CN106970024A - Gauge detection distance-finding method and system based on camera and controllable stroboscopic light source - Google Patents
Gauge detection distance-finding method and system based on camera and controllable stroboscopic light source Download PDFInfo
- Publication number
- CN106970024A CN106970024A CN201710158503.5A CN201710158503A CN106970024A CN 106970024 A CN106970024 A CN 106970024A CN 201710158503 A CN201710158503 A CN 201710158503A CN 106970024 A CN106970024 A CN 106970024A
- Authority
- CN
- China
- Prior art keywords
- camera
- light source
- controllable
- train
- stroboscopic light
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/01—Arrangements or apparatus for facilitating the optical investigation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/70—Circuitry for compensating brightness variation in the scene
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/90—Arrangement of cameras or camera modules, e.g. multiple cameras in TV studios or sports stadiums
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/95—Computational photography systems, e.g. light-field imaging systems
- H04N23/951—Computational photography systems, e.g. light-field imaging systems by using two or more images to influence resolution, frame rate or aspect ratio
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2201/00—Features of devices classified in G01N21/00
- G01N2201/10—Scanning
- G01N2201/102—Video camera
Abstract
The invention discloses a kind of gauge detection distance-finding method based on camera and controllable stroboscopic light source and system, under train driving state, adjust the flash frequency of CMOS cameras frame frequency and stroboscopic light source, controllable stroboscopic light source is carried out multiple stroboscopic in CMOS camera exposure cycles same frame in, make measured object cycle imaging in the same two field picture of CMOS cameras less than CMOS viewing field of camera 1/3rd；Fourier transformation is carried out to cycle image, amplitude reaches maximum at specific frequency in amplitude spectral curve, frequency corresponding to the maximum repeats the pixel quantity between being imaged for measured object in cycle strobe, pixel quantity measured object in cycle strobe repeated between being imaged is multiplied with each length in pixels of camera, obtains the spacing between multiimage；Calculated according to camera imaging principle and obtain measured object distance, the distance-finding method measurement accuracy and accuracy are high, be suitable for railway high speed on-line checking.
Description
Technical field
The present invention relates to railway clearance detection technique field, and in particular to one kind is based on CMOS area array cameras and controllable flash
The gauge detection distance-finding method and system of light source.
Background technology
Railway clearance detection enjoys railway relevant departments to pay much attention to always, and safety of railway traffic and conevying efficiency are carried
Height plays vital effect.The means of testing of current gauge detection mainly includes the detection of contact gauge and contactless limit
Detect on boundary.
The early stage of gauge detection, domestic and international gauge detection range-measurement system generally uses contact, mainly there is mechanical tentacle formula
And testing stand pattern.Both approaches are all by a detection car framework installed above, in operation, by recording frame
Frame is contacted with surrounding objects realizes that gauge is detected, detection speed is very low, has a strong impact on railway normal operation.
Contactless gauge detects that simplest method of testing is, come measurement distance, to be surveyed with theodolite using electromagnetic positioning
Measuring angle.And this detection method needs collection in worksite higher density measuring point, labor intensive time.
Laser ranging is, based on one of measuring method being most widely used in optical non-contact detection technology, can to divide
For pulse ranging, continuous wave phase difference ranging and interferometric method laser ranging.Radar and supersonic sounding technology are surveyed with laser pulsed
It is to be using ultrasonic wave or the aerial spread speed of electromagnetic wave, it is known that measuring the ripple in hair away from more similar in principle
The time that object is reflected is run into after penetrating, calculating obtains relative distance.The core of these methods is to returning to luminous intensity
In identification, gauge detection process, camera data is big with respect to saltus step, sensitivity is too low, it is impossible to reach the accuracy and precision of ranging
Requirement.
Conventional machine vision ranging is mainly the side that object dimensional geological information is obtained by multiple image (general two width)
Method, in computer vision system, with twin camera, obtains two width digital pictures of surrounding scenes simultaneously from different perspectives, by
The 3D shape of computer reconstruction surrounding scenes and position.General stereoscopic vision detection technology algorithm based on homotopy mapping is answered
Miscellaneous, matching is difficult, frame per second finite sum data amount of calculation is huge, can be obtained preferably in the case where low speed, detection object are fixed
As a result, but be difficult to meet at a high speed and the changeable requirement of environment.Range of triangle gauge detection method have compared with high measurement accuracy and compared with
The advantage of close circumferential sampled point and higher longitudinal sample frequency, but gauge detection speed is very fast, vertical and horizontal measuring point
More, data volume is larger, and the switching speed of liquid crystal board each unit can not meet requirement, causes to be difficult to high speed on-line checking.
The content of the invention
The technical problems to be solved by the invention are to overcome the shortcomings of to mention with defect that there is provided one in background above technology
Plant and be suitable for railway high speed on-line checking, measurement accuracy height and the good gauge inspection based on camera and controllable stroboscopic light source of accuracy
Survey distance-finding method and system.
In order to solve the above technical problems, technical scheme proposed by the present invention is：
A kind of gauge detection distance-finding method based on camera and controllable stroboscopic light source, comprises the following steps：
S1：Under train driving state, the flash frequency of regulation CMOS area array cameras frame frequency and controllable stroboscopic light source, making can
Control stroboscopic light source and carry out multiple stroboscopic in the same frame in of CMOS area array cameras exposure cycles, and the corresponding train of the frame duration is transported
It is dynamic to be considered as at the uniform velocity, and then make the measured object less than CMOS area array cameras visual field 1/3rd in the same two field picture of CMOS area array cameras
Middle cycle imaging；
S2：Based on measured object, brightness is larger under light source irradiation and the cycle is imaged, and Fourier is carried out to cycle image
Amplitude reaches maximum at conversion, the specific frequency in amplitude spectral curve, and the frequency corresponding to the maximum of points is stroboscopic week
In phase measured object repeat be imaged between pixel quantity, by cycle strobe measured object repeat be imaged between pixel quantity with
CMOS area array cameras length in pixels is multiplied, and produces the spacing d between same object multiimage in a two field picture；
S3：Calculated according to camera imaging principle and obtain measured object distance, calculation formula is as follows：
In formula, h is object distance, and unit is mm；F is the focal length of CMOS area array cameras, and unit is mm；V is train running speed,
Unit is m/s；F is controllable light source strobing frequency, and unit is fps；D be in a two field picture between same object multiimage between
Away from unit is mm.
Traditional distance-finding method (stereoscan distance-finding method such as based on principle of triangulation and based on TOF (during flight
Between) the machine vision distance-finding method of principle) requirement of the online gauge detection of railway high speed can not be met.Due to gauge detection speed
Comparatively fast, more using the stereoscan distance-finding method vertical and horizontal measuring point based on principle of triangulation, data volume is larger, liquid crystal
The switching speed of plate each unit can not meet requirement, cause to be difficult to monitor on-line at a high speed.And in the machine based on TOF principles
In visual token method, the data stability of CCD (charge coupled cell) line-scan digital camera and sensitivity are difficult to meet measuring accuracy
Requirement.The present invention is using CMOS (metal oxide semiconductor device) as imaging sensor, and CMOS has into compared with CCD
As process is fast, integrated level is high, small power consumption, fast reading speed the advantages of, the detection of railway high speed online gauge is met well and is surveyed
Away from the requirement that speed and integrated level etc. are read for imaging sensor.The present invention by adjust CMOS area array cameras frame frequency and can
The flash frequency of stroboscopic light source is controlled, controllable stroboscopic light source is carried out multiple frequency in the same frame in of CMOS area array cameras exposure cycles
Dodge, so make measured object less than CMOS area array cameras visual field 1/3rd in the same two field picture of CMOS area array cameras the cycle into
Picture, then Fourier transformation is carried out to cycle image, the pixel quantity between measured object repetition imaging in cycle strobe is obtained,
And then the spacing d in a two field picture between same object multiimage is drawn, and object distance is then calculated according to camera imaging principle, should
Method measuring speed is fast, measurement accuracy is high, accuracy is good, is suitable for railway high speed on-line checking, disclosure satisfy that 200km/h reality
The online gauge detection of car.And the shortcoming of tunnel and Metro Clearance can only be detected by overcoming conventional laser ranging, can be in open-wire line
Upper progress gauge detection；Irradiated, round-the-clock can detected using controllable stroboscopic light source, overcome traditional gauge detection by ambient light
Influence is big, the shortcoming of poor anti jamming capability.
Further, in step S3, train running speed V is measured by the following method to be obtained：
S301：Train line timing diagram picture is gathered by two linear array CCD cameras being set up in parallel along train length direction,
The visual angle of two linear array CCD cameras, which exists, to be overlapping, and successively passes twice through two linear array CCD cameras collection figure simultaneously
Picture, and same camera successively twice acquired image exist it is overlapping；
S302：The partly overlapping image gathered twice according to priority, to separate unit linear array CCD camera between twice successively
Circulation related operation is carried out every the timing diagram picture gathered in Δ t, the number of pixels of train movement in Δ t is obtained；
S303：Circulation phase is carried out based on overlapping visual angle in synchronization acquired image to two linear array CCD cameras
Computing is closed, obtained result produces the reality that each pixel of linear array CCD camera is represented except the distance between axles of two linear array CCD cameras
Distance；
S304：The number of pixels that train is moved in period Δ t is multiplied with the actual range that each pixel is represented to be produced
The distance that linear array CCD camera is moved in Δ t；
S305：Linear array CCD camera is moved in Δ t distance divided by time Δ t, produce the speed of service V of train.
Existing speed-measuring method mainly has laser velocimeter, GPS to test the speed, radar velocity measurement and locomotive etc..Laser velocimeter is deposited
The shortcoming that laser equipment need to remain static in equipment price costliness, test process；GPS, which tests the speed, has satellite-signal on mountain
The sections such as area, tunnel receive difficult, the low problem of rate accuracy；Radar velocity measurement exist radar angular installation and debugging requirement it is higher,
The problems such as test process error is larger；It is difficult and the data-interface of locomotive is more special, it is contemplated that the security of railway transportation
In data syn-chronization to gauge to be detected.The present invention gathers the timing diagram picture of train line by two linear array CCD cameras, uses
Circular correlation algorithm obtains the number of pixels that train is moved within certain period, then obtains with circular correlation algorithm each pixel and represent
Actual range, it is final to calculate the speed of service for obtaining train.The influence of train motion, topography and geomorphology etc. is the process eliminate,
Rate accuracy is high, meets the requirement of the online gauge detection of railway high speed.
Further, in step S302 and step S303, circulation is carried out to the timing diagram picture that linear array CCD camera is gathered related
Before computing, high-pass filtering processing first is carried out to image, the low frequency part in image is filtered out.When two image sequence identical portions
Point gray scale it is more gentle, and different part greyscale transformation it is more violent when, result of calculation will produce error.Therefore, right
Linear array images first carry out high-pass filtering before carrying out circulation related operation to image, low frequency part in image are filtered, to enter one
Step improves the degree of accuracy of result of calculation.
Further, in step S301, linear array CCD camera collection train line timing diagram as when active used shine
Mingguang City source is infrared light supply, and the infrared light supply is the line source that width is not less than 5mm, and the length of infrared light supply covers two lines
The summation of array CCD camera visual field scope.In order to not influence the operation safety of rail track, velocity-measuring system is used as from infrared light supply
Active illumination light source.
Further, in step S1, the flash frequency of the controllable stroboscopic light source is more than the CMOS area array cameras frame per second
Three times, controllable stroboscopic light source in CMOS area array cameras field of view wait cycle stroboscopic.Ensure to deposit in cmos image sensor
The equidistant measured object imaging more than 3.
Further, the CMOS area array cameras and the controllable stroboscopic light source are arranged on the end of train, controllable flash
Light source covers the full visual angle of CMOS area array cameras.
Further, the controllable stroboscopic light source is feux rouges, blue light or purple LED stroboscopic light source.Due to CMOS faces battle array phase
The spectral response of machine is that 700nm or so reaches peak value in wavelength, when wavelength is less than 700nm, under spectral responsivity gradually subtracts, wavelength
Spectral responsivity declines very fast during more than 700nm.From wavelength actively frequency is used as 640-750nm red-light LED stroboscopic light source
Flash source meets ranging requirement, and may influence railway signal in view of red, yellow, green, also can select wavelength close
Blue light or purple light as LED stroboscopic light sources color.
It is corresponding with the above method, perform being glistened based on camera and controllable frequency for the above method invention additionally discloses a kind of
The gauge detection range-measurement system in source.
Compared with prior art, the advantage of the invention is that：
(1) gauge of the invention detection distance-finding method measuring speed is fast, measurement accuracy is high, accuracy good, is suitable for railway
High speed on-line checking, and overcome conventional laser ranging and can only detect the shortcoming of tunnel and Metro Clearance, can be on open-wire line
Carry out gauge detection；Irradiated, round-the-clock can detected using controllable stroboscopic light source, overcome traditional gauge detection by environment shadow
Ring the big, shortcoming of poor anti jamming capability.
(2) present invention gathers the timing diagram picture of train line by two linear array CCD cameras, using circular correlation algorithm most
The speed of service for obtaining train is calculated eventually.The speed-measuring method eliminates the influence of train motion, topography and geomorphology etc., rate accuracy
Height, meets the requirement that the online gauge of railway high speed detects ranging.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is the accompanying drawing used required in technology description to be briefly described, it should be apparent that, drawings in the following description are the present invention
Some embodiments, for those of ordinary skill in the art, on the premise of not paying creative work, can also basis
These accompanying drawings obtain other accompanying drawings.
Fig. 1 is the grey scale curve of CMOS area array cameras target imaging rows.
Fig. 2 is the amplitude curve that Fourier transformation is done to gray level image in Fig. 1.
Fig. 3 is camera imaging principle schematic.
Fig. 4 is the structured flowchart of cmos image sensor.
Fig. 5 is two dimensional image related operation schematic diagram.
Fig. 6 is two CCD line-scan digital camera combination infrared imagery technique measuring principle schematic diagrames.
Fig. 7 is that CCD line-scan digital camera combination infrared imagery techniques test the speed flow chart.
Fig. 8 is the schematic diagram calculation that CCD line-scan digital cameras single pixel point characterizes actual range.
Fig. 9 is the row image cycle correlation computations flow chart of CCD line-scan digital cameras two.
Figure 10 is the first row line array CCD grey scale curve figure in circular correlation algorithm simulation analysis.
Figure 11 is the second row line array CCD grey scale curve figure in circular correlation algorithm simulation analysis.
Figure 12 is conjugated following for calculating for directly progress after two row image Fourier transformations in circular correlation algorithm simulation analysis
Ring correlated results figure.
Figure 13 is progress conjugation calculating after two row data Fourier transformation high-pass filterings in circular correlation algorithm simulation analysis
Circulation correlated results figure.
Embodiment
For the ease of understanding the present invention, more complete is made to the present invention below in conjunction with Figure of description and preferred embodiment
Face, meticulously describe, but protection scope of the present invention is not limited to embodiment in detail below.
Unless otherwise defined, the implication that all technical terms used hereinafter are generally understood that with those skilled in the art
It is identical.Technical term used herein is intended merely to describe the purpose of specific embodiment, is not intended to the limitation present invention
Protection domain.
Unless otherwise specified, various raw material, instrument and equipment used in the present invention etc. can be purchased by market
It can buy or can be prepared by existing method.
Embodiment
Gauge of the present invention based on camera and controllable stroboscopic light source detects a kind of embodiment of distance-finding method, gauge detection
Distance-finding method is primarily adapted for use in the high speed on-line checking of railway clearance, and object distance data are provided for gauge detection.The distance-finding method
Mainly include the following steps that：
First under train driving state, the flash frequency of regulation CMOS area array cameras frame per second and controllable stroboscopic light source makes
Controllable stroboscopic light source carries out multiple stroboscopic in the same frame in of CMOS area array cameras exposure cycles, and by the corresponding train of frame duration
Motion is considered as at the uniform velocity, and then makes the measured object less than CMOS area array cameras visual field 1/3rd in the same frame figure of CMOS area array cameras
The cycle is imaged as in.
Being then based on measured object, brightness is larger under light source irradiation and the cycle is imaged, and Fourier is carried out to cycle image
Amplitude reaches maximum at conversion, the specific frequency in amplitude spectral curve, and the frequency corresponding to the maximum of points is stroboscopic week
Measured object repeats the pixel quantity between being imaged in phase.That is, under the premise of exposure is not reaching to saturation, being known by image
Not, Fourier transformation is carried out to repeating imaged rows image, takes certain bandpass filtering, amplitude spectral curve is shown, due to gray scale
Imaging moiety gray value under light source irradiation is repeated in image larger, and with certain mechanical periodicity, it is special in amplitude spectral curve
Determine amplitude at frequency and reach maximum, the corresponding frequency of point is the pixel quantity that target repeats between being imaged in cycle strobe.
Fig. 1 is the grey scale curve of target imaging row, and Fig. 2 is the amplitude curve that Fourier transformation is done to gray level image in Fig. 1.By stroboscopic week
The pixel quantity that measured object is repeated between being imaged in phase is multiplied with the distance represented by each pixel of CMOS area array cameras, produces one
Spacing d in two field picture between same object multiimage.
Calculated further according to camera imaging principle and obtain measured object distance, calculation formula is as follows：
In the formula, h is object distance, and unit is mm；F is the focal length of CMOS area array cameras, and unit is mm；V is train operation speed
Degree, unit is m/s；F is controllable light source strobing frequency, and unit is fps；D is between same object multiimage in a two field picture
Spacing, unit is mm.
Camera imaging principle is as shown in figure 3, as seen from Figure 3, when train speed is identical, h1、h2The object at place in the picture into
The spacing d of picture1、d2Difference, by image recognition and processing, draws imaging spacing d, with reference to train speed V, light source flash of light frequency
Rate F and camera focus f, you can calculate object point to the distance of video camera.
The range-measurement system mainly includes imaging optical subsystem and high rate burst communication system.Wherein imaging optical subsystem
Mainly include high-power LED array, illumination frequency regulating system, brightness control system and camera lens and filtering apparatus, it is high-power
LED array be used for ensure it is round-the-clock detection and photosensitive sensor can receive abundance illumination；Illumination frequency is adjusted
System is according to the speed of service of train, and the flash frequency of LED/light source is adjusted in real time makes the frame frequency of itself and image mutually coordinate, it is ensured that every
In one two field picture, same object point repeats imaging at least more than three；Brightness control system is used for adjusting LED intensity of illumination,
Intensity of illumination is changed with the change of environment, such as in dark, illumination brightness is increased, so as to increase image objects
Contrast, is easy to image recognition；Camera lens and filtering apparatus are mainly used to increase the light-inletting quantity of photosensitive sensor while filtering influence
The light composition of imaging.High rate burst communication system includes high-speed area array cmos image sensor, camera frame frequency regulating system and figure
As capture card, wherein high-speed area array cmos image sensor is used for sensing light intensity, converts light energy into magnitude of voltage output；Camera frame
Frequency regulating system adjusts the frame frequency of camera according to train speed, to adapt to the collection requirement of two field picture；Image pick-up card is main
For storage image, and do further image procossing.
During high speed acquisition is carried out, resolution ratio, shutter form and output form to imaging sensor etc. have higher
It is required that.The serial CMOS digital image sensors of MT9P031 of 5,000,000 pixels of the present embodiment selection APTINA companies production.Remove
It can meet outside resolution ratio, shutter and way of output requirement, the imaging sensor also has remarkable low optical property；Can be with
Programming carries out the control in terms of gain, frame rate, frame sign, exposure, and frame per second can reach 53fps；During high speed acquisition, wink
Between expose, dark current is extremely low, can be neglected；It is two-wire serial and interface is relatively simple.The structure of cmos image sensor
Block diagram is as shown in Figure 4.
Because CMOS spectral response in wavelength is that 700nm or so reaches peak value, when wavelength is less than 700nm, spectral response
Degree is gradually reduced, and spectral responsivity declines very fast when wavelength is more than 700nm.In sunshine visible spectrum, purple light is averagely per nm
The minimum 0.943Wm of radiancy of wavelength-2·nm-1, secondly it is the feux rouges that wavelength is 640-750nm, average every nm wavelength
Irradiation level is 1.225Wm-2·nm-1, obtain both purple lights ratio range by calculating for 0.53-0.64, feux rouges both it
It is 0.78-0.82 than scope, it is clear that feux rouges is more suitable for LED/light source color, in the present embodiment, preferably with red LED frequency
Flash source is used as active light source.May influence railway signal in view of red, yellow, green, can also from blue light or
Purple light as LED light source colour.
It is the basis for ensureing accurate ranging to the accurate test of train speed, in the present embodiment, train running speed V is excellent
Choosing is adopted measurement with the following method and obtained：Using two CCD line-scan digital camera combination infrared light supplies, the timing diagram gathered to CCD camera
Calculated as carrying out circulation relevant treatment, finally give the speed of service of train.
The correlation computations of image include carrying out autocorrelation calculation and the cross-correlation calculation to two images to image.Auto-correlation
All it is the similarity degree for describing two signals with cross-correlation, its size represents intensity of variation of each signal on space-time,
The maximum position of correlation, the similarity degree of two signals is maximum.Assuming that the sequence x (n) and y (n) of two finite energies, mutually
Pass is calculated as：
M is the delay variable of time or displacement, and it is negative to be moved to the left, and is moved right as just, above formula represents x (n) sequences
Motionless, y (n) sequences are to left m times or displacement.
According to the relativity of time-shifting, above formula is equivalent to：
Above formula represents that y (n) sequences are motionless, and x (n) sequences are to right translation m times or displacement, therefore the cross-correlation that two formulas are obtained
Sequence is equal.
X (n)=y (n) under special circumstances, above-mentioned two formula has reformed into the auto-correlation of x (n) sequences, i.e.,
Or
In signal, one dimensional image and two dimensional image, x (n) and y (n) are typically limited sequence, and especially length is N
Sequence x (n) and y (n) when meeting condition n ＜ 0 or n >=N, x (n) and y (n) each element are zero, cross-correlation and auto-correlation
It is calculated as：
In formula, i=m during m ＞ 0, i=0 during k=0 and m ＜ 0, it is clear that auto-correlation rxx(n) maximum is obtained at m=0.
When obtaining maximum using correlation function, the maximum property of two sequence similarity degrees contains same characteristic features to two
Image sequence carry out correlation computations, obtained peak change be image displacement.In two dimensional image, it is
Raising arithmetic speed, chooses a range of pixel in image and constitutes a region, referred to as feature templates T (referring to Fig. 5) allows
T is translated in image whole region S, and the pixel of overlay area is designated as S (i, j) in feature templates T moving process, and (i, j) is fixed
Justice is the coordinate of the pixel in S (i, j) upper left corner.Feature templates T and overlay area S (i, j) similarity degree can use two sequences
The sum of squares of deviations of row is weighed, i.e.,
Correlation function is obtained after image binaryzation：
Find the distance of the difference between primary template position and R (i, j) peak, as characteristic area movement.
During two dimensional image correlation computations, characteristic area can be determined by image procossing, more convenient, but surveyed
In the case that examination speed is higher, it is desirable to which related operation efficiency is very high, current existing technology is extremely difficult to this requirement.This hair
Bright use linear CCD sensor continuous acquisition image, carries out circulation related operation by one dimensional image, can obtain two images
Displacement, finally calculates translational speed.
It is to carry out opening up for some cycles by the cycle of N to the data y (n) that finite sequence length is N to prolong to circulate correlation, then
Shift in one direction, every time a mobile position, due to periodically, being moved to the left a position and being equivalent to right translation N-
1 position.N number of value of (0, N-1) is finally taken just to obtain N number of sequential value after cyclic shift.I.e.
rxy(m)=∑ x (n) y ((n))N
Specific calculation process is as follows：
It is known
X (n)=[x1, x2…xn-1, xn] y (n)=[y1, y2…yn-1, yn]
Y (n) progress period expansions are obtained：
Y (n)=[... y1, y2…yn-1, yn, y1, y2…yn-1, yny1, y2…yn-1, yn…]
X (n)=[x1, x2…xn-1, xn]
Then
rxy(0)=x1y1+x2y2+…xn-1yn-1+xnyn
rxy(1)=x1y2+x2y3+…xn-1yn+xny1
rxy(n-2)=x1yn-1+x2yn+…xn-1yn-3+xnyn-3
rxy(n-1)=x1yn+x2y1+…xn-1yn-2+xnyn-1
It is related as with the form expression circulation of matrix
Calculate the row vector for obtaining a N point length, i.e. rxy(m) m=0,1 ..., (N-1).
In this paper scheme that tests the speed, two CCD line-scan digital cameras are arranged along vehicle lengthwise direction, with the accurate cloth of constant spacing
Put point-blank, camera perspective, which exists, to be overlapping (as shown in Figure 6).The frequency acquisition of line-scan digital camera is controlled, by same
Trigger signal is walked, CCD line-scan digital cameras collection circuit image, circulation correlation computations are calibrated simultaneously with reference to two camera optical axis spacing
Carve the sign distance of single pixel point；Circulation correlation computations are carried out to one camera time sequence image data, obtained in data collection cycle
The number of pixels of train movement, the number of pixels moved by sign distance and the train in collection period of single pixel point, which is calculated, to be arranged
The car speed of service.
The flow that tests the speed of CCD line-scan digital camera combination infrared imagery techniques is as shown in fig. 7, wherein g1(n, t) and g2(n, t) point
Not Wei the data that are gathered in t of camera 1 and camera 2, Δ t is the collection period of line-scan digital camera, PiCorrelation is circulated for ith
Pixel quantity, D is the distance of shaft centers of two line-scan digital cameras, and s is train displacement in the collection period of line-scan digital camera, and V transports for train
Scanning frequency degree.
As seen from Figure 7, the essence for the scheme that tests the speed is exactly one camera in t, the sequential image cycle of t+ Δs t collection
Correlation obtains the number of pixels of train movement in the Δ t times；Double camera is after the image cycle related operation that synchronization is gathered
Obtained result produces the actual range that each pixel of camera is represented except the distance between axles of two cameras；Mobile number of pixels with it is every
The product for the actual range that individual pixel is represented is the distance that camera is moved in Δ t, and circulates related operation to two cameras
Results averaged may finally calculate train running speed further to improve computational accuracy.
The computational methods for the actual range that each pixel is represented are specific as follows：
As shown in figure 8, the image apart from the object that projection centre H length is x in two linear array CCD cameras is respectively
A1B1With A2B2If μ is the length of each pixel, and total number of pixels of every CCD camera is N0, A1B1With A2B2Include N number of picture
Element, there is A1B1=A2B2=N μ, then the actual range k that single pixel is characterized0=x/N.Theorem is projected according to perspective, single pixel point is characterized
The ratio relation that actual range can be converted into camera focus f and object point to focal length H is calculated.However, in actual fortune
During dynamic, the train operation posture moment changes, and H will not also immobilize.To solve this problem, the present invention passes through side
Case is designed and geometrical relationship derives the actual range that the single pixel point not influenceed by train motion is characterized, and detailed process is such as
Under：
Obtained by projecting principle thoroughly
In formula
A1B1=N μ
Can be in the hope of
The distance that then each pixel is represented as
It is imaged jointly, it is necessary to meet in order that two linear array CCD cameras have
x≥0
I.e.
From circulation relative theory, double camera is resulting after the data that synchronization is gathered carry out circulation correlation
As a result P0Meet following relation
P0+ N=N0
To sum up
It can be seen that, any instant single pixel point characterize actual range be two line-scan digital camera axis space Ds (distance of shaft centers) with it is same
The related gained pixel P of image cycle of the camera of moment two collection0Ratio, D is the two camera axis hearts away from immobilizing, eliminate
The influence of train motion, improves rate accuracy.
What deserves to be explained is, the present invention not is directly to use Circular correlation method when testing the speed, because directly using circulation
Correlation there is problem in logic realization, it is necessary to much multipliers, and need multiple counting cycles.Can not only computing be influenceed to imitate
Rate, it is also possible to computational accuracy can be reduced.Therefore, the present invention is changed related original calculation principle is circulated, and is obtained more
Easily algorithm, to ensure test the speed efficiency and rate accuracy.
By the property of correlation function, the related maximum reflection of circulation be two signals time difference, i.e., ought wherein one
Individual signal moves the corresponding position of maximum, and the degree of correlation of two signals is maximum, similarity highest.But it is related in circulation
Calculate during linear array images, have a problem that, when the gray scale of two image sequence same sections is more gentle, and it is different
Part greyscale transformation is more violent, and result of calculation will produce error.Therefore, linear array images are carried out needing before circulation related operation
High-pass filtering is carried out, by low frequency part Partial filtration in image, the degree of accuracy of result of calculation is improved, it is assumed that
Make gn=y-n, then gnFourier transformation be
In addition
Obtained according to convolution theorem
From above formula, the related operation of two row data is equivalent to carry out two row data after Fourier transformation (FFT), its
The result of middle data line is multiplied by after the conjugation of another data line, then carries out inverse Fourier transform (IFFT)；This process is in logic
In realization, it is only necessary to can easier be realized using FFT cores, as a result of circular correlation algorithm, Fourier's change is being carried out
Line period expansion need not be entered when changing again.The row image cycle correlation computations flow of line array CCD two is as shown in Figure 9.
In order to verify the feasibility of circulation correlation, early stage is developed in speed measuring equipment, the two row data that line array CCD is gathered are entered
The related simulation analysis of circulation of having gone.4096 points of line array CCD is sampled as, data bit width is 8, and pixel size is 5 μm of 5 μ m.
In data analysis process, which part data are exported on request, 4mm gathered datas are separated by same direction, and carried out image
Processing is with calculating.
From calculating analysis, because two pixel elements are closely joined together, the centre distance between them is
It is 5 μm for the dimensions length of pixel elements, therefore two row data difference 4mm is equivalent to moving 800 pixels, that is, circulate
After correlation computations, circulation correlation curve should be maximum in the value at pixel 800.As shown in figures 10-13, Figure 12 is by Figure 10 and figure
11 curves are directly multiplied by the conjugation of another data line with the result of wherein data line, then carry out in Fu after carrying out Fourier transformation
The circulation correlation curve that leaf inverse transformation is directly obtained.As seen from Figure 12, peak point is occurred in that in 800 location of pixels, still
Due to the influence of low frequency background, the point is not circulation maximum correlation point.Figure 13 is to be filtered after Fourier transformation by high pass
Ripple eliminates the curve obtained after background influence, and maximum appears in 800 location of pixels, fully demonstrates circulation correlation in twin-line array
Practicality in CCD camera combination infrared light supply imaging technique velocity test scheme.
The preferred embodiments of the present invention are the foregoing is only, are not intended to limit the invention, for the skill of this area
For art personnel, the present invention can have various modifications and variations.Within the spirit and principles of the invention, that is made any repaiies
Change, equivalent substitution, improvement etc., should be included in the scope of the protection.
Claims (8)
1. a kind of gauge detection distance-finding method based on camera and controllable stroboscopic light source, comprises the following steps：
S1：Under train driving state, the flash frequency of regulation CMOS area array cameras frame frequency and controllable stroboscopic light source makes controllable frequency
Flash source carries out multiple stroboscopic in the same frame in of CMOS area array cameras exposure cycles, and the corresponding train motion of the frame duration is regarded
For at the uniform velocity, and then make the measured object less than CMOS area array cameras visual field 1/3rd all in the same two field picture of CMOS area array cameras
Phase is imaged；
S2：Based on measured object, brightness is larger under light source irradiation and the cycle is imaged, and Fourier transformation is carried out to cycle image,
Amplitude reaches maximum at specific frequency in amplitude spectral curve, and the frequency corresponding to the maximum of points is quilt in cycle strobe
The pixel quantity surveyed between thing repetition imaging, the pixel quantity between being imaged and CMOS faces gust are repeated by measured object in cycle strobe
Camera pixel length is multiplied, and produces the spacing d between same object repetition imaging in a two field picture；
S3：Calculated according to camera imaging principle and obtain measured object distance, calculation formula is as follows：
In formula, h is object distance, and unit is mm；F is the focal length of CMOS area array cameras, and unit is mm；V is train running speed, unit
For m/s；F is controllable light source strobing frequency, and unit is fps；D is the spacing between same object multiimage in a two field picture,
Unit is mm.
2. the gauge detection distance-finding method according to claim 1 based on camera and controllable stroboscopic light source, it is characterised in that：
In step S3, train running speed V is measured by the following method to be obtained：
S301：Train line timing diagram picture, two are gathered by two linear array CCD cameras being set up in parallel along train length direction
The visual angle of linear array CCD camera, which exists, to be overlapping, and successively passes twice through two linear array CCD cameras while gathering image, and
Same camera successively twice acquired image exist it is overlapping；
S302：According to image overlapping in priority twice acquired image, separate unit linear array CCD camera is spaced twice in priority
Period Δ t in the timing diagram picture of collection carry out circulation related operation, obtain the pixel that train in period Δ t is moved
Number；
S303：Circulation related operation is carried out in the timing diagram picture that synchronization is gathered to two linear array CCD cameras, through circulating phase
The result obtained after computing is closed except the distance between axles of two linear array CCD cameras, the reality that each pixel of linear array CCD camera is represented is produced
Distance；
S304：The number of pixels that train is moved in period Δ t is multiplied with the actual range that each pixel is represented to be produced then
Between in section Δ t linear array CCD camera movement distance；
S305：Linear array CCD camera is moved in period Δ t distance divided by time Δ t, produce the speed of service V of train.
3. the gauge detection distance-finding method according to claim 2 based on camera and controllable stroboscopic light source, it is characterised in that：
In step S302 and step S303, before the timing diagram picture progress circulation related operation gathered to linear array CCD camera, first to image
High-pass filtering processing is carried out, the low frequency part in image is filtered out.
4. the gauge detection distance-finding method according to claim 2 based on camera and controllable stroboscopic light source, it is characterised in that：
In step S301, linear array CCD camera collection train line timing diagram as when active illumination light source used be infrared light supply,
The infrared light supply is the line source that width is not less than 5mm, and the length of infrared light supply covers two linear array CCD camera field ranges
Summation.
5. the gauge detection distance-finding method according to claim 1 based on camera and controllable stroboscopic light source, it is characterised in that：
In step S1, the flash frequency of the controllable stroboscopic light source is more than three times of the CMOS area array cameras frame per second, controllable frequency flash of light
Cycle stroboscopic is waited in CMOS area array cameras field of view in source.
6. the gauge detection distance-finding method according to claim 1 based on camera and controllable stroboscopic light source, it is characterised in that：
The CMOS area array cameras and the controllable stroboscopic light source are arranged on the end of train, controllable stroboscopic light source covering CMOS faces battle array phase
The full visual angle of machine.
7. the detection distance-finding method of the gauge based on camera and controllable stroboscopic light source according to any one of claim 1-6,
It is characterized in that：The controllable stroboscopic light source is feux rouges, blue light or purple LED stroboscopic light source.
8. a kind of detection of the gauge based on camera and controllable stroboscopic light source for performing any methods described of the claims 1 to 7
Range-measurement system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710158503.5A CN106970024B (en) | 2017-03-16 | 2017-03-16 | Distance measuring method and system are detected based on the limit of camera and controllable stroboscopic light source |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710158503.5A CN106970024B (en) | 2017-03-16 | 2017-03-16 | Distance measuring method and system are detected based on the limit of camera and controllable stroboscopic light source |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106970024A true CN106970024A (en) | 2017-07-21 |
| CN106970024B CN106970024B (en) | 2019-10-15 |
Family
ID=59329020
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710158503.5A Active CN106970024B (en) | 2017-03-16 | 2017-03-16 | Distance measuring method and system are detected based on the limit of camera and controllable stroboscopic light source |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106970024B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109640145A (en) * | 2018-12-24 | 2019-04-16 | 郑州畅想高科股份有限公司 | A kind of driving drilling method and device for track train |
| CN110030988A (en) * | 2019-03-01 | 2019-07-19 | 中国航空工业集团公司北京长城航空测控技术研究所 | A kind of multi-beacon high-speed synchronous recognition methods for high dynamic pose measurement |
| WO2021008052A1 (en) * | 2019-07-17 | 2021-01-21 | 南昌欧菲生物识别技术有限公司 | Lens accuracy calibration method, apparatus and device for 3d photographic module |
| CN112492227A (en) * | 2020-12-15 | 2021-03-12 | 维沃移动通信有限公司 | Exposure method, camera module and electronic equipment |
| CN113762122A (en) * | 2021-08-31 | 2021-12-07 | 中船重工鹏力（南京）大气海洋信息系统有限公司 | Raindrop detection algorithm based on stroboscopic photograph |
| CN114235837A (en) * | 2021-12-20 | 2022-03-25 | 重庆大学 | LED packaging surface defect detection method, device, medium and equipment based on machine vision |
| CN114513589A (en) * | 2020-10-06 | 2022-05-17 | 联发科技股份有限公司 | Image acquisition method and related image acquisition system |
| CN114979427A (en) * | 2022-03-30 | 2022-08-30 | 天津大学 | High-speed synchronous stroboscopic rotor blade tip image acquisition device and method |
| CN116723599A (en) * | 2023-08-09 | 2023-09-08 | 中国计量科学研究院 | Construction method and device of stroboscopic standard light source capable of predicting stroboscopic characteristics |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103139483A (en) * | 2011-11-30 | 2013-06-05 | 宝山钢铁股份有限公司 | Light source strobe control method in time delayed and integration (TDI) charge couple device (CCD) uniform imaging under the condition of non-uniform motion |
| CN104048970A (en) * | 2014-06-19 | 2014-09-17 | 樊晓东 | High-speed detection system and method of tunnel defects |
-
2017
- 2017-03-16 CN CN201710158503.5A patent/CN106970024B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103139483A (en) * | 2011-11-30 | 2013-06-05 | 宝山钢铁股份有限公司 | Light source strobe control method in time delayed and integration (TDI) charge couple device (CCD) uniform imaging under the condition of non-uniform motion |
| CN104048970A (en) * | 2014-06-19 | 2014-09-17 | 樊晓东 | High-speed detection system and method of tunnel defects |
Non-Patent Citations (4)
| Title |
|---|
| KRONAUER R E 等: "Reorganization and diversification of signal in vision", 《IEEE TRANS ON SMC》 * |
| 司徒国强 等: "隧道限界检测中车体振动补偿问题的研究", 《自动化仪表》 * |
| 李鹏 等: "铁路限界检测可控频闪光源测距方法", 《铁道科学与工程学报》 * |
| 王浩: "铁路隧道限界检测仪的研究和设计", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109640145A (en) * | 2018-12-24 | 2019-04-16 | 郑州畅想高科股份有限公司 | A kind of driving drilling method and device for track train |
| CN109640145B (en) * | 2018-12-24 | 2021-08-06 | 郑州畅想高科股份有限公司 | Driving practicing method and device for rail train |
| CN110030988A (en) * | 2019-03-01 | 2019-07-19 | 中国航空工业集团公司北京长城航空测控技术研究所 | A kind of multi-beacon high-speed synchronous recognition methods for high dynamic pose measurement |
| WO2021008052A1 (en) * | 2019-07-17 | 2021-01-21 | 南昌欧菲生物识别技术有限公司 | Lens accuracy calibration method, apparatus and device for 3d photographic module |
| CN114513589A (en) * | 2020-10-06 | 2022-05-17 | 联发科技股份有限公司 | Image acquisition method and related image acquisition system |
| CN112492227B (en) * | 2020-12-15 | 2022-02-08 | 维沃移动通信有限公司 | Exposure method, camera module and electronic equipment |
| CN112492227A (en) * | 2020-12-15 | 2021-03-12 | 维沃移动通信有限公司 | Exposure method, camera module and electronic equipment |
| CN113762122A (en) * | 2021-08-31 | 2021-12-07 | 中船重工鹏力（南京）大气海洋信息系统有限公司 | Raindrop detection algorithm based on stroboscopic photograph |
| CN113762122B (en) * | 2021-08-31 | 2023-10-13 | 中船鹏力(南京)大气海洋信息系统有限公司 | Raindrop detection algorithm based on stroboscopic photo |
| CN114235837A (en) * | 2021-12-20 | 2022-03-25 | 重庆大学 | LED packaging surface defect detection method, device, medium and equipment based on machine vision |
| CN114979427A (en) * | 2022-03-30 | 2022-08-30 | 天津大学 | High-speed synchronous stroboscopic rotor blade tip image acquisition device and method |
| CN116723599A (en) * | 2023-08-09 | 2023-09-08 | 中国计量科学研究院 | Construction method and device of stroboscopic standard light source capable of predicting stroboscopic characteristics |
| CN116723599B (en) * | 2023-08-09 | 2023-10-27 | 中国计量科学研究院 | Construction method and device of stroboscopic standard light source capable of predicting stroboscopic characteristics |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106970024B (en) | 2019-10-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106970024B (en) | Distance measuring method and system are detected based on the limit of camera and controllable stroboscopic light source | |
| Huang et al. | Limitation and improvement of PIV: Part I: limitation of conventional techniques due to deformation of particle image patterns | |
| US10302424B2 (en) | Motion contrast depth scanning | |
| Maas et al. | Particle tracking velocimetry in three-dimensional flows: Part 1. Photogrammetric determination of particle coordinates | |
| US5905568A (en) | Stereo imaging velocimetry | |
| CN104574393B (en) | A kind of three-dimensional pavement crack pattern picture generates system and method | |
| CN109791696A (en) | It is positioned and is mapped simultaneously using event video camera | |
| US20120327287A1 (en) | Method and system for producing image frames using quantum properties | |
| CN106772430A (en) | The single pixel photon counting 3-D imaging system and method approached based on multiresolution wavelet | |
| CN107462741B (en) | Moving object speed and acceleration measuring device | |
| CN103472256A (en) | Flow two-dimensional velocity profile measuring method and device based on planar array CCD spatial filter | |
| CN101218602B (en) | Image processor and environment information observing device | |
| CN103900494A (en) | Homologous point rapid matching method used for binocular vision three-dimensional measurement | |
| CN108132058A (en) | Digital Photogrammetric System on-line displacement measurement calibrates for error device and method | |
| CN108957024A (en) | A kind of method, apparatus and electronic equipment of tachometric survey | |
| CN102914261B (en) | Non-contact thermal target size measurement system and method | |
| CN105222901A (en) | Based on the real-time quantitative phase detection device of light intensity transmission equation | |
| WO2006006250A1 (en) | Fluid flow measurement system, fluid flow measurement method, and computer program | |
| CN113466489A (en) | Single-camera particle image velocimetry method with low particle density | |
| CN115901178B (en) | System and method for measuring wave resonance flow field characteristics among multi-body marine structures | |
| Mendelsohn | Automated pavement crack detection: An assessment of leading technologies | |
| CN115824170A (en) | Method for measuring ocean waves by combining photogrammetry and laser radar | |
| Loktev et al. | Image Blur Simulation for the Estimation of the Behavior of Real Objects by Monitoring Systems. | |
| CN105890575A (en) | Method and device for visually identifying artificial intelligent machine under common and special environments | |
| Merzkirch | Particle image velocimetry |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |