丽娟,吕文乐. 基于BP神经网络的儿童乘员头部损伤预测模型及评估参数研究[J]. 汽车工程, 2024, 46(2): 329-336.
[3]	邓斌, 李维汉, 吴迪, 张冰战, 赵韩. 基于KFESO的四轮主动转向积分滑模控制[J]. 汽车工程, 2024, 46(1): 100-108.
[4]	李勇滔,孙晨旭,郑伟光,许恩永,李育方,王善超. 基于毫米波雷达与视觉融合的碰撞预警[J]. 汽车工程, 2023, 45(9): 1666-1676.
[5]	胡明辉,朱广曜,刘长贺,唐国峰. 考虑迟滞特性的卡尔曼滤波和门控循环单元神经网络的锂离子电池SOC联合估计[J]. 汽车工程, 2023, 45(9): 1688-1701.
[6]	管欣,仲昭辉,詹军,奚腾龙,叶昊,高深圳,成健,廖世辉,蔡均. 基于卷积神经网络的汽车操纵稳定性试验类型识别方法[J]. 汽车工程, 2023, 45(9): 1765-1771.
[7]	王明,唐小林,杨凯,李国法,胡晓松. 考虑预测风险的自动驾驶车辆运动规划方法[J]. 汽车工程, 2023, 45(8): 1362-1372.
[8]	陈国强,申正义,孙利,支梦帆,李彤. 基于BP神经网络优化遗传算法的智能座舱感性意象预测[J]. 汽车工程, 2023, 45(8): 1479-1488.
[9]	吴骁, 史文库, 陈志勇. 基于交互式多模型卡尔曼滤波的主动悬架控制[J]. 汽车工程, 2023, 45(7): 1200-1211.
[10]	孙晓强, 王玉麟, 胡伟伟, 蔡英凤, 陈龙, Wong Pak Kin. 基于轮胎分段仿射辨识模型的车辆行驶状态估计策略研究[J]. 汽车工程, 2023, 45(7): 1212-1221.
[11]	张雷, 关可人, 丁晓林, 郭鹏宇, 王震坡, 孙逢春. 基于图像识别与动力学融合的路面附着系数估计方法[J]. 汽车工程, 2023, 45(7): 1222-1234.
[12]	浦震峰, 唐亮, 上官文斌, 王伟玮, 蒋开洪. 基于多传感器信息的汽车低速车速估计方法[J]. 汽车工程, 2023, 45(7): 1235-1243.
[13]	韩勇,林旭洁,黄红武,蔡鸿瑜,罗金镕,李燕婷. 典型汽车碰撞事故场景中行人运动轨迹预测方法[J]. 汽车工程, 2023, 45(6): 1022-1030.
[14]	陈建华,徐中明,张志飞. 基于轴距预瞄的非匀速工况车辆悬架状态估计[J]. 汽车工程, 2023, 45(6): 1040-1049.
[15]	赵霞,李朝,付锐,葛振振,王畅. 基于深度卷积-Tokens降维优化视觉Transformer的分心驾驶行为实时检测[J]. 汽车工程, 2023, 45(6): 974-988.