CN103531028A - Vehicle detection method based on linear sound and vibration sensor array - Google Patents
Vehicle detection method based on linear sound and vibration sensor array Download PDFInfo
- Publication number
- CN103531028A CN103531028A CN201310449792.6A CN201310449792A CN103531028A CN 103531028 A CN103531028 A CN 103531028A CN 201310449792 A CN201310449792 A CN 201310449792A CN 103531028 A CN103531028 A CN 103531028A
- Authority
- CN
- China
- Prior art keywords
- signal
- sensor array
- sensor
- vehicle
- speed
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Traffic Control Systems (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention provides a vehicle detection method based on a linear sound and vibration sensor array. The method comprises the following steps that 1, a sensor array is distributed and arranged; 2, the short time energy enveloping line of array signals is calculated; 3, the enveloping line of the sensor array signals are subjected to deviation overlapping; 4, interference signals and mistake correlation signals are deleted; 5, the vehicle signal detection and speed detection and the like are carried out. The similarity of the enveloping lines on the sensor array and the relevance between the speed and direction of the vehicle target are utilized, the sound or vibration signals can be singly adopted for carrying out vehicle detection, and the sound and vibration signals can be adopted in a combined way for detection.
Description
Technical field
The present invention relates to a kind of vehicle checking method, particularly the vehicle checking method based on vehicle sounds or vibration signal.The road surface moving vehicle that the method can be used for Unattended Ground Sensor network detects, and can fast detecting go out the vehicle moving along road, and provide speed and direction.
Background technology
Sound and vibration signal are a kind of means always that on modern battlefield, information obtains, and have passive detection, good concealment, safety coefficient is high, can long time continuous working, be not subject to the features such as landform and illumination condition restriction, there is comprehensive, round-the-clock detectivity.It is to grasp situation of battlefield that vehicle target based on sound and vibration signal is surveyed, and understands a kind of important means that enemy and we troops distribute, the method demand that is also widely used in fields such as anti-terrorism, security protections.
The method that vehicle detection method based on sound and vibration signal adopts is conventionally: first extract the feature of signal, and then signal characteristic and neighbourhood noise relatively or by signal characteristic and known sample storehouse are compared, and then detect signals of vehicles.Conventional sound and vibration signal feature have amplitude, energy, envelope, power spectrum, zero-crossing rate, entropy, wavelet character, Mei Er (Mel) cepstrum etc.As Huadong Wu, in " Vehicle Sound Signature Recognition by Frequency Vector Principal Component Analysis " that Mel Siegel and Pradeep Khosla deliver on " the TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT " in October, 1999, propose to utilize the power spectrum principal component analysis (PCA) of signal to extract feature, then adopt the eigenface method (eigenfaces method method) in Identification of Images, it is carried out to suitable modification, then vehicle acoustical signal is carried out to pattern-recognition, Wu Zongfu, Yang Rushu, in " the acoustic target intelligence system based on knowledge " that Li Xiaofeng etc. deliver on " computer measurement and control " of the 3rd phase the 18th volume in 2010, propose to adopt and get the knowledge base that Mel cepstrum coefficient (MFCC) parameter attribute builds acoustic target information, acoustic target MFCC characteristic parameter is mapped as to bianry image and carries out template matches knowledge method for distinguishing, Duan Hanqing, proposes the vehicle vibration signal detection algorithm based on Power Spectrum Distribution and negentropy in " vehicle vibration signal detection algorithm research under strong noise background " that all think of Yue are delivered for 2012 in " radio engineering ".Utilize the signal detection method right with noise characteristic diversity ratio relative simple, but be easily subject to noise, false drop rate is very high; And detection method based on pattern-recognition, as statistical recognition methods, function method of identification, logic identification method, Fuzzy Recognition, Grey Identification, neural network recognization method etc., can not only detect vehicle target, can also identify type of vehicle, but pattern-recognition need to there is a large amount of sample event bases just can reach desirable detection effect.To detect based on feature difference or the moving vehicle based on pattern-recognition the movement velocity that all can only carry out vehicle detection and can not provide target vehicle.
Summary of the invention
The object of the invention is to provide a kind of simple and reliable vehicle detection method based on linear voice, shock sensor array, the method utilizes vehicle target in the relevance of the similarity of sensor array coenvelope line and speed, direction, both can adopt separately sound or vibration signal to carry out vehicle detection, also can combine and adopt sound and vibration signal to detect, not only can whether there is moving vehicle by sound vibrations input, can also detect speed and the direction of moving vehicle.
Technical solution of the present invention is:
A vehicle detection method for linear voice, shock sensor array, its special character is: comprise the following steps:
1] sensor array is laid:
Along a road arrangement N sensor of intending detecting, N>=3, form a N unit linear sensor array parallel with road bearing of trend, and the spacing of sensor is got 5~20 meters, obtains relative coordinate or the absolute coordinates (x of sensor array
0, x
1..., x
n);
2] the short-time energy envelope of computing array signal:
For signal x (t), the computing formula that defines its short-time energy envelope is as follows:
In formula: l is frame length, t=t
0+ δ l, δ l is that frame moves, t
0for signal initial time;
The signal of N unit linear sensor array record is x
0(t), x
1(t) ..., x
n(t), according to formula (1), calculate respectively, obtain corresponding signal short-time energy envelope e
0(t), e
1(t) ..., e
n(t);
3] envelope of sensor array signal is offset to stack:
The vehicle moving along road can only be along road bearing of trend, sets a direction for just, and reverse speed, for negative, is set the top speed V of vehicle movement, the speed v ∈ of vehicle to be detected [V, V],
First select the envelope signal of any one sensor of N unit linear sensor array as reference signal, according to predefined one group of velocity amplitude (v
1, v
2..., v
m), according to δ t=r/v, carry out array signal short-time energy envelope skew stack respectively, obtain one group of skew superposition value:
In formula: the number that N is sensor array, r
ifor the distance of reference sensor to i sensor, v
kcar speed for hypothesis;
4] undesired signal and mistake correlation signal are rejected:
Skew stack maximum value to each short-time energy envelope
divided by each sensor signal e
i(t+r
i/ v
k) variance STD[e
i(t+r
i/ v
k)], obtain a new characteristic quantity:
5] signals of vehicles detects and speed detection:
To the sensor array of laying as step 1, according to 2,3,4 steps, describe, be offset frame by frame stack, obtain one group of maximum value
with corresponding speed v
k, according to experimental verification, when
time can think that this testing result is a vehicle target, its movement velocity is v corresponding in formula (3)
k.
2, the vehicle detection method based on linear voice, shock sensor array according to claim 1, it is characterized in that: the first linear sensor array of described N is comprised of N sound transducer, or by N shock sensor, formed, or constituted jointly by sound transducer and shock sensor.
Tool of the present invention has the following advantages:
1, the present invention utilizes vehicle target in the relevance of the similarity of sensor array coenvelope line and speed, direction, both can adopt separately sound or vibration signal to carry out vehicle detection, also can combine and adopt sound and vibration signal to detect.
2, whether the present invention not only can have moving vehicle by sound vibrations input, can also detect speed and the direction of moving vehicle.
3, sensor array of the present invention is laid simple.
4, invention adopts the short-time energy feature of signal, and characteristic signal is stable, and algorithm of target detection is simple.
5, the present invention has effectively utilized the relevance of array signal, and target detection rate is high, and false alarm rate is low.
6, the present invention can effectively distinguish the intensive situation about passing through of a plurality of vehicles.
Accompanying drawing explanation
Fig. 1 is the schematic diagram of array laying of the present invention and practical application.
Fig. 2 is the schematic diagram that signal divides frame.
Fig. 3 is that a plurality of vehicles are through out-of-date linear sensor array signal.
Fig. 4 utilizes the testing result of the present invention to the measured signal shown in Fig. 3.
Embodiment
Step of the present invention is as follows:
The first step, sensor array is laid.
Along road arrangement N(N>=3 of intending detecting) individual sensor, form a linear sensor array parallel with road bearing of trend, obtain relative coordinate or the absolute coordinates (x of sensor array
0, x
1..., x
n), according to experimental verification, for common road traveling vehicle, the spacing of sensor get 5~20 meters proper.
Second step, the short-time energy envelope of computing array signal.
Short-time energy is a kind of expression mode of signal intensity, and to square being directly proportional of signal amplitude, the short-time energy envelope of signal has reflected the time dependent process of signal intensity.For signal x (t), the computing formula that defines its short-time energy envelope is as follows:
In formula: l is frame length, t=t
0+ δ l, δ l is that frame moves, t
0for signal initial time.
The signal of N unit sensor array record is x
0(t), x
1(t) ..., x
n(t), according to formula (1), calculate respectively, obtain corresponding signal envelope e
0(t), e
1(t) ..., e
n(t).
The 3rd step, is offset stack to the envelope of sensor array signal.
Because the intensity of vehicle sounds, vibration signal and the distance between vehicle and sensor are inversely proportional to, if be v along the car speed of road driving, at t, constantly pass through sensor x
0, envelope signal e
0(t) be a maximum value, sensor x
0to sensor x
idistance be r
i, vehicle is through sensor x
itime be t+r
i/ v, envelope signal e
i(t+r
i/ v) be maximum value, according to time delay δ t=r/v, the short-time energy envelope of all the sensors be offset to stack, must obtain a maximum value
According to general knowledge, the movement velocity of vehicle is inevitable in certain scope, conventionally as the maximal rate < 200km/h of car, truck maximal rate < 120km/h, the vehicle moving along road can only be along road bearing of trend, sets a direction for just, and reverse speed is for negative, for example, when road is East and West direction, suppose eastwards for canonical is westwards for negative.The top speed V that sets vehicle movement, the speed v ∈ of vehicle to be detected [V, V], sets one group of speed :-V<v<sub TranNum="155">1</sub><v<sub TranNum="156">2</sub><...<v<sub TranNum="157">n</sub><V, must have a speed v<sub TranNum="158">k</sub>approach most true velocity v.
Array signal envelope when stack skew, first select the envelope signal of any one sensor of array as reference signal, according to predefined one group of velocity amplitude (v
1, v
2..., v
m), according to δ t=r/v, carry out array signal short-time energy envelope skew stack respectively, obtain one group of skew superposition value:
Wherein: the number that N is sensor array, r
ifor the distance of reference sensor to i sensor, v
kcar speed for hypothesis.
The 4th step, undesired signal and mistake correlation signal are rejected.
In formula (2)
obtain maximum value and have three kinds of situation: 1, e
i(t+r
i/ v
k) be all maximum value and signal similar; 2, e
i(t+r
i/ v
k) be all maximum value but signal is dissimilar; 3, e
i(t+r
i/ v
k) be not maximum value entirely but in these superposed signals, have a very large undesired signal.In the region of laying at linear sensor array, the variation of vehicle-state and noise can be ignored, and the waveform that same vehicle target lists at sensor array is similar.If the first situation, the variance between signal is very little, and the signal variance in latter two situation is larger, therefore the skew stack maximum value to each short-time energy envelope
variance STD[e divided by each sensor signal ei (t+ri/vk)
i(t+r
i/ v
k)], obtain a new characteristic quantity:
This new characteristic quantity has merged signals of vehicles envelope extreme value and two features of similarity of linear sensor array record, can effectively suppress the interference of singular value to pack result.
The 5th step, signals of vehicles detects and speed detects.
To the sensor array of laying as step 1, according to 2,3,4 steps, describe, be offset frame by frame stack, obtain one group of maximum value
with corresponding speed v
k, according to experimental verification, when
time can think that this testing result is a vehicle target, its movement velocity is v corresponding in formula (3)
k.Because testing result comprises time and velocity information simultaneously, even if therefore have, a plurality of vehicles are intensive to be passed through, and the method still can utilize the speed of target to detect and distinguish.
The embodiment of this patent is described below in conjunction with a vehicle detection example
1, shown in Fig. 1, in the section of needs monitoring, lay 9 yuan of linear acoustic sensor array, transducer spacing 20m.
2, Fig. 3 is 3 vehicles being recorded to of sensor array waveforms of process successively.According to method shown in Fig. 2, divide frame respectively, frame length 1s, frame moves 0.25s, then according to formula (1), calculates the short-time energy envelope that calculates respectively 9 sensors.
3, select the reference signal that first sensor is skew stack, from the direction of nine sensors of first sensor to for just, set the velocity range [3232] detecting, the m/s of unit, sets one group of speed with the step-length of 0.5m/s: (32 ,-31.5-31,31,3.15,32), then according to formula (2), carry out envelope skew stack successively, obtain one group
data, and then right according to formula (3)
revise, obtain
fig. 4 is
two-dimensional distribution, X-axis is signal time, the inverse that Y-axis is speed, each point represents the result of carrying out envelope skew stack based on its corresponding time and speed.
4, find out
be greater than 4 testing result and corresponding time and speed thereof, as can be seen from Figure 4, when 51s, 57s and 70s, there are three cars with the speed of 20m/s, 15m/s and 18m/s, to pass through respectively the position of sensor 1, direction is all the direction from sensor 1 to sensor 9.
Claims (2)
1. the vehicle detection method based on linear voice, shock sensor array, is characterized in that: comprise the following steps:
1] sensor array is laid:
Along a road arrangement N sensor of intending detecting, N>=3, form a N unit linear sensor array parallel with road bearing of trend, and the spacing of sensor is got 5~20 meters, obtains relative coordinate or the absolute coordinates (x of sensor array
0, x
1..., x
n);
2] the short-time energy envelope of computing array signal:
For signal x (t), the computing formula that defines its short-time energy envelope is as follows:
In formula: l is frame length, t=t
0+ δ l, δ l is that frame moves, t
0for signal initial time;
The signal of N unit linear sensor array record is x
0(t), x
1(t) ..., x
n(t), according to formula (1), calculate respectively, obtain corresponding signal short-time energy envelope e
0(t), e
1(t) ..., e
n(t); 3] envelope of sensor array signal is offset to stack:
The vehicle moving along road can only be along road bearing of trend, sets a direction for just, and reverse speed, for negative, is set the top speed V of vehicle movement, the speed v ∈ of vehicle to be detected [V, V],
First select the envelope signal of any one sensor of N unit linear sensor array as reference signal, according to predefined one group of velocity amplitude (v
1, v
2..., v
m), according to δ t=r/v, carry out array signal short-time energy envelope skew stack respectively, obtain one group of skew superposition value:
In formula: the number that N is sensor array, r
ifor the distance of reference sensor to i sensor, v
kcar speed for hypothesis;
4] undesired signal and mistake correlation signal are rejected:
Skew stack maximum value to each short-time energy envelope
divided by each sensor signal e
i(t+r
i/ v
k) variance STD[e
i(t+r
i/ v
k)], obtain a new characteristic quantity:
5] signals of vehicles detects and speed detection:
To the sensor array of laying as step 1, according to 2,3,4 steps, describe, be offset frame by frame stack, obtain one group of maximum value
with corresponding speed v
k, according to experimental verification, when
time can think that this testing result is a vehicle target, its movement velocity is v corresponding in formula (3)
k.
2. the vehicle detection method based on linear voice, shock sensor array according to claim 1, it is characterized in that: the first linear sensor array of described N is comprised of N sound transducer, or by N shock sensor, formed, or constituted jointly by sound transducer and shock sensor.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310449792.6A CN103531028A (en) | 2013-09-27 | 2013-09-27 | Vehicle detection method based on linear sound and vibration sensor array |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310449792.6A CN103531028A (en) | 2013-09-27 | 2013-09-27 | Vehicle detection method based on linear sound and vibration sensor array |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103531028A true CN103531028A (en) | 2014-01-22 |
Family
ID=49932999
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310449792.6A Pending CN103531028A (en) | 2013-09-27 | 2013-09-27 | Vehicle detection method based on linear sound and vibration sensor array |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103531028A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104332162A (en) * | 2014-11-25 | 2015-02-04 | 武汉大学 | Audio signal recognition system for vehicle |
CN104851301A (en) * | 2015-05-22 | 2015-08-19 | 重庆交通大学 | Vehicle parameter identification method based on deceleration strip sound analysis |
CN109087512A (en) * | 2018-08-20 | 2018-12-25 | 中北大学 | A kind of overload of vehicle dynamic monitoring method based on distributed shock sensor array |
CN109141598A (en) * | 2018-08-20 | 2019-01-04 | 中北大学 | A kind of vehicle dynamic overload monitoring system based on distributed shock sensor array |
CN109472973A (en) * | 2018-03-19 | 2019-03-15 | 国网浙江桐乡市供电有限公司 | A kind of real-time traffic methods of exhibiting and system based on voice recognition |
CN112710989A (en) * | 2020-12-30 | 2021-04-27 | 中国人民解放军32212部队 | Tank armored vehicle sound vibration artificial intelligence detection positioning system |
CN115019521A (en) * | 2022-05-19 | 2022-09-06 | 河北工业大学 | Method and system for determining vehicle speed |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6384740B1 (en) * | 2001-07-30 | 2002-05-07 | Khaled A. Al-Ahmed | Traffic speed surveillance and control system |
US6535143B1 (en) * | 1998-04-08 | 2003-03-18 | Kabushiki Kaisha Kenwood | Vehicle detection system |
CN101256715A (en) * | 2008-03-05 | 2008-09-03 | 中科院嘉兴中心微系统所分中心 | Multiple vehicle acoustic signal based on particle filtering in wireless sensor network |
CN101266717A (en) * | 2008-04-25 | 2008-09-17 | 北京科技大学 | A car detection recognition system and method based on MEMS sensor |
CN102622881A (en) * | 2012-03-19 | 2012-08-01 | 深圳市锐明视讯技术有限公司 | Method and device for detecting vibration |
-
2013
- 2013-09-27 CN CN201310449792.6A patent/CN103531028A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6535143B1 (en) * | 1998-04-08 | 2003-03-18 | Kabushiki Kaisha Kenwood | Vehicle detection system |
US6384740B1 (en) * | 2001-07-30 | 2002-05-07 | Khaled A. Al-Ahmed | Traffic speed surveillance and control system |
CN101256715A (en) * | 2008-03-05 | 2008-09-03 | 中科院嘉兴中心微系统所分中心 | Multiple vehicle acoustic signal based on particle filtering in wireless sensor network |
CN101266717A (en) * | 2008-04-25 | 2008-09-17 | 北京科技大学 | A car detection recognition system and method based on MEMS sensor |
CN102622881A (en) * | 2012-03-19 | 2012-08-01 | 深圳市锐明视讯技术有限公司 | Method and device for detecting vibration |
Non-Patent Citations (1)
Title |
---|
陶良小,等: "基于线性声传感器阵列信号包络线叠加的运动车辆检测方法", 《国家安全地球物理丛书(七)—地球物理与核探测》, 31 December 2011 (2011-12-31), pages 379 - 384 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104332162A (en) * | 2014-11-25 | 2015-02-04 | 武汉大学 | Audio signal recognition system for vehicle |
CN104851301A (en) * | 2015-05-22 | 2015-08-19 | 重庆交通大学 | Vehicle parameter identification method based on deceleration strip sound analysis |
CN104851301B (en) * | 2015-05-22 | 2017-01-25 | 重庆交通大学 | Vehicle parameter identification method based on deceleration strip sound analysis |
CN109472973A (en) * | 2018-03-19 | 2019-03-15 | 国网浙江桐乡市供电有限公司 | A kind of real-time traffic methods of exhibiting and system based on voice recognition |
CN109472973B (en) * | 2018-03-19 | 2021-01-19 | 国网浙江桐乡市供电有限公司 | Real-time traffic display method based on voice recognition |
CN109087512A (en) * | 2018-08-20 | 2018-12-25 | 中北大学 | A kind of overload of vehicle dynamic monitoring method based on distributed shock sensor array |
CN109141598A (en) * | 2018-08-20 | 2019-01-04 | 中北大学 | A kind of vehicle dynamic overload monitoring system based on distributed shock sensor array |
CN112710989A (en) * | 2020-12-30 | 2021-04-27 | 中国人民解放军32212部队 | Tank armored vehicle sound vibration artificial intelligence detection positioning system |
CN115019521A (en) * | 2022-05-19 | 2022-09-06 | 河北工业大学 | Method and system for determining vehicle speed |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103531028A (en) | Vehicle detection method based on linear sound and vibration sensor array | |
CN105574537B (en) | Dangerous driving behavior testing and evaluation method based on multisensor | |
US11373523B2 (en) | Method for vehicle detection based on single-axis geomagnetic sensor | |
US9733350B2 (en) | Vehicle radar control | |
CN102866391B (en) | Multi-target detection method based on short-time Fourier transform and fractional Fourier transform | |
CN103761889B (en) | A kind of front vehicles effective target defining method of preventing collision | |
CN103946715B (en) | Sound Sources Detection system | |
CN103728599B (en) | The method of false targets interference is suppressed with the active radar and passive radar net of other place configure | |
CN111582160B (en) | Vehicle track extraction method and device based on distributed optical fiber sensing system | |
Xu et al. | Tire slip angle estimation based on the intelligent tire technology | |
CN106154332A (en) | A kind of borehole microseismic ripple event first arrival recognition methods in length and breadth | |
CN106596149A (en) | Method for monitoring and diagnosing flat wheel state of rail vehicle based on VMD | |
CN105862935B (en) | A kind of damnification recognition method for retaining wall structure system | |
Kawakatsu et al. | Deep sensing approach to single-sensor vehicle weighing system on bridges | |
CN103177584B (en) | Vehicle speed detection method based on enumeration probe | |
CN104851301B (en) | Vehicle parameter identification method based on deceleration strip sound analysis | |
CN107274679A (en) | Vehicle identification method, device, equipment and computer-readable recording medium | |
Xu et al. | Seismic target recognition based on parallel recurrent neural network for unattended ground sensor systems | |
CN102830421B (en) | Method for identifying redundancies and assembly of satellite-borne electronic device | |
CN106846900A (en) | A kind of Roadside Parking position piezoelectric cable laying structure and the method for parking space state detection | |
CN106353819A (en) | In-well three-component microseism first arrival pickup method | |
Lan et al. | Recognition of moving ground targets by measuring and processing seismic signal | |
CN106168675A (en) | Borehole microseismic P/S ripple event recognition method and device | |
Kubo et al. | Design of ultra low power vehicle detector utilizing discrete wavelet transform | |
CN106371082B (en) | Linear speed pull-off jamming recognition methods |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20140122 |