CN110211183A - The multi-target positioning system and method for big visual field LED lens attachment are imaged based on single - Google Patents
The multi-target positioning system and method for big visual field LED lens attachment are imaged based on single Download PDFInfo
- Publication number
- CN110211183A CN110211183A CN201910511605.XA CN201910511605A CN110211183A CN 110211183 A CN110211183 A CN 110211183A CN 201910511605 A CN201910511605 A CN 201910511605A CN 110211183 A CN110211183 A CN 110211183A
- Authority
- CN
- China
- Prior art keywords
- sample
- target
- point
- image
- positioning
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/12—Edge-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
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- 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
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
-
- 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
- G06T2207/30152—Solder
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
- Manipulator (AREA)
Abstract
The invention discloses the multi-target positioning system and method for a kind of big visual field LED lens attachment based on single imaging, which includes large area image acquisition, target positioning, position correction.The present invention also provides a kind of multi-target orientation methods.This method include first standard sample processing method with second~processing method of N number of sample.The processing step of standard sample: acquisition aluminum substrate image, setting target and Mark point region, image segmentation, BLOB analysis, feature extraction and target position, and the positioning of Mark dot center and its line midpoint and tilt angle obtain；Second~N number of sample handling procedure: acquisition image, point location in the positioning of Mark dot center and its straightway obtain the offset of sample angle and displacement, construct rigid body translation matrix, carry out affine transformation to the multiple target point of standard sample.System and method for provided by the present invention has locating speed is fast, full-automatic, prosthetic teaching intervention can be achieved obtain target point, uninterrupted continuous production can be achieved.
Description
Technical field:
The present invention relates to Multi-Target Locating Technology fields, and in particular to the big visual field LED aluminum base plate based on single imaging is saturating
The multi-target positioning system apparatus and method of mirror attachment pad.The present invention proposes a kind of multi-target orientation method, and in particular to one
Kind centralized positioning and its alignment technique.Of the invention is specifically related to show as, and the R. concomitans of system and device and method are realized big
The Multi-target position of the LED lens attachment pin pad of visual field.
Background technique:
With the development of industry, the Multi-target position identification technology in industrial applicability becomes the market demand.The progress of science and technology,
Artificial human eye fixation and recognition target is gradually eliminated, but fixation and recognition technology of today is still not mature enough, and domestic industry is used
In the region Multi-target position problem that there is also operating errors is big, detection accuracy is small, operation duration is low.Especially at present
The Multi-target position function of domestic dispenser, there is that system structure is complicated, precision is not high, stability is poor, operating error
Greatly, the phenomenon that detection accuracy is small, operation duration is low.
In the utilization of industry, for imaging acquisition environment under the conditions of LED illumination, industrial camera holds PCB aluminum substrate pad
Capable imaging acquisition of taking pictures, but realize that Multi-target position identification technology in region is perfect not enough, current Multi-target position identifies skill
Art, there is positioning, target zone is small, stability is poor, the working time is short, detection speed is slow, detection accuracy is low, algorithm calculation amount
Big deficiency.
Chinese patent [CN109201413A] discloses a kind of vision positioning dispenser system, including camera shooting mechanism, dispenser
Structure, three axis mechanisms, computer and controller.The invention also discloses a kind of vision positioning dispensing method, include the following steps: pair
Plates part shoots photo, reads shooting picture and establishes image coordinate system, determines the transformational relation between each coordinate system, right
Plates part waits for that dispensing regional graphics carry out Threshold segmentation, extracts the dispensing position to dispensing region；By dispensing position data
It is transmitted to controller, three axle systems of control run to designated position and complete dispensing.There is detection speed for the system slowly, orientation range
Smaller, the higher shortcoming of imaging acquisition frequency.
Chinese patent [CN105964492A] is related to a kind of automatic dispensing machine, including pedestal, positioning element, point glue component with
Main panel lid, told positioning element mainly include positioning plate, the secondary positioning plate for being set to told positioning plate upper end, are set to and are told
It the adjusting rod of secondary positioning plate side, setting and tells time locking piece of positioning front edge of board, be set to and tell secondary positioning plate upper end
Magnetic orientation seat, the connecting plate for being set to told master positioning plate lower end, is set to institute at the locating piece being set to inside told magnetic orientation
The rodless cylinder told the guide rail of connecting plate lower end, be set to told positioning plate lower end is set to the buffering of told positioning plate lower end
Device.But the detection accuracy of this method is lower, higher cost, system structure are more complicated.
Chinese patent [CN202102970U] discloses a kind of relay glue dispenser with positioning device, including body, machine
Feed rail, feed rail side are equipped with supporting plate to the installation of body upper end again, and supporting plate upper end is equipped with the branch perpendicular to feed rail
Seat, support upper end are slidably fitted with carrying pawl, limit plate are equipped between feed rail and supporting plate, and limit plate upper end is slidably installed
There are several positioning columns to stretch up and down, positioning column is fixedly arranged at the front end with elastic cushion, and carrying grabs front end equipped with several card slots, and limit plate is high
Degree, which is higher than to carry, grabs, and carrying, which is grabbed, stretches above feed rail, carries out Multi-target position using positioning column.The positioning accurate of the system
Accuracy is lower, system structure is complex, operation duration is lower.
European patent [EP2066166A1] relates to the side that variation is thermally expanded in a kind of positioning system and compensation positioning system
Method and component mounter, injecting type dispenser, including positioning system.The positioning system includes along the elongated of axis (X) extension
The shape of beam, central sill is changed during positioning system operation due to thermal expansion.The system includes movably suspending
Positioning unit on crossbeam, for provide the mobile motor of positioning unit along axis (X) and be installed on positioning unit or with
Integrated work head.A kind of elongate reference element that shape is fixed is provided, has along axis (X) and is arranged essentially parallel to beam
It is longitudinal to extend.At multiple positions along axis (X), reference distance between measurement and positioning unit and reference element.Positioning unit
Position fixing process in measure resulting reference distance compensate thermal expansion.The system cost is higher, location efficiency is poor, detection is smart
It spends lower.
In conclusion about the multi-target positioning system in industrial applicability, however it remains various deficiencies, especially
In the target positioning of single imaging, always it is difficult to be widely used in industrial production line.And this invention address that creating one kind
The multi-target positioning system apparatus and method of big visual field LED aluminum base plate lens attachment pad based on single imaging, system of the present invention
It unites and the problems such as being used in combination, efficiency, error, precision, stability when target positioning will be solved, realizes big visual field with method
The Multi-target position of LED lens attachment pin pad.
Summary of the invention:
It is an object of that present invention to provide a kind of multi-target positioning system device, the system and device by camera collecting part,
Mechanism position portion, detection identification division composition.The system is in the case where stablizing light environment, and industrial camera is to LED aluminum base plate
Lens pad carries out Image Acquisition.Image-capturing platform mainly includes industrial camera (120), LED aluminum base plate pad (121), passes
Sensor (122), guide rail (123), the working environment of camera imaging platform are as follows:
The imaging object is to be based on aluminum substrate (121) surface region, and guide rail (123) is then positioned at aluminum substrate (121)
Front, sensor (122) occupy the side of aluminum substrate (121), and industrial camera (120) is vertically fixed on aluminum substrate (121)
Surface；The transmitting of guide rail (123) realizes that aluminum substrate (121) travel forward, and aluminum substrate (121) moves to sensor (122) sense
When answering range, when sensor (122) induction is based on aluminum substrate (121) target position, the sensor camera issues image and adopts
The signal instruction of collection.
The system and device is based on camera and acquires image, by carrying out Multi-target position to the image of acquisition, to obtain
The centre coordinate of multiple targets.System and device of the invention is combined with each other with method of the invention, solves positioning system
The problem that locating speed in the process of running is slow, the degree of automation is low, system stability is insufficient
The positioning step of the multi-target orientation method of big visual field LED lens attachment pad based on single imaging is specifically such as
Under:
Step 1 carries out Image Acquisition to PCB aluminum substrate pad by industrial camera, obtains the picture number of first sample
According to；
Original image is split as RGB channel and the channel HSV by step 2；
Step 3, edge detection original image RGB channel and the channel HSV, the optimal channel of comparative selection degree；
The ROI region setting of step 4, the setting of the ROI region of target point and two Mark points；
Step 5, based on best threshold method with the ROI region image of a target point, Mark point is divided using average gray
ROI region image, realize image segmentation；
Step 6 executes connection characteristic Blob analysis；
Step 7, according to the feature extractions such as different areas, length, width, rectangular degree, circularity welding disking area and carry out it is more
The centralized positioning of a target；
Step 8 determines Mark dot center coordinate according to gray average；
Step 9, the multiple coordinates of targets Value Data (X for saving point glue pointk, Yk), wherein subscript K indicates k-th target point；
The pixel coordinate of multiple coordinates of targets Value Datas of step 10, point glue point is converted to the mechanical of robot motion and sits
Mark；
Step 11, the tilt angle theta for obtaining built straightway0, and save；
Step 12, the midpoint coordinates (X for calculating two Mark point straightways0, Y0), and save tilt angle and straightway
Midpoint coordinates leaves it at that for the Processing Algorithm process of first sample (standard sample)；
Step 13, acquisition second~N number of sample (current sample) image；
Current picture is split as RGB channel and the channel HSV by step 14；
Step 15, the channel that maximum-contrast is selected according to target contrast；
The Mark point region that step 16, basis have been set in first sample, carries out the choosing of ROI area-of-interest
It selects；
Step 17, to two Mark point regions, Mark point region subgraph is divided using the method based on average gray；
Step 18 executes connected region characteristic Blob analysis, and feature extraction Mark point target region according to area；
Step 19, the centre coordinate that Mark point is determined according to the coordinate average value of object pixel；
Step 20, for second~N number of sample, using simplified positioning strategy.By the coordinates of targets of first sample,
And the position control otherness (i.e. angular deviation and offset) of current sample and first sample, it can obtain fast and reliablely
The dispensing point target of current sample is sat.Current sample image is acquired, centralized positioning is carried out to selected two Mark point, is obtained
The even straightway of the line of centres of Mark point, calculates the tilt angle theta of straightwayi；
Step 21, obtain second~N number of sample (current sample) image two Mark point line straightways midpoint
Coordinate (Xi, Yi)。
Step 22 is positioned according to the Mark of present image (second~N number of sample) and first sample (standard sample) and is tied
The otherness of fruit, obtains the difference (being indicated with Δ Θ) of tilt angle, and obtains deviant (the wherein X of straightway midpoint coordinates
The offset in direction is indicated with Δ X, the offset of Y-direction is indicated with Δ Y).It can be stated with mathematic(al) representation are as follows:
Wherein: subscript 0 is expressed as the label of first sample, and i is expressed as second~label of N number of sample.
Step 23, the offset based on center point coordinate and angular deviation (Δ X, Δ Y, Δ Θ), can configurations rigid body
Transformation matrix；
Step 24, the coordinates of targets to multiple glue points of first sample, carry out two-dimensional linear affine transformation, are worked as
Multiple dispensing point target coordinates of preceding sample, and pixel coordinate is converted to the mechanical coordinate of robot motion.To quickly and
Steadily realize the high speed positioning of the coordinates of targets of current sample point glue point.
System and device of the invention and method only need Polaroid to obtain hundreds of or even thousands of a targets and position
Point, the process can be realized in 0.1 second.It is extremely long caused raw that the apparatus and method not only overcome artificial teaching method time-consuming
Producing line is interrupted completely, and is overcome and be repeatedly imaged required for the big multiple image mosaic technology in despot face, to cause needs
Multiple image mosaic to be further introduced into image mosaic error, and causes the interruption of production.Therefore system has to have and determine
Bit rate is fast, positioning accuracy is higher, the good feature of stability.Compared with prior art, it has the advantages that
(1) fast the showing themselves in that of present system locating speed only needs to carry out Polaroid to obtain visual field
Width W is greater than 800mm, and multiple targets of the height H greater than 600mm of visual field position.Wherein LED, which is mounted, several hundred on aluminum substrate
Dispensing position (target land), it might even be possible to realize thousands of a target points, and imaging precision is up to 0.1mm.In Image Acquisition
When, acquisition speed is high, it is only necessary to which 30~50ms can be transferred to image data the memory headroom of computer.In image analysis
When, about 50ms is only needed for first sample image processing analysis；The target positioning time of second~N number of sample is shorter, consumption
When be less than 10ms.Therefore, the present invention only needs 0.1 second Image Acquisition that super large breadth can be completed and up to thousands of a targets
It is disposable to be accurately positioned, it can be achieved that production line continuously uninterruptedly produces.It when consuming above set limit and needs to stop production for a long time compared to traditional
Artificial teaching, which obtains goal approach and existing still time-consuming and the small field of view figure of production disruption is needed to acquire, combines multiframe figure
As the object localization method of splicing, method proposed by the invention is greatly improved in speed.
(2) for the first time, imaging precision is up to 0.1mm for high the showing themselves in that of positioning accuracy that the present invention works.To guarantee above-mentioned skill
Camera used by art condition is the side using big despot face and the Polaroid positioning of high-speed industrial camera of high-resolution fixation
Method.Positioning manual operation error brought by subjective factor, fatigue factor that this method avoids artificial teaching operation from introducing etc.；Separately
Outside, it and avoids image split-joint method transmission mechanism movement is needed to drive camera mechanical brought by the different position imagings
Position error, the stitching algorithm error that merging algorithm for images introduces.Therefore, the invention avoids manual operation error, machinery are fixed
Position error, stitching algorithm error etc., have the characteristics that positioning accuracy is high.
(3) what stability of the present invention was good shows themselves in that the method for the present invention carries out target using first sample (standard sample)
The processing (current sample) of positioning, and second~N number of sample uses method for correcting position, and the stability of system can be improved.Tool
Body is realized as follows；Target positioning is carried out to first sample (standard sample), and realizes two Mark point locations of standard sample, and
Determine the point line midpoint Mark and line section tilt angle；For second~processing (current sample) of N number of sample it is only necessary to
Two Mark points are positioned, and determine the point line midpoint Mark and line section tilt angle.It is compared, obtains with first sample
Angular deviation, shift offset are obtained, to construct rigid body translation matrix, affine change is carried out to hundreds of target points of standard sample
It changes, can be obtained second~aiming spot of N number of sample, treating capacity and the time of image data can be greatly reduced.The most
It crucially avoids since the subtle differences of product lead to the missing of target positioning or decoy occur, causes to position mistake,
It is unstable using system.Therefore, the present invention improves system high-speed quantity-produced stability.
Detailed description of the invention:
Fig. 1, illustrate for the multi-target positioning system device of the big visual field LED lens attachment of the invention based on single imaging
Figure
Fig. 2, the target object distribution schematic diagram to be positioned
Fig. 3, the target point schematic diagram that region is extracted for original image
Fig. 4, the target point schematic diagram that region is extracted for secondary image
Fig. 5, track schematic diagram of making a connection for the positioning target point of original image and secondary image
Fig. 6, multi-target orientation method better embodiment flow chart is acquired for single image of the present invention
It Fig. 7, is n times Image Acquisition multi-target orientation method better embodiment flow chart of the present invention
As shown in the figure are as follows: 120- industrial camera, 121-PCB aluminum substrate pad, 122- sensor, 123- pipeline
Specific embodiment:
For a better understanding of the present invention, it is further described with reference to the accompanying drawing to of the present invention, but the present invention
Embodiment it is without being limited thereto.
The multi-target positioning system of big visual field LED aluminum base plate lens attachment pad disclosed by the invention based on single imaging
Device includes camera imaging collecting part, mechanism position portion, detection identification division.
Show as Fig. 1 is said, camera system imaging circumstances main body frame of the invention includes: industrial camera (120), PCB aluminium base
Plate pad (121), sensor (122), guide rail (123).The imaging of camera system includes: that Image Acquisition target is based on aluminum substrate
(121) surface region, and guide rail (123) is then located at the front of aluminum substrate (121), sensor (122) occupy aluminum substrate (121)
Side, industrial camera (120) is vertically fixed on the surface of aluminum substrate (121)；Aluminum substrate is realized in the transmission of pipeline (123)
(121) it travels forward, when aluminum substrate (121) moves to sensor (122) induction range, the sensor (122) of side incudes base
When aluminum substrate (121) target position, it is transmitted to industrial camera (120) and implements image pick-up signal instruction, complete camera imaging
Acquisition.
As shown in Fig. 2, positioning target object distribution schematic diagram of the invention includes: that knowledge is set in right position identification point 131, left position
Other point 133, image center 132, target area 134 to be positioned.
Mechanism position portion described in multi-target positioning system is based on imaging acquisition region according to the present invention；Reading is taken pictures
The region of imaging carries out edge detection to the region of imaging, splits out RGB channel and the channel HSV, selected by contrast best
Channel selects threshold value according to feature extraction, completes the segmentation of area image, converted by digital image morphology, rank region
And focus target is calculated, picture position identification point is extracted in positioning, according to centre of location point coordinate and rectilinear end slope.
Coordinate pair ratio of the identification detection and localization according to this system based on secondary imaging and original image；Obtain original image
After centre coordinate, camera carries out secondary acquisition of taking pictures to positional shift aluminum substrate, carries out edge detection to acquisition imaging region, real
The target area image again now deviated on aluminum substrate is extracted, and positioning obtains secondary image centre coordinate, calculates original image and two
Coordinate position deviation of making a connection between secondary image realizes metro planning, executes attachment pad motion command.
As shown in Figures 6 and 7, the multi-target orientation method of the big visual field LED lens attachment pad based on single imaging, is pressed
Following steps complete Multi-target position, include the following steps:
S201, Image Acquisition is carried out to PCB aluminum substrate pad by industrial camera, obtains the picture number of first sample
According to；
S202, original image is split as to RGB channel and the channel HSV；
S203, edge detection original image RGB channel and the channel HSV, the optimal channel of comparative selection degree；
S204, the ROI region setting of target point and the ROI region setting of two Mark points；
S205, based on best threshold method with the ROI region image of a target point, using average gray segmentation Mark point
ROI region image realizes image segmentation；
S206, connection characteristic Blob analysis is executed；
S207, it according to the feature extractions such as different areas, length, width, rectangular degree, circularity welding disking area and carries out multiple
The centralized positioning of target；
S208, Mark dot center coordinate is determined according to gray average；
S209, the multiple coordinates of targets Value Data (X for saving point glue pointk, Yk), wherein subscript K indicates k-th target point；
S210, the multiple coordinates of targets Value Datas for putting glue point pixel coordinate be converted to the mechanical coordinate of robot motion；
S211, the tilt angle theta for obtaining built straightway0, and save；
S212, the midpoint coordinates (X for calculating two Mark point straightways0, Y0), and save in tilt angle and straightway
Point coordinate, leaves it at that for the Processing Algorithm process of first sample (standard sample)；
S301, acquisition second~N number of sample (current sample) image；
S302, current picture is split as RGB channel and the channel HSV；
S303, the channel that maximum-contrast is selected according to target contrast；
The Mark point region that S304, basis have been set in first sample, carries out the selection of ROI area-of-interest；
S305, to two Mark point regions, Mark point region subgraph is divided using the method based on average gray；
S306, connected region characteristic Blob analysis, and feature extraction Mark point target region according to area are executed；
S307, the centre coordinate that Mark point is determined according to the coordinate average value of object pixel；
S308, for second~N number of sample, using simplified positioning strategy.By the coordinates of targets of first sample, with
And the position control otherness (i.e. angular deviation and offset) of current sample and first sample, it can be worked as fast and reliablely
The dispensing point target of preceding sample is sat.Current sample image is acquired, centralized positioning is carried out to selected two Mark point, is connected
The straightway of the line of centres of a Mark point, calculates the tilt angle theta of straightwayi；
S309, the image for obtaining second~N number of sample (current sample) the midpoints of two Mark point line straightways sit
Mark (Xi, Yi)。
S310, according to the Mark positioning result of present image (second~N number of sample) and first sample (standard sample)
Otherness, obtain the difference (being indicated with Δ Θ) of tilt angle, and obtain deviant (the wherein side X of straightway midpoint coordinates
To offset indicated with Δ X, the offset of Y-direction is indicated with Δ Y).It can be stated with mathematic(al) representation are as follows:
Wherein: subscript 0 is expressed as the label of first sample, and i is expressed as second~label of N number of sample.
S311, the offset based on center point coordinate and angular deviation (Δ X, Δ Y, Δ Θ), can the change of configurations rigid body
Change matrix；
S312, the coordinates of targets to multiple glue points of first sample, carry out two-dimensional linear affine transformation, obtain current
Multiple dispensing point target coordinates of sample, and pixel coordinate is converted to the mechanical coordinate of robot motion.To quickly and steady
Surely the high speed positioning of the coordinates of targets of current sample point glue point is realized.
Through described in above step, the present invention realizes the Image Acquisition and target signature of LED aluminum base plate lens attachment pad
It extracts, and using multi-target orientation method as core, the positioning system of multiple targets for industrial LED lens attachment pad.
Above-mentioned implementation of the invention is only illustrative to have annotated the principle of the present invention, not to embodiments of the present invention into
Row limits.For fields obtain technical staff, it can also make on the basis of the above description other various forms of
It changes.Therefore, any modification done within the spirit and principles of the present invention, it is equivalent the behaviors such as to improve, should all by comprising
In in the protection scope of the claims in the present invention.
Claims (5)
1. the multi-target positioning system device of the big visual field LED aluminum base plate lens attachment pad based on single imaging, feature exist
In the system includes large area image acquisition, Multi-target position, position correction；Wherein single imaging and big visual field,
It is characterized in that, only needs to carry out the Polaroid LED that can obtain and mount several hundred a dispensing positions (target welderings on aluminum substrate
Disk) centre coordinate, it might even be possible to realize thousands of a target points.And imaging viewing field is huge, and the W of visual field is greater than 800mm, visual field
Height H be greater than 600mm.In addition, imaging precision is up to 0.1mm；To guarantee that camera used by above-mentioned technical conditions is use
Big despot face and high-resolution high-speed industrial camera；Compared to traditional when consuming above set limit and the artificial teaching that needs to stop production for a long time obtains
Goal approach and target that is existing still time-consuming and needing the small field of view figure acquisition of production disruption that multiple image is combined to splice
Localization method, method proposed by the invention are greatly improved in speed, only need super large breadth can be completed within 0.1 second
Image Acquisition and up to thousands of a targets disposable accurate positioning, it can be achieved that production line continuously uninterruptedly produces.
2. large area image acquisition as described in claim 1, it is characterised in that only need to be imaged 1 time, the visual field of imaging
Width W and height H be can be adjusted according to the size of sample, the width of the size of maximum sample is highly greater than greater than 800mm
600mm, and the speed of Image Acquisition is fast, it is only necessary to and the memory that 30~50ms can be transferred to image data in computer is empty
Between.
3. Multi-target position as described in claim 1, it is characterised in that only large area image need to can be completed within 0.1 second and be acquired
LED aluminum base plate lens attachment pad on several hundred or even thousands of quantity dispensing positions target land, 0.1 second when
In, precision is up to 0.1mm and completes big view field image acquisition and several hundred or even thousands of glue point (target welderings next time
Disk) positioning.
4. position correction as described in claim 1, it is characterised in that as long as by being carried out to first sample (standard sample)
Target positioning, and realize two Mark point locations of standard sample, and determine the point line midpoint Mark and line section tilt angle；Needle
To second~processing of N number of sample, it is only necessary to two Mark points are positioned, and determine that the point line midpoint Mark and line section tilt
Angle；It is compared by sample second~N number of sample with first sample, angular deviation, shift offset is obtained, to construct
Rigid body translation matrix carries out affine transformation to hundreds of target points of standard sample, can be obtained second~target of N number of sample
Point position, can greatly reduce treating capacity and the time of image data, and and avoid since the subtle differences of product cause
There is decoy in the missing of target positioning.Finally improve the system high-speed quantity-produced speed of service and stability.
5. the multi-target orientation method of the big visual field LED lens attachment pad based on single imaging, this method are matched with the system
The R. concomitans that bulk cargo is set, which comprises the steps of:
Step 1 carries out Image Acquisition to PCB aluminum substrate pad by industrial camera, obtains the image data of first sample；
Original image is split as RGB channel and the channel HSV by step 2；
Step 3, edge detection original image RGB channel and the channel HSV, the optimal channel of comparative selection degree；
The ROI region setting of step 4, the setting of the ROI region of target point and two Mark points；
Step 5, based on best threshold method with the ROI region image of a target point, using the ROI of average gray segmentation Mark point
Area image realizes image segmentation；
Step 6 executes connection characteristic Blob analysis；
Step 7 according to the feature extractions such as different areas, length, width, rectangular degree, circularity welding disking area and carries out multiple mesh
Target centralized positioning；
Step 8 determines Mark dot center coordinate according to gray average；
Step 9, the multiple coordinates of targets Value Data (X for saving point glue pointk, Yk), wherein subscript K indicates k-th target point；
The pixel coordinate of multiple coordinates of targets Value Datas of step 10, point glue point is converted to the mechanical coordinate of robot motion；
Step 11, the tilt angle theta for obtaining built straightway0, and save；
Step 12, the midpoint coordinates (X for calculating two Mark point straightways0, Y0), and save the midpoint of tilt angle and straightway
Coordinate leaves it at that for the Processing Algorithm process of first sample (standard sample)；
Step 13, acquisition second~N number of sample (current sample) image；
Current picture is split as RGB channel and the channel HSV by step 14；
Step 15, the channel that maximum-contrast is selected according to target contrast；
The Mark point region that step 16, basis have been set in first sample, carries out the selection of ROI area-of-interest；
Step 17, to two Mark point regions, Mark point region subgraph is divided using the method based on average gray；
Step 18 executes connected region characteristic Blob analysis, and feature extraction Mark point target region according to area；
Step 19, the centre coordinate that Mark point is determined according to the coordinate average value of object pixel；
Step 20, for second~N number of sample, using simplified positioning strategy.By the coordinates of targets of first sample, and
The position control otherness (i.e. angular deviation and offset) of current sample and first sample can obtain current fast and reliablely
The dispensing point target of sample is sat.Current sample image is acquired, centralized positioning is carried out to selected two Mark point, the company of acquisition
The straightway of the line of centres of Mark point, calculates the tilt angle theta of straightwayi；
Step 21, obtain second~N number of sample (current sample) image two Mark point line straightways midpoint coordinates
(Xi, Yi)。
Step 22, according to present image (second~N number of sample) and the Mark positioning result of first sample (standard sample)
Otherness, obtains the difference (being indicated with Δ Θ) of tilt angle, and obtains deviant (the wherein X-direction of straightway midpoint coordinates
Offset indicated with Δ X, the offset of Y-direction is indicated with Δ Y).It can be stated with mathematic(al) representation are as follows:
Wherein: subscript 0 is expressed as the label of first sample, and i is expressed as second~label of N number of sample.
Step 23, the offset based on center point coordinate and angular deviation (Δ X, Δ Y, Δ Θ), can configurations rigid body translation
Matrix；
Step 24, the coordinates of targets to multiple glue points of first sample, carry out two-dimensional linear affine transformation, obtain current sample
Multiple dispensing point target coordinates of product, and pixel coordinate is converted to the mechanical coordinate of robot motion.To quickly and stable
Realize the high speed positioning of the coordinates of targets of current sample point glue point in ground.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910511605.XA CN110211183B (en) | 2019-06-13 | 2019-06-13 | Multi-target positioning system based on single-imaging large-view-field LED lens mounting |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910511605.XA CN110211183B (en) | 2019-06-13 | 2019-06-13 | Multi-target positioning system based on single-imaging large-view-field LED lens mounting |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110211183A true CN110211183A (en) | 2019-09-06 |
| CN110211183B CN110211183B (en) | 2022-10-21 |
Family
ID=67792641
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910511605.XA Active CN110211183B (en) | 2019-06-13 | 2019-06-13 | Multi-target positioning system based on single-imaging large-view-field LED lens mounting |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110211183B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110930455A (en) * | 2019-11-29 | 2020-03-27 | 深圳市优必选科技股份有限公司 | Positioning method, positioning device, terminal equipment and storage medium |
| CN111257346A (en) * | 2020-02-20 | 2020-06-09 | 清华大学 | PCB positioning device and method based on projection filtering |
| CN111558939A (en) * | 2020-05-06 | 2020-08-21 | 珠海格力智能装备有限公司 | Valve body assembling method, system, device, storage medium and processor |
| CN113034418A (en) * | 2019-12-05 | 2021-06-25 | 中国科学院沈阳自动化研究所 | Circuit board identification and pad/chip rapid positioning method for electronic industry |
| CN113304966A (en) * | 2021-04-26 | 2021-08-27 | 深圳市世宗自动化设备有限公司 | Dynamic dispensing compensation method and device, computer equipment and storage medium thereof |
Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5233669A (en) * | 1990-11-22 | 1993-08-03 | Murata Manufacturing Co., Ltd. | Device for and method of detecting positioning marks for cutting ceramic laminated body |
| JP2003208616A (en) * | 2002-01-17 | 2003-07-25 | Fujitsu Ltd | Image recognition device and program |
| US20090180085A1 (en) * | 2008-01-15 | 2009-07-16 | Kirtas Technologies, Inc. | System and method for large format imaging |
| US20100092079A1 (en) * | 2008-10-14 | 2010-04-15 | Joshua Victor Aller | Target and method of detecting, identifying, and determining 3-d pose of the target |
| CN101750017A (en) * | 2010-01-18 | 2010-06-23 | 战强 | Visual detection method of multi-movement target positions in large view field |
| CN103308007A (en) * | 2013-05-24 | 2013-09-18 | 华南理工大学 | System and method for measuring coplanarity of integrated circuit (IC) pins through multistage reflection and raster imaging |
| CN104680570A (en) * | 2015-03-24 | 2015-06-03 | 东北大学 | Action capturing system and method based on video |
| CN106504262A (en) * | 2016-10-21 | 2017-03-15 | 泉州装备制造研究所 | A kind of small tiles intelligent locating method of multiple features fusion |
| CN106570904A (en) * | 2016-10-25 | 2017-04-19 | 大连理工大学 | Multi-target relative posture recognition method based on Xtion camera |
| US20170242235A1 (en) * | 2014-08-18 | 2017-08-24 | Viewsiq Inc. | System and method for embedded images in large field-of-view microscopic scans |
| CN107578431A (en) * | 2017-07-31 | 2018-01-12 | 深圳市海思科自动化技术有限公司 | A kind of Mark points visual identity method |
| CN108827316A (en) * | 2018-08-20 | 2018-11-16 | 南京理工大学 | Mobile robot visual orientation method based on improved Apriltag label |
| CN108966500A (en) * | 2018-08-07 | 2018-12-07 | 向耀 | The pcb board of view-based access control model tracking is secondary and multiple accurate drilling method |
| CN109410229A (en) * | 2018-08-27 | 2019-03-01 | 南京珂亥韧光电科技有限公司 | Multiple target lens position and male and fomale(M&F) know method for distinguishing |
| WO2019080229A1 (en) * | 2017-10-25 | 2019-05-02 | 南京阿凡达机器人科技有限公司 | Chess piece positioning method and system based on machine vision, storage medium, and robot |
| KR101975209B1 (en) * | 2018-12-14 | 2019-05-07 | 서정원 | Position tracing system using Affine transformation |
| CN109785324A (en) * | 2019-02-01 | 2019-05-21 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of large format pcb board localization method |
-
2019
- 2019-06-13 CN CN201910511605.XA patent/CN110211183B/en active Active
Patent Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5233669A (en) * | 1990-11-22 | 1993-08-03 | Murata Manufacturing Co., Ltd. | Device for and method of detecting positioning marks for cutting ceramic laminated body |
| JP2003208616A (en) * | 2002-01-17 | 2003-07-25 | Fujitsu Ltd | Image recognition device and program |
| US20090180085A1 (en) * | 2008-01-15 | 2009-07-16 | Kirtas Technologies, Inc. | System and method for large format imaging |
| US20100092079A1 (en) * | 2008-10-14 | 2010-04-15 | Joshua Victor Aller | Target and method of detecting, identifying, and determining 3-d pose of the target |
| CN101750017A (en) * | 2010-01-18 | 2010-06-23 | 战强 | Visual detection method of multi-movement target positions in large view field |
| CN103308007A (en) * | 2013-05-24 | 2013-09-18 | 华南理工大学 | System and method for measuring coplanarity of integrated circuit (IC) pins through multistage reflection and raster imaging |
| US20170242235A1 (en) * | 2014-08-18 | 2017-08-24 | Viewsiq Inc. | System and method for embedded images in large field-of-view microscopic scans |
| CN104680570A (en) * | 2015-03-24 | 2015-06-03 | 东北大学 | Action capturing system and method based on video |
| CN106504262A (en) * | 2016-10-21 | 2017-03-15 | 泉州装备制造研究所 | A kind of small tiles intelligent locating method of multiple features fusion |
| CN106570904A (en) * | 2016-10-25 | 2017-04-19 | 大连理工大学 | Multi-target relative posture recognition method based on Xtion camera |
| CN107578431A (en) * | 2017-07-31 | 2018-01-12 | 深圳市海思科自动化技术有限公司 | A kind of Mark points visual identity method |
| WO2019080229A1 (en) * | 2017-10-25 | 2019-05-02 | 南京阿凡达机器人科技有限公司 | Chess piece positioning method and system based on machine vision, storage medium, and robot |
| CN108966500A (en) * | 2018-08-07 | 2018-12-07 | 向耀 | The pcb board of view-based access control model tracking is secondary and multiple accurate drilling method |
| CN108827316A (en) * | 2018-08-20 | 2018-11-16 | 南京理工大学 | Mobile robot visual orientation method based on improved Apriltag label |
| CN109410229A (en) * | 2018-08-27 | 2019-03-01 | 南京珂亥韧光电科技有限公司 | Multiple target lens position and male and fomale(M&F) know method for distinguishing |
| KR101975209B1 (en) * | 2018-12-14 | 2019-05-07 | 서정원 | Position tracing system using Affine transformation |
| CN109785324A (en) * | 2019-02-01 | 2019-05-21 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of large format pcb board localization method |
Non-Patent Citations (4)
| Title |
|---|
| QIUSHENG ZHONG ET AL.: "A Fast Coplanarity Inspection System for Double-Sides IC Leads Using Single Viewpoint", 《LECTURE NOTES IN ARTIFICIAL INTELLIGENCE》 * |
| QIUSHENG ZHONG ET AL.: "Feature-Based Object Location of IC Pins by Using Fast Run Length Encoding BLOB Analysis", 《IEEE TRANSACTIONS ON COMPONENTS PACKAGING AND MANUFACTURING TECHNOLOGY》 * |
| 周前飞等: "机载光电成像平台的多目标自主定位系统研究", 《光学学报》 * |
| 钟球盛: "面向IC封装的在线检测理论与技术研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110930455A (en) * | 2019-11-29 | 2020-03-27 | 深圳市优必选科技股份有限公司 | Positioning method, positioning device, terminal equipment and storage medium |
| CN110930455B (en) * | 2019-11-29 | 2023-12-29 | 深圳市优必选科技股份有限公司 | Positioning method, positioning device, terminal equipment and storage medium |
| CN113034418A (en) * | 2019-12-05 | 2021-06-25 | 中国科学院沈阳自动化研究所 | Circuit board identification and pad/chip rapid positioning method for electronic industry |
| CN113034418B (en) * | 2019-12-05 | 2023-10-13 | 中国科学院沈阳自动化研究所 | Circuit board identification and bonding pad/chip rapid positioning method for electronic industry |
| CN111257346A (en) * | 2020-02-20 | 2020-06-09 | 清华大学 | PCB positioning device and method based on projection filtering |
| CN111257346B (en) * | 2020-02-20 | 2022-02-25 | 清华大学 | PCB positioning device and method based on projection filtering |
| CN111558939A (en) * | 2020-05-06 | 2020-08-21 | 珠海格力智能装备有限公司 | Valve body assembling method, system, device, storage medium and processor |
| CN113304966A (en) * | 2021-04-26 | 2021-08-27 | 深圳市世宗自动化设备有限公司 | Dynamic dispensing compensation method and device, computer equipment and storage medium thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110211183B (en) | 2022-10-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110211183A (en) | The multi-target positioning system and method for big visual field LED lens attachment are imaged based on single | |
| CN100582657C (en) | Three-dimensional microcosmic appearance inclined scanning method and apparatus | |
| CN101753775B (en) | Imaging device | |
| CN102761708B (en) | A kind of linear array CCD image scan method | |
| CN109405755B (en) | Large-size pipe plate aperture and pitch measuring device and measuring method | |
| CN103350281A (en) | Automatic focusing device and automatic focusing method for laser marking machine | |
| CN109201413A (en) | A kind of vision positioning dispenser system and its method | |
| CN106403828A (en) | Monorail contact line remain height measurement method based on checkerboard calibration and monorail contact line remain height measurement system thereof | |
| CN108344693B (en) | Automatic welding-oriented visual measurement method for misalignment of sheet welding seam | |
| CN102230758A (en) | Method and device for testing tracking stability precision during travel of self-propelled gun | |
| CN102636378A (en) | Machine vision orientation-based biochip microarrayer and positioning method thereof | |
| CN110044262A (en) | Contactless precision measuring instrument and measurement method based on image super-resolution rebuilding | |
| US20210035359A1 (en) | Method for Improving Modeling Speed of Digital Slide Scanner | |
| CN108489830A (en) | Crack propagation measuring system | |
| CN102780860A (en) | Linear array CCD (charge coupled device) image scanning method | |
| CN103228107A (en) | Position correction device and position correction method using same | |
| CN202512009U (en) | Biological chip microarrayer based on machine vision positioning | |
| CN116560062B (en) | Microscope focusing anti-collision control method | |
| CN112381888A (en) | Dynamic compensation method for H-shaped steel cutting path | |
| CN111260561A (en) | Rapid multi-graph splicing method for mask defect detection | |
| CN204389406U (en) | Sapphire crystallization defect and surface imperfection Systems for optical inspection | |
| CN108989690A (en) | A kind of line-scan digital camera multiple labeling point focusing method, device, equipment and storage medium | |
| CN106556350B (en) | The measuring method and a kind of microscope of microscopic slide curved surface height value | |
| CN110726402A (en) | Laser point vision guiding method of non-orthogonal shafting laser total station | |
| CN215868286U (en) | Machine vision teaching experiment platform of linear array scanning type |
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 |