CN108226163A - A kind of block automated detection method - Google Patents
A kind of block automated detection method Download PDFInfo
- Publication number
- CN108226163A CN108226163A CN201711440418.4A CN201711440418A CN108226163A CN 108226163 A CN108226163 A CN 108226163A CN 201711440418 A CN201711440418 A CN 201711440418A CN 108226163 A CN108226163 A CN 108226163A
- Authority
- CN
- China
- Prior art keywords
- connected domain
- block
- detection
- area
- 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
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/02—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness
- G01B11/022—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness by means of tv-camera scanning
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/26—Measuring arrangements characterised by the use of optical techniques for measuring angles or tapers; for testing the alignment of axes
- G01B11/27—Measuring arrangements characterised by the use of optical techniques for measuring angles or tapers; for testing the alignment of axes for testing the alignment of axes
- G01B11/272—Measuring arrangements characterised by the use of optical techniques for measuring angles or tapers; for testing the alignment of axes for testing the alignment of axes using photoelectric detection means
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8854—Grading and classifying of flaws
- G01N2021/8867—Grading and classifying of flaws using sequentially two or more inspection runs, e.g. coarse and fine, or detecting then analysing
- G01N2021/887—Grading and classifying of flaws using sequentially two or more inspection runs, e.g. coarse and fine, or detecting then analysing the measurements made in two or more directions, angles, positions
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8854—Grading and classifying of flaws
- G01N2021/8874—Taking dimensions of defect into account
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of block automated detection methods, include the following steps:Block elastic tape notch detection;Bound edge detects;Arc cavity detection inside elastic tape;Five golden plates position and eccentricity detecting;Rosin joint and fried black detection;Burn-through detects;Scratch detection;Bottom plate eccentricity detecting;Elastic tape side damage testing.
Description
Technical field
The present invention relates to new energy battery cap detection devices, are related to lithium battery, and in particular to a kind of block automation inspection
Survey method.
Background technology
Lithium battery is applied more and more in the industry-by-industry of society.Lithium battery cap is pushed up above lithium battery
One parts at end, are the important components in lithium battery, and the safety that lithium battery can be protected to use has anti-mistake
The characteristics of pressure, explosion-proof, anti-overflow.Lithium battery cap needs to carry out full inspection to it before assembly.
Invention content
It is an object of the present invention to provide a kind of block automated detection methods, from different angles, different intensity of illumination feelings
Under shape, Image Acquisition is carried out to block, using image procossing, detection is carried out at the same time to the front of block, reverse side, side, is reached
The effect of full inspection.
In order to achieve the goal above, the technical solution adopted by the present invention is:A kind of block automated detection method, including such as
Lower step:
A), block elastic tape notch detection:To image zooming-out edge, with reference to the radius value of block, the outer circle of block is obtained
Profile;Excircle configuration is interrupted, splits into multiple sub- profiles;By concyclic sub- contour fitting into circle;According to block outer circle, paint
Mask artwork processed;Mask artwork makes the difference with block binaryzation region, obtains notch pixel;The jagged connected domain of detection institute;If some
The area of notch connected domain is more than 20 pixels, then alarms;
B), bound edge detects:Based on experience value, bound edge ROI region is set;Extract bound edge image;Edge detection;To all
Edge connected domain is analyzed, and the connected domain that distance is less than to 10 pixels connects;To the connected domain after obtained connection
It is analyzed, is detected whether with notch;Notch is oriented, and carries out dimension analysis;
C), arc cavity detection inside elastic tape:On the basis of five golden plate circles positioning, the position of connecting plate is found
It puts;Based on experience value, the ROI positions of arc area are set;In ROI positions, by binaryzation, size is more than the connection of threshold value
Region;Handle the relationship of arcuate structure and connecting plate;To each arcuate structure, two sides of length direction are extracted;To two
The width of side is analyzed, and counts the width value at each position.If some position width value, with the threshold value of setting not
Meet, then early warning;
D), the positioning of five golden plates and eccentricity detecting:Block image interior zone is extracted;Obtain five golden plate region exteriors
Profile;
E), rosin joint and fried black detection:According to the center location of five golden plates, ROI region is set;Extract spot area image;
Binaryzation, connected domain detection, filter area are less than the connected domain of 5 pixels;Solder joint number is counted, if number is less than or waits
In 3, directly alarm;Find four connected domains of area maximum;It calculates the center position of four connected domains and each connects
The area in logical domain;Largest connected domain area, the ratio value with minimum connected domain area, more than 2 times, then alarms;It finds most upper, most
Under, most left, most right four connected domains;The distance on four sides of four connected domain compositions is calculated, if maximum distance, than most narrow spacing
From long more than 3 pixels, then alarm;The distance of opposite side is calculated, if the length of two opposite side distances is more than 3 pixels, is alarmed;
F), burn-through detects:Binaryzation obtains highlight regions;Connected domain extraction is carried out to binary image;Find area most
Big connected domain, as detection object;To area, largest connected domain region, extracts gray-scale map again;Detect spot;As long as
The presence of spot is just alarmed;
G), scratch detection:According to the exterior contour of block, according to concentric rule, bottom plate inner circle detection zone is set;Meter
Calculate bottom plate inner circle gradient magnitude;Grads threshold is set, to gradient magnitude image, carries out binaryzation;Binary image is connected
Logical domain detection;Find the connected domain of area maximum;Calculate the minimum enclosed rectangle of the connected domain;If the long side of the connected domain
Length is more than threshold value, this threshold value is 30 pixels, then early warning, there are cuts on bottom plate at this time;
H), bottom plate eccentricity detecting:Bottom plate Image Edge-Detection;Circular fit;It is right according to the radius empirical value of bottom plate outer circle
Round positioning result is filtered, and obtains bottom plate outer circle;Detect crescent structure;Within crescent structure, curved line is detected
Item forms partition board inner circle;Partition board inner circle positions;Partition board inner circle central point is calculated, to the distance of bottom plate outer circle central point;If away from
It is more than threshold value from value, this threshold value is 5 pixels, then illustrates bottom plate bias, early warning;
I), elastic tape side damage testing:Elastic tape side pixel region is positioned from image;Calculate gradient magnitude;Ladder
Degree figure carries out binaryzation;Connected domain is detected, finds the connected domain of area maximum;The area in largest connected domain is more than 20 pixels,
Then illustrate that there is cut in side, alarm.
Further, the step c) is exchanged with step d) processes.
Further, setting ROI region is and five golden plate concentrics, radius 50 in step e).
Further, binary-state threshold is 100 in step f).
Further, it is partition board outer circle notch detection the step of detection crescent structure in step h);Based on experience value,
Obtain the camber line of crescent structure;Make masked areas.
Further, in step c) inside elastic tape notch is to determine the step of the detection of arc cavity;Obtain notch end
The point midway of point line;The five golden plate centers of circle with midpoint are connected, obtain straight line L;Two notch endpoints are crossed, two is done and puts down
Row straight line, they are parallel with straight line L;The length of connecting plate is set based on experience value;Setting connecting plate region, upper and lower both sides,
It is overlapped with the straight line for crossing endpoint, the left and right sides is set based on experience value.
Further, in step c) the step of the relationship of processing arcuate structure and connecting plate it is all arcuate structures of statistics
Length;Find complete arcuate structure;On the basis of complete arcuate structure, the position of remaining arcuate structure is determined;
By theoretical position, make comparisons with practical connected domain, by all connected domains it is corresponding with theoretical position on.
Further, the step of being extracted in step d) to block image interior zone is to obtain the outermost of block
After portion's circular contour, justified using exterior contour as masked areas, filter out the non-block pixel in image;To block inner area
Domain uses 100 as threshold value, carries out binarization operation;To the foreground area in binary picture phase, maximum connected domain is found, is made
For the corresponding connected domain of five golden plates.
Further, the step of five golden plate region exterior profiles is obtained in step d) is to five golden plate connected domains, is carried out
The operation of profile is extracted, obtains outer profile;Profile interrupts, and obtains profile segment;Concyclic Contours connection;The wide round plan of hardware web wheel
It closes;Five golden plate eccentricity detectings.
Further, the step of hardware web wheel exterior feature circular fit is in all profile fragments, finds longest
One;To the longest profile fragment, circular fit is carried out;Five final golden plate positioning results.
The technical effects of the invention are that:The present invention proposes a kind of automated detection method, using multiple cameras, never
With angle, under different intensity of illumination situations, Image Acquisition is carried out to block, using image procossing, to the positive, anti-of block
Face, side are carried out at the same time detection, judge whether current block belongs to abnormal, achieve the effect that full inspection.
Description of the drawings
Fig. 1 is the image of five golden plate side of block;
Fig. 2 is the image of block bottom plate side;
Fig. 3 is to use 100 as threshold value to block interior zone in step d), carries out the image after binarization operation;
Fig. 4 is that five golden plate connected domains are extracted the operation of profile, obtain the image of outer profile in step d);
Fig. 5 is that profile interrupts in step d), obtains the image of profile segment;
Fig. 6 is the image of concyclic Contours connection in step d);
Fig. 7 is the image of hardware web wheel exterior feature circular fit in step d);
Fig. 8 is side image when burn-through detects;
Image when Fig. 9 is scratch detection.
Specific embodiment
With reference to attached drawing, specific steps of the invention:
A), block elastic tape notch detection
Since block is that elastic tape makes, the probability that notch occurs is higher.The present invention examines with the following method
Survey block notch:
A1), to image zooming-out edge, with reference to the radius value of block, the excircle configuration of block is obtained.
A2), excircle configuration is interrupted, splits into multiple sub- profiles, reference frame is that this little profile meets circular arc characteristic.
A3), by concyclic sub- contour fitting into circle.
A4), according to block outer circle, mask artwork is drawn.
A5), mask artwork and block binaryzation region make the difference, and obtain notch pixel.
A6 the jagged connected domain of), detection institute.
A7) if, the area of some notch connected domain be more than 20 pixels, alarm.
B), bound edge detects
There is a circle bound edge in the inside of elastic tape.This circle binding structure is to be come out with mechanical compaction, therefore energy above
Enough see decorative pattern.When pressing process when something goes wrong, there is irregular phenomenon in the decorative pattern of bound edge.
Specific detection method is as follows:
B1), based on experience value, bound edge ROI region is set.
B2 bound edge image), is extracted.
B3), edge detection.
B4), all edge connected domains are analyzed, the connected domain of (distance is less than 10 pixels) closer to the distance is connected
It picks up and.
B5), the connected domain after obtained connection is analyzed, is detected whether with notch.
B6 notch), is oriented, and carries out dimension analysis.
C), arc cavity detection inside elastic tape:
On the elastic tape jacket layer of part block, there is arc hollow structure.These structures are for radiating.
But in process, it is likely that because unbalance stress, and then elastic tape is caused to be pullled, arcuate structure becomes
Shape.The present invention uses following method, detects the arcuate structure of deformation:
C1), on the basis of five golden plate circles positioning, the position of connecting plate is found.
C11) since connecting plate intersects with five golden plates, the excircle configurations of five golden plates be not it is complete, centainly
There are a notches.The notch is exactly the position of connecting plate.
C12 the point midway of notch endpoint line) is obtained.
C13) the five golden plate centers of circle with midpoint are connected, obtain straight line L.
C14 two notch endpoints) are crossed, do two parallel lines, they are parallel with straight line L.
C15 the length of connecting plate) is set based on experience value.
C16) setting connecting plate region, upper and lower both sides are overlapped with the straight line for crossing endpoint.The left and right sides is set based on experience value
It is fixed.
C2), based on experience value, the ROI positions of arc area are set.
C3), in ROI positions, by binaryzation, size is more than the connected region of threshold value.(connected domain width is more than 10
Pixel)
C4 the relationship of arcuate structure and connecting plate), is handled.
The width value of connecting plate is more than the distance between two adjacent arcuate structures.Therefore connecting plate at least can be with 1 arc
Shape structure has overlapping.If some arcuate structure, connected plate interrupts, then needs to reconnect two portions of this arcuate structure
Point.If connecting plate has overlapping with two arcuate structures simultaneously, need individually to extract each arcuate structure.
Specific practice is as follows:
C41 the length of all arcuate structures) is counted.
C42) find complete arcuate structure (length value meets empirical value).Complete arcuate structure at least there are two.
C43) on the basis of complete arcuate structure, the position of remaining arcuate structure is determined.
C44) by theoretical position, make comparisons with practical connected domain, by all connected domains it is corresponding with theoretical position on.
C5), to each arcuate structure, two sides of length direction are extracted.
C6), the width of two sides is analyzed, counts the width value at each position.If some position is wide
Angle value is not met, then early warning with the threshold value (40 pixel) of setting.
D), the positioning of five golden plates and eccentricity detecting
D1), block image interior zone is extracted
D11) after the most external circular contour of block is obtained, it is possible to using exterior contour circle as masked areas, mistake
Filter the non-block pixel in image.
D12) to block interior zone, 100 are used as threshold value, between range of luminance values is 0-255 here, selects 100 to do two
Value threshold value carries out binarization operation.As shown in Figure 3.
D13) to the foreground area in binary picture phase, maximum connected domain is found, as the corresponding connected domain of five golden plates.
D2 five golden plate region exterior profiles), are obtained
D21) to five golden plate connected domains, the operation of profile is extracted, obtains outer profile.Total red lines institute such as Fig. 4
Show.
D22), profile interrupts, and obtains profile segment
By five golden plate exterior contours, according to whether circular arc characteristic is met into Break Row, multiple sub-line items are split into.Such as Fig. 5
It is shown.
D23), concyclic Contours connection
Due to the exterior contour corresponding to five golden plate border circular areas, it is likely that be broken into a plurality of sub-line section, therefore have must
Split in advance is carried out, the multiple sub-line sections for meeting concyclic characteristic are integrated into a line segment.
As shown in fig. 6, the fragmentation lines of border circular areas are merged.
D24), hardware web wheel exterior feature circular fit
D241) in all profile fragments, longest one is found.
D242) to the longest profile fragment, circular fit is carried out.
D243) five final golden plate positioning results are as shown in Figure 7.
D25), five golden plate eccentricity detecting
By the center location of five golden plates, the center location with block is compared, the distance between calculating.
Here, step c) can exchange detection with step d) processes.
E), rosin joint and fried black detection:
In block of the present invention, solder joint be using laser burn come.This process is it is easy to appear rosin joint and fries
Black situation.When rosin joint occurs, solder joint can not show a candle to that normal solder joint is black, and brightness value is higher.It can be carried out using this feature
Rosin joint detects.When occurring fried black, it is possible to bond pad locations offset or the increased number of phenomenon of solder joint occur.
Specific practice is as follows:
E1), according to the center location of five golden plates, setting ROI region is (with five golden plate concentrics, radius 50, in image
50) pixel value is.
E2 spot area image), is extracted.
E3), binaryzation, connected domain detection, filter area are less than the connected domain of 5 pixels.
E4 solder joint number), is counted.If number is less than or equal to 3, directly alarm.
E5 four connected domains of area maximum), are found.
E6), the area of the center position of four connected domains of calculating and each connected domain.
E7), largest connected domain area, the ratio value with minimum connected domain area, more than 2 times, then alarms.
E8), find it is most upper, most under, most left, most right four connected domains.
E9 the distance on four sides of four connected domain compositions), is calculated.If maximum distance, more than 3 pictures longer than minimum range
Element is then alarmed.
E10 the distance of opposite side), is calculated.If the length of two opposite side distances is more than 3 pixels, alarm.
F), burn-through detects:
Since solder joint is all to use laser irradiation, directly it is welded on partition board.When this situation of generation burn-through, partition board can quilt
It burns.At this point, LED light is irradiated from side, punch position is due to that can not reflect LED lamplight, it will apparent black hole occur.Such as
Shown in Fig. 8, side image when being burn-through detection.By detecting whether that there are black holes, it can be determined that whether the block is normal.
Detect specific practice during burn-through hole:
F1) binaryzation obtains highlight regions (binary-state threshold 100)
Due to being irradiated from side using light, highlighted pixel can be formed in area of diaphragm.And other are not obtained
The region of irradiation, then brightness is very low.System employs 6 blocks of glass and is reflected in total, therefore can see in the picture multiple
Highlight regions.
F2 connected domain extraction) is carried out to binary image.
F3 the connected domain of area maximum) is found, as detection object.
F4) the largest connected domain region to area, extracts gray-scale map again.
F5 spot) is detected.
F6) if there are one or more spot presence, just alarm.The situation can be assumed that as wrong presence.
G), scratch detection
The surface of bottom plate side, it is possible to cut occur.As shown in Figure 9:
With the following method, the detection of cut is carried out:
G1), according to the exterior contour of block, according to concentric rule, bottom plate inner circle detection zone is set.
G2 bottom plate inner circle gradient magnitude), is calculated.
G3 Grads threshold), is set, to gradient magnitude image, carries out binaryzation.
G4 connected domain detection), is carried out to binary image.
G5 the connected domain of area maximum), is found.
G6 the minimum enclosed rectangle of the connected domain), is calculated.
G7) if, the length of the long side of the connected domain be more than threshold value (30 pixels), early warning.Exist on bottom plate at this time
Cut.
H), bottom plate eccentricity detecting:
Relative to partition board, also there is eccentric possibility in the position of bottom plate.The present invention uses following method, and it is inclined to carry out bottom plate
The heart detects:
H1), bottom plate Image Edge-Detection.
H2), circular fit.
H3), according to the radius empirical value of bottom plate outer circle, round positioning result is filtered, obtains bottom plate outer circle.
H4 crescent structure), is detected.
H41) partition board outer circle notch detection.
H42) based on experience value, the camber line of crescent structure is obtained.
H43 masked areas) is made.
H5), within crescent structure, arcuate line is detected.(partition board inner circle)
H6), partition board inner circle positions.
H7 partition board inner circle central point), is calculated, to the distance of bottom plate outer circle central point.
H8) if, distance value be more than threshold value (5 pixels), illustrate bottom plate bias, early warning.
I), elastic tape side damage testing:
The side of elastic tape, due to by mechanical presses, it is also possible to phenomena such as damaged, impression occur.The present invention passes through to lid
Cap carries out side shooting, then carries out damage testing on this image.
Specific method is as follows:
I1 elastic tape side pixel region), is positioned from image.
I2 gradient magnitude), is calculated.
I3), gradient map carries out binaryzation.
I4 connected domain), is detected, finds the connected domain of area maximum.
I5), the area in largest connected domain is more than 20 pixels, then illustrates that there is cut in side, is alarmed.
Claims (10)
1. a kind of block automated detection method, includes the following steps:
A), block elastic tape notch detection:To image zooming-out edge, with reference to the radius value of block, the excircle configuration of block is obtained;
Excircle configuration is interrupted, splits into multiple sub- profiles;By concyclic sub- contour fitting into circle;According to block outer circle, drafting is covered
Mould figure;Mask artwork makes the difference with block binaryzation region, obtains notch pixel;The jagged connected domain of detection institute;If some notch
The area of connected domain is more than 20 pixels, then alarms;
B), bound edge detects:Based on experience value, bound edge ROI region is set;Extract bound edge image;Edge detection;To all edges
Connected domain is analyzed, and the connected domain that distance is less than to 10 pixels connects;Connected domain after obtained connection is carried out
Analysis, detects whether with notch;Notch is oriented, and carries out dimension analysis;
C), arc cavity detection inside elastic tape:On the basis of five golden plate circles positioning, the position of connecting plate is found;Root
According to empirical value, the ROI positions of arc area are set;In ROI positions, by binaryzation, size is more than the connected region of threshold value;
Handle the relationship of arcuate structure and connecting plate;To each arcuate structure, two sides of length direction are extracted;To two sides
Width is analyzed, and counts the width value at each position.If some position width value, does not meet with the threshold value of setting,
Then early warning;
D), the positioning of five golden plates and eccentricity detecting:Block image interior zone is extracted;Obtain five golden plate region exterior wheels
It is wide;
E), rosin joint and fried black detection:According to the center location of five golden plates, ROI region is set;Extract spot area image;Two-value
Change, connected domain detection, filter area is less than the connected domain of 5 pixels;Solder joint number is counted, if number is less than or equal to 3
It is a, directly alarm;Find four connected domains of area maximum;The center position of four connected domains of calculating and each connection
The area in domain;Largest connected domain area, the ratio value with minimum connected domain area, more than 2 times, then alarms;It finds most upper, most
Under, most left, most right four connected domains;The distance on four sides of four connected domain compositions is calculated, if maximum distance, than most narrow spacing
From long more than 3 pixels, then alarm;The distance of opposite side is calculated, if the length of two opposite side distances is more than 3 pixels, is alarmed;
F), burn-through detects:Binaryzation obtains highlight regions;Connected domain extraction is carried out to binary image;Find area maximum
Connected domain, as detection object;To area, largest connected domain region, extracts gray-scale map again;Detect spot;As long as spot
Presence, just alarm;
G), scratch detection:According to the exterior contour of block, according to concentric rule, bottom plate inner circle detection zone is set;Calculate bottom
Plate inner circle gradient magnitude;Grads threshold is set, to gradient magnitude image, carries out binaryzation;Connected domain is carried out to binary image
Detection;Find the connected domain of area maximum;Calculate the minimum enclosed rectangle of the connected domain;If the length of the long side of the connected domain
More than threshold value, this threshold value is 30 pixels, then early warning, there are cuts on bottom plate at this time;
H), bottom plate eccentricity detecting:Bottom plate Image Edge-Detection;Circular fit;According to the radius empirical value of bottom plate outer circle, to circle
Positioning result is filtered, and obtains bottom plate outer circle;Detect crescent structure;Within crescent structure, arcuate line, shape are detected
Into partition board inner circle;Partition board inner circle positions;Partition board inner circle central point is calculated, to the distance of bottom plate outer circle central point;If distance value
More than threshold value, this threshold value is 5 pixels, then illustrates bottom plate bias, early warning;
I), elastic tape side damage testing:Elastic tape side pixel region is positioned from image;Calculate gradient magnitude;Gradient map
Carry out binaryzation;Connected domain is detected, finds the connected domain of area maximum;The area in largest connected domain is more than 20 pixels, then says
There is cut in bright side, alarm.
2. a kind of block automated detection method according to claim 1, it is characterised in that:The step c) and step
D) process is exchanged.
3. a kind of block automated detection method according to claim 1 or 2, it is characterised in that:Setting ROI in step e)
Region is and five golden plate concentrics, radius 50.
4. a kind of block automated detection method according to claim 1 or 2, it is characterised in that:Binaryzation in step f)
Threshold value is 100.
5. a kind of block automated detection method according to claim 1 or 2, it is characterised in that:The detection moon in step h)
The step of thread form structure is partition board outer circle notch detection;Based on experience value, the camber line of crescent structure is obtained;Make mask regions
Domain.
6. a kind of block automated detection method according to claim 1 or 2, it is characterised in that:Elastic tape in step c)
The step of internal arc cavity detection, is to determine notch;Obtain the point midway of notch endpoint line;By the five golden plate centers of circle in
Point connects, and obtains straight line L;Two notch endpoints are crossed, do two parallel lines, they are parallel with straight line L;Based on experience value
Set the length of connecting plate;Connecting plate region is set, upper and lower both sides are overlapped with the straight line for crossing endpoint, and the left and right sides is according to warp
Test value setting.
7. a kind of block automated detection method according to claim 1 or 2, it is characterised in that:Processing arc in step c)
The step of shape structure and the relationship of connecting plate is the length for counting all arcuate structures;Find complete arcuate structure;Complete
On the basis of arcuate structure, the position of remaining arcuate structure is determined;By theoretical position, make comparisons with practical connected domain, it will
On all connected domains are corresponding with theoretical position.
8. a kind of block automated detection method according to claim 1 or 2, it is characterised in that:To block in step d)
The step of image interior zone extracts is after the most external circular contour of block is obtained, using exterior contour circle as covering
Diaphragm area filters out the non-block pixel in image;To block interior zone, 100 are used as threshold value, carry out binarization operation;
To the foreground area in binary picture phase, maximum connected domain is found, as the corresponding connected domain of five golden plates.
9. a kind of block automated detection method according to claim 1 or 2, it is characterised in that:Five are obtained in step d)
The step of golden plate region exterior profile is to five golden plate connected domains, extracts the operation of profile, obtains outer profile;Profile is beaten
It is disconnected, obtain profile segment;Concyclic Contours connection;Hardware web wheel exterior feature circular fit;Five golden plate eccentricity detectings.
10. a kind of block automated detection method according to claim 9, it is characterised in that:The hardware web wheel is wide
The step of circular fit is in all profile fragments, finds longest one;To the longest profile fragment, round plan is carried out
It closes;Five final golden plate positioning results.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711440418.4A CN108226163B (en) | 2017-12-27 | 2017-12-27 | Automatic cap detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711440418.4A CN108226163B (en) | 2017-12-27 | 2017-12-27 | Automatic cap detection method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108226163A true CN108226163A (en) | 2018-06-29 |
CN108226163B CN108226163B (en) | 2020-09-11 |
Family
ID=62648069
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711440418.4A Expired - Fee Related CN108226163B (en) | 2017-12-27 | 2017-12-27 | Automatic cap detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108226163B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109632007A (en) * | 2019-01-17 | 2019-04-16 | 北京理工大学 | A kind of edge point extracting method and gear high-precision vision measuring system |
CN110715621A (en) * | 2018-07-12 | 2020-01-21 | 大族激光科技产业集团股份有限公司 | Detection method, device and equipment |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103345632A (en) * | 2013-06-25 | 2013-10-09 | 宁波成电泰克电子信息技术发展有限公司 | Method for detecting surface scratch defects of battery tail end |
CN105548201A (en) * | 2016-01-15 | 2016-05-04 | 浙江野马电池有限公司 | Battery welding cap visual-inspection method |
CN106251352A (en) * | 2016-07-29 | 2016-12-21 | 武汉大学 | A kind of cover defect inspection method based on image procossing |
CN206742294U (en) * | 2017-05-02 | 2017-12-12 | 深圳市东恒新能源科技有限公司 | Battery cap and battery |
CN107492604A (en) * | 2017-08-02 | 2017-12-19 | 江苏镭腾能源科技有限公司 | A kind of dynamic lithium battery explosion-proof cap component |
CN108154506A (en) * | 2017-12-27 | 2018-06-12 | 合肥市雅视智能科技有限公司 | A kind of hex nut automatic testing method |
-
2017
- 2017-12-27 CN CN201711440418.4A patent/CN108226163B/en not_active Expired - Fee Related
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103345632A (en) * | 2013-06-25 | 2013-10-09 | 宁波成电泰克电子信息技术发展有限公司 | Method for detecting surface scratch defects of battery tail end |
CN105548201A (en) * | 2016-01-15 | 2016-05-04 | 浙江野马电池有限公司 | Battery welding cap visual-inspection method |
CN106251352A (en) * | 2016-07-29 | 2016-12-21 | 武汉大学 | A kind of cover defect inspection method based on image procossing |
CN206742294U (en) * | 2017-05-02 | 2017-12-12 | 深圳市东恒新能源科技有限公司 | Battery cap and battery |
CN107492604A (en) * | 2017-08-02 | 2017-12-19 | 江苏镭腾能源科技有限公司 | A kind of dynamic lithium battery explosion-proof cap component |
CN108154506A (en) * | 2017-12-27 | 2018-06-12 | 合肥市雅视智能科技有限公司 | A kind of hex nut automatic testing method |
Non-Patent Citations (3)
Title |
---|
何程涛等: "基于HALCON的电池盖帽检测", 《数据通信》 * |
殷苏民等: "一种基于机器视觉的电池涂胶缺陷检测方法", 《机械制造》 * |
陈观应: "基于机器视觉的干电池缺陷并行检测方法研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110715621A (en) * | 2018-07-12 | 2020-01-21 | 大族激光科技产业集团股份有限公司 | Detection method, device and equipment |
CN110715621B (en) * | 2018-07-12 | 2021-07-02 | 大族激光科技产业集团股份有限公司 | Detection method, device and equipment |
CN109632007A (en) * | 2019-01-17 | 2019-04-16 | 北京理工大学 | A kind of edge point extracting method and gear high-precision vision measuring system |
CN109632007B (en) * | 2019-01-17 | 2020-12-04 | 北京理工大学 | Edge point extraction method and gear high-precision vision measurement system |
Also Published As
Publication number | Publication date |
---|---|
CN108226163B (en) | 2020-09-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20230368361A1 (en) | Screen defect detection method, apparatus, and electronic device | |
CN103376263B (en) | Scolding tin soaks shaping inspection technique, automatic shape inspection apparatus and base plate inspection system | |
CN110070524A (en) | A kind of intelligent terminal panel visual fault detection system | |
CN108226163A (en) | A kind of block automated detection method | |
CN107052086A (en) | Stamping parts surface defect detection apparatus and detection method based on 3D vision | |
KR101898947B1 (en) | Defect detection method and defect detection system | |
CN107633502B (en) | Target center identification method for automatic centering of shaft hole assembly | |
CN108154506B (en) | Automatic hexagonal nut detection method | |
JP2003279334A (en) | Three-dimensional measuring apparatus, filter lattice moire plate and illuminating means | |
CN106226316A (en) | A kind of single camera wide visual field vision thread detecting device and detection method thereof | |
CN105005161B (en) | Mistake proofing crimping production method, structure and the display panel of display panel | |
CN101531258A (en) | Machine vision-based instant noodle seasoning packet automatic detection instrument and method | |
CN108827597B (en) | Light spot uniformity detection method and detection system of structured light projector | |
CN106770367A (en) | A kind of FPC solder joints detection method | |
CN109580644B (en) | Visual inspection equipment for automobile parts | |
JP6691541B2 (en) | Method and apparatus for checking tires on a production line | |
CN112361991A (en) | Three-dimensional scanning method and device, computer equipment and storage medium | |
CN109406519A (en) | A kind of detection device and method of the special-shaped irregular solder joint of inductance element pin | |
CN104764752B (en) | Solder joint tin connection detection method and device | |
CN104483319A (en) | A sandwich biscuit defect detecting device and a method | |
JP2004219108A (en) | Method and apparatus for inspecting irregularities in film thickness of colored film | |
JP2011226814A (en) | Surface defect inspection device and surface defect inspection method | |
CN106770322A (en) | Calibration point depth detection method and temperature controller appearance detecting method | |
CN108291857B (en) | Apparatus and method for analysing tyres | |
CN103134427B (en) | Annulus Parts Recognition method |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200911 Termination date: 20211227 |