CN114240916A - Multi-polarized light point cloud data fusion method and device for appearance state of steel rail - Google Patents
Multi-polarized light point cloud data fusion method and device for appearance state of steel rail Download PDFInfo
- Publication number
- CN114240916A CN114240916A CN202111583998.9A CN202111583998A CN114240916A CN 114240916 A CN114240916 A CN 114240916A CN 202111583998 A CN202111583998 A CN 202111583998A CN 114240916 A CN114240916 A CN 114240916A
- Authority
- CN
- China
- Prior art keywords
- polarization
- steel rail
- point cloud
- data
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 156
- 239000010959 steel Substances 0.000 title claims abstract description 156
- 238000007500 overflow downdraw method Methods 0.000 title claims abstract description 25
- 230000010287 polarization Effects 0.000 claims abstract description 157
- 230000004927 fusion Effects 0.000 claims abstract description 62
- 238000000034 method Methods 0.000 claims abstract description 27
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 23
- 238000004590 computer program Methods 0.000 claims description 29
- 238000005070 sampling Methods 0.000 claims description 10
- 238000003860 storage Methods 0.000 claims description 9
- 230000011218 segmentation Effects 0.000 claims description 4
- 238000010586 diagram Methods 0.000 description 14
- 238000005259 measurement Methods 0.000 description 14
- 238000001514 detection method Methods 0.000 description 7
- 238000011156 evaluation Methods 0.000 description 6
- 238000000227 grinding Methods 0.000 description 6
- 238000003384 imaging method Methods 0.000 description 6
- 230000003287 optical effect Effects 0.000 description 6
- 230000006870 function Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000005299 abrasion Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 239000011800 void material Substances 0.000 description 2
- 238000000149 argon plasma sintering Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000004451 qualitative analysis Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/143—Segmentation; Edge detection involving probabilistic approaches, e.g. Markov random field [MRF] modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30136—Metal
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Probability & Statistics with Applications (AREA)
- Software Systems (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses a multi-polarized light point cloud data fusion method and a device for the appearance state of a steel rail, wherein the method comprises the following steps: merging the read polarization camera data into steel rail profile data; dividing the steel rail profile data into a plurality of areas according to the coordinates of the rail vertex points and the coordinates of the rail distance points; fitting the steel rail profile data of the plurality of regions by using a random sample consensus (RANSAC) algorithm to obtain fitting curves of the plurality of regions; performing curve fusion on the fitted curves of the plurality of regions to obtain an optimal steel rail profile fitted curve; and measuring the steel rail profile by using the optimal steel rail profile fitting curve. The invention effectively solves the problem of local underexposure of the laser section of the steel rail.
Description
Technical Field
The invention relates to the technical field of structured light imaging, in particular to a method and a device for fusing multi-polarized light point cloud data of a steel rail appearance state.
Background
This section is intended to provide a background or context to the embodiments of the invention that are recited in the claims. The description herein is not admitted to be prior art by inclusion in this section.
The method has the advantages that the appearance state of the steel rail is measured on line, the health state of the steel rail is known in time, the outline and abrasion change trend of the steel rail are analyzed, and the scratch defect of the top surface of the steel rail is detected, so that the method is an important means for railway operation and maintenance. The steel rail profile measurement refers to the comparison of the actually measured steel rail profile data with the steel rail standard profile data to obtain indexes of the steel rail such as side grinding, vertical grinding, angle grinding and total abrasion, and provides a data base for the detection of the surface damage of the steel rail. The line structured light vision measurement based on triangulation can acquire three-dimensional section data of the steel rail outline at high speed in real time, is a typical non-contact optical measurement method, becomes a mainstream dynamic measurement mode of the steel rail outline and top surface scratch, and is widely used in the railway detection industry. However, when the method is applied to a railway site, factors such as irregularity of the surface of the steel rail, surface dirt, corrosion, light bands of wheel-rail contact surfaces and the like can influence the light scattering direction and energy intensity of the surface of the steel rail, so that abnormal energy distribution is caused, and the imaging quality is seriously influenced. For example, due to long-term contact friction with wheels, the surface of a steel rail is very smooth, the specular reflection effect of light is significant, and the specific expression is that most of the energy of incident light is distributed near the specular reflection direction, only a small amount of diffuse reflection light is collected by a camera, so that image information obtained by the camera is very weak, an underexposure (namely underexposure) phenomenon occurs, the image contrast and the confidence coefficient of the center of a light bar are both low, the accurate acquisition of the outline data of the steel rail is influenced, and further, the phenomenon that partial-area holes (data loss) occur in the steel rail after three-dimensional reconstruction is caused. This situation causes difficulty in ensuring the accuracy and stability of the rail profile measurement, and affects subsequent profile analysis and rail top surface defect detection results. The adjustment of the exposure time is a common solution, and although the problem of under exposure of the area of the light band is solved, the over exposure of other areas can be caused to cause the over thickness of the light bar, which causes a new problem.
Based on the information correlation and complementarity between the polarized images, some scholars propose various polarized image fusion methods, which mainly comprise: frequency domain fusion, spatial domain fusion, and the like. In the frequency domain, some scholars use discrete wavelet transform to decompose the image into low-frequency and high-frequency parts with different scales, and determine wavelet coefficients of the fused image by taking the wavelet coefficients in the low-frequency and high-frequency images as characteristic quantities. Some scholars propose a polarization image fusion algorithm based on two-dimensional discrete wavelet transform, which is used for enhancing image details and improving the visual effect of an image; in a spatial domain, some scholars provide a polarization image fusion method based on feature analysis, and the method determines a fusion weight value according to gray features, texture features and shape features of an image, fuses the image and is used for solving the problem that detailed information is lost when a polarization parameter image is calculated. Recently, image fusion methods based on deep learning are called research hotspots, such as deep fuse, fusion gan, etc., but most of these networks are directed to natural scenes with abundant colors and texture features, and the data volume of laser light bars is small and the abundant textures and color features are lacked, so these methods are not suitable for laser polarized light bar image fusion.
How to determine the fusion weight is the key to improve the fusion quality of the multi-polarized light image. Some scholars introduce a light strip reliability evaluation mechanism, determine the fusion weight of the source image according to the light strip reliability evaluation mechanism, and the light strip reliability evaluation takes characteristic quantities such as light strip width, gray scale, light strip average residual square sum and the like as evaluation indexes. For each source image, the light bar confidence for each column is calculated. The total pixel intensity of the light bars or the width of the light bars are independently selected as evaluation indexes to calculate the reliability of the polarized imaging of the light bars, and the optimal fusion effect is obtained by continuously adjusting the weight of each component image; the two evaluation indexes can be comprehensively used for calculating the light bar reliability to obtain the image fusion weight. The method effectively overcomes the problem of light reflection on the surface of the steel rail, but still has some problems to be solved when the method is applied to the dynamic measurement process of the profile system:
(1) the fusion strategy and weight calculation lack systematic quantitative analysis, and depend on qualitative analysis results and artificial experience threshold values excessively;
(2) the light bar changes along with the running of the train, and the unpredictable change of the light bar can be caused by the abrasion of the steel rail, the interference of sunlight and foreign matters, and the like, so that the width and the brightness of the light bar are selected as the judgment standards of the fusion weight, and the light bar is difficult to adapt to the complex and changeable line state of the whole railway;
(3) the time cost of fusion calculation of a plurality of polarization images is large, and the real-time performance of a measurement system is influenced.
Disclosure of Invention
The embodiment of the invention provides a multi-polarized light point cloud data fusion method for a steel rail appearance state, which is used for solving the problem of insufficient local exposure of a laser section image in steel rail outline measurement of non-polarized structured light projection, and comprises the following steps:
merging the read polarization camera data into steel rail profile data;
dividing the steel rail profile data into a plurality of areas according to the coordinates of the rail vertex points and the coordinates of the rail distance points;
fitting the steel rail profile data of the plurality of regions by using a random sample consensus (RANSAC) algorithm to obtain fitting curves of the plurality of regions;
performing curve fusion on the fitted curves of the plurality of regions to obtain an optimal steel rail profile fitted curve;
and measuring the steel rail profile by using the optimal steel rail profile fitting curve.
The embodiment of the invention also provides a multi-polarization light spot cloud data fusion device for the appearance state of the steel rail, which is used for solving the problem of insufficient local exposure of a laser section image in the steel rail profile measurement of non-polarized structured light projection, and comprises the following components:
the data merging module is used for merging the read polarization camera data into steel rail profile data;
the region division module is used for dividing the steel rail profile data into a plurality of regions according to the rail vertex coordinates and the rail distance point coordinates;
the fitting module is used for fitting the steel rail profile data of the multiple regions by using a random sampling consistency RANSAC algorithm to obtain fitting curves of the multiple regions;
the fusion module is used for carrying out curve fusion on the fitting curves of the plurality of areas to obtain an optimal steel rail profile fitting curve;
and the measuring module is used for measuring the steel rail profile by utilizing the optimal steel rail profile fitting curve.
The embodiment of the invention also provides computer equipment which comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein the processor executes the computer program to realize the multi-polarized light point cloud data fusion method of the steel rail appearance state.
The embodiment of the invention also provides a computer readable storage medium, wherein a computer program is stored in the computer readable storage medium, and when the computer program is executed by a processor, the multi-polarization light point cloud data fusion method for the appearance state of the steel rail is realized.
The embodiment of the invention also provides a computer program product, which comprises a computer program, and when the computer program is executed by a processor, the multi-polarization light point cloud data fusion method for the appearance state of the steel rail is realized.
In the embodiment of the invention, aiming at the problem of insufficient local exposure of a laser section image faced by the rail profile measurement of the non-polarized structured light projection, the read polarized camera data is combined into the rail profile data; dividing the steel rail profile data into a plurality of areas according to the coordinates of the rail vertex points and the coordinates of the rail distance points; fitting the steel rail profile data of the plurality of regions by using a random sample consensus (RANSAC) algorithm to obtain fitting curves of the plurality of regions; performing curve fusion on the fitted curves of the plurality of regions to obtain an optimal steel rail profile fitted curve; the optimal steel rail profile fitting curve is utilized to measure the steel rail profile, the fused image effectively solves the problem of local underexposure of the steel rail laser section, the contrast ratio of the optical strips, the central confidence coefficient of the optical strips and the image quality are improved, and the phenomenon of cavities does not occur on the steel rail after three-dimensional reconstruction.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts. In the drawings:
FIG. 1 is a first flowchart of a multi-polarized light point cloud data fusion method for the appearance state of a steel rail according to an embodiment of the present invention;
FIG. 2 is a flow chart of a multi-polarized light point cloud data fusion method for the appearance state of a steel rail according to an embodiment of the present invention;
FIG. 3 is a flow chart of a multi-polarized light point cloud data fusion method for the appearance state of a steel rail according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a steel rail profile data acquisition principle based on multi-polarization light fusion in the embodiment of the invention;
FIG. 5 is a four-way polarization component image, a linear polarization angle image, a linear polarization degree image, and a Stokes parameter image according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of a fused point cloud data of segmented RANSAC results and a schematic diagram of segmented fitting results;
FIG. 7 shows three-dimensional reconstruction results before and after polarization fusion of the light bar image of the measured steel rail and corresponding partial enlarged views;
FIG. 8 is a first structural block diagram of a multi-polarized light point cloud data fusion device for the appearance state of a steel rail according to an embodiment of the present invention;
fig. 9 is a structural block diagram of a multi-polarized light point cloud data fusion device for the appearance state of a steel rail in the embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the embodiments of the present invention are further described in detail below with reference to the accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.
Fig. 1 is a flow chart of a first method for fusing multi-polarization point cloud data of an appearance state of a steel rail in the embodiment of the invention, as shown in fig. 1, the method includes:
step 101: merging the read polarization camera data into steel rail profile data;
step 102: dividing the steel rail profile data into a plurality of areas according to the coordinates of the rail vertex points and the coordinates of the rail distance points;
step 103: fitting the steel rail profile data of the plurality of regions by using a random sample consensus (RANSAC) algorithm to obtain fitting curves of the plurality of regions;
step 104: performing curve fusion on the fitted curves of the plurality of regions to obtain an optimal steel rail profile fitted curve;
step 105: and measuring the steel rail profile by using the optimal steel rail profile fitting curve.
In an embodiment of the present invention, the polarization camera data includes a four-directional polarization component image, a linear polarization angle image, a linear polarization degree image, and a stokes parameter image;
as shown in fig. 2, step 101 merges the read polarized camera data into steel profile data, including:
step 201: respectively calculating light bar centers corresponding to the four-direction polarization component image, the linear polarization angle image, the linear polarization degree image and the Stokes parameter image;
step 202: calculating to obtain corresponding profile polarization point cloud data according to corresponding light strip centers;
step 203: and arranging and combining the corresponding profile polarization point cloud data into steel rail profile data according to the point cloud coordinates.
Specifically, for 9 polarization component images, four-way polarization component image I0，I45，I90，I135Linear polarization angle image IAoPLinear polarization degree image IDoPStokes parametric image S0，S1，S2Respectively calculating the centers of the light bars to obtain corresponding profile polarization point cloud data, and recording the data as P0，P45，P90，P135，PAoP，PDoP，S′0，S′1，S′2. The method comprises the following steps of calculating to obtain corresponding profile polarization point cloud data according to corresponding light bar centers, wherein the corresponding profile polarization point cloud data can be calculated by a method in the prior art.
Arranging and combining 9 profile polarization point cloud data into profile data P according to point cloud coordinates, and recording the profile data P as profile data P
P＝∪(P0,P45,P90,P135,PAoP,PDoP,S′0,S′1,S′2)；
Wherein, P represents the data of the steel rail profile; u (·) represents point cloud coordinate data merging operation; p0Image I representing a polarization component with a polarization direction of 0 DEG0Corresponding profile polarization point cloud data; p45Image I showing polarization component with polarization direction of 45 DEG45Corresponding profile polarization point cloud data; p90Representing the polarization directionImage I of 90 ° polarized component90Corresponding profile polarization point cloud data; p135Image I showing polarization component with polarization direction of 135 DEG135Corresponding profile polarization point cloud data; pAoPImage I representing linear polarization angleAoPCorresponding profile polarization point cloud data; pDoPRepresenting images I of linear polarizationDoPCorresponding profile polarization point cloud data; s'0First component image S representing a Stokes parametric image0Corresponding profile polarization point cloud data; s'1Second component image S representing a Stokes parametric image1Corresponding profile polarization point cloud data; s'2Third component image S representing a Stokes parametric image2And corresponding profile polarization point cloud data.
In the embodiment of the invention, the profile is divided into 5 areas according to the rail top track pitch of the steel rail. Dividing the steel rail profile data into a plurality of areas according to the coordinates of the rail vertex points and the coordinates of the rail distance points according to the following formula:
wherein, Pt,
Rail profile data respectively representing the divided 5 regions; segment (-) represents a region segmentation operation; p represents the data of the steel rail profile; t isx,yRepresenting the coordinates of the rail vertex; gx,yRepresenting the gauge point coordinates.
In the embodiment of the invention, a random sample consensus (RANSAC) algorithm is executed on each region to obtain a best-fit polynomial equation curve of each region. Fitting the steel rail profile data of the plurality of regions by using a random sampling consistency RANSAC algorithm according to the following formula to obtain fitting curves of the plurality of regions:
wherein, Ct,
RANSAC fitting polynomial curves respectively representing the divided 5 regions; xit,
Respectively representing the inner point threshold values of the divided 5 regions, wherein the sampled values smaller than the inner point threshold values participate in fitting, and the sampled values larger than the inner point threshold values are used as noise for filtering; m ist,
Respectively representing the sampling iteration times of the divided 5 regions; n ist,
And respectively representing the optimal polynomial fitting powers of the divided 5 regions, wherein the values can be obtained by selecting a plurality of typical profile data from the actual road, constructing a global optimization model and obtaining the optimal polynomial fitting powers by a least square method.
In the embodiment of the invention, a curve C is fitted to 5 sectionst,
The steel rail is spliced into a complete steel rail half-section profile. And (3) carrying out curve fusion on the fitted curves of the plurality of regions according to the following formula to obtain an optimal steel rail profile fitted curve:
c represents an optimal steel rail profile fitting curve obtained after fusion of a plurality of polarization point cloud data; ct,
RANSAC fitting polynomial curves respectively representing the divided 5 regions; stilling (. cndot.) represents a curve fusion operation.
In the embodiment of the present invention, as shown in fig. 3, the method further includes:
step 301: and judging whether the fitting curves of the multiple regions meet the requirements, if not, adjusting RANSAC algorithm parameters, and fitting again. The parameters may include a threshold, an iteration number, an optimal polynomial fitting power, and the like.
The invention provides a multi-polarized light point cloud data fusion method for the appearance state of a steel rail, aiming at the problem of insufficient local exposure of a laser section image in the steel rail outline measurement of non-polarized structured light projection. A point cloud data fusion algorithm of a four-way polarization component image, a Stokes parameter image, a linear polarization angle image and a linear polarization degree image is constructed based on a subsection RANSAC algorithm, the fused image effectively solves the problem of local underexposure of a steel rail laser section, the contrast ratio and the central confidence coefficient of a light bar and the image quality of the light bar are improved, and the phenomenon that a hole does not appear on a steel rail after three-dimensional reconstruction is avoided.
DETAILED DESCRIPTION OF EMBODIMENT (S) OF INVENTION
A Genie Nano M2450 Polarized model polarization camera was selected, which is manufactured by Teledyne Dalsa, Canada and has good polarization imaging performance. The camera is integrated into a steel rail profile detection system, polarization fusion point cloud data of the steel rail profile is obtained according to the processes shown in fig. 1 to 4, and a steel rail laser section four-way polarization component image and a stokes parameter image S shown in fig. 5 can be obtained0、S1、S2Linear polarization angle image IAoPAnd linear polarization degree image IDoLP. Wherein, (a) in fig. 5 represents a 0 ° direction polarization component image; (b) representing a 135 ° directional polarization component image; (c) representing a 45 ° directional polarization component image; (d) a 90 ° directional polarization component image; (e) representing stokes parametric images S0(ii) a (f) Representing stokes parametric images S1(ii) a (g) Representing stokes parametric images S2(ii) a (h) Image I representing linear polarization angleAoP(ii) a (i) Representing images I of linear polarizationDoP。
The polarization components are fused by adopting a subsection random sampling consistency polarization point cloud data fusion algorithm, and the result is shown in fig. 6. Wherein (a) represents fused point cloud data; (b) the piecewise fitting results are shown.
Selecting a steel rail near a Kyoto line descending K346 Qinghe city station, respectively measuring the profile after the steel rail grinding vehicle carries out primary grinding and secondary grinding, comparing the profile obtained by the detection system with the minipro profile, and finally confirming the system precision. The profile obtained by the system adopts the fusion algorithm provided by the project, and because the number of the collected points of the system is different from the Miniprof number of the points, the distance between the directly calculated point pairs cannot be directly used for judging the profile difference, so that the profile needs to be subjected to smoothing treatment and then is dispersed, the corresponding point pairs are ensured to be compared, and the profile difference is really judged. The comparative results are shown in Table 1.
TABLE 1 comparison of profile measurement accuracy before and after fusion
| Profile | Maximum value | Minimum value | Mean value of | Value of 95 quantile |
| Before fusion | 0.2587 | 0.0019 | 0.0542 | 0.1230 |
| After fusion | 0.1202 | 0.0011 | 0.0363 | 0.0800 |
As shown in Table 1, the difference of 95 quantile values of the fused system detection profile and the profile measured by the 0.02 millimeter precision contact type detection device (miniprof) is controlled within 0.1mm, namely the polarization fusion method provided by the invention effectively improves the precision of the profile measurement system, and the system measurement precision is superior to 0.1 mm.
In order to further verify the three-dimensional imaging effect of the steel rail top surface fused by the multi-polarization light polarization, a section of steel rail which is easy to be abnormally exposed is placed on the electric control translation stage. The linear structure light profile scanning device is built, the laser section images of the steel rail are obtained at equal intervals at 2mm sampling intervals by using the polarization camera, and the measured steel rail is subjected to three-dimensional reconstruction through processes of light bar center extraction, calibration and the like. Fig. 7 shows three-dimensional reconstruction results of the measured steel rail light bar image before polarization fusion and after polarization fusion, respectively, where (a) represents the result before polarization fusion; (b) represents a partial enlargement of (a); (c) after representing polarization fusion; (d) shows a partial enlargement of (c). It can be seen that, due to abnormal exposure of the total intensity image before polarization fusion, a void (data loss) phenomenon occurs in a partial region of the rail after three-dimensional reconstruction, while the imaging quality of the image after polarization fusion is obviously improved, the rail after three-dimensional reconstruction of the fused image does not have the void phenomenon, and the reconstruction result can still better reflect the real condition of the rail.
The embodiment of the invention also provides a multi-polarization light spot cloud data fusion device for the appearance state of the steel rail, which is described in the following embodiment. The principle of solving the problems by the device is similar to the multi-polarized light point cloud data fusion method of the appearance state of the steel rail, so the implementation of the device can refer to the implementation of the multi-polarized light point cloud data fusion method of the appearance state of the steel rail, and repeated parts are not repeated.
Fig. 8 is a first structural block diagram of a multi-polarized light point cloud data fusion device for an appearance state of a steel rail in an embodiment of the present invention, and as shown in fig. 8, the multi-polarized light point cloud data fusion device for an appearance state of a steel rail includes:
the data merging module 02 is used for merging the read polarization camera data into steel rail profile data;
the region dividing module 04 is used for dividing the steel rail profile data into a plurality of regions according to the rail vertex coordinates and the rail distance point coordinates;
the fitting module 06 is used for fitting the steel rail profile data of the plurality of regions by using a random sampling consistency RANSAC algorithm to obtain fitting curves of the plurality of regions;
the fusion module 08 is used for performing curve fusion on the fitted curves of the multiple regions to obtain an optimal rail profile fitted curve;
and the measuring module 10 is used for measuring the rail profile by using the optimal rail profile fitting curve.
In an embodiment of the present invention, the polarization camera data includes a four-directional polarization component image, a linear polarization angle image, a linear polarization degree image, and a stokes parameter image;
the data merging module is specifically configured to:
respectively calculating light bar centers corresponding to the four-direction polarization component image, the linear polarization angle image, the linear polarization degree image and the Stokes parameter image;
calculating to obtain corresponding profile polarization point cloud data according to corresponding light strip centers;
and arranging and combining the corresponding profile polarization point cloud data into steel rail profile data according to the point cloud coordinates.
In the embodiment of the invention, the read polarization camera data are combined into the steel rail profile data according to the following formula:
P＝∪(P0,P45,P90,P135,PAoP,PDoP,S′0,S′1,S′2)；
wherein, P represents the data of the steel rail profile; u (·) represents point cloud coordinate data merging operation; p0Image I representing a polarization component with a polarization direction of 0 DEG0Corresponding profile polarization point cloud data; p45Representing a polarization with a polarization direction of 45 degVibration component image I45Corresponding profile polarization point cloud data; p90Image I showing polarization component with 90 ° polarization direction90Corresponding profile polarization point cloud data; p135Image I showing polarization component with polarization direction of 135 DEG135Corresponding profile polarization point cloud data; pAoPImage I representing linear polarization angleAoPCorresponding profile polarization point cloud data; pDoPRepresenting images I of linear polarizationDoPCorresponding profile polarization point cloud data; s'0First component image S representing a Stokes parametric image0Corresponding profile polarization point cloud data; s'1Second component image S representing a Stokes parametric image1Corresponding profile polarization point cloud data; s'2Third component image S representing a Stokes parametric image2And corresponding profile polarization point cloud data.
In the embodiment of the invention, the rail profile data is divided into a plurality of areas according to the coordinates of the rail vertex and the coordinates of the rail distance point according to the following formula:
wherein, Pt,
Rail profile data respectively representing the divided 5 regions; segment (-) represents a region segmentation operation; p represents the data of the steel rail profile; t isx,yRepresenting the coordinates of the rail vertex; gx,yRepresenting the gauge point coordinates.
In the embodiment of the invention, the rail profile data of a plurality of regions are fitted by using a random sample consensus (RANSAC) algorithm according to the following formula to obtain fitting curves of the plurality of regions:
wherein, Ct,
RANSAC fitting polynomial curves respectively representing the divided 5 regions; xit,
Inner point thresholds respectively representing the divided 5 regions; m ist,
Respectively representing the sampling iteration times of the divided 5 regions; n ist,
Respectively representing the best polynomial fit powers of the divided 5 regions.
In the embodiment of the invention, the fitted curves of a plurality of areas are subjected to curve fusion according to the following formula to obtain the optimum rail profile fitted curve:
wherein C represents an optimal rail profile fitting curve; ct,
RANSAC fitting polynomial curves respectively representing the divided 5 regions; stilling (. cndot.) represents a curve fusion operation.
In the embodiment of the present invention, as shown in fig. 9, the method further includes:
and the judging module 12 is configured to judge whether fitting curves of the multiple regions meet requirements, and if not, adjust RANSAC algorithm parameters and perform fitting again.
The embodiment of the invention also provides computer equipment which comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein the processor executes the computer program to realize the multi-polarized light point cloud data fusion method of the steel rail appearance state.
The embodiment of the invention also provides a computer readable storage medium, wherein a computer program is stored in the computer readable storage medium, and when the computer program is executed by a processor, the multi-polarization light point cloud data fusion method for the appearance state of the steel rail is realized.
The embodiment of the invention also provides a computer program product, which comprises a computer program, and when the computer program is executed by a processor, the multi-polarization light point cloud data fusion method for the appearance state of the steel rail is realized.
In the embodiment of the invention, aiming at the problem of insufficient local exposure of a laser section image faced by the rail profile measurement of the non-polarized structured light projection, the read polarized camera data is combined into the rail profile data; dividing the steel rail profile data into a plurality of areas according to the coordinates of the rail vertex points and the coordinates of the rail distance points; fitting the steel rail profile data of the plurality of regions by using a random sample consensus (RANSAC) algorithm to obtain fitting curves of the plurality of regions; performing curve fusion on the fitted curves of the plurality of regions to obtain an optimal steel rail profile fitted curve; the optimal steel rail profile fitting curve is utilized to measure the steel rail profile, the fused image effectively solves the problem of local underexposure of the steel rail laser section, the contrast ratio of the optical strips, the central confidence coefficient of the optical strips and the image quality are improved, and the phenomenon of cavities does not occur on the steel rail after three-dimensional reconstruction.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (17)
1. A multi-polarization light point cloud data fusion method for the appearance state of a steel rail is characterized by comprising the following steps:
merging the read polarization camera data into steel rail profile data;
dividing the steel rail profile data into a plurality of areas according to the coordinates of the rail vertex points and the coordinates of the rail distance points;
fitting the steel rail profile data of the plurality of regions by using a random sample consensus (RANSAC) algorithm to obtain fitting curves of the plurality of regions;
performing curve fusion on the fitted curves of the plurality of regions to obtain an optimal steel rail profile fitted curve;
and measuring the steel rail profile by using the optimal steel rail profile fitting curve.
2. The steel rail appearance state multi-polarized light point cloud data fusion method according to claim 1, wherein the polarized camera data comprises a four-way polarized component image, a linear polarization angle image, a linear polarization degree image and a stokes parameter image;
merging the read polarized camera data into steel profile data, comprising:
respectively calculating light bar centers corresponding to the four-direction polarization component image, the linear polarization angle image, the linear polarization degree image and the Stokes parameter image;
calculating to obtain corresponding profile polarization point cloud data according to corresponding light strip centers;
and arranging and combining the corresponding profile polarization point cloud data into steel rail profile data according to the point cloud coordinates.
3. The method for fusing the multi-polarization-light point cloud data of the appearance state of the steel rail according to claim 1, wherein the read polarization camera data are merged into the steel rail profile data according to the following formula:
P＝∪(P0,P45,P90,P135,PAoP,PDoP,S′0,S′1,S′2)；
wherein, P represents the data of the steel rail profile; u (·) represents point cloud coordinate data merging operation; p0Image I representing a polarization component with a polarization direction of 0 DEG0Corresponding profile polarization point cloud data; p45Image I showing polarization component with polarization direction of 45 DEG45Corresponding profile polarization point cloud data; p90Image I showing polarization component with 90 ° polarization direction90Corresponding profile polarization point cloud data; p135Image I showing polarization component with polarization direction of 135 DEG135Corresponding profile polarization point cloud data; pAoPImage I representing linear polarization angleAoPCorresponding profile polarization point cloud data; pDoPRepresenting images I of linear polarizationDoPCorresponding profile polarization point cloud data; s'0First component image S representing a Stokes parametric image0Corresponding profile polarization point cloud data; s'1Second component image S representing a Stokes parametric image1Corresponding profile polarization point cloud data; s'2Third component image S representing a Stokes parametric image2And corresponding profile polarization point cloud data.
4. The method for fusing the multi-polarized light point cloud data of the appearance state of the steel rail as claimed in claim 1, wherein the steel rail profile data is divided into a plurality of areas according to the coordinates of the rail vertex and the coordinates of the rail distance point according to the following formula:
wherein, Pt,
Rail profile data respectively representing the divided 5 regions; segment (-) represents a region segmentation operation; p represents the data of the steel rail profile; t isx,yRepresenting the coordinates of the rail vertex; gx,yRepresenting the gauge point coordinates.
5. The method for fusing the multi-polarization light point cloud data of the appearance state of the steel rail as claimed in claim 1, wherein the rail profile data of the plurality of regions is fitted by using a random sample consensus (RANSAC) algorithm according to the following formula to obtain fitting curves of the plurality of regions:
wherein, Ct,
RANSAC fitting polynomial curves respectively representing the divided 5 regions; xit,
Inner point thresholds respectively representing the divided 5 regions; m ist,
Respectively representing the sampling iteration times of the divided 5 regions; n ist,
Respectively representing the best polynomial fit powers of the divided 5 regions.
6. The method for fusing the multi-polarized light point cloud data of the appearance state of the steel rail according to claim 1, wherein the fitted curves of the plurality of areas are subjected to curve fusion according to the following formula to obtain an optimal steel rail profile fitted curve:
wherein C represents an optimal rail profile fitting curve; ct,
RANSAC fitting polynomial curves respectively representing the divided 5 regions; stilling (. cndot.) represents a curve fusion operation.
7. The method for fusing the multi-polarization-light point cloud data of the appearance state of the steel rail according to claim 1, further comprising:
and judging whether the fitting curves of the multiple regions meet the requirements, if not, adjusting RANSAC algorithm parameters, and fitting again.
8. A multi-polarization light spot cloud data fusion device for appearance states of steel rails is characterized by comprising:
the data merging module is used for merging the read polarization camera data into steel rail profile data;
the region division module is used for dividing the steel rail profile data into a plurality of regions according to the rail vertex coordinates and the rail distance point coordinates;
the fitting module is used for fitting the steel rail profile data of the multiple regions by using a random sampling consistency RANSAC algorithm to obtain fitting curves of the multiple regions;
the fusion module is used for carrying out curve fusion on the fitting curves of the plurality of areas to obtain an optimal steel rail profile fitting curve;
and the measuring module is used for measuring the steel rail profile by utilizing the optimal steel rail profile fitting curve.
9. The steel rail appearance state multi-polarization light point cloud data fusion device according to claim 8, wherein the polarization camera data comprise a four-way polarization component image, a linear polarization angle image, a linear polarization degree image and a stokes parameter image;
the data merging module is specifically configured to:
respectively calculating light bar centers corresponding to the four-direction polarization component image, the linear polarization angle image, the linear polarization degree image and the Stokes parameter image;
calculating to obtain corresponding profile polarization point cloud data according to corresponding light strip centers;
and arranging and combining the corresponding profile polarization point cloud data into steel rail profile data according to the point cloud coordinates.
10. The steel rail appearance state multi-polarization light spot cloud data fusion device according to claim 8, wherein the read polarization camera data is merged into steel rail profile data according to the following formula:
P＝∪(P0,P45,P90,P135,PAoP,PDoP,S′0,S′1,S′2)；
wherein, P represents the data of the steel rail profile; u (·) represents point cloud coordinate data merging operation; p0Image I representing a polarization component with a polarization direction of 0 DEG0Corresponding profile polarization point cloud data; p45Image I showing polarization component with polarization direction of 45 DEG45Corresponding profile polarization point cloud data; p90Image I showing polarization component with 90 ° polarization direction90Corresponding profile polarization point cloud data; p135Image I showing polarization component with polarization direction of 135 DEG135Corresponding profile polarization point cloud data; pAoPImage I representing linear polarization angleAoPCorresponding profile polarization point cloud data; pDoPRepresenting images I of linear polarizationDoPCorresponding profile polarization point cloud data; s'0First component image S representing a Stokes parametric image0Corresponding profile polarization point cloud data; s'1Second component image S representing a Stokes parametric image1Corresponding profile polarization point cloud data; s'2Third component image S representing a Stokes parametric image2And corresponding profile polarization point cloud data.
11. The rail appearance state multi-polarization light spot cloud data fusion device according to claim 8, wherein the rail profile data is divided into a plurality of areas according to the rail vertex coordinates and the gauge point coordinates according to the following formula:
wherein, Pt,
Rail profile data respectively representing the divided 5 regions; segment (-) represents a region segmentation operation; p represents the data of the steel rail profile; t isx,yRepresenting the coordinates of the rail vertex; gx,yRepresenting the gauge point coordinates.
12. The device for fusing the multi-polarization light spot cloud data of the appearance state of the steel rail as claimed in claim 8, wherein the rail profile data of the plurality of regions is fitted by using a random sample consensus (RANSAC) algorithm according to the following formula to obtain a fitted curve of the plurality of regions:
wherein, Ct,
RANSAC fitting polynomial curves respectively representing the divided 5 regions; xit,
Inner point thresholds respectively representing the divided 5 regions; m ist,
Respectively representing the sampling iteration times of the divided 5 regions; n ist,
Respectively representing the best polynomial fit powers of the divided 5 regions.
13. The steel rail appearance state multi-polarization light point cloud data fusion device according to claim 8, wherein the fitting curves of the plurality of regions are subjected to curve fusion according to the following formula to obtain an optimal steel rail profile fitting curve:
wherein C represents an optimal rail profile fitting curve; ct,
RANSAC fitting polynomial curves respectively representing the divided 5 regions; stilling (. cndot.) represents a curve fusion operation.
14. The steel rail appearance state multi-polarization light spot cloud data fusion device according to claim 8, further comprising:
and the judging module is used for judging whether the fitting curves of the multiple regions meet the requirements, and if not, adjusting RANSAC algorithm parameters and fitting again.
15. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the multi-polarized light point cloud data fusion method for the appearance state of the steel rail according to any one of claims 1 to 7 when executing the computer program.
16. A computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the multi-polarized light point cloud data fusion method for the appearance state of the steel rail according to any one of claims 1 to 7.
17. A computer program product, characterized in that the computer program product comprises a computer program, and the computer program is executed by a processor to realize the multi-polarized light point cloud data fusion method for the appearance state of the steel rail according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111583998.9A CN114240916A (en) | 2021-12-22 | 2021-12-22 | Multi-polarized light point cloud data fusion method and device for appearance state of steel rail |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111583998.9A CN114240916A (en) | 2021-12-22 | 2021-12-22 | Multi-polarized light point cloud data fusion method and device for appearance state of steel rail |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114240916A true CN114240916A (en) | 2022-03-25 |
Family
ID=80761586
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111583998.9A Pending CN114240916A (en) | 2021-12-22 | 2021-12-22 | Multi-polarized light point cloud data fusion method and device for appearance state of steel rail |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114240916A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114663438A (en) * | 2022-05-26 | 2022-06-24 | 浙江银轮智能装备有限公司 | Track detection method, system, apparatus, storage medium and computer program product |
| CN116862976A (en) * | 2023-06-08 | 2023-10-10 | 中铁第四勘察设计院集团有限公司 | Rail center line extraction method and system based on unmanned aerial vehicle laser point cloud |
| CN117966529A (en) * | 2024-03-28 | 2024-05-03 | 中铁第四勘察设计院集团有限公司 | Virtual-real combination-based steel rail polishing control method and equipment |
-
2021
- 2021-12-22 CN CN202111583998.9A patent/CN114240916A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114663438A (en) * | 2022-05-26 | 2022-06-24 | 浙江银轮智能装备有限公司 | Track detection method, system, apparatus, storage medium and computer program product |
| CN116862976A (en) * | 2023-06-08 | 2023-10-10 | 中铁第四勘察设计院集团有限公司 | Rail center line extraction method and system based on unmanned aerial vehicle laser point cloud |
| CN116862976B (en) * | 2023-06-08 | 2024-04-02 | 中铁第四勘察设计院集团有限公司 | Rail center line extraction method and system based on unmanned aerial vehicle laser point cloud |
| CN117966529A (en) * | 2024-03-28 | 2024-05-03 | 中铁第四勘察设计院集团有限公司 | Virtual-real combination-based steel rail polishing control method and equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114240916A (en) | Multi-polarized light point cloud data fusion method and device for appearance state of steel rail | |
| RU2743020C2 (en) | System and method for monitoring railway wheels | |
| Zhou et al. | Rail profile measurement based on line-structured light vision | |
| CN109389639B (en) | Method and device for extracting center of laser stripe of steel rail outline in dynamic driving environment | |
| CN117237368B (en) | Bridge crack detection method and system | |
| US20160033368A1 (en) | System and method for analyzing tire tread parameters | |
| CN112651968B (en) | Wood board deformation and pit detection method based on depth information | |
| CN103411533B (en) | Structured light self-adaptation multiexposure method | |
| CN104296680B (en) | Particle-reinforced titanium-based composite material grinding surface quality evaluation method | |
| CN110473187B (en) | Object-oriented line scanning three-dimensional pavement crack extraction method | |
| Shahabi et al. | Simulation and measurement of surface roughness via grey scale image of tool in finish turning | |
| CN108725511A (en) | The real time position modification method of rail corrugation string measuring point | |
| CN112881042A (en) | Wheel set detection system, method and medium based on laser profile | |
| Peng et al. | A new pixel matching method using the modulation of shadow areas in online 3D measurement | |
| Yang et al. | High dynamic range fringe pattern acquisition based on deep neural network | |
| CN114581805A (en) | Coating roller surface defect detection method adopting 3D line laser profile technology | |
| Usamentiaga et al. | Flatness measurement using two laser stripes to remove the effects of vibrations | |
| Chunfu et al. | Research on the line point cloud processing method for railway wheel profile with a laser profile sensor | |
| Wu et al. | Research of pavement topography based on NURBS reconstruction for 3D structured light | |
| Wang et al. | A study of error correction method based on selective focus for curved rail profile measurement | |
| Yan et al. | Adaptive mask generating algorithm based on the fuzzy set theory for the weighted least-squares phase unwrapping | |
| Iglesias et al. | Application of computer vision techniques to estimate surface roughness on wood-based sanded workpieces | |
| Li et al. | An on-line measurement method of the medium thickness steel plate based on structured light vision sensor | |
| CN106023150A (en) | Real-time objective quality evaluation method for digital hologram noise interference | |
| Yan et al. | Online measurement system of slab front-end bending in hot rough rolling based on line structured light vision |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |