CN110400348A - The unmanned vibration equipment of the twin-rotor housing of view-based access control model is rotated to detection, scaling method - Google Patents
The unmanned vibration equipment of the twin-rotor housing of view-based access control model is rotated to detection, scaling method Download PDFInfo
- Publication number
- CN110400348A CN110400348A CN201910554219.9A CN201910554219A CN110400348A CN 110400348 A CN110400348 A CN 110400348A CN 201910554219 A CN201910554219 A CN 201910554219A CN 110400348 A CN110400348 A CN 110400348A
- Authority
- CN
- China
- Prior art keywords
- twin
- steering angle
- rotor housing
- unmanned
- 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.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 72
- 238000000034 method Methods 0.000 title claims abstract description 64
- 238000001914 filtration Methods 0.000 claims abstract description 16
- 230000004438 eyesight Effects 0.000 claims description 28
- 230000008569 process Effects 0.000 claims description 17
- 238000004364 calculation method Methods 0.000 claims description 16
- 238000012937 correction Methods 0.000 claims description 14
- 238000000605 extraction Methods 0.000 claims description 9
- 239000003550 marker Substances 0.000 claims description 9
- 230000000877 morphologic effect Effects 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 238000007689 inspection Methods 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 4
- 230000007797 corrosion Effects 0.000 claims description 3
- 238000005260 corrosion Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000009434 installation Methods 0.000 abstract description 8
- 238000005516 engineering process Methods 0.000 abstract description 4
- 230000016776 visual perception Effects 0.000 abstract description 3
- 238000003384 imaging method Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 239000000284 extract Substances 0.000 description 6
- 238000006243 chemical reaction Methods 0.000 description 5
- 238000010276 construction Methods 0.000 description 5
- 230000000007 visual effect Effects 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 239000002184 metal Substances 0.000 description 2
- 238000012536 packaging technology Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- HUTDUHSNJYTCAR-UHFFFAOYSA-N ancymidol Chemical compound C1=CC(OC)=CC=C1C(O)(C=1C=NC=NC=1)C1CC1 HUTDUHSNJYTCAR-UHFFFAOYSA-N 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000007596 consolidation process Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000005764 inhibitory process Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000010355 oscillation Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000009490 roller compaction Methods 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/80—Geometric correction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—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
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/168—Segmentation; Edge detection involving transform domain methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/66—Analysis of geometric attributes of image moments or centre of gravity
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- 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
- G06T2207/20048—Transform domain processing
- G06T2207/20061—Hough transform
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geometry (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The unmanned vibration equipment of twin-rotor housing that the present invention provides a kind of view-based access control model is rotated to detection, scaling method, using the dedicated three proofings camera of engineering, cooperation is arranged in the color characteristic at the crossbeam of twin-rotor housing unmanned vibration equipment wheel front, the detection of the unmanned vibration equipment wheel steering angle of twin-rotor housing is realized first with visual perception, then numerical filtering and calibration are carried out as the angle to obtained by, improves the stability and precision of steering angle detection.The problems such as this technology is innovated in physical structure and Method And Principle, is avoided in the prior art because of sensor installation, signal fluctuation, and GPS signal the drifts about inaccurate unstable problem of bring twin-rotor housing unmanned equipment steering angle detection.
Description
Technical field
The invention belongs to the unmanned fields of engineering machinery, are related to image processing techniques, especially a kind of view-based access control model
The unmanned vibration equipment of twin-rotor housing is rotated to detection, scaling method.
Background technique
In recent years, being constantly progressive with artificial intelligence and unmanned technology, modern project machinery is all towards automatic
Change, is information-based, intelligent direction fast development.Twin-rotor housing Working machine is focused on building the road to build a dam to play in equal constructions
The effect wanted, but the working environment built a dam of often building the road is very severe, and site operation personnel need to carry out for a long time it is heavy, again
The unmanned equipment driving operation of multiple high-intensitive twin-rotor housing, on the one hand since work persistently dry as dust leads to construction personnel
Working efficiency reduces, and roller compaction construction quality cannot ensure, on the other hand to meet compacting job requirement in construction operation,
Exciting agency generates high vibration can bring the discomfort on body to operator., in order to ensure construction personnel's health, together
When reduce the unmanned equipment of twin-rotor housing operating cost and improve its working efficiency, the unmanned equipment of twin-rotor housing is unmanned
Realization have a very important significance.
By taking road roller as an example.At present in the research of the unmanned technology of road roller, the accurate course changing control of vehicle is one
The very crucial research work of item.Realize that the accurate course changing control of the vehicle of subsequent time must just be detected by particular technology
To the steering angle of current time vehicle, then in conjunction with the information such as the course of default desired trajectory and other sensors, position into
Row, which calculates, to be corrected and provides current track.Therefore the detection of the steering angle of current time vehicle is most important.
For three-wheel or four-wheel car, the steering angle for having the front-wheel of turning function is exactly required steering angle
Degree, but for articulated steering structure this for road roller, preceding vibration with aftercarriage is connected by central hinged cross axle
It connects.Therefore the vehicle steering angle detected here is the steering angle of vibrating wheels before road roller.
To the steering angle of vibrating wheels before road roller, there are mainly two types of methods in the prior art.
One is mechanical angle measurement method, by before and after road roller car body hinged place cooperated using angular transducer
Specific mounting bracket and link mechanism, link mechanism one end connect angular transducer, and the other end passes through particular stent and vibration
Surrounding crossbeam consolidation is taken turns, wherein angular transducer is installed on hinged place by bracket, and hinged place and aftercarriage are rigidly connected.Pressure
Road machine is in steering procedure, so that it may the steering angle of vibrating wheels is provided by angular transducer.But this mechanical angle is surveyed
Amount method is highly prone to angular transducer installation error, the installation of mechanical structure be not connected firmly lean on, link mechanism and front vehicle body
Be consolidated with sensor internal precision components damage under the conditions of gap, long-term work, sensor signal fluctuation the problems such as it is dry
It disturbs, causes the detection of vibrating wheels steering angle inaccurate.
Another kind is GPS angle calculation method, and the orientation and positioning of vehicle may be implemented in one group of double antenna GPS, therefore makes
With two groups of double antenna GPS vibrating wheels corner can be obtained by resolving two groups of course angles.Specific implementation method is to press first
An antenna GPS is respectively placed in car body corresponding position before and after the machine of road, forms one group of double antenna GPS, in this way can when road roller turns to
To obtain course angle S1, direction along two GPS antennas of front and back car body line.Then longitudinally total on road roller aftercarriage
Line arranges one group of double GPS antenna, and an available course angle S2, no matter when all longitudinally conllinear with aftercarriage direction is, this
Sample can be the steering angle of current time vibrating wheels by the angle in the resolving direction S1 and the direction S2.This method is at high cost
It is high, it needs to arrange two groups of double antenna GPS, is also highly prone to the influence of the installation site of GPS antenna, and GPS itself is easy
The problems such as existing satellite signal receiving is unstable, signal fluctuation, drift can all cause the detection of vibrating wheels steering angle inaccurate.
Summary of the invention
It is an object of the invention in place of overcome the deficiencies in the prior art, provide a kind of twin-rotor housing of view-based access control model nobody
The method that steer vibrating wheels turn to detection and its calibration, by pattern distortion correction, pretreatment and feature extraction, in conjunction with filter
Wave algorithm can stablize the steering detection for realizing the unmanned vibration equipment wheel of twin-rotor housing, then by under the record true corner of multiple groups
Vision measurement value for making calibration scale, realized using interpolation method in the unmanned vibration equipment wheel of twin-rotor housing according to calibration scale
When steering, available calibrated steering angle compensates for precision problem existing for vision-based detection.
The technical proposal for solving the technical problem of the invention is:
A kind of unmanned vibration equipment wheel steering detection method of the twin-rotor housing of view-based access control model, steps are as follows:
(1) original image acquires;
(2) region of interest ROI is extracted;
(3) distortion correction is carried out to extraction ROI region;
(4) according to color Threshold segmentation characteristic area;
(5) morphological image process;
(6) Hough loop truss or feature contour detection;
(7) if previous step is Hough loop truss, vibrating wheels steering angle is calculated;If previous step is feature contour
Detection then first calculates feature contour mass center and calculates vibrating wheels steering angle again;
(8) steering angle numerical filtering;
(9) stablize output steering angle.
Moreover, the steering angle that step (9) obtains is demarcated, calibrated steering angle is obtained.
Moreover, the calibration is to acquire true steering angle according to the hinged angle transducer of the unmanned equipment of true twin-rotor housing
Degree establishes true steering angle and the mapping relations between steering angle that step (9) obtains, and formation calibration scale looks into calibration scale
Numerical intervals are looked for, realize geometric calibration using data interpolations.
Moreover, the interpolation method be define vision-based detection go out vibrating wheels steering angle be α, note temp_ α be it is to be calibrated in
The area of a room, wherein α=α/5 temp_, then execute downward floor operation to temp_ α, obtain one and represent less than or be equal to temp_
The maximum integer of α is denoted as insert_A, while remembering insert_B=insert_A+1, at this time [insert_A*5, insert_B*
5] interpolation section where being steering angle α that vision-based detection provides, remembers that corresponding f (insert_A*5) is insert_A*5 correspondence
The unmanned equipment steering angle numerical value of true twin-rotor housing, corresponding f (insert_B*5) is insert_B*5 corresponding true double
The unmanned equipment steering angle numerical value of cylinder body carries out interpolation operation, the following institute of formula according to this interpolation section and corresponding true value
Show:
Moreover, the acquisition of original image described in step (1) is that industrial camera is fixed on to the unmanned equipment vehicle of twin-rotor housing
Body transverse direction center position carries out shockproof processing using shockproof hard spring buffer or setting Shockproof rubber gasket.
Moreover, carrying out the inner parameter and outside that distortion correction is combining camera to extraction ROI region described in step (3)
Parameter realizes that the distortion correction of camera is really schemed using formula img (U, V)=imgD (Ud, Vd) from fault image imgD
As img.
Moreover, morphological image process described in step (5) be unexpected characteristic area is first eliminated using corrosion treatment, then
The desired character region left using expansion process.
Moreover, the method for calculating vibrating wheels steering angle by Hough loop truss are as follows: in the unmanned vibration equipment of twin-rotor housing
It is consistent with beam width that along the unmanned equipment longitudinal midline of twin-rotor housing two length and width dimensions have been arranged symmetrically on wheel front beam
Circle marker object, two maximum characteristic circles of symmetrical size are obtained by Hough loop truss, according to two center of circle image coordinate,
The slope for calculating two center of circle image coordinate lines, is denoted as k1, takes image x direction horizontal axis as reference line, remembers with reference to line slope
For k2, k2 0, θ are the angle of two lines, and calculation formula is as follows:
Steering angle when the unmanned vibration equipment wheel of twin-rotor housing does not turn to is set as 90 °, what vision was calculated
Vibrating wheels steering angle α is calculated by following formula:
+ 90 ° of α=θ.
Moreover, detecting the method for calculating vibrating wheels steering angle by feature contour are as follows: in the unmanned equipment vibration of twin-rotor housing
Two length and width dimensions and beam width one have been arranged symmetrically along the unmanned equipment longitudinal midline of twin-rotor housing on driving wheel front beam
The marker of cause is first found out the marker profile information for including in image, is then traversed to all profiles, and calculates each
The image moment of a profile obtains the centroid position (Cx, Cy) of profile:
Wherein Cx is the abscissa of profile mass center on the image;Cy is the ordinate of profile mass center on the image;M00 is 0
Rank image moment represents profile area encompassed area;M10 represents the 1 rank image moment of the point on profile in the x direction;M01 generation
The 1 rank image moment of point in y-direction on table skeleton;
The slope for going out two center-of-mass coordinate lines according to two feature contour centroid calculations for stablizing output, is denoted as k1, here
It takes image x direction horizontal axis as reference line, is denoted as k2, k2 0 with reference to line slope, remember that θ is the angle of two lines here, calculate public
Formula is as follows:
Steering angle when the unmanned vibration equipment wheel of twin-rotor housing does not turn to is set as 90 °, then vision is calculated
Vibrating wheels steering angle α can be calculated by following formula:
+ 90 ° of α=θ.
Moreover, the steering angle numerical filtering is using recurrence average filter method.
The advantages and positive effects of the present invention are:
The present invention realizes that the unmanned vibration equipment of twin-rotor housing is rotated to detection using visual perception.Compared with prior art,
The present invention goes to implement steering angle detection from a kind of completely new angle.Specifically, using industrial three proofings (waterproof, shockproof, dust-proof)
Vision camera can ensure the acquisition of image stabilization, in physical structure and receiving side signal face not by precision of equipment installation, satellite
Dropout, signal fluctuation and drift etc. influence;In terms of algorithm process, feature is detected by vision algorithm, it can be tentatively real
Existing reliable and stable detection is detected using high-precision steering angle may be implemented after filtering and calibration.To sum up, the present invention exists
It is innovated on physical structure and Method And Principle, effectively reduces the equipment cost of engineering developme, avoid in the prior art
Not because of the unmanned equipment steering angle detection of the problems such as sensor is installed, signal fluctuation, and GPS signal drifts about bring twin-rotor housing
Accurate unstable problem, brings great convenience for unmanned engineering development.
Detailed description of the invention
Fig. 1 is the detection method flow diagram in the embodiment of the present invention 1;
Fig. 2 is the detection method flow diagram in the embodiment of the present invention 2;
Fig. 3 is the scaling method flow diagram in the embodiment of the present invention 3;
Fig. 4 is functional module connection figure of the present invention.
Specific embodiment
For the purposes, technical schemes and advantages of the disclosure are more clearly understood, below in conjunction with specific embodiment, and reference
Attached drawing is described in further detail the disclosure.It is understood that specific embodiment described herein is used only for solving
Release the present invention, rather than limitation of the invention.It also should be noted that illustrate only for ease of description, in attached drawing with
The relevant part of the present invention rather than entire infrastructure.
Embodiment 1
Fig. 1, which is that a kind of unmanned road roller twin-rotor housing for view-based access control model that the embodiment of the present invention 1 provides is unmanned, to be set
Standby vibrating wheels steering detection method flow diagram, this method can be by the unmanned vibration equipment wheel of twin-rotor housing of view-based access control model
Detection device is turned to execute.This method specifically includes as follows:
1, original image acquires.
Image capture device is at least arranged a set of using industrial camera, is fixed on road roller cross-car center position.By
Vibration can be brought in the motion process of road roller, is prevented using shockproof hard spring buffer or setting Shockproof rubber gasket
Shake processing.The acquisition angles of image capture device are adjusted, guarantee vibration wheel of road roller and front beam in the feelings of the left-right rotation limit
Under condition, remain to all clearly be shown in visual field.
2, area-of-interest (ROI) extracts.
The selection of area-of-interest (Region of Interest, ROI), general there are two types of situations: ROI known to a) is scheming
Position as in;B) Location-Unknown of ROI in the picture.Position known case is belonged to for ROI according to the present invention extraction,
Here we extract ROI for the collected original image of institute in previous step.For an image of acquisition, generally to scheme
The picture upper left corner is coordinate origin, is established image pixel coordinates system (uOv coordinate system), and note image level direction is u direction, remembers image
Vertical direction is the direction v.Set picture size: for w as the width of image, h is the height of image.Then in a coordinate system, a sub-picture
Four angular coordinate is (0,0), (w, 0), (w, h), (0, h).
It is surrounded since the real structure of road roller is vibrating wheels periphery by a circle metal structure, in this circle metal structure just
Front is a crossbeam.And the steering of crossbeam and the steering of vibrating wheels are consistent.The principle that we extract ROI is: finding
One guarantee vibration wheel of road roller front beam is under left and right turn limiting case, and feature to be detected can also completely be shown on crossbeam
Region as ROI.Providing ROI region width under this principle is w ', a height of h ', 4 coordinates are (0,0), (w ', 0),
(w',h'),(0,h').Extracting ROI method is directly to cut to original image according to ROI four angular coordinate, and then obtain me
The area-of-interest (ROI) that needs.
3, distortion correction is carried out to extraction ROI region.
The imaging process of image collecting device such as camera etc. is substantially the conversion of several coordinate systems.It first will be in space
Point is transformed into " camera coordinates system " by " world coordinate system ", is then projected into imaging plane, namely " image physical coordinates again
Data on imaging plane are finally transformed into the plane of delineation, namely " image pixel coordinates system " again by system ".But due to lens
The accuracy of manufacture and the deviation of packaging technology can introduce distortion, lead to the distortion of original image.
The distortion of camera lens is divided into two class of radial distortion and tangential distortion.Radial distortion is along lens radius directional spreding
Distortion, producing cause are light in the local than being more bent by paracentral place of separate lens centre, and this distortion is general
Lead in cheap camera lens show it is more obvious.Tangential distortion be due to lens itself and camera sensor plane (imaging plane) or
The plane of delineation it is not parallel and generate, such case mostly be that the installation deviation being adhered in lens module due to lens is caused.
The principle of distortion correction: the undistorted coordinate (U, V) under conventional images pixel coordinate system (uOv coordinate system) is described in detail,
It has been fallen in after radial distortion and tangential distortion on (Ud, the Vd) of uOv coordinate system.I.e. that is, true picture img and distortion
Relationship between image imgD are as follows: img (U, V)=imgD (Ud, Vd).
Our obtained ROI regions of previous step are the images after distortion, and the image not distorted will be by abnormal
Varying model derives its mapping relations.Relationship between true figure img and fault image imgD are as follows: img (U, V)=imgD (Ud,
Vd).All img (U, V) can be found out by this relationship.During this, need such as to open by camera calibration method
Positive friend's standardization obtains the inner parameter of camera, and the installation site of combining camera obtains the external parameter of camera.And then combine phase
The inner parameter and external parameter of machine realize the distortion correction of camera.
Because U and V are integers for piece image, because U and V represents the pixel coordinate position of composition image.
During correction is orthoscopic image img (U, V) from the image imgD (Ud, Vd) after distortion, can be calculated (Ud,
Vd) the case where being frequently not integer, it is therefore desirable to be solved using interpolation method, such as arest neighbors interpolation and bilinear interpolation method
Deng achievable.
4, according to color Threshold segmentation characteristic area.
In the present invention, two length and width dimensions and the consistent red marker object of beam width are mainly had chosen, along pressure road
Machine crossbeam middle line is symmetrically arranged on crossbeam.Since road roller intrinsic colour is yellow, select here red as feature face
Color.It is also an option that other representational colors, such as white, blue etc. other than red.Here to characteristic color
It is required that are as follows: there is obvious differentiation with road roller and its component intrinsic colour;To have obviously with road environment color around road roller
It distinguishes.Vision carries out color space conversion when detecting, to the image of image acquisition device first, by rgb color space
It is transformed into HSV color space.Again by the red color threshold value under setting HSV, and then red area is extracted, and to non-red color area
Domain pixel value is set to 0.This completes the purposes that characteristic area is split according to color threshold value.
5, morphological image process.
It, can be due to the surrounding working environment of road roller in the image that previous step obtains in the image processing process of this example
And the influence of natural lighting etc., lead to will appear a series of unexpected features being extracted, therefore we need to carry out figure
As Morphological scale-space, unexpected characteristic area is eliminated using first corrosion, image shape is realized in the desired character region that reflation leaves
State opening operation reaches smooth features region contour, goes the effect of the unwanted areas such as flash removed.
6, Hough loop truss.
As described in preceding 4, it has been arranged symmetrically on vibration wheel of road roller front beam along road roller longitudinal midline two long
Wide size and beam width consistent red marker object use the scheme of Hough loop truss in this example, therefore it is circle that marker, which is arranged,
Shape red marker object, and circular diameter is not more than road roller beam width.The image that image acquisition device arrives passes through aforementioned step
Suddenly, the bianry image of available only circular feature, and then carry out Hough loop truss, after adjusting parameter available two it is right
The maximum characteristic circle of the size of title realizes feature detection.Ginseng process this example is adjusted not repeat.
7, vibrating wheels steering angle is calculated.
According to the Hough loop truss of previous step, two characteristic circles of available stable output.According to Hough loop truss
Output, can directly obtain the image coordinate in the center of circle of two characteristic circles.According to two center of circle image coordinate, two circles can be calculated
The slope of heart image coordinate line, is denoted as k1, takes image x direction horizontal axis as reference line here, is denoted as k2 with reference to line slope, shows
Right k2 is 0.Here note θ is the angle of two lines, and calculation formula is as follows:
Cambered value again carried out after arctangent computation to tan θ be converted to angle value to can be obtained the numerical value of θ, setting is pressed
Steering angle when road machine vibration wheel does not turn to is 90 °, and maximum steering range is positive and negative 35 ° namely 55 °~125 °.Note is left
Switch to " bear ", turn right as " just ".The vibrating wheels steering angle α that then vision is calculated can be calculated by following formula:
+ 90 ° of α=θ
8, steering angle numerical filtering.
After the vision-based detection of abovementioned steps, the vibrating wheels steering angle that can be detected, but the angle of visual feedback
Value itself is obtained by the angle calcu-lation for being then based on the circle center line connecting that Hough loop truss provides and image X-direction.And Hough circle
Detection can have a series of unstable detections, can bring about the fluctuation of output angle angle value, it is therefore necessary to using numerical filtering
Mode just can guarantee the stability of angle output.The present embodiment uses recurrence average filter method (also known as moving average filter
Method) Lai Shixian vibrating wheels steering angle stablize output.
Detailed description are as follows for recurrence average filter method, and the steering angle that visual perception continues output is considered as and is continuously adopted
The N number of sampled value continuously obtained is regarded as a queue by sample value, and the length of queue is fixed as N, samples a new data every time
It is put into tail of the queue, and throws away a data (first in first out) of original head of the queue, N number of data in queue are carried out arithmetic average
Operation, obtaining new filter result can be used.The advantages of this filtering method, is there is good inhibition to PERIODIC INTERFERENCE
Effect, smoothness is high, the system suitable for the higher-order of oscillation.
According to the realization of above-mentioned steps, the unmanned vibration equipment of twin-rotor housing that view-based access control model may be implemented in this example rotate to
Detection.
Embodiment 2
Fig. 2 is that a kind of unmanned vibration equipment of twin-rotor housing for view-based access control model that the embodiment of the present invention 2 provides is rotated to inspection
Method flow schematic diagram is surveyed, the technical solution of the present embodiment is based on above-described embodiment, further, to step 6 Hough circle
Detection calculates vibrating wheels steering angle with step 7 and is improved and optimized.This method specifically includes as follows:
1, original image acquires.
2, area-of-interest (ROI) extracts.
3, distortion correction is carried out to extraction ROI region.
4, according to color Threshold segmentation characteristic area.
5, morphological image process.
6, feature contour detects.
Profile is the characteristic feature of image object, can simply be considered to connect together continuous point (boundary of ining succession)
Curve, color or gray scale having the same.Profile is very useful in shape analysis and the detection of object and identification.Feature contour
For more accurate contour identification during detection, it is necessary to carry out binary conversion treatment to original image, be connected using marginal point
Level difference, extract and be located at the set that the high region point set of structure feature is constituted, this part point set is probably object
Profile.The output of feature contour detection is a list, contains all profiles detected, each profile is by series of points group
At.
Feature contour detection is more stable encirclement character shape compared to the advantage of Hough loop truss, because camera
Installation shooting angle makes the circular feature being arranged on road roller crossbeam be imaged as ellipse, this results in carrying out Hough circle
When detection, it is difficult to which the shake for bringing angle to detect is being shaken always in stable encirclement character shape, the Hough loop truss center of circle.And make
It with feature detection, is not just influenced by feature shape itself, the result of output contour detecting that can be stable.It also implies that
Character shape can also need not be circle, can be set to arbitrary shape.It may also can go out in feature detection process simultaneously
The profile of existing some missing inspections, this can be filtered by the way that closed outline region area is arranged.To sum up, stable wheel may be implemented
Exterior feature detection.
7, feature contour centroid calculation.
In image procossing, the related fieldss such as computer vision, image moment is some particular weights of image pixel intensity
Average value.The profile information for including in image is found out by step 6, each profile that this example detects is an enclosed area
Domain.Then all profiles are traversed, and calculates the image moment of each profile, so that it may obtain the centroid position of object
(Cx, Cy):
Wherein Cx is the abscissa of profile mass center on the image;Cy is the ordinate of profile mass center on the image;M00 is 0
Rank image moment represents profile area encompassed area;M10 represents the 1 rank image moment of the point on profile in the x direction;M01 generation
The 1 rank image moment of point in y-direction on table skeleton.According to this method, available extracted feature contour mass center, and
It can show on the image, to subsequent calculating mass center line.
8, vibrating wheels steering angle is calculated.
According to the detection of the feature contour of previous step and centroid calculation, two feature contours and phase of available stable output
Answer the mass center of profile.The slope of two center-of-mass coordinate lines, note can be calculated according to two feature contour mass centers for stablizing output
For k1, takes image x direction horizontal axis as reference line here, be denoted as k2 with reference to line slope, it is clear that k2 0.Here note θ is two lines
Angle, calculation formula is as follows:
Cambered value again carried out after arctangent computation to tan θ be converted to angle value to can be obtained the numerical value of θ, setting is pressed
Steering angle when road machine vibration wheel does not turn to is 90 °, and maximum steering range is positive and negative 35 ° namely 55 °~125 °.Note is left
Switch to " bear ", turn right as " just ".The vibrating wheels steering angle α that then vision is calculated can be calculated by following formula:
+ 90 ° of α=θ
9, steering angle numerical filtering.
After the vision-based detection of abovementioned steps, the vibrating wheels steering angle that can be detected, but the angle of visual feedback
What the mass center line of the feature contour of value Detection and Extraction by being then based on itself and the angle calcu-lation of image X-direction obtained.Though
Relatively stable numerical value output has may be implemented in right feature detection, it is contemplated that the reliability of enhancing data, still
Numerical filtering is carried out to visual output angle.The present embodiment comes using recurrence average filter method (also known as moving average filter method)
That realizes vibrating wheels steering angle stablizes output.
According to the realization of above-mentioned steps, the unmanned vibration equipment of twin-rotor housing for the realization view-based access control model that this example can be stable
It rotates to detection.
Embodiment 3
Fig. 3 is a kind of unmanned vibration equipment wheel direction indicator of twin-rotor housing for view-based access control model that the embodiment of the present invention 3 provides
Determine method flow schematic diagram, this method can be rotated by the unmanned vibration equipment of twin-rotor housing of view-based access control model to caliberating device Lai
It executes, which can be realized in the form of hardware and/or software.This method specifically includes as follows:
1, calibration scale is made according to setting corner magnitude and vision actual measurement corner.
The actual rotational angle range of vibration wheel of road roller is that have physical limit, if being denoted as middle position when not rotating with vibrating wheels
Corresponding 90 °, then the limiting value of left and right turn is respectively 55 ° and 125 °, therefore the present embodiment when making calibration scale with regard to foundation
This angle range, the angle provided by vision-based detection, manual operation road roller turns to, steady in vision-based detection steering angle respectively
It is scheduled on 55 °, 60 °, 65 °, 70 °, 75 °, 80 °, 85 °, 90 °, 95 °, 100 °, 105 °, 110 °, 115 °, 120 °, 125 ° of these angles
When neighbouring, the true corner value of measurement vibration wheel of road roller at this time is gone using angle measurement tool, and record respectively.Form a system
Corresponding angle is arranged, production calibration scale is completed.
2, it inputs vision-based detection vibration wheel of road roller steering angle and reads calibration scale storage calibration value.
Output valve is turned to according to the vibrating wheels of embodiment 1 and the available vision-based detection of embodiment 2, this angle value is made
For the input of calibration.Wherein calibration scale can store as specific format, such as .xml .json .yml etc..In calibration process
Read calibration scale.
3, interpolation method, which is realized, demarcates and exports calibrated corner magnitude.
The calibration scale read is combined to realize interpolation method calibration, the following institute of process of interpolation method calibration according to the angle value of input
It states.First according to the setting of calibration scale, be divided into 5 ° between two calibration values, therefore in practical Interpolation Process, according to embodiment 1 with
Embodiment 2, it is α that we, which define the vibrating wheels steering angle that vision-based detection goes out, and note temp_ α is intermediate quantity to be calibrated, wherein temp_ α
Then=α/5 execute downward floor operation to temp_ α, available one represents less than or whole equal to the maximum of temp_ α
Number, is denoted as insert_A, while remembering insert_B=insert_A+1, and [insert_A*5, insert_B*5] is to regard at this time
Feel the place steering angle α detected and provided in interpolation section.Remember that corresponding f (insert_A*5) is insert_A*5 corresponding true
Road roller steering angle numerical value, corresponding f (insert_B*5) they are the corresponding true road roller steering angle numerical value of insert_B*5, according to
This interpolation section and corresponding true value, so that it may carry out interpolation operation, formula is as follows:
According to the above method, vision-based detection road roller steering angle calibrating function may be implemented.
Embodiment 4
The method of the present invention can include Image Acquisition by sequentially connected multiple Implement of Function Module, the functional module
Module, pattern distortion rectification module, detection processing module, feature calculation module, numerical filtering module and data demarcating module.
Image capture module is mainly used for providing the initial data source for carrying out vision-based detection.It is set by adjusting Image Acquisition
Standby acquisition angles guarantee vibration wheel of road roller and front beam in the case where the left-right rotation limit, remain to all clear displays
In in visual field.And realize the function that image is sent to subsequent module.
Pattern distortion rectification module is mainly used for correcting the pattern distortion from acquisition device input.Image collecting device is such as
The imaging process of camera etc. is substantially the conversion of several coordinate systems.The point in space is transformed by " world coordinate system " first
" camera coordinates system " is then projected into imaging plane, namely " image physical coordinates system " again, finally again will be on imaging plane
Data be transformed into the plane of delineation, namely " image pixel coordinates system ".But due to the lens accuracy of manufacture and packaging technology
Deviation can introduce distortion, lead to the distortion of original image.Therefore pattern distortion rectification module utilizes the inside and outside of image collecting device
The distortion correction of parameter combination coordinate system conversion formula realization image pixel.
Detection processing module is extracted, edge is examined mainly for the treatment of the image after distortion correction by color characteristic threshold value
The image processing methods such as survey, Morphological scale-space, contours extract are applied in combination, and finally obtain desired contour feature stablizes inspection
It surveys, provides initial data for subsequent feature calculation.
Feature calculation module and numerical filtering module are mainly used for the initial data exported according to detection processing module, carry out
Extraneous features based on contour area are rejected, and characteristic point line slope calculates, steering angle angle calculation etc., and use filtering algorithm
(such as recursion mean filter, Kalman filtering and its variant etc.) carries out numerical filtering, removes because erroneous detection, shake bring numerical value are jumped
Become, realizes the stable detection and output of steering angle.
Data scaling module is mainly used for finding vision-based detection steering angle value pass corresponding with the true corner value of road roller
System determines calibration relationship, the numerical value calibration of vision-based detection steering angle, Ke Yishi is realized in conjunction with interpolation method by making calibration scale
The now stable reliable true steering angle value of road roller of output.
What has been described above is only a preferred embodiment of the present invention, it is noted that for those of ordinary skill in the art
For, under the premise of not departing from inventive concept, various modifications and improvements can be made, these belong to protection of the invention
Range.
Claims (10)
1. a kind of unmanned vibration equipment wheel steering detection method of the twin-rotor housing of view-based access control model, it is characterised in that: steps are as follows:
(1) original image acquires;
(2) region of interest ROI is extracted;
(3) distortion correction is carried out to extraction ROI region;
(4) according to color Threshold segmentation characteristic area;
(5) morphological image process;
(6) Hough loop truss or feature contour detection;
(7) if previous step is Hough loop truss, vibrating wheels steering angle is calculated;If previous step is feature contour inspection
It surveys, then first calculates feature contour mass center and calculate vibrating wheels steering angle again;
(8) steering angle numerical filtering;
(9) stablize output steering angle.
2. the unmanned vibration equipment wheel steering detection method of the twin-rotor housing of view-based access control model according to claim 1, special
Sign is: the steering angle that step (9) obtains being demarcated, calibrated steering angle is obtained.
3. the unmanned vibration equipment wheel steering detection method of the twin-rotor housing of view-based access control model according to claim 2, special
Sign is: the calibration is to acquire true steering angle according to the hinged angle transducer of the unmanned equipment of true twin-rotor housing, is built
The mapping relations that true steering angle is found between steering angle that step (9) obtains form calibration scale, search numerical value to calibration scale
Geometric calibration is realized using data interpolations in section.
4. the unmanned vibration equipment wheel steering detection method of the twin-rotor housing of view-based access control model according to claim 3, special
Sign is: the interpolation method be define vision-based detection go out vibrating wheels steering angle be α, note temp_ α be intermediate quantity to be calibrated,
Then wherein α=α/5 temp_ execute downward floor operation to temp_ α, obtain one and represent less than or most equal to temp_ α
Big integer is denoted as insert_A, while remembering insert_B=insert_A+1, and [insert_A*5, insert_B*5] i.e. at this time
Interpolation section where the steering angle α provided for vision-based detection remembers that corresponding f (insert_A*5) is insert_A*5 corresponding true
The real unmanned equipment steering angle numerical value of twin-rotor housing, corresponding f (insert_B*5) are the corresponding true twin-rotor housing of insert_B*5
Unmanned equipment steering angle numerical value carries out interpolation operation according to this interpolation section and corresponding true value, and formula is as follows:
5. the unmanned vibration equipment wheel steering detection method of the twin-rotor housing of view-based access control model according to claim 1, special
Sign is: the acquisition of original image described in step (1) is that industrial camera is fixed on to the unmanned equipment cross-car of twin-rotor housing
Center position carries out shockproof processing using shockproof hard spring buffer or setting Shockproof rubber gasket.
6. the unmanned vibration equipment wheel steering detection method of the twin-rotor housing of view-based access control model according to claim 1, special
Sign is: carrying out the inner parameter and external parameter reality that distortion correction is combining camera to extraction ROI region described in step (3)
The distortion correction of existing camera obtains true picture img from fault image imgD using formula img (U, V)=imgD (Ud, Vd).
7. the unmanned vibration equipment wheel steering detection method of the twin-rotor housing of view-based access control model according to claim 1, special
Sign is: morphological image process described in step (5) is first to eliminate unexpected characteristic area using corrosion treatment, then use swollen
It is swollen to handle the desired character region left.
8. the unmanned vibration equipment wheel steering detection method of the twin-rotor housing of view-based access control model according to claim 1, special
Sign is: the method for calculating vibrating wheels steering angle by Hough loop truss are as follows: horizontal before the unmanned vibration equipment wheel of twin-rotor housing
Two length and width dimensions and the consistent circle of beam width have been arranged symmetrically along the unmanned equipment longitudinal midline of twin-rotor housing on beam
Marker obtains two maximum characteristic circles of symmetrical size by Hough loop truss and is calculated according to two center of circle image coordinate
The slope of two center of circle image coordinate lines, is denoted as k1, takes image x direction horizontal axis as reference line, is denoted as k2, k2 with reference to line slope
It is the angle of two lines for 0, θ, calculation formula is as follows:
Steering angle when the unmanned vibration equipment wheel of twin-rotor housing does not turn to is set as 90 °, the vibration that vision is calculated
Wheel steering angle α is calculated by following formula:
+ 90 ° of α=θ.
9. the unmanned vibration equipment wheel steering detection method of the twin-rotor housing of view-based access control model according to claim 1, special
Sign is: detecting the method for calculating vibrating wheels steering angle by feature contour are as follows: before the unmanned vibration equipment wheel of twin-rotor housing
Two length and width dimensions and the consistent mark of beam width have been arranged symmetrically along the unmanned equipment longitudinal midline of twin-rotor housing on crossbeam
Will object is first found out the marker profile information for including in image, is then traversed to all profiles, and calculates each profile
Image moment, obtain the centroid position (Cx, Cy) of profile:
Wherein Cx is the abscissa of profile mass center on the image;Cy is the ordinate of profile mass center on the image;M00 is 0 rank figure
As square, profile area encompassed area is represented;M10 represents the 1 rank image moment of the point on profile in the x direction;M01 represents wheel
The 1 rank image moment of point in y-direction on exterior feature;
The slope for going out two center-of-mass coordinate lines according to two feature contour centroid calculations for stablizing output, is denoted as k1, takes figure here
The picture direction x horizontal axis is denoted as k2, k2 0 with reference to line slope as reference line, remembers that θ is the angle of two lines here, calculation formula is such as
Under:
Steering angle when the unmanned vibration equipment wheel of twin-rotor housing does not turn to is set as 90 °, then the vibration that vision is calculated
Driving wheel steering angle α can be calculated by following formula:
+ 90 ° of α=θ.
10. the unmanned vibration equipment wheel steering detection method of the twin-rotor housing of view-based access control model according to claim 1, special
Sign is: the steering angle numerical filtering is using recurrence average filter method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910554219.9A CN110400348B (en) | 2019-06-25 | 2019-06-25 | Method for detecting and calibrating steering of vibrating wheel of double-cylinder unmanned equipment based on vision |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910554219.9A CN110400348B (en) | 2019-06-25 | 2019-06-25 | Method for detecting and calibrating steering of vibrating wheel of double-cylinder unmanned equipment based on vision |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110400348A true CN110400348A (en) | 2019-11-01 |
CN110400348B CN110400348B (en) | 2022-12-06 |
Family
ID=68323519
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910554219.9A Active CN110400348B (en) | 2019-06-25 | 2019-06-25 | Method for detecting and calibrating steering of vibrating wheel of double-cylinder unmanned equipment based on vision |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110400348B (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103255755A (en) * | 2013-04-28 | 2013-08-21 | 河海大学 | Lossless method for fast evaluating filling compaction quality of soil building stones in real time and evaluating device thereof |
CN103321128A (en) * | 2013-07-03 | 2013-09-25 | 中联重科股份有限公司 | Method, system and engineering machine for preventing sinking of vibrating wheels of vibratory roller |
WO2013151266A1 (en) * | 2012-04-04 | 2013-10-10 | Movon Corporation | Method and system for lane departure warning based on image recognition |
CN103850241A (en) * | 2014-02-20 | 2014-06-11 | 天津大学 | Earth and rockfill dam milling excitation frequency and excitation force real-time monitoring system and monitoring method |
CN104503240A (en) * | 2014-12-23 | 2015-04-08 | 福建船政交通职业学院 | Ergonomic dynamic design method based on chaotic recognition |
CN204530394U (en) * | 2015-03-17 | 2015-08-05 | 广东华路交通科技有限公司 | A kind of road roller data collecting system |
-
2019
- 2019-06-25 CN CN201910554219.9A patent/CN110400348B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2013151266A1 (en) * | 2012-04-04 | 2013-10-10 | Movon Corporation | Method and system for lane departure warning based on image recognition |
CN103255755A (en) * | 2013-04-28 | 2013-08-21 | 河海大学 | Lossless method for fast evaluating filling compaction quality of soil building stones in real time and evaluating device thereof |
CN103321128A (en) * | 2013-07-03 | 2013-09-25 | 中联重科股份有限公司 | Method, system and engineering machine for preventing sinking of vibrating wheels of vibratory roller |
CN103850241A (en) * | 2014-02-20 | 2014-06-11 | 天津大学 | Earth and rockfill dam milling excitation frequency and excitation force real-time monitoring system and monitoring method |
CN104503240A (en) * | 2014-12-23 | 2015-04-08 | 福建船政交通职业学院 | Ergonomic dynamic design method based on chaotic recognition |
CN204530394U (en) * | 2015-03-17 | 2015-08-05 | 广东华路交通科技有限公司 | A kind of road roller data collecting system |
Non-Patent Citations (2)
Title |
---|
WANG LINBING等: "智能路面发展与展望", 《中国公路学报》 * |
喻东晓等: "沥青路面施工质量自动化远程监控技术实践研究", 《公路交通科技(应用技术版)》 * |
Also Published As
Publication number | Publication date |
---|---|
CN110400348B (en) | 2022-12-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109696663B (en) | Vehicle-mounted three-dimensional laser radar calibration method and system | |
CN110009681B (en) | IMU (inertial measurement unit) assistance-based monocular vision odometer pose processing method | |
CN110244321B (en) | Road passable area detection method based on three-dimensional laser radar | |
CN109059954B (en) | Method and system for supporting high-precision map lane line real-time fusion update | |
CN105974940B (en) | Method for tracking target suitable for aircraft | |
CN110929710B (en) | Method and system for automatically identifying meter pointer reading based on vision | |
CN115407357B (en) | Low-harness laser radar-IMU-RTK positioning mapping algorithm based on large scene | |
CN110031829B (en) | Target accurate distance measurement method based on monocular vision | |
CN107229063A (en) | A kind of pilotless automobile navigation and positioning accuracy antidote merged based on GNSS and visual odometry | |
CN109993800A (en) | A kind of detection method of workpiece size, device and storage medium | |
CN107481315A (en) | A kind of monocular vision three-dimensional environment method for reconstructing based on Harris SIFT BRIEF algorithms | |
CN112987065B (en) | Multi-sensor-integrated handheld SLAM device and control method thereof | |
CN105222788A (en) | The automatic correcting method of the aircraft course deviation shift error of feature based coupling | |
CN110189314A (en) | Automobile instrument panel image position method based on machine vision | |
CN110223354A (en) | A kind of Camera Self-Calibration method based on SFM three-dimensional reconstruction | |
CN106056121A (en) | Satellite assembly workpiece fast-identification method based on SIFT image feature matching | |
CN114485654A (en) | Multi-sensor fusion positioning method and device based on high-precision map | |
CN116844147A (en) | Pointer instrument identification and abnormal alarm method based on deep learning | |
CN114719873A (en) | Low-cost fine map automatic generation method and device and readable medium | |
CN115265493A (en) | Lane-level positioning method and device based on non-calibrated camera | |
CN112729318A (en) | AGV fork truck is from moving SLAM navigation of fixed position | |
CN105303564A (en) | Tower type crane load stereo pendulum angle vision detection method | |
CN110400348A (en) | The unmanned vibration equipment of the twin-rotor housing of view-based access control model is rotated to detection, scaling method | |
CN110415299B (en) | Vehicle position estimation method based on set guideboard under motion constraint | |
CN114429469A (en) | Heading machine body pose determination method and system based on three-laser-spot target |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |