基于异常声音的隧道交通事故检测方法

马庆禄; 付冰琳; 马恋; 李杨梅

doi:10.3963/j.jssn.1674-4861.2023.01.004

基于异常声音的隧道交通事故检测方法

doi: 10.3963/j.jssn.1674-4861.2023.01.004

1.
重庆交通大学交通运输学院重庆 400074
2.
宁夏交投高速公路有限公司银川 750000

基金项目:

国家重点研发计划项目 2018YFB1600200

重庆市研究生科研创新项目 CYS21356

详细信息

通讯作者:
马庆禄（1980—），博士，教授. 研究方向：智能交通与安全. E-mail：mql360@qq.com

中图分类号: U412.6
计量
- 文章访问数: 688
- HTML全文浏览量: 309
- PDF下载量: 28
- 被引次数: 0
出版历程
- 收稿日期: 2022-03-07
- 网络出版日期: 2023-05-13

A Method for Detecting Traffic Accidents on Highway Tunnel Sections Based on Abnormal Sound

1.
School of Traffic & Transportation, Chongqing Jiaotong University, Chongqing 400074, China
2.
Ningxia Expressway Co., LTD, Yinchuan 750000, China

摘要

摘要: 针对公路隧道内交通事故的动态感知问题，在传统检测方法的基础上引入声学检测理论与方法，研究基于异常声音检测的隧道交通事故智能检测方法。通过分析短时能量（short term energy，STE）和梅尔倒谱系数（Mel-scale frequency cepstral coefficients，MFCC）检测方法在事故段特征表征以及精度干扰方面的缺陷，提出1种改进的融合特征MFCCE研究隧道环境下的交通事故检测。提取STE和MFCC特征并使用主成分分析（principal component analysis，PCA）进行特征融合得到新的融合特征MFCCE。以真实行车事故数据为基础，构建包含刹车与碰撞声的2段隧道噪声实验样本数据，分别对应早高峰时段（07：00—08：00）及平峰时段（12：00—13：00）的行车条件对隧道内的事故环境进行模拟分析，利用端点检测对所提方法进行验证并与其余2种方法进行对比分析。使用Pearson简单相关系数法作为最终的评价方法，通过该方法计算得到的相关系数r对比三种检测结果与原始样本的正相关相性。实验结果表明：STE在平峰及早高峰时段的相关系数分别为0.933和0.988；MFCC在平峰及早高峰时段的相关系数均为0.998；而无论在平峰还是早高峰时段，MFCCE的相关系数（0.999）均高于另外其他2种检测方法。MFCCE的平均相关系数相比于其他2种检测方法（STE、MFCC）分别提高了3.95%和1.00%。
- 交通安全 /
- 交通事故 /
- 融合特征 /
- 端点检测 /
- 隧道安全
Abstract: In response to the need of effectively detecting traffic accidents in highway tunnel sections, a novel acoustic detection method is introduced, so as to study an intelligent way for detecting traffic accidents in tunnels based on abnormal sound. By analyzing the issues of using Short-Term Energy (STE) and Mel-scale Frequency Cepstral Coefficients (MFCC) in identifying accident sections and interfering with precision, a modified fusion feature MFCCE is proposed to detect traffic accidents in tunnel sections. The new fusion feature of the MFCCE is obtained by extracting STE and MFCC features in virtue of Principal Component Analysis (PCA) to conduct feature fusion. Based on an observed traffic accident dataset, a sample dataset of noise experiments in two tunnels containing braking and collision sounds is developed, which corresponds to the traffic scenario of morning peak hours (from 07:00 to 08:00) and regular hours (from 12:00 to 13:00) respectively. Then an endpoint detection method is utilized to validate the proposed method, which is then compared with the other two methods (STE and MFCC). The Pearson correlation coefficient is determined as the final evaluation method, through which correlation coefficients r is used to compare the positive correlation of the three test results with the original samples. Experimental results show that the correlation coefficients of STE are 0.933 and 0.988 in the regular and morning peak hours respectively; the correlation coefficients of MFCC are 0.998 in both regular and morning peak hours, while the correlation coefficient of MFCCE (0.999) is higher than the other two detection methods in both regular and morning peak hours. The average correlation coefficients of MFCCE are 3.95% and 1.00% higher than the other two detection methods, respectively.
- road safety /
- traffic accident /
- multiple features fusion /
- endpoint detection /
- tunnel safety

HTML全文

图 1 改进谱减法流程对比

Figure 1. Comparison of improved spectral subtraction process

下载: 全尺寸图片幻灯片

图 2 实际数据构建

Figure 2. Actual data construction

下载: 全尺寸图片幻灯片

图 3 隧道环境噪声的采集

Figure 3. Collecting tunnel ambient noise

下载: 全尺寸图片幻灯片

图 4 4个时段采样数据的原始波形示意图

Figure 4. The original waveform diagram of the sampled data in four periods

下载: 全尺寸图片幻灯片

图 5 实际数据构建

Figure 5. Actual data construction

下载: 全尺寸图片幻灯片

图 6 样本数据降噪前后对比图

Figure 6. Comparison of sample data before and after noise reduction

下载: 全尺寸图片幻灯片

图 7 平峰及高峰时段3种检测方法结果

Figure 7. The results of three detection methods in flat peak period and peak period

下载: 全尺寸图片幻灯片

图 8 样本检测方法结果对比图

Figure 8. Sample detection method results comparison diagram

下载: 全尺寸图片幻灯片

表 1 实际数据与合成样本的参数对比

Table 1. Comparison of actual data with synthetic samples

对比参数	实际数据	合成数据	匹配度/%
声时历程/s	10.3	10.2	99.03
频率/Hz	48 000	48 000	100
响度/sone	31.62	29.47	93.20
尖锐度/acum	1.78	1.62	91.01
粗糙度/asper	0.99	0.90	90.91
波动度/vacil	0.40	0.36	90.00

下载: 导出CSV

表 1 平峰时段检测结果对比

Table 1. Comparison of three detection results in flat peak period 单位: s

算法	t_s^b	t_e^b	t_s^c	t_e^c
REAL	3.50	4.50	5.00	7.50
STE	3.31	4.71	4.93	7.25
MFCC	3.46	4.58	4.98	8.19
MFCCE	3.46	4.58	4.99	7.40
注：t_s^b，t_e^b为刹车段的开始时刻与结束时刻；t_s^c，t_e^c为碰撞段的开始时刻与结束时刻；REAL为原始样本。

下载: 导出CSV

表 2 早高峰时段检测结果对比

Table 2. Comparison of detection results during morning peak hours 单位: s

算法	t_s^b	t_e^b	t_s^c	t_e^c
REAL	3.50	4.50	5.00	7.50
STE	3.61	4.74	5.00	6.58
MFCC	3.46	4.58	5.00	8.05
MFCCE	3.63	4.73	5.00	7.41

下载: 导出CSV

表 3 3种检测方法Pearson简单相关系数分析对比

Table 3. Pearson simple correlation coefficient analysis and comparison of three detection methods

算法	r₁	r₂	p₁	p₂
STE	0.933	0.988	0.007	0.012
MFCC	0.998	0.998	0.002	0.002
MFCCE	0.999	0.999	0.001	0.001
注：r₁，p₁为平峰时段的相关系数及概率p值；r₂，p₂为高峰时段的相关系数及概率p值。

下载: 导出CSV

参考文献(25)

[1]	李雪玲. 高速公路长隧道车辆运行速度分析与安全措施研究[D]. 西安: 长安大学, 2012. LI X L. Analysis of vehicle operating speed and study on safety measures of freeway long tunnel[D]. Xi'an: Chang'an University, 2012. (in Chinese)
[2]	张璇, 唐进君, 黄合来, 等. 山区高速公路隧道路段与开放路段的事故影响因素分析[J]. 交通信息与安全, 2022, 40 (3): 10-18. doi: 10.3963/j.jssn.1674-4861.2022.03.002 ZHANG X, TANG J J, HUANG H L, et al. An analysis of influential factors of crashes at tunnels and open sections of mountainous freeways[J]. Journal of Transport Information and Safety, 2022, 40(3): 10-18. (in Chinese) doi: 10.3963/j.jssn.1674-4861.2022.03.002
[3]	陈湘生, 徐志豪, 包小华, 等. 中国隧道建设面临的若干挑战与技术突破[J]. 中国公路学报, 2020, 33(12): 1-14. doi: 10.3969/j.issn.1001-7372.2020.12.001 CHEN X S, XU Z H, BAO X H, et al. Challenges and technological breakthroughs in tunnel construction in China[J]. China Journal of Highway and Transport, 2020, 33(12): 1-14. (in Chinese) doi: 10.3969/j.issn.1001-7372.2020.12.001
[4]	徐志刚, 李金龙, 赵祥模, 等. 智能公路发展现状与关键技术[J]. 中国公路学报, 2019, 32(8): 1-24. doi: 10.19721/j.cnki.1001-7372.2019.08.001 XU Z G, LI J L, ZHAO X M, et al. A review on intelligent road and its related key technologies[J]. China Journal of Highway and Transport, 2019, 32(8): 1-24. (in Chinese) doi: 10.19721/j.cnki.1001-7372.2019.08.001
[5]	马庆禄, 邹政, 刘丰杰. 基于行车声音端点检测的交通量统计[J]. 科学技术与工程, 2020(4): 1676-1683. doi: 10.3969/j.issn.1671-1815.2020.04.058 MA Q L, ZOU Z, LIU F J. Traffic statistics based on the endpoint detection of driving acoustic signals[J]. Science Technology and Engineering, 2020(4): 1676-1683. (in Chinese) doi: 10.3969/j.issn.1671-1815.2020.04.058
[6]	胡永. 地下隧道单向双车道双洞交通事故准确检测[J]. 计算机仿真, 2019, 36(6): 155-159. doi: 10.3969/j.issn.1006-9348.2019.06.031 HU Y. One-way two-lane double tunnel tunnel traffic event detection simulation[J]. Computer Simulation, 2019, 36(6): 155-159. (in Chinese) doi: 10.3969/j.issn.1006-9348.2019.06.031
[7]	阳国清, 莫鸿强, 李文, 等. 连续语音流中咳嗽信号的端点检测[J]. 生物医学工程学杂志, 2010, 27(3): 544-547, 555. https://www.cnki.com.cn/Article/CJFDTOTAL-SWGC201003015.htm YANG G Q, MO H Q, LI W, et al. The endpoint detection of cough signal in continuous speech[J]. Journal of Biomedical Engineering, 2010, 27(3): 544-547, 555. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-SWGC201003015.htm
[8]	高晗, 裴玉龙. 基于车辆噪音时域特征的交通量统计方法[J]. 公路交通科技, 2008(4): 113-116, 121. doi: 10.3969/j.issn.1002-0268.2008.04.023 GAO H, PEI Y L. Statistic method on traffic volume based on vehicle's time-domain noise features[J]. Journal of Highway and Transportation Research and Development, 2008(4): 113-116, 121. (in Chinese) doi: 10.3969/j.issn.1002-0268.2008.04.023
[9]	何小华. 基于车辆运行噪声的交通量在线统计算法研究[D]. 长春: 东北师范大学, 2008. HE X H. Research of online statistical algorithm about traffic volume based on running vehicles noise[D]. Changchun: Northeast Normal University, 2008. (in Chinese)
[10]	马庆禄, 邹政. 1种识别重叠噪声的交通量检测算法[J]. 计算机应用研究, 2020, 37(4): 1069-1072, 1080. https://www.cnki.com.cn/Article/CJFDTOTAL-JSYJ202004022.htm MA Q L, ZOU Z. Traffic detection algorithm for identify overlapping noise[J]. Application Research of Computers, 2020, 37(4): 1069-1072, 1080. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JSYJ202004022.htm
[11]	NOORALAHIYAN A Y, KIRBY H R, MCKEOWN D. Vehicle classification by acoustic signature[J]. Mathematical & Computer Modelling, 1998, 27(9-11): 205-214.
[12]	LEFEBVRE N, CHEN X D, BEAUSEROY P, et al. Traffic flow estimation using acoustic signal[J]. Engineering Applications of Artificial Intelligence, 2017, 64: 164-171. doi: 10.1016/j.engappai.2017.05.019
[13]	张宏睿, 马秀荣, 单云龙. 车站运行列车异音检测方法[J]. 计算机应用与软件, 2019, 36(8): 130-137, 171. ZHANG H R, MA X R, SHAN Y L. Abnormal sound detection method of trains running at stations[J]. Computer Applications and Software, 2019, 36(8): 130-137, 171. (in Chinese)
[14]	张璐璐, 陈耀武, 蒋荣欣. 智能监控前端系统中异常声音检测的实现[J]. 计算机工程, 2014, 40(1): 218-221, 227. ZHANG L L, CHEN Y W, JIANG R X. Implementation of abnormal sound detection in intelligent surveillance front-end system[J]. Computer Engineering, 2014, 40(1): 218-221, 227. (in Chinese)
[15]	王若平, 李仁仁, 陈达亮, 等. 基于改进小波包去噪与梅尔倒谱系数的低信噪比交通环境声音识别[J]. 科学技术与工程, 2019, 19(36): 290-295. https://www.cnki.com.cn/Article/CJFDTOTAL-KXJS201936058.htm WANG R P, LI R R, CHEN D L, et al. Low signal to noise ratio traffic environment acoustic recognition based on improved wavelet packet denoising and Mel cepstrum coefficient[J]. Science Technology and Engineering, 2019, 19 (36): 290-295. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-KXJS201936058.htm
[16]	吕霄云, 王宏霞. 基于MFCC和短时能量混合的异常声音识别算法[J]. 计算机应用, 2010, 30(3): 796-798. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJY201003063.htm LYU X Y, WANG H X. Abnormal audio recognition algorithm based on mfcc and short-term energy[J]. Journal of Computer Applications, 2010, 30(3): 796-798. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JSJY201003063.htm
[17]	毛锦, 李林聪, 刘凯, 等. 无人驾驶汽车行车环境下鲁棒性声学特征提取算法[J]. 中国公路学报, 2019, 32(6): 169-175. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGGL201906018.htm MAO J, LI L C, LIU K, et al. Robust acoustic feature extraction algorithm for driving environment of driverless cars[J]. China Journal of Highway and Transport, 2019, 32(6): 169-175. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZGGL201906018.htm
[18]	韦娟, 张芃楠, 岳凤丽, 等. 基于PSO-PF算法的SVM识别方法及其在异常声音中的应用[J]. 北京邮电大学学报, 2019, 42(3): 58-63. https://www.cnki.com.cn/Article/CJFDTOTAL-BJYD201903009.htm WEI J, ZHANG P N, YUE F L, et al. Recognition and application of abnormal sound via SVM based on PSO-PF[J]. Journal of Beijing University of Posts and Telecommunications, 2019, 42(3): 58-63. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-BJYD201903009.htm
[19]	VIVEK T, SHIVKUMAR K, RAGHURAM K et al. Vehicular traffic density state estimation based on cumulative road acoustics[J]. IEEE Transactions on Intelligent Transportation Systems, 2012, 13(3): 1156-1166.
[20]	AMRIT S, CHIRANJEEV K. An approach to detect the accident in VANETs using acoustic signal[J]. Applied Acoustics, 2020(163): 107205.
[21]	胡涛, 张超, 程炳, 等. 卷积神经网络在异常声音识别中的研究[J]. 信号处理, 2018, 34(3): 357-367. https://www.cnki.com.cn/Article/CJFDTOTAL-XXCN201803013.htm HU T, ZHANG C, CHENG B, et al. Research on abnormal audio event detection based on convolutional neural networks[J]. Journal of Signal Processing, 2018, 34(3): 357-367. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-XXCN201803013.htm
[22]	KAUR A, SOOD N, AGGARWAL N, et al. Traffic state detection using smartphone based acoustic sensing[J]. Journal of Intelligent & Fuzzy Systems, 2017, 32(4): 3159-3166.
[23]	JADHAV P, SAWARKAR S D, PETE D J. Roadside acoustic signals based road traffic density estimation[C]. 4^th International Conference on Computing Communication Control and Automation(ICCUBEA), Pune, India: IEEE, 2018.
[24]	SU Y, ZHANG K, WANG J, et al. Environment sound classification using a two-stream CNN based on decision-level fusion[J]. Sensors, 2019, 19(7): 1733.
[25]	柯炜, 张铭, 张铁成. 1种利用分布式传声器阵列的声源三维定位方法[J]. 声学学报, 2017, 42(3): 361-369. https://www.cnki.com.cn/Article/CJFDTOTAL-XIBA201703013.htm KE W, ZHANG M, ZHANG T C. Three-dimensional sound source localization using distributed microphone arrays[J]. Act Acistoca, 2017, 42(3): 361-369. https://www.cnki.com.cn/Article/CJFDTOTAL-XIBA201703013.htm