CN100462678C - Steel rail near laser visual dynamic measuring device and method - Google Patents
Steel rail near laser visual dynamic measuring device and method Download PDFInfo
- Publication number
- CN100462678C CN100462678C CNB2005101237250A CN200510123725A CN100462678C CN 100462678 C CN100462678 C CN 100462678C CN B2005101237250 A CNB2005101237250 A CN B2005101237250A CN 200510123725 A CN200510123725 A CN 200510123725A CN 100462678 C CN100462678 C CN 100462678C
- Authority
- CN
- China
- Prior art keywords
- rail
- measurement
- profile
- coordinate
- laser
- 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.)
- Expired - Fee Related
Links
Images
Abstract
The disclosed measuring device is composed of laser vision sensor, computer, image acquisition card in high speed, calibrating target drone and measuring software. The method includes steps: after imaging system for rail section images light plane projected from laser projector, the image acquisition card captures rail section images and sands them to computer. Processing the collected images, the computer picks up image coordinates of sectional feature contour, as well as calculates out vertical abrasion, side abrasion and total abrasion based on measured model. The invention is capable of measuring sectional contour of rail in single side, vertical abrasion, side abrasion and parameter of total abrasion through single laser vision sensor.
Description
Technical field
The invention belongs to field of measuring technique, relate to a kind of noncontact dynamic measurement method of steel rail wearing integrative parameter.
Background technology
The situation and the security of railway are closely related, and be particularly more important for bullet train.The regular detection of rail is very important for rationalization scheme and the low maintenance cost of maintenance.Help to formulate the better maintaining timetable at rail abrasion with being out of shape to measure in early days, remake emergency treatment when observing unsafe condition to avoid waiting, on the other hand, safeguard in advance to make limited manpower and instrument and equipment resource obtain better utilization, thereby reduced cost.In addition, correct maintenance has reduced because of the bad noise pollution that is produced that contacts between wheel and the rail.The greatly developing of China railways speed-raising in recent years and heavy haul transport, passenger-cargo freight volume and rate of traffic flow increase substantially, and cause rail wear serious day by day.How efficiently and accurately to measure the rail wear value is an important subject of domestic and international railroad always.
In order to guarantee railway operation safety, require under the condition that does not influence train running, will at least once overhaul every month to rail.During the State's Ninth Five-Year Plan period, China railways rail mileage reaches more than 68000 kilometer, does not have suitable measurement means so far.For a long time, domestic measurement to rail wear all is to adopt hand dipping, detects its section loss with special-purpose slide calliper rule artificial sampling.This mode inefficiency expends a large amount of manpower and materials, and is in the highly dangerous.Owing in measurement, introduced gauger's human factor inevitably, directly influenced precision and the reliability measured.Hand dipping can not adapt to the development of China railways cause.Therefore, a kind of self-operated measuring unit of rail wear fast and accurately of development has very big realistic meaning.
Domestic and international many research institutions and scholar study the detection method of rail wear, and have successfully developed various measurement mechanisms, according to the difference of detection mode, roughly can be divided into contact and non-contact detection.The special-purpose gauge measurement of rail wear based of the MiniProf device of U.S. GreenWood company development adopts contact mode point-to-point measurement rail profile, compares with nominal contour then.Wroclaw University of Science and Technology has developed similar contact high precision rail profile measurement of wear system.At home, railway Weights and Measures Bureau in Harbin has developed a kind of similar external slide calliper rule formula measurement of rail wear based device.Coutact type steel rail measurement of wear The measuring precision height, but only be fit to static measurement, complicated operation, efficiency of measurement is low, uses when being mainly used in laboratory study rail anti-wear performance, is not suitable for on-line measurement.
North America and European developed country just utilize since phase early 1980s that computing machine carries out automatically, the accurately research of measurement of rail wear based.The eighties Australia takes the lead in having developed LitesLice solid state sensor rail profile profile automatic measurement system.Come into operation in Australia, the U.S., Canada and Britain etc. in succession after 1984, and constantly perfect, become the most general rail profile abrasion automatic measurement system of use in the world gradually.The defective of this method is that adopting image coboundary and left hand edge is positioning reference, therefore requires survey sensor can not have bigger vibration to exist, thereby relatively is suitable for static measurement.In fact, the survey sensor that is installed on the locomotive certainly exists bigger vibration, has influence on the precision and the reliability of measurement result thus.In addition, the both sides of adopting two cover survey sensors to measure same rail are respectively arranged abroad, a side measurement result that adopts not abrasion is as benchmark, though this method solves the influence of vibration to measurement result, but the cost height, and need two unifications of overlapping the survey sensor coordinate systems, complicated operation.At present, the Rail Abrasion Detection System system of domestic research is basically with similar abroad, and because measurement coordinate system arrives the not well solution of conversion links of standard (design) coordinate system, therefore the device of practical study only is used for static experimental study, and the automatic measurement that does not have to solve rail wear under the dynamic condition.
Summary of the invention
Technical matters to be solved by this invention is: provide a kind of precision higher, measuring speed is fast, and single-sensor can be measured the online dynamic real-time measurement method and system device of comprehensive parameters such as the vertically abrasion of one-sided rail, side wear and total abrasion simultaneously.To reduce the volume of measuring equipment, reduce the measuring equipment cost, improve efficiency of measurement, improve operability and convenience that its through engineering approaches is used.
Technical solution of the present invention is: the steel rail near laser visual dynamic method, use the steel rail near laser visual dynamic device to carry out the kinetic measurement of comprehensive parameters, the laser vision sensor 7 that said steel rail near laser visual dynamic device is made up of laser projecting apparatus 3 and rail profile imaging system 6, computing machine, high speed image acquisition board, demarcation target and Survey Software are formed; Laser projecting apparatus 3 is made up of semiconductor laser 1 and optical plane generating optics system 2; Rail profile imaging system 6 is made up of ccd video camera 4 and filter system 5, and the optical axis confluent angle of the optical axis of rail profile imaging system 6 and laser projecting apparatus 3 is 45 °～60 °; Laser vision sensor 7 is fixed on the car body of train bottom, with the distance of tested rail 10 be 300～700mm; The optical plane 8 of laser projecting apparatus 3 projections is perpendicular to tested rail 10 and parallel with rail profile; The imaging system of rail profile imaging system 6 can observe projection light plane 8 and intersect rail head and the web of the rail feature contour 9 that forms with tested rail 10 surfaces; Said demarcation target is a two dimensional surface, the unique point that sets in advance is arranged on the target plane, the black square that arranged is arranged on the target plane, the quantity of square is 4～100, the length of side of square is 3～50mm, and its precision is 0.01～0.05mm, and the square spacing is 3～50mm, its precision is 0.01～0.05mm, and the summit of choosing each square is a unique point; It is characterized in that the concrete steps of measurement are:
1, sets up the camera model and the measurement model of laser vision sensor 7, calibrating camera parameters and sensor construction parameter;
1.1, on optical plane, set up two-dimensional measurement coordinate system o
m-x
my
m, initial point o
mBe arbitrfary point on the optical plane, x
mAxle and y
mAxle is mutually vertical and be positioned at optical plane;
1.2, the structural parameters of the inner parameter of calibrating camera and laser vision sensor and be saved in computing machine;
1.3, set up standard rail two-dimensional design coordinate system o
n-x
ny
n, be y with the rail profile axis of symmetry
nAxle, the flange of rail is x
nAxle, the intersection point of two of web of the rail part great circle and roundlet centers of circle and circle center line connecting and roundlet is the alignment characteristics point on the selection standard rail, calculates the design coordinate of alignment characteristics point respectively, and is saved in computing machine, prepares against measuring phases and calls;
2, measure rail wear;
2.1, mileometer is connected on the computing machine by serial ports;
2.2, the established standards profile with measure the profile alignment parameter, set rail profile minimum length threshold T
L, it is shown that by the contour feature point numerical table span is 40～100; Rail head profile and web of the rail profile spacing minimum threshold T
D, it is represented that by the pixel coordinate of contour images span is 15～30 pixels, representative value 20 pixels; Preserve setup parameter, call in order to follow-up measurement;
2.3, process of measurement monitors the serial data state, when receiving startup command that serial ports sends and mileage, preserves mileage and also controls video camera and take the image that a width of cloth comprises the rail profile feature contour, is called measurement image;
2.4, extract the image coordinate of rail profile feature contour in the measurement image, do not satisfy length greater than T if extract two
LFeature contour, perhaps the spacing of two feature contours is less than T
D, then turn back to step 2.3, again according to com-state, the rail measurement image is gathered in control; If find two correct rail feature contours,, utilize the measurement coordinate of the image coordinate calculating rail profile feature contour of feature contour then according to the measurement model of laser vision sensor 7;
2.5, according to the measurement coordinate of rail profile feature contour, will measure profile and be divided into rail head, web of the rail great circle and web of the rail roundlet three parts automatically.By the feature contour of web of the rail great circle of cutting apart and the little circular portion of the web of the rail, adopt radius constrained procedure match web of the rail great circle and web of the rail roundlet respectively, little radius of circle is 20mm, big radius of circle is got 400mm, be called the 60kg alignment pattern, big radius of circle is got 350mm, is called the 50kg alignment pattern;
2.6, adopt 60kg alignment pattern and 50kg alignment pattern respectively, calculate the measurement coordinate of alignment characteristics point, by wherein the measurement coordinate and the design coordinate of at least 2 alignment characteristics points, calculate measurement coordinate system and arrive the two-dimensional transform that designs coordinate system;
2.7, according to the conversion of the measurement coordinate system under two kinds of patterns, obtained to the design coordinate system, respectively with the measurement coordinate conversion of rail profile rail head part in the design coordinate system, direct and standard rail is compared, and calculates vertical abrasion, side wear and total abrasion of rail;
2.8, two kinds of alignment patterns, 3 wearing valves down relatively, choose smaller value as last measurement result, and obtain the rail type of tested rail according to the alignment pattern that is adopted, the mileage number of measurement result, rail type and correspondence is saved in the measurement result file;
2.9, send to measure to serial ports and to finish signal;
2.10, repeated for 2.3～2.9 steps, measure next time.
Advantage of the present invention is:
The first, first realized that single laser vision sensor measures one-sided rail profile and vertical abrasion, side wear simultaneously and always wear away multiparameter, reduced equipment cost;
The second, automatically cut apart rail head, web of the rail great circle and the web of the rail roundlet of measuring the rail profile profile, realized under the dynamic condition in-site measurement rail wear;
Three, with the rail profile web of the rail great circle and the roundlet center of circle and derive and be characterized as the alignment characteristics point, obtain measuring basis automatically, realized that rail measures profile and aim at the automatic of design profile, avoided in the train running process vibration to the influence of measurement;
Four, realized that tested rail model is discerned automatically and the mileage record of abrasion actual detected point;
Five, the measuring system apparatus structure is simple, cost is low, automaticity is high, and measuring accuracy height, speed are fast, and measuring process is simple.
Six, the steel rail near laser visual measuring system obtains the through engineering approaches application.
Description of drawings
Fig. 1 is a steel rail near laser visual dynamic system basic principle schematic.Among Fig. 1, the 1st, semiconductor laser, the 2nd, the optical plane generating optics system, the 3rd, by 1 and 2 laser projecting apparatuss of forming, the 4th, ccd video camera, the 5th, filter system, the 6th, by the 4 and 5 rail profile imaging systems of forming, the 7th, by 3 and 6 laser vision sensors that constitute, the 8th, optical plane, the 10th, tested rail, 9 is 8 and 10 crossing rail profile feature contours 9 that form.
Fig. 2 is the measurement mathematical model of laser vision sensor.The 4th, ccd video camera, the 3rd, laser projecting apparatus, the 7th, by 4 and 3 laser vision sensors of forming, the 204th, the plane of delineation, the 8th, optical plane, the 206th, testee.
Fig. 3 is the foundation of rail design two-dimensional coordinate system and the selection synoptic diagram of alignment characteristics point and measurement of wear point.
Fig. 4 be project the feature striation on rail surface and collection comprise rail profile feature contour image.
Fig. 5 is the transformation calculations that measurement coordinate system arrives the design coordinate system.
Fig. 6 is " steel rail near laser visual dynamic detection system " software interface.
Fig. 7 is that the right rail of 50kg is measured contour images and extracted feature contour.
Fig. 8 is the coordinate of the right rail profile of 50kg in measurement coordinate system and design coordinate system.
Fig. 9 measures profile to aim at back and abrasion result of calculation with design profile.
Figure 10 is continuous 540 the measuring error curve maps of same rail wear.
Figure 11 is the Survey Software process flow diagram.
Embodiment
Below the present invention is described in further details.The present invention combines modern laser and computer vision technique, uses single laser vision sensor, under dynamic condition, has realized one-sided rail profile profile, vertical abrasion, side wear and always worn away the on-the-spot of multiparameter measuring automatically.
Steel rail near laser visual dynamic apparatus system ultimate principle of the present invention as shown in Figure 1, laser vision sensor 7, computing machine, high speed image acquisition board and Survey Software etc. that it is made up of laser projecting apparatus 3 and rail profile imaging system 6 are formed.Laser projecting apparatus 3 is made up of semiconductor laser 1 and optical plane generating optics system 2; Rail profile imaging system 6 is made up of ccd video camera 4 and filter system 5; The optical axis confluent angle of the optical axis of rail profile imaging system 6 and laser projecting apparatus 3 is 45 °～60 °; Laser vision sensor 7 is fixed on the car body of train bottom, with the distance of tested rail 10 be 300～700mm.The device of fixed laser vision sensor 7 is existing fixed device or similar simple fixation mechanism.
Adjust laser projecting apparatus 3, make projection light plane 8 perpendicular to tested rail 10 and parallel with rail profile; Adjust rail profile imaging system 6, guarantee that imaging system can observe projection light plane 8 and intersect rail head and the web of the rail feature contour 9 that forms with tested rail 10 surfaces; Tighten up laser projecting apparatus 3 and imaging system 6 respectively.The adjusting mechanism of laser projecting apparatus 3 and rail profile imaging system 6 can adopt general various two-dimentional machinery adjustment and tighten up mechanism.
Said demarcation target is a two dimensional surface, and the unique point that sets in advance is arranged on the target plane.The black square that arranged is arranged on the target plane, the quantity of square are 4～100, and the length of side of square is 3～50mm, its precision is 0.01～0.05mm, the square spacing is 3～50mm, and its precision is 0.01～0.05mm, and the summit of choosing each square is a unique point.
The measurement mechanism same form two covers of the present invention are installed in train bottom, the abrasion of measuring two rail of left and right sides respectively.
The principle of work of steel rail near laser visual detection of dynamic is: project optical plane 8 by high-intensity laser projecting apparatus 3, the inboard that optical plane 8 is aimed at tested rail 10 is shone and perpendicular to rail.Optical plane 8 intersects with rail 10, form rail cross sectional feature profile 9, after 6 imagings of rail profile imaging system, collect calculator memory by high speed image acquisition board, image is carried out the image coordinate that Treatment Analysis obtains the rail profile feature contour, according to the mathematical model of survey sensor, try to achieve the two-dimensional coordinate of real profile in measurement coordinate system of rail profile.The measurement coordinate of rail profile is aimed at automatically with the design coordinate,, calculated vertically abrasion, side abrasion and total abrasion according to the regulation of measurement of wear.
Referring to Figure 11, the major function of " steel rail near laser visual dynamic detection system " software comprises: the real-time collection of rail profile feature contour image; Automatic " sub-pix " precision of feature contour image coordinate is extracted; The camera lens distortion correction; The rail surface profile measurement; The parameter calibration of measuring system; Rail type (left and right sides rail and specification 50kg and 60kg) is identification automatically; Measuring basis is set up automatically, rail surface measurement profile is aimed at the automatic of nominal contour; The automatic calculating of vertical abrasion and side wear; Measure abrasion and functions such as mileage is corresponding.
The laser vision sensor mathematical model.
The mathematical model of the laser vision sensor of being made up of ccd video camera 4 and laser projecting apparatus 37 as shown in Figure 2.The video camera three-dimensional coordinate is o
c-x
cy
cz
c, the image coordinate that image is sat plane 204 is o
u-x
uy
uOn optical plane 8 with 1 o
mBe initial point, set up three-dimensional reference frame o
m-x
my
mz
m, o wherein
m-x
my
mBe the two-dimensional measurement coordinate system, the equation of optical plane 8 under reference frame is:
z
m＝0 [1]
If the three-dimensional reference coordinate of any 1 P is (x on the optical plane
m, y
m, z
m), the respective two-dimensional measurement coordinate system is (x
m, y
m), the perspective projection point of some P on the plane of delineation 204 is p, the image coordinate of p is (x
u, y
u), the normalized image coordinate is (x
n, y
n), then camera model can show be:
ρ wherein, λ is a scale factor, A is the inner parameter matrix of video camera 302, f
x, f
yBe the effective focal length on x, the y direction, (u
0, v
0) be the principal point coordinate.
Be 3 * 3 quadrature rotation matrixs,
Be 3 * 1 translation vectors.
With
Represented reference frame o
m-x
my
mz
mTo camera coordinate system O
c-x
cy
cz
cConversion.
Can be got by formula [1]～[3], the normalized image coordinate with the measurement transformation of coordinates is:
Formula [4] shows that the relation between the optical plane and the plane of delineation can pass through one 3 * 3 matrix representation.Formula [1]～[4] have constituted the mathematical model of the laser vision sensor on the theorem in Euclid space.Can calculate the normalized image coordinate according to formula [3] by the image coordinate of putting on the optical plane,, calculate the measurement coordinate of its correspondence then according to formula [4].
The scaling method of laser vision sensor model parameter is referring to the patent of invention " a kind of structured light vision sensor scaling method based on plane target drone " of Zhou Fuqiang etc., application number: 03142658.1, and open day: on January 19th, 2005.
According to the measurement model and the measuring principle of the present invention of laser vision sensor 7, steel rail near laser visual dynamic is divided into the Measuring System Models parameter calibration and measures two stages according to model.
The demarcation concrete steps of Measuring System Models parameter are as follows:
1, setting laser vision sensor.Laser vision sensor 7 is to be made up of rail profile imaging system 6 and laser projecting apparatus 3.After the optical axis of the optical axis of rail profile imaging system 6 and laser projecting apparatus 3 set by at 45～60 a ° angle, its relative position remained unchanged after tightening up.Laser projecting apparatus 3 is made of semiconductor laser 1 and optical plane generating optics system 2, and the optical maser wavelength of semiconductor laser is infrared laser or visible laser, and optical plane generating optics system 2 is made of collimated light path and cylindrical mirror.Laser projecting apparatus can adopt commercial product, a word laser instrument of producing as Xi etc.The rail profile imaging system is made up of ccd video camera 4 and filter system 5, and the peak wavelength of filter system is identical with optical maser wavelength.Laser vision sensor 7 is fixed on the car body of train bottom, with the distance of tested rail 10 be 300～700mm.
2, adjust laser projecting apparatus 3, make projection light plane 8 perpendicular to tested rail 10 and parallel with rail profile; Adjust rail profile imaging system 6, guarantee that imaging system can observe projection light plane 8 and intersect rail head and the web of the rail feature contour 9 that forms with tested rail 10 surfaces; Tighten up laser projecting apparatus 3 and imaging system 6 respectively.
3, said demarcation target is a two dimensional surface, and the unique point that sets in advance is arranged on the target plane.The black square that arranged is arranged on the target plane, the quantity of square are 4～100, and the length of side of square is 3～50mm, its precision is 0.01～0.05mm, the square spacing is 3～50mm, and its precision is 0.01～0.05mm, and the summit of choosing each square is a unique point.
4, on optical plane, set up two-dimensional measurement coordinate system o
m-x
my
m, initial point o
mBe arbitrfary point on the optical plane, x
mAxle and y
mAxle is mutually vertical and be positioned at optical plane;
5, open the power supply of sensor, start " steel rail near laser visual dynamic detection system " software, the initialisation image capture card.
6, the camera parameters of Calibration of Laser vision sensor and sensor construction parameter, scaling method and step are referring to the patent of invention " a kind of structured light vision sensor scaling method based on plane target drone " of Zhou Fuqiang etc., application number: 03142658.1, open day: on January 19th, 2005.Calibrating parameters is saved in computing machine, calls in order to measuring phases.
7, as shown in Figure 3, set up standard rail two-dimensional design coordinate system o
nx
ny
n, be y with the rail axis of symmetry
nAxle, the flange of rail is x
nAxle.Web of the rail part great circle on the AB section contour representation standard rail (radius of 50kg rail is 350mm, and the radius of 60kg rail is 400mm) is called web of the rail great circle contour segment.BC section contour representation web of the rail part roundlet (radius is 20mm) is called web of the rail roundlet contour segment.EF section profile is the rail head part, is called the rail head contour segment.The G point is the side wear measurement point, and the H point is vertical measurement of wear point.According to the regulation of " railway track maintenance rule ", the G point is positioned at standard section apart from rail surface 16mm place down, and the H point is positioned at wide 1/3 place of rail surface (apart from the standard action edge).Choose 2 centers of circle of web of the rail great circle and web of the rail roundlet and (use D
1And D
2Represent) and circle center line connecting D
1D
2(use D with the intersection point of roundlet
3And D
4) be the alignment characteristics point, calculate the design coordinate of the alignment characteristics point of 50kg and 60kg rail respectively, and be saved in the computing machine, call in order to measuring phases.
The parameter of measuring system only need be demarcated once, demarcated measuring system after, just can carry out the rail wear detection of dynamic.The concrete measuring process of rail wear is as follows:
8, mileometer is fixed on the train appropriate location, and is connected on the computing machine by serial ports.
9, under train movement or static two kinds of situations, can start " abrasion detection of dynamic " function of Survey Software.
10, set the rail design profile and measure the profile alignment parameter.Set rail profile minimum length (showing) threshold value T by the contour feature point numerical table
L, span is 40～100, representative value is about 50.Rail head profile and web of the rail profile spacing (pixel coordinate by contour images is represented) minimum threshold T
D, span is 15～30 pixels, representative value 20 pixels.The minimum length of setting web of the rail great circle is T
BCL(shown that by the contour feature point numerical table it is little to guarantee to handle the length that obtains than real image, representative value is 70), the maximum length of web of the rail roundlet is T
SCL(showing that by the contour feature point numerical table representative value is 30).Above alignment parameter can obtain by Flame Image Process according to the rail image of actual photographed, and during setting, to obtain data little by about 5 than actual.Set roundlet key point distance parameter T
14, T
14=‖ D
1D
4‖-r
BC, r
BCBe the big radius of circle of the web of the rail.For 50kg rail, T
14=3.51mm is for 60kg rail, T
14=3.07mm.Preserve setup parameter to computing machine, call in order to follow-up measurement.
11, process of measurement monitors the serial data state, and when startup command that receives the serial ports transmission and mileage, process of measurement is preserved mileage and controlled video camera simultaneously and take the image that a width of cloth comprises the rail profile feature contour, is called measurement image.
12, extract the image coordinate of rail profile feature contour in the measurement image,, utilize the two-dimensional coordinate of image coordinate calculated characteristics profile under measurement coordinate system of feature contour, be called the measurement coordinate of feature contour according to the sensor measurement model.The image coordinate extraction algorithm of feature contour is referring to the article " a kind of symmetrical curve structure detector " [An UnbiasedDetector of Curvilinear Structures] of Carsten Steger, IEEE periodical " pattern analysis and machine intelligence ", 20 (2), the 113rd～125 page, 1998 years.[IEEE?Transactions?on?PatternAnalysis?and?Machine?Intelligence，20(2)，February1998]。As shown in Figure 4, laser projecting apparatus projection light plane has two at the feature contour that the rail surface forms, and if the AC section profile and the EF section profile of difference corresponding diagram 4 are the length T of AC and EF profile
L, perhaps the distance between the image coordinate that the A point is corresponding with the C point is less than T
D, then show no rail contour feature in the image, return step 11, again according to com-state, the rail measurement image is gathered in control.Normal condition obtains two contour features that satisfy preseting length threshold value and distance threshold, respectively the image coordinate of two profiles is pressed row-coordinate ordering, and the row-coordinate of profile starting point is less is called the rail head profile, and its length is N
Head, another is called web of the rail profile, and length is N
WaistCompare the starting point of rail head profile and the row coordinate of rearmost point, if begin the point range coordinate greater than rearmost point, then rail is right rail, otherwise is left rail.
13,,, calculate the two-dimensional measurement coordinate of web of the rail profile according to the measurement model of the sensor of formula [3] and [4] expression by the image coordinate of web of the rail profile.
14, according to the two-dimensional measurement coordinate of web of the rail profile, web of the rail profile is divided into web of the rail great circle and web of the rail roundlet two parts automatically, extracts alignment characteristics point D
1～D
4The two-dimensional measurement coordinate.
The ratio that web of the rail great circle outline data accounts for whole great circle by there not being constraint match circle, obtains radius of circle and centre coordinate less than 1/3, and error is big.Therefore, the present invention supposes the standard value when radius of circle is design, adopts radius constraint nonlinear optimization method match circle, and main purpose is the central coordinate of circle that obtains circle.The objective function of nonlinear optimization is:
(x
M0, y
M0) being the match central coordinate of circle, r is a known radius, N is for participating in the point quantity of match circle.
By T
BCLThe two-dimensional measurement coordinate fitting web of the rail great circle of individual web of the rail point obtains great circle center of circle D
1The two-dimensional measurement coordinate.T from the web of the rail
BCL+ 1 point begins, and is calculated as follows:
As satisfied (N
Waist-T
SCL/ 2)〉d
iT
14The time, stop to calculate.Claim that the point of this moment is a roundlet key feature point, its coordinate is an i web of the rail point.If do not satisfy condition, then turned back to for the 12nd step, again according to com-state, the rail measurement image is gathered in control.
With i-T
SCL/ 2 to i+T
SCLThe two-dimensional measurement coordinate of/2 web of the rail point adopts about beam radius match web of the rail roundlet, tries to achieve roundlet center of circle D
2The two-dimensional measurement coordinate.By straight line D
1D
2With the equation of roundlet, try to achieve D
3And D
4The two-dimensional measurement coordinate.
15, by the partial contour unique point two-dimensional measurement coordinate of web of the rail great circle and web of the rail roundlet, 14 described methods are calculated measurement coordinate system to the standard design transformation of coordinates set by step.Little radius of circle is 20mm, and big radius of circle is got 400mm, is called the 60kg alignment pattern, and big radius of circle is got 350mm, is called the 50kg alignment pattern.Adopt 60kg alignment pattern and 50kg alignment pattern respectively, calculate alignment characteristics point D
1～D
4The two-dimensional measurement coordinate, by at least two alignment characteristics points wherein, calculate the two-dimensional transform of measurement coordinate system to the design coordinate system.
As shown in Figure 5, measurement coordinate system o
mx
my
mTo standard design coordinate system o
nx
ny
nBetween conversion comprise the rotation and translation.Great circle center of circle D
1With roundlet center of circle D
2Coordinate under conventional coordinates is known, and the coordinate under the measurement coordinate system is obtained in step 19, then vector in the measurement coordinate system
With corresponding vector in the design coordinate system
Between angle theta represent rotation between two coordinate systems.Promptly
Measurement coordinate system is transformed to the design coordinate system:
By 4 pairs of alignment characteristics points, according to formula [8], employing linear least square method is tried to achieve the translation between two coordinate systems.Obtain transformation parameter θ, t between two coordinate systems thus
xAnd t
y
16, according to the conversion of the measurement coordinate system of under two kinds of patterns, obtaining to the design coordinate system, respectively with the measurement coordinate conversion of rail profile rail head part in the design coordinate system, direct and standard rail is compared, and obtains vertical abrasion, side wear and total abrasion of rail.
According to the regulation of " railway track maintenance rule ", the vertical abrasion W of rail
vMeasure side wear W in rail surface 1/3 place (apart from the standard action edge)
hMeasure at the 16mm place down at standard tread (pressing standard section).
Total abrasion are: W
s=W
v+ W
h/ 2.
17, two kinds of alignment patterns, 3 wearing valves are down relatively got smaller value as last measurement result, and are discerned the rail type automatically according to the alignment pattern that is adopted, and the mileage number of measurement result, rail type and the present rail correspondence of surveying is saved in the file.
18, send measurement to serial ports and finish signal.
19, repeated for 11～18 steps, measure next time.
Embodiment
Propose measurement mechanism and measuring method according to the present invention, designed the practical laser vision sensor, rail has been carried out on-the-spot kinetic measurement, the software interface of measuring system as shown in Figure 6.
Fig. 7 has provided the contour feature image that a width of cloth comprises the measurement image of 50kg rail and extracts.Fig. 8 is at the rail profile profile of measurement coordinate system and transforms to the profile that designs under the coordinate system; Fig. 9 measures profile and design profile for after aiming at.
Measure at last to wear away and be respectively: the vertical abrasion is 1.9mm, and the side abrasion are-0.1mm that total the abrasion is that 1.8mm, rail type are 50R (the right rail of 50kg).At the PIV3.0 computing machine, the single measurement time is 25.5ms.
When Figure 10 stops for train,, carry out 540 times continuously and measure the rail surface abrasion reproducibility error curve that obtains same fixedly rail surface.The RMS error is respectively: 0.10mm, 0.04mm, 0.10mm.
Claims (1)
1, steel rail near laser visual dynamic method, use the steel rail near laser visual dynamic device to carry out the kinetic measurement of comprehensive parameters, laser vision sensor (7), computing machine, high speed image acquisition board, demarcation target and Survey Software that said steel rail near laser visual dynamic device is made up of laser projecting apparatus (3) and rail profile imaging system (6) are formed; Laser projecting apparatus (3) is made up of semiconductor laser (1) and optical plane generating optics system (2); Rail profile imaging system (6) is made up of ccd video camera (4) and filter system (5), and the optical axis confluent angle of the optical axis of rail profile imaging system (6) and laser projecting apparatus (3) is 45 °～60 °; Laser vision sensor (7) is fixed on the car body of train bottom, with the distance of tested rail (10) be 300～700mm; The optical plane (8) of laser projecting apparatus (3) projection is perpendicular to tested rail (10) and parallel with rail profile; The imaging system of rail profile imaging system (6) can observe projection light plane (8) and intersect rail head and the web of the rail feature contour (9) that forms with tested rail (10) surface; Said demarcation target is a two dimensional surface, the unique point that sets in advance is arranged on the target plane, the black square that arranged is arranged on the target plane, the quantity of square is 4～100, the length of side of square is 3～50mm, and its precision is 0.01～0.05mm, and the square spacing is 3～50mm, its precision is 0.01～0.05mm, and the summit of choosing each square is a unique point; It is characterized in that the concrete steps of measurement are:
1.1, set up the camera model and the measurement model of laser vision sensor (7), calibrating camera parameters and sensor construction parameter;
1.1.1, on optical plane, set up two-dimensional measurement coordinate system o
m-x
my
m, initial point o
mBe arbitrfary point on the optical plane, x
mAxle and y
mAxle is mutually vertical and be positioned at optical plane;
1.1.2, the structural parameters of the inner parameter of calibrating camera and laser vision sensor and be saved in computing machine;
1.1.3, set up standard rail two-dimensional design coordinate system o
n-x
ny
n, be y with the rail profile axis of symmetry
nAxle, the flange of rail is x
nAxle, the intersection point of two of web of the rail part great circle and roundlet centers of circle and circle center line connecting and roundlet is the alignment characteristics point on the selection standard rail, calculates the design coordinate of alignment characteristics point respectively, and is saved in computing machine, prepares against measuring phases and calls;
1.2, measure rail wear;
1.2.1, mileometer is connected on the computing machine by serial ports;
1.2.2, the established standards profile with measure the profile alignment parameter, set rail profile minimum length threshold T
L, it is shown that by the contour feature point numerical table span is 40～100; Rail head profile and web of the rail profile spacing minimum threshold T
D, it is represented that by the pixel coordinate of contour images span is 15～30 pixels, representative value 20 pixels; Preserve setup parameter, call in order to follow-up measurement;
1.2.3, process of measurement monitors the serial data state, when receiving startup command that serial ports sends and mileage, preserves mileage and also controls video camera and take the image that a width of cloth comprises the rail profile feature contour, is called measurement image;
1.2.4, extract the image coordinate of rail profile feature contour in the measurement image, do not satisfy length greater than T if extract two
LFeature contour, perhaps the spacing of two feature contours is less than T
D, then turn back to step 1.2.3, again according to com-state, the rail measurement image is gathered in control; If find two correct rail feature contours,, utilize the measurement coordinate of the image coordinate calculating rail profile feature contour of feature contour then according to the measurement model of laser vision sensor (7);
1.2.5, according to the measurement coordinate of rail profile feature contour, profile be will measure and rail head, web of the rail great circle and web of the rail roundlet three parts will be divided into automatically, feature contour by web of the rail great circle of cutting apart and the little circular portion of the web of the rail, adopt radius constrained procedure match web of the rail great circle and web of the rail roundlet respectively, little radius of circle is 20mm, and big radius of circle is got 400mm, is called the 60kg alignment pattern, big radius of circle is got 350mm, is called the 50kg alignment pattern;
1.2.6, adopt 60kg alignment pattern and 50kg alignment pattern respectively, calculate the measurement coordinate of alignment characteristics point, by wherein the measurement coordinate and the design coordinate of at least 2 alignment characteristics points, calculate measurement coordinate system and arrive the two-dimensional transform that designs coordinate system;
1.2.7, according to the conversion of the measurement coordinate system under two kinds of patterns, obtained to the design coordinate system, respectively with the measurement coordinate conversion of rail profile rail head part in the design coordinate system, direct and standard rail is compared, and calculates vertical abrasion, side wear and total abrasion of rail;
1.2.8,3 wearing valves under two kinds of alignment patterns relatively, choose smaller value as last measurement result, and obtain the rail type of tested rail according to the alignment pattern that is adopted, measurement result, rail type and corresponding mileage number are saved in the measurement result file;
1.2.9, send to measure to serial ports and to finish signal;
1.2.10, repeat 1.2.3～1.2.9 step, measure next time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005101237250A CN100462678C (en) | 2005-11-22 | 2005-11-22 | Steel rail near laser visual dynamic measuring device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005101237250A CN100462678C (en) | 2005-11-22 | 2005-11-22 | Steel rail near laser visual dynamic measuring device and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1776364A CN1776364A (en) | 2006-05-24 |
| CN100462678C true CN100462678C (en) | 2009-02-18 |
Family
ID=36766004
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005101237250A Expired - Fee Related CN100462678C (en) | 2005-11-22 | 2005-11-22 | Steel rail near laser visual dynamic measuring device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100462678C (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109059775A (en) * | 2016-08-30 | 2018-12-21 | 大连民族大学 | The method of Edge extraction |
| ES2749734A1 (en) * | 2018-09-21 | 2020-03-23 | Univ Madrid Carlos Iii | Mobile 3D imaging device using a flat laser beam (Machine-translation by Google Translate, not legally binding) |
Families Citing this family (32)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100480627C (en) | 2007-10-26 | 2009-04-22 | 北京航空航天大学 | Steel rail wearing integrative parameter vehicle-mounted dynamic measuring device and method |
| CN101949116B (en) * | 2010-08-06 | 2012-02-08 | 中国科学院长春光学精密机械与物理研究所 | Method for repairing steel rail with laser |
| CN102310869A (en) * | 2011-06-01 | 2012-01-11 | 王平 | Trolley for testing profile of steel rail |
| CN104180767B (en) * | 2013-05-24 | 2017-05-10 | 武汉铁路局武汉大型养路机械运用检修段 | steel-rail wear measurement device based on machine vision and grinding strategy method thereof |
| CN104417583A (en) * | 2013-08-29 | 2015-03-18 | 中国铁道科学研究院铁道建筑研究所 | Steel rail milling and grinding machine operation accuracy real-time detecting device |
| CN103630088B (en) * | 2013-11-06 | 2017-01-04 | 北京市地铁运营有限公司 | High accuracy tunnel cross-section detection method based on bidifly light belt and device |
| CN104179101A (en) * | 2014-07-22 | 2014-12-03 | 上海工程技术大学 | Method for accurately repairing metro curve rail corrugation |
| CN104359444B (en) * | 2014-11-10 | 2017-01-18 | 上海同儒机电科技有限公司 | Rail outline detection method and system |
| CN104359421B (en) * | 2014-11-10 | 2017-05-24 | 上海同儒机电科技有限公司 | Rail outline detection system and method |
| CN105675272A (en) * | 2014-11-20 | 2016-06-15 | 中国航空工业第六八研究所 | Detection method for judging whether dynamic characteristic of pendulum parts is good or bad |
| CN104442917A (en) * | 2014-12-31 | 2015-03-25 | 北京二七轨道交通装备有限责任公司 | Track detection method |
| CN105091778B (en) * | 2015-04-28 | 2017-10-24 | 长春机械科学研究院有限公司 | Obform guide rail one side characteristic detection method and device based on line-structured laser |
| CN105172839A (en) * | 2015-09-28 | 2015-12-23 | 成都多极子科技有限公司 | Full-automatic detection system for train guide rail contours |
| CN105423924B (en) * | 2015-11-17 | 2018-07-27 | 北京交通大学 | The On-line Measuring Method and system of sound barrier |
| CN105480256B (en) * | 2015-11-20 | 2018-06-22 | 武汉滨湖电子有限责任公司 | A kind of high-speed railway rail measuring device and measuring method |
| CN106080661A (en) * | 2016-05-27 | 2016-11-09 | 电子科技大学 | Train guide rail profile undulatory wear measuring method |
| CN108177660B (en) * | 2016-08-30 | 2020-07-14 | 大连民族大学 | Steel rail abrasion detection method with laser image processing step |
| CN106643545B (en) * | 2016-10-09 | 2018-10-26 | 湖南大学 | Laser displacement technology measures the calibration method of rail profile |
| CN108195291B (en) * | 2018-01-03 | 2020-05-05 | 中山大学 | Moving vehicle three-dimensional detection method and detection device based on differential light spots |
| CN108917591A (en) * | 2018-04-27 | 2018-11-30 | 中国铁道科学研究院集团有限公司基础设施检测研究所 | Rail profile autoegistration method and device under a kind of dynamic environment |
| CN109017867B (en) * | 2018-08-01 | 2021-05-25 | 湖南大学 | Dynamic measuring method for rail corrugation |
| CN109540027A (en) * | 2018-10-11 | 2019-03-29 | 伍波 | A kind of quick 3D vision laser measurement localization method |
| CN109798844B (en) * | 2018-12-28 | 2020-07-28 | 北京交通大学 | Rail profile automatic registration method in dynamic measurement of rail abrasion |
| CN110645908A (en) * | 2019-08-16 | 2020-01-03 | 广州瑞松北斗汽车装备有限公司 | Vehicle body appearance defect detection method and detection system |
| CN110926373A (en) * | 2019-12-10 | 2020-03-27 | 中南大学 | Structured light plane calibration method and system under railway foreign matter detection scene |
| CN111003018B (en) * | 2019-12-27 | 2021-02-26 | 郑州铁路职业技术学院 | System and method for dynamically detecting track condition of high-speed railway |
| CN111882599A (en) * | 2020-07-31 | 2020-11-03 | 北斗国铁(武汉)智能科技有限公司 | Data processing method and device for steel rail abrasion and storage medium |
| CN112414318B (en) * | 2020-11-10 | 2022-09-30 | 杭州申昊科技股份有限公司 | Steel rail abrasion rapid measurement algorithm based on structured light |
| CN112683195B (en) * | 2020-12-07 | 2022-10-25 | 中国铁道科学研究院集团有限公司基础设施检测研究所 | Steel rail longitudinal calibration method and device |
| CN112902869B (en) * | 2021-01-21 | 2022-10-25 | 中国铁道科学研究院集团有限公司基础设施检测研究所 | Method and device for adjusting laser plane of rail profile measuring system |
| CN113432630B (en) * | 2021-05-13 | 2023-02-17 | 国网河北省电力有限公司衡水供电分公司 | Distributed optical fiber sensing monitoring system |
| CN114046744A (en) * | 2021-11-19 | 2022-02-15 | 武汉利德测控技术有限公司 | 2D steel rail profile intelligent detection system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001304824A (en) * | 2000-04-19 | 2001-10-31 | Tokimec Inc | Portable rail section measuring apparatus |
| CN1566906A (en) * | 2003-06-11 | 2005-01-19 | 北京航空航天大学 | Construction optical visual sense transducer calibration method based on plane targets |
-
2005
- 2005-11-22 CN CNB2005101237250A patent/CN100462678C/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001304824A (en) * | 2000-04-19 | 2001-10-31 | Tokimec Inc | Portable rail section measuring apparatus |
| CN1566906A (en) * | 2003-06-11 | 2005-01-19 | 北京航空航天大学 | Construction optical visual sense transducer calibration method based on plane targets |
Non-Patent Citations (4)
| Title |
|---|
| An Unbiased Detector of Curvilinear Structures. Carsten Strger.IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE,Vol.20 No.2. 1998 |
| An Unbiased Detector of Curvilinear Structures. Carsten Strger.IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE,Vol.20 No.2. 1998 * |
| 实时动态检测钢轨磨耗系统研究. 纪淑波，蒋本和，靳文瑞.光电技术应用，第19卷第4期. 2004 |
| 实时动态检测钢轨磨耗系统研究. 纪淑波，蒋本和，靳文瑞.光电技术应用，第19卷第4期. 2004 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109059775A (en) * | 2016-08-30 | 2018-12-21 | 大连民族大学 | The method of Edge extraction |
| CN109059775B (en) * | 2016-08-30 | 2020-11-24 | 大连民族大学 | Steel rail abrasion detection method with image edge extraction step |
| ES2749734A1 (en) * | 2018-09-21 | 2020-03-23 | Univ Madrid Carlos Iii | Mobile 3D imaging device using a flat laser beam (Machine-translation by Google Translate, not legally binding) |
| WO2020058557A1 (en) * | 2018-09-21 | 2020-03-26 | Universidad Carlos Iii De Madrid | Mobile device for 3d image generation by means of a planar laser beam |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1776364A (en) | 2006-05-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100462678C (en) | Steel rail near laser visual dynamic measuring device and method | |
| CN101571379B (en) | Method for measuring diameter and straightness accuracy parameters of seamless round steel pipe | |
| US6909514B2 (en) | Wheel profile inspection apparatus and method | |
| CN100480627C (en) | Steel rail wearing integrative parameter vehicle-mounted dynamic measuring device and method | |
| CN107607044B (en) | A kind of hollow on worn tread detection method based on laser displacement sensor | |
| CN101694084B (en) | Ground on-vehicle mobile detecting system | |
| CN101532827A (en) | Deviation correction method for measurement of rail wear based on laser vision | |
| CN102749061B (en) | Steel rail abrasion measuring method based on dynamic template | |
| Sun et al. | Motion deviation rectifying method of dynamically measuring rail wear based on multi-line structured-light vision | |
| CN105292180A (en) | Non-contact wheel set dimension online detection method and device based on various sensors | |
| CN107421502B (en) | A kind of railway freight-car limit automatic measurement method | |
| CN105133471B (en) | Linear structured light pavement surface detection system-based pavement depth image production method | |
| CN103499302A (en) | Camshaft diameter online measuring method based on structured light visual imaging system | |
| CN109017867B (en) | Dynamic measuring method for rail corrugation | |
| CN100999220A (en) | On-line detection method and device for vehicle wheel set diameter | |
| CN103075976A (en) | Measuring method for dynamic envelope line of high-speed train | |
| CN102564335A (en) | Method for measuring deformation of large-scale tunnel | |
| CN102997863A (en) | Direct detection system for surface-shape errors in full-aperture optical aspheric surfaces | |
| CN109443245A (en) | A kind of multi-line structured light vision measuring method based on homography matrix | |
| CN105292181A (en) | Online wheel set dimension detecting method and device based on two kinds of sensors | |
| CN101957182A (en) | Large-caliber high-gradient optical mirror surface on-line measuring system | |
| CN102944188A (en) | Calibration method of spot scanning three-dimensional topography measuring system | |
| CN104501719A (en) | Method for determining axial line of train wheel set | |
| CN110243293A (en) | Section of jurisdiction faulting of slab ends device for fast detecting and method based on structure light and machine vision | |
| CN105292182A (en) | Wheel set size on-line detection method and device based on various sensors |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090218 Termination date: 20121122 |