[1]李坤伦,魏泽发,宋焕生.基于SqueezeNet卷积神经网络的车辆颜色识别[J].长安大学学报(自然科学版),2020,40(4):109-116.
 LI Kun lun,WEI Ze fa,SONG Huan sheng.Vehicle color recognition based on SqueezeNet[J].Journal of Chang’an University (Natural Science Edition),2020,40(4):109-116.
点击复制

基于SqueezeNet卷积神经网络的车辆颜色识别()
分享到:

长安大学学报(自然科学版)[ISSN:1006-6977/CN:61-1281/TN]

卷:
第40卷
期数:
2020年4期
页码:
109-116
栏目:
交通工程
出版日期:
2020-07-15

文章信息/Info

Title:
Vehicle color recognition based on SqueezeNet
作者:
李坤伦魏泽发宋焕生
(1. 长安大学 教育技术与网络中心,陕西 西安 710064; 2. 长安大学 电子与控制工程学院,陕西 西安 710064; 3. 长安大学 信息工程学院,陕西 西安 710064)
Author(s):
LI Kunlun12 WEI Zefa1 SONG Huansheng13
(1. Educational Technology and Network Center, Changan University, Xian 710064, Shaanxi, China;2. School of Electronic and Control Engineering, Changan University, Xian 710064, Shaanxi, China;3. School of Information Engineering, Changan University, Xian 710064, Shaanxi, China)
关键词:
交通工程卷积神经网络智能交通车辆颜色识别SqueezeNet
Keywords:
traffic engineering convolutional neural network intelligent transportation vehicle color recognition SqueezeNet
文献标志码:
A
摘要:
为实现车辆颜色的精准识别,辅助现有车牌识别方法在智能交通系统中完成对车辆身份的精确认证,通过分析现有解决方案和探究卷积神经网络在实际应用中的问题,提出基于轻量级卷积神经网络SqueezeNet实现对车辆颜色识别的方法。轻量级卷积神经网络SqueezeNet的参数量是AlexNet网络结构的1/50,同时还能保证精度,避免由于网络结构复杂、参数量大造成的大规模计算和较高的计算机硬件需求,提升了模型的可移植性。选取车辆颜色识别(Vehicle_Color_Recognition)数据集作为研究基础,针对数据集进行了扩充和增强处理。以SqueezeNet为基准探究了特征融合对分类结果的影响,通过对比试验确定将fire7/concat输出特征图和fire9/concat输出特征图相融合。研究结果表明:轻量级卷积神经网络SqueezeNet在保证模型大小2.9 M、单张测试时间15 ms的基础上,训练精度为96.28%,而AlexNet的模型大小为227 M、单张测试时间24 ms、训练精度为96.18%;在实现同等精度的前提下,轻量级卷积神经网络SqueezeNet训练得到的模型更适合移植到如现场可编程门列阵(FPGA)这种开发板上,同时在服务器上的处理速度也更快;融合后的模型最终的分类结果提升为96.48%。利用轻量级卷积神经网络SqueezeNet进行车辆颜色识别可以较好地应用在智能交通系统中,并在一定程度上解决目前车牌识别的难点。
Abstract:
In order to accurate identification of vehicle colors which can assist existing license plate recognition system and enable accurate identification of vehicles in intelligent transportation system, existing solutions were analyzed, and the problems of the practical applications of convolutional neural network (CNN) were explored. A method of realizing vehicle color recognition by using light weight convolutional neural network (LCNN) was proposed. The parameter quantity of SqueezeNet is one fiftieth of that of AlexNet network structure, and it can also ensure the accuracy, which avoids the largescale calculation and high computer hardware requirements caused by the complex network structure and large parameter quantity, and therefore improves the portability of the model. The vehicle color recognition data set was selected as the research basis, and the data set was expanded and enhanced. Based on SqueezeNet, the impact of feature fusion on classification results was explored. The output feature map of fire7/concat and the output feature map of fire9/concat were fused. The results show that SqueezeNet has a training accuracy of 96.28% on the basis of guaranteed model size of 2.9 M and a single sheet test time of 15 ms, while that of AlexNet is 227 M, 24 ms and 96.18%. It can be seen that, under the premise of achieving the same accuracy, the model trained by the lightweight convolutional neural network SqueezeNet is more suitable to be transplanted to the development board such as fieldprogrammable aate array (FPGA), and the processing speed on the server is faster. Therefore, the lightweight convolutional neural network SqueezeNet, which is used to identify vehicle colors, can be better applied in the intelligent transportation system and solve, to a certain extent, current problems and drawbacks. 3 tabs, 7 figs, 23 refs.

相似文献/References:

[1]王建伟,李娉,高洁,等.中国交通运输碳减排区域划分[J].长安大学学报(自然科学版),2012,32(01):0.
[2]李曙光,周庆华.具有破坏排队的离散时间动态网络装载算法[J].长安大学学报(自然科学版),2012,32(01):0.
[3]凌海兰,郗恩崇.基于随机波动条件的公交客运量预测模型[J].长安大学学报(自然科学版),2012,32(01):0.
[4]田娥,肖庆,陆小佳,等.安全驾驶的横向安全预警报警阈值的确定[J].长安大学学报(自然科学版),2012,32(01):0.
[5]侯贻栋,赵炜华,魏 朗,等.驾驶人空间距离判识规律心理学分析[J].长安大学学报(自然科学版),2012,32(03):86.
 HOU Yi-dong,ZHAO Wei-hua,WEI Lang,et al.Analysis on psychology in cognitive distance about drivers[J].Journal of Chang’an University (Natural Science Edition),2012,32(4):86.
[6]赵跃峰,张生瑞,魏 华.隧道群路段运行速度特性分析[J].长安大学学报(自然科学版),2012,32(06):67.
 ZHAO Yue-feng,ZHANG Sheng-rui,WEI hua.Operating speed characteristics of tunnel group section[J].Journal of Chang’an University (Natural Science Edition),2012,32(4):67.
[7]林 杉,许宏科,刘占文.一种高速公路隧道交通流元胞自动机模型[J].长安大学学报(自然科学版),2012,32(06):73.
 LIN Shan,XU Hong-ke,LIU Zhan-wen.One cellular automaton traffic flow model for expressway tunnel[J].Journal of Chang’an University (Natural Science Edition),2012,32(4):73.
[8]刘俊德,徐 兵,梁永东,等.交通事故下高速公路行车安全评估[J].长安大学学报(自然科学版),2012,32(06):78.
 LIU Jun-de,XU bing,LIANG Yong-dong,et al.Traffic safety assessment of expressway in the accident[J].Journal of Chang’an University (Natural Science Edition),2012,32(4):78.
[9]芮海田,吴群琪,赵跃峰,等.公路建设对区域经济发展的影响分析——以陕西省为例[J].长安大学学报(自然科学版),2012,32(06):83.
 RUI Hai-tian,WU Qun-qi,ZHAO Yue-feng,et al.Influence of highway construction on regional economy development——taking Shaanxi as an example[J].Journal of Chang’an University (Natural Science Edition),2012,32(4):83.
[10]彭 辉,续宗芳,韩永启,等.城市群城际运输结构配置客流分担率模型[J].长安大学学报(自然科学版),2012,32(02):91.
 PENG Hui,XU Zong-fang,HAN Yong-qi,et al.Sharing ratios model of passenger flows in intercity transportation structure configuration among urban agglomeration[J].Journal of Chang’an University (Natural Science Edition),2012,32(4):91.

更新日期/Last Update: 2020-07-31