CN105928484A - Elevator guide rail automatic measurement system based on binocular vision - Google Patents
Elevator guide rail automatic measurement system based on binocular vision Download PDFInfo
- Publication number
- CN105928484A CN105928484A CN201610184490.4A CN201610184490A CN105928484A CN 105928484 A CN105928484 A CN 105928484A CN 201610184490 A CN201610184490 A CN 201610184490A CN 105928484 A CN105928484 A CN 105928484A
- Authority
- CN
- China
- Prior art keywords
- binocular vision
- feature
- guide rail
- matching
- elevator guide
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B21/00—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant
- G01B21/22—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant for measuring angles or tapers; for testing the alignment of axes
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B21/00—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant
- G01B21/30—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant for measuring roughness or irregularity of surfaces
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
For the purpose of solving problems including high labour intensity and low efficiency and precision for elevator guide rail manual measurement, an elevator guide rail automatic measurement system based on binocular vision is designed, and the system is formed by a hardware system and a software system. For the problems of low contrast ratio, few feature points and the like on elevator guide rail surface, a laser is used for projecting grids onto the to-be-measured workpiece surface, so that the workpiece surface has certain recognition features, and an RANSAC algorithm is used for extracting feature grids; a SURF algorithm is used for extracting and matching image feature points, and space coordinates of the feature points are calculated according to a camera calibration result; and form and location tolerance of the elevator guide rail is calculated according to the space coordinates of the feature points. The experimental result shows that the system has advantages of reasonable model selection, good soft stability and high efficiency, the measurement efficiency and precision of the elevator guide rail is improved, and the system has good practical value.
Description
Technical field
The present invention relates to the system of a kind of technical field of image processing, a kind of cage guide automatic measurement system based on binocular vision.
Background technology
The crudy of cage guide, is the key factor affecting elevator ride quality, is to ensure that elevator safety, reliable a, important performance indexes of stable operation.In GBJ310 88 " standard for quality inspection and assessment of erection works of elevators ", the requirement to lift rail perpendicularity is every 5m tolerance 0.7mm, and the method for inspection is " messenger wire, dipstick metering inspection ".The method of this " messenger wire, dipstick metering inspection " slide rail verticality has continued to use many decades at elevator industry, and its advantage is that survey tool, measurer are simple, and measurement data is directly perceived；Shortcoming is that certainty of measurement varies with each individual, and is difficult to realize measurement data and automatically gathers, inefficiency.In recent years, elevator industry occurs in that the laser plummet for guide rail detection, high-precision laser beam is used to replace hanging vertical line, accuracy of detection is made to increase, measuring principle and technique are essentially identical with " plumbing method ", but this method needs on each detection position by manually measuring one by one, and certainty of measurement is the highest, and measurement data is by manually reading, the interface capability lacking with there is powerful disposal ability computer now.Additionally, by the flatness of manual measurement cage guide with the depth of parallelism takes time and effort especially, degree of accuracy is low.
In recent years, machine vision technique quickly grows, and the application in industrial nondestructive testing is more and more universal, such as industrial part detection, the assembling detection of automobile zero/accessory, and industry character recognition etc..The detection automatically for cage guide that develops rapidly of machine vision technique provides good theoretical basis.
Summary of the invention
The present invention is directed to deficiencies of the prior art, propose a kind of cage guide automatic measurement system based on binocular vision.The system hardware of the present invention is stable, and algorithm robustness is good, it is possible to adapts to the adverse circumstances of cage guide manufacturing shop, automatically measures for cage guide, have preferable practical value.
The present invention is achieved by the following technical solutions:
The present invention includes: binocular vision hardware system and binocular vision software system.
Described binocular vision hardware system, including light source, camera and camera lens.
Described binocular vision software system, calculates and form and position tolerance calculating with characteristic matching, characteristic point coordinate including camera calibration, feature extraction.
Compared with prior art, the invention has the beneficial effects as follows: effectively accelerate the speed of cage guide detection, add the precision of detection, have good repeatability, there is the interface capability good with computer.
Detailed description of the invention
Below embodiments of the invention are elaborated: the present embodiment is implemented under premised on technical solution of the present invention, give detailed embodiment and process, but protection scope of the present invention is not limited to following embodiment.
The present embodiment includes: binocular vision hardware system and binocular vision software system.
Described binocular vision hardware system, including light source, camera and camera lens.
Described light source, specifically: light source is different according to its lighting system, can be divided into illumination and front illumination dorsad, and backlight usually obtains higher contrast, and front illumination can obtain the surface information of detection object.In order to preferably extract the surface character of workpiece, native system uses front illumination.Owing to cage guide surface contrast is low, characteristic point is few, use laser instrument by Grid Projection to surface of the work to be measured herein, make this surface possess the recognizable texture and feature determined.
Described camera, specifically: native system uses collection with low cost, principle simple CCD industrial camera carries out image.Binocular vision system has two kinds of image pickup modes, is respectively three-dimensional parallel model and three-dimensional convergence pattern.If two ccd video cameras are according to the parallel installation of optical axis, then constitute three-dimensional parallel model；If two video cameras, install according to two optical axises are angled so that two optical axises converge on target object simultaneously, then constitute three-dimensional convergence pattern.That three-dimensional parallel model is demarcated is convenient, calculate simplicity, but for the less video camera in visual angle, when the spacing of two video cameras, i.e. baseline are certain and object distance video camera is certain, and video camera cannot obtain object complete image and cause blind area simultaneously.But three-dimensional convergence pattern binocular camera, can be by adjusting the angle between two optical axises so that two video cameras converge in target, dead zone-eliminating, and then obtain high-quality image simultaneously.Therefore, about native system, two video cameras use three-dimensional convergence pattern to place.
Described camera lens, specifically: the basic optical performance of camera lens has focal length, resolution and the depth of field etc..Wherein, focal length determines that the topmost parameter of camera lens, and the computing formula of focal distance f is:
In formula, v is the target surface size of CCD chip；V is field size；D is operating distance.
In native system, the size of cage guide is 133mm × 85mm, is field size；Image sensor size is 5.632mm × 5.632mm, is chip target surface size；Operating distance is 300mm.The diagonal line value utilizing field size and chip size substitutes into formula (1), and obtaining focal length is 15.13mm, and therefore, native system selects the LM16JC10M camera lens that KAWA company produces, and this lens focus is 16 ± 3mm.
Described binocular vision software system, calculates and form and position tolerance calculating with characteristic matching, characteristic point coordinate including camera calibration, feature extraction.
Described camera calibration, specifically uses the Jean-Yves Bouguet camera calibration workbox on MATLAB platform.In binocular solid calibration process, left video camera projection matrix M1 is:
On the basis of left video camera, the rightest video camera projection matrix M2 is:
Assuming the most successfully to determine in two width images, picture point A1 and picture point A2 are the corresponding point in two images of 1 A in space.Simultaneously, it is also assumed that obtained projection matrix M1 and M2 of two video cameras by camera calibration.Then have:
In formula (4) and formula (5), (u1,v1, 1) and (u2,v2, 1) and it is respectively A1 and A2 homogeneous coordinates in the picture；(X, Y, Z, 1) is A point homogeneous coordinates in world coordinate system；It is respectively MkThe i-th row jth column element.By the Z in formula (4) and formula (5)C1And ZC2Eliminate, obtain four linear equation about X, Y, Z:
Can be obtained by analytical geometry knowledge, the plane equation in space is linear equation, and two plane equation simultaneous are the equation of the two plane intersection line, then the geometric meaning of formula (6) and formula (7) is through O1A1 with through the straight line of O2A2.Again because A is the intersection point of O1A1 and O2A2, necessarily meet formula (6) and formula (7) simultaneously.Therefore, the three-dimensional coordinate (X, Y, Z) of A point is obtained by simultaneous formula (6) and formula (7).But, the linear equation of these 4 simultaneous only comprises (X, Y, Z) 3 variablees.This is because assume that A1 and A2 is the corresponding point of same point A in space before.In actual applications, due to the existence of data noise, method of least square is generally used to solve X, Y, Z, i.e. the three-dimensional coordinate of spatial point.
Described feature extraction includes with characteristic matching: without the feature extracting and matching of laser network projection picture with there is the feature extracting and matching of laser network projection picture.
The feature extracting and matching of the described picture of projection without laser network, specifically: native system uses SURF algorithm to carry out the feature extracting and matching without laser network projection picture.The basic procedure of concrete SURF algorithm is as shown in Figure 1.First experimental image is gathered, and on the basis of original image, set up integral image, asked for the response image of Hessian matrix determinant by different size box Filtering Template integral image, response image uses 3D non-maxima suppression, asks for the speckle of various different scale.Then utilize the integral image formed when feature point detection to generate SURF Feature Descriptor, finally carry out images match.As shown in Figures 2 and 3, for using SURF algorithm to carry out bottom surface and the result of end face characteristic matching respectively.Characteristic of correspondence point circle in two width figures is marked, and is connected with straight line.By SURF algorithm, multipair match point can be obtained.
The described feature extracting and matching having laser network projection picture, specifically: owing to substantial amounts of error hiding can be produced when using SURF algorithm to carry out and have laser network projection images match, therefore, the present invention uses RANSAC algorithm to extract grid search-engine, and then carries out images match.As shown in Figure 4 and Figure 5, for using RANSAC algorithm to carry out the result of grid search-engine extraction.After laser network is extracted, calculate straight line intersection point each other as characteristic point.Owing to laser network has relative defined location relation, it is only necessary to carry out characteristic matching according to the relative position of characteristic point.
Described characteristic point coordinate calculates, and specifically: through images match, can obtain multipair match point.The coordinate of match point is substituted into formula (6) and (7), just can obtain the space coordinates of characteristic point.
Described form and position tolerance calculates and includes that flatness and the depth of parallelism calculate.
Described flatness calculates, specifically: the normal equation of plane is
Ax+By+Cz+D=0 (8)
The coordinate being calculated n (n ＞ 3) individual point is xi、yi、zi, (i=1 ..., n), the distance of each point to plane is
Constructing according to flat normal equation (8) and formula (9) with minor function, this function is the relational expression that least square central plane should meet
The calculating of described flatness, specifically: define according to parallelism tolerance in GB/T1182-2008 national standard, in the face of the definition of datum plane parallelism tolerance band is that tolerance range is that spacing is equal to tolerance value t, is parallel to the region that two parallel planes of datum plane are limited.The normal direction of plane on the basis of direction, size is two plane-parallel spacing.Cage guide bottom surface is as reference element, it is necessary to first reject flatness error.Method of least square Evaluation plane degree error is to meet the assessment criteria of minimal condition method, and the housing face tried to achieve after using the method can regard ideal plane as.After obtaining datum plane, calculate each characteristic point on guide rail end face and, to the distance of datum plane, then obtain the difference of ultimate range and minimum range, just obtain the depth of parallelism of guide rail end face opposite rail bottom surface.
Implementation result
As shown in Figure 6, for system flow chart, according to this flow process, the flatness error of cage guide bottom surface and the end face parallelism error relative to bottom surface just can be obtained.In experiment herein, take pictures to from camera and calculate form and position tolerance, be the most time-consumingly 8.5 seconds.The flatness tolerance of the bottom surface recorded is 0.11, and the result that three coordinate measuring machine records is 0.15, and result is close.End face is 0.032 relative to the parallelism tolerance of bottom surface, and the result that three coordinate measuring machine records is 0.041, and result is close.
The present embodiment, compared to manual detection, has higher detection speed and precision, and has good repeatability, it is to avoid the incidental error that manual measurement brings；Meanwhile, this system has the interface capability good with computer, develops and to improve space huge, can effectively replace manual measurement.
Accompanying drawing explanation
The basic flow sheet of Fig. 1 SURF algorithm；
Fig. 2 SURF algorithm end face matching result；
Fig. 3 SURF algorithm bottom surface matching result；
Fig. 4 RANSAC algorithm end face feature extraction result；
Fig. 5 RANSAC algorithm bottom surface feature extraction result；
Fig. 6 system flow chart.
Claims (7)
1. a cage guide automatic measurement system based on binocular vision, soft including binocular vision hardware system and binocular vision
Part system.
Described binocular vision hardware system, including light source, camera and camera lens.
Described binocular vision software system, calculates with characteristic matching, characteristic point coordinate including camera calibration, feature extraction.
Cage guide automatic measurement system based on binocular vision the most according to claim 1, is characterized in that, described light
Source, specifically: light source is different according to its lighting system, can be divided into illumination and front illumination dorsad, and backlight is usually
Obtain higher contrast, and front illumination can obtain the surface information of detection object.In order to preferably extract the surface of workpiece
Feature, native system uses front illumination.Owing to cage guide surface contrast is low, characteristic point is few, laser instrument is used to be thrown by grid
Shadow, to surface of the work to be measured, makes this surface possess the recognizable texture and feature determined.
Cage guide automatic measurement system based on binocular vision the most according to claim 1, is characterized in that, described phase
Machine, specifically: using collection with low cost, that principle simple CCD industrial camera carries out image, left and right two video camera is adopted
Place with three-dimensional convergence pattern.
Cage guide automatic measurement system based on binocular vision the most according to claim 1, is characterized in that, described phase
Machine is demarcated, specifically: use the Jean-Yves Bouguet camera calibration workbox on MATLAB platform to demarcate.
Cage guide automatic measurement system based on binocular vision the most according to claim 1, is characterized in that, described spy
Levy extraction and characteristic matching, including: without laser network projection picture feature extracting and matching with have laser network projection as
Feature extracting and matching.
Feature extraction the most according to claim 5 and characteristic matching, is characterized in that, the described picture of projection without laser network
Feature extracting and matching, specifically: system use SURF algorithm carry out the feature extraction without laser network projection picture and
Join.First gather experimental image, and on the basis of original image, set up integral image, amassed by different size box Filtering Template
Partial image asks for the response image of Hessian matrix determinant, uses the suppression of 3D non-maximum, ask for various on response image
The speckle of different scale.Then the integral image formed when feature point detection is utilized to generate SURF Feature Descriptor, the most laggard
Row images match.
Feature extraction the most according to claim 5 and characteristic matching, is characterized in that, described has laser network projection picture
Feature extracting and matching, specifically: big owing to using SURF algorithm can produce when carrying out having laser network projection images match
The error hiding of amount, therefore, the present invention uses RANSAC algorithm to extract grid search-engine, and then carries out images match.Right
After laser network extracts, calculate straight line intersection point each other as characteristic point.Owing to laser network has the most true
Fixed position relationship, it is only necessary to carry out characteristic matching according to the relative position of characteristic point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610184490.4A CN105928484B (en) | 2016-03-28 | 2016-03-28 | Cage guide automatic measurement system based on binocular vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610184490.4A CN105928484B (en) | 2016-03-28 | 2016-03-28 | Cage guide automatic measurement system based on binocular vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105928484A true CN105928484A (en) | 2016-09-07 |
| CN105928484B CN105928484B (en) | 2019-05-17 |
Family
ID=56840373
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610184490.4A Active CN105928484B (en) | 2016-03-28 | 2016-03-28 | Cage guide automatic measurement system based on binocular vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105928484B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108458707A (en) * | 2018-01-22 | 2018-08-28 | 西南科技大学 | Work robot autonomic positioning method and its positioning system under more Suspended pipeline scenes |
| CN111381215A (en) * | 2020-03-25 | 2020-07-07 | 中国科学院地质与地球物理研究所 | Phase correction method and meteor position acquisition method |
| CN113343473A (en) * | 2021-06-18 | 2021-09-03 | 广东工业大学 | Type selection method for two-rail goods elevator guide rail |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102003938A (en) * | 2010-10-11 | 2011-04-06 | 中国人民解放军信息工程大学 | Thermal state on-site detection method for large high-temperature forging |
| EP2165915A3 (en) * | 2008-09-23 | 2011-04-13 | VolkerRail Nederland BV | Monitoring a turnout of a railway or tramway line |
| CN103217100A (en) * | 2013-03-29 | 2013-07-24 | 南京工业大学 | Online binocular vision measurement device for carriage of large bus |
| CN104330041A (en) * | 2014-09-30 | 2015-02-04 | 中铁山桥集团有限公司 | Measuring method for track switch steel rail member drill hole dimension |
| CN104408772A (en) * | 2014-11-14 | 2015-03-11 | 江南大学 | Grid projection-based three-dimensional reconstructing method for free-form surface |
| CN105043350A (en) * | 2015-06-25 | 2015-11-11 | 闽江学院 | Binocular vision measuring method |
-
2016
- 2016-03-28 CN CN201610184490.4A patent/CN105928484B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2165915A3 (en) * | 2008-09-23 | 2011-04-13 | VolkerRail Nederland BV | Monitoring a turnout of a railway or tramway line |
| CN102003938A (en) * | 2010-10-11 | 2011-04-06 | 中国人民解放军信息工程大学 | Thermal state on-site detection method for large high-temperature forging |
| CN103217100A (en) * | 2013-03-29 | 2013-07-24 | 南京工业大学 | Online binocular vision measurement device for carriage of large bus |
| CN104330041A (en) * | 2014-09-30 | 2015-02-04 | 中铁山桥集团有限公司 | Measuring method for track switch steel rail member drill hole dimension |
| CN104408772A (en) * | 2014-11-14 | 2015-03-11 | 江南大学 | Grid projection-based three-dimensional reconstructing method for free-form surface |
| CN105043350A (en) * | 2015-06-25 | 2015-11-11 | 闽江学院 | Binocular vision measuring method |
Non-Patent Citations (3)
| Title |
|---|
| 卢选民等: "一种改进的基于SURF的快速图像匹配算法研究", 《敦煌研究》 * |
| 唐坚刚等: "双目立体视觉测量中的特征点快速匹配算法", 《信息技术》 * |
| 汪眩紫: "轨道平整度检测系统的设计与研究", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108458707A (en) * | 2018-01-22 | 2018-08-28 | 西南科技大学 | Work robot autonomic positioning method and its positioning system under more Suspended pipeline scenes |
| CN108458707B (en) * | 2018-01-22 | 2020-03-10 | 西南科技大学 | Autonomous positioning method and positioning system of operating robot in multi-pendulous pipeline scene |
| CN111381215A (en) * | 2020-03-25 | 2020-07-07 | 中国科学院地质与地球物理研究所 | Phase correction method and meteor position acquisition method |
| CN113343473A (en) * | 2021-06-18 | 2021-09-03 | 广东工业大学 | Type selection method for two-rail goods elevator guide rail |
| CN113343473B (en) * | 2021-06-18 | 2022-06-14 | 广东工业大学 | Type selection method for two-rail goods elevator guide rail |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105928484B (en) | 2019-05-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109741405B (en) | Depth information acquisition system based on dual structured light RGB-D camera | |
| CN106595519B (en) | A kind of flexible 3 D contour measuring method and device based on laser MEMS projection | |
| CN110296667B (en) | High-reflection surface three-dimensional measurement method based on line structured light multi-angle projection | |
| CN106595528A (en) | Digital speckle-based telecentric microscopic binocular stereoscopic vision measurement method | |
| CN109341668B (en) | Multi-camera measuring method based on refraction projection model and light beam tracking method | |
| TW201405092A (en) | Three-dimensional image measuring apparatus | |
| CN107816942A (en) | A kind of planar dimension measurement method based on cross structure light vision system | |
| CN112629441A (en) | 3D curved surface glass contour scanning detection method and system based on line structured light dynamic vision | |
| CN108805870A (en) | A kind of detection method of the connector with needle stand | |
| CN113310433A (en) | Virtual binocular stereo vision measuring method based on line structured light | |
| CN105928484B (en) | Cage guide automatic measurement system based on binocular vision | |
| CN112595236A (en) | Measuring device for underwater laser three-dimensional scanning and real-time distance measurement | |
| CN107271445B (en) | Defect detection method and device | |
| WO2018171031A1 (en) | Method for matching feature points of three-camera group, measurement method and three-dimensional detection apparatus | |
| CN110044266B (en) | Photogrammetry system based on speckle projection | |
| CN103234483B (en) | A kind of detection method of parallelism of camera chip and device | |
| CN208012553U (en) | A kind of cylinder inner wall detecting system | |
| CN112556580B (en) | Method, device, system, electronic device and storage medium for measuring three-dimensional size | |
| CN109506562A (en) | A kind of Binocular vision photogrammetry device for the detection of solar wing spreading lock depth | |
| CN108180825B (en) | A kind of identification of cuboid object dimensional and localization method based on line-structured light | |
| CN110657750B (en) | Detection system and method for passivation of cutting edge of cutter | |
| Zhang et al. | Freight train gauge-exceeding detection based on three-dimensional stereo vision measurement | |
| CN209820423U (en) | Auxiliary target screen device for rapid calibration of laser plane space equation | |
| CN112414316B (en) | Strain gauge sensitive grid size parameter measuring method | |
| CN112184790B (en) | Object size high-precision measurement method based on depth camera |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |