CN102519359B - Method for detecting label of polyethylene terephthalate (PET) bottle - Google Patents
Method for detecting label of polyethylene terephthalate (PET) bottle Download PDFInfo
- Publication number
- CN102519359B CN102519359B CN201110409840.XA CN201110409840A CN102519359B CN 102519359 B CN102519359 B CN 102519359B CN 201110409840 A CN201110409840 A CN 201110409840A CN 102519359 B CN102519359 B CN 102519359B
- Authority
- CN
- China
- Prior art keywords
- label
- straight line
- bounds
- gradient
- point
- 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.)
- Active
Links
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a method for detecting a label of a polyethylene terephthalate (PET) bottle, which comprises the steps of: inputting the shot label image into a computer, solving a left margin and a right margin of the label, carrying out statistics on the position of a central line of an upper margin and a lower margin within the range of the left margin and the right margin of the label, judging whether the label has up and down excursion and misplacement of a bonding surface or not, and judging whether the label has left and right excursion or not by matching a template. The method can be used for detecting the label of the PET bottle in real time in a continuous way on a high-speed automatic production line, and the detection efficiency is high.
Description
Technical field
The invention belongs to test technique automatic field, especially relate to a kind of detection method of PET bottle label.
Background technology
In bottled drink production run, the label in the packing of product all needs to face the strict demand of industry at present.Need the beverage label detecting can mainly be divided into cover mark and labeling two classes, for cover mark, trademark plastering machine above beverage bottle directly by label cover on body, label after label heating is shunk and is fixed on body, generally can there is not the situation that label joining place pastes askew dislocation in such label, need the height that mainly contains label position of detection and the problems such as skew of left and right.For labeling, labelling machine blocks out a complete label and is attached on beverage bottle, and when labeling completes, the adhesive surface of label all can be unified outwardly and face a direction, therefore, except height and the left and right skew of label position, also may there is the problem of adhesive surface dislocation in labeling.The label problem that may occur for these, all carries out quality control by visual inspection in the past and determines, detection efficiency is low.
Summary of the invention
The object of the invention is: provide a kind of and can in high-speed automated production line, carry out the detection method of the uninterrupted PET bottle label detecting in real time.
Technical scheme of the present invention is: comprise the following steps:
(1) by the label image input computing machine photographing.
(2) ask the border, left and right of label, the method is: one, label is asked to gradient in the horizontal direction, and find out partial gradient maximum of points binaryzation; Its two, the summation that the figure that binaryzation is obtained carries out on vertical direction obtains gradient accumulation curve; Its three, the peaked suitable percentage size in curve of usining, as threshold value, finds the Local Extremum that is greater than the leftmost side and the rightmost side of this threshold value in curve, obtains the left and right boundary position of label.
(3) in the bounds of label left and right, add up label up-and-down boundary position of center line, judgement label has or not skew up and down and adhesive surface to have dislocation-free:
(1) in the bounds of label left and right, at vertical direction, ask label gradient, find out partial gradient maximum of points binaryzation;
(2) in the vertical direction is asked the mid point of Dian He lower side point in top side in gradient point;
(3) Point Set obtaining is carried out to fitting a straight line;
(4) the straight line that matching obtains can represent that label is in the position at vertical direction place, and the normal place of demarcating compares, and as label, is offset up and down basis for estimation;
(5) Point Set is being simulated after the longest straight line, the point of all center lines in the effective coverage of this straight line of cancellation, at remaining Point Set matching second straight line;
(6) if being greater than the distance of threshold value and two straight lines, counting on second straight line be greater than certain threshold value, as adhesive surface dislocation basis for estimation.
(4), in the bounds of label left and right, by template matches, judge that label has or not left and right skew:
(1) get standard label figure, by method described in step (two), obtain, behind border, standard label left and right, in bounds, getting the image of suitable size area as standard form, and the distance that record standard template arrives border, left and right is as gauged distance;
(2) tags detected is found to border, label left and right by method described in step (two);
(3) in the bounds of tags detected left and right, search the matching area of standard form, by matching area position, as label left and right, be offset basis for estimation.
The invention has the beneficial effects as follows: 1, owing to adopting, based on edge center line, add up the upper-lower position of determining label, utilized preferably the gradient information of image, reduced the impact that illumination variation is brought testing result, simultaneously, the method of center line statistics and fitting a straight line has overcome has an X-rayed the deformation that changes the up-and-down boundary causing, utilize center line characteristic greatly to simplify computation complexity, and further eliminate the impact that the noise points such as hot spot bring, make testing result more accurate; What 2, for the left and right offset detection of label, will adopt is the template matching method based on frequency domain, and computing velocity is very fast; 3, in high-speed automated production line, PET bottle label is carried out to uninterrupted detection in real time, detection efficiency is high, and speed realizes soon and easily, makes production line keep the quality standard of height.
Accompanying drawing explanation
Fig. 1 is process flow diagram of the present invention;
Fig. 2 is label up-and-down boundary center line schematic diagram;
Fig. 3 is that adhesive surface dislocation markers is signed lower boundary center line schematic diagram.
Embodiment
In conjunction with process flow diagram of the present invention shown in Fig. 1, describe the present invention, the detection method of PET bottle label comprises the following steps:
Step (one): by the label image input computing machine photographing.
Step (two): ask the border, left and right of label, method is as follows: one, label is asked to gradient in the horizontal direction, find out partial gradient maximum of points binaryzation; Its two, the summation that the figure that binaryzation is obtained carries out on vertical direction obtains gradient accumulation curve; Its three, the peaked suitable percentage size in curve of usining, as threshold value, finds the Local Extremum that is greater than the leftmost side and the rightmost side of this threshold value in curve, obtains the left and right boundary position of label.
Step (three): add up label up-and-down boundary position of center line in the bounds of label left and right, have or not skew up and down according to position of center line judgement label, can also judge that in labeling situation adhesive surface has dislocation-free.Statistics label up-and-down boundary center method is as follows:
(1) no matter be square body or circular body, in the label image obtaining, should be point-blank on the mid point sets theory of up-and-down boundary, as shown in Figure 2;
(2) obtain behind border, label left and right, in bounds, at vertical direction, ask label gradient and find out partial gradient maximum of points binaryzation;
(3) in the vertical direction is asked the mid point of Dian He lower side point in top side in gradient point;
(4) impact bringing in order to get rid of noise spot, carries out fitting a straight line to the Point Set obtaining; The straight line that matching obtains can represent that label is in the position at vertical direction place, and the normal place of demarcating compares and can judge whether label is not inconsistent the skew up and down of standard;
(5) for the situation of labeling, if there is dislocation at adhesive surface, center line can fragment into two sections of obvious left and right in theory, as shown in Figure 3;
(6) for detecting the problem whether adhesive surface dislocation occurs, Point Set is being simulated after the longest straight line to the point of all center lines in the effective coverage of this straight line of cancellation trial matching second straight line in remaining Point Set; If counting on second straight line is greater than the distance of threshold value and two straight lines and is greater than certain threshold value, think the problem of adhesive surface dislocation of having occurred.
Step (four): judge that by template matches label has or not left and right drift condition in the bounds of label left and right, method is as follows:
(1) get standard label figure, by method described in step (two), obtain, behind border, standard label left and right, in bounds, getting the image of suitable size area as standard form, and the distance that record standard template arrives border, left and right is as gauged distance;
(2), to tags detected, by method described in step (two), find border, label left and right;
(3) in bounds, search the matching area of standard form;
(4) if can not find matching area or coupling occupy left and right frontier distance and compare gap with gauged distance and be greater than certain threshold value, think that the skew on left and right directions has occurred label.
Claims (1)
1. a detection method for PET bottle label, is characterized in that, comprises the following steps:
(1) by the label image input computing machine photographing;
(2) ask the border, left and right of label, the method is: one, label is asked to gradient in the horizontal direction, and find out partial gradient maximum of points binaryzation; Its two, the summation that the figure that binaryzation is obtained carries out on vertical direction obtains gradient accumulation curve; Its three, the peaked suitable percentage size in curve of usining, as threshold value, finds the Local Extremum that is greater than the leftmost side and the rightmost side of this threshold value in curve, obtains the left and right boundary position of label;
(3) in the bounds of label left and right, add up label up-and-down boundary position of center line, judgement label has or not skew up and down and adhesive surface to have dislocation-free:
in the bounds of label left and right, at vertical direction, ask label gradient, find out partial gradient maximum of points binaryzation;
in the vertical direction is asked the mid point of Dian He lower side point in top side in gradient point;
the straight line that matching obtains can represent that label is in the position at vertical direction place, and the normal place of demarcating compares, and as label, is offset up and down basis for estimation;
point Set is being simulated after the longest straight line, and the point of all center lines in the effective coverage of this straight line of cancellation, at remaining Point Set matching second straight line;
if counting on second straight line is greater than the distance of setting value I and two straight lines and is greater than setting value II, as adhesive surface dislocation basis for estimation;
(4), in the bounds of label left and right, by template matches, judge that label has or not left and right skew:
get standard label figure, by method described in step (two), obtain, behind border, standard label left and right, in bounds, getting the image of suitable size area as standard form, and the distance that record standard template arrives border, left and right is as gauged distance;
tags detected is found to border, label left and right by method described in step (two);
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201110409840.XA CN102519359B (en) | 2011-12-12 | 2011-12-12 | Method for detecting label of polyethylene terephthalate (PET) bottle |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201110409840.XA CN102519359B (en) | 2011-12-12 | 2011-12-12 | Method for detecting label of polyethylene terephthalate (PET) bottle |
Publications (2)
Publication Number | Publication Date |
---|---|
CN102519359A CN102519359A (en) | 2012-06-27 |
CN102519359B true CN102519359B (en) | 2014-03-19 |
Family
ID=46290364
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201110409840.XA Active CN102519359B (en) | 2011-12-12 | 2011-12-12 | Method for detecting label of polyethylene terephthalate (PET) bottle |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102519359B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103279979B (en) * | 2013-05-03 | 2015-08-19 | 山东明佳包装检测科技有限公司 | Based on the detection method of the nearly cylinder bottle label of 3D modeling |
CN109724547A (en) * | 2019-02-28 | 2019-05-07 | 泸州金能移动能源科技有限公司 | A kind of photovoltaic module dislocation on-line measuring device |
CN112541894B (en) * | 2020-12-11 | 2023-12-08 | 无锡雷德环保设备有限公司 | Product deformation detection method |
CN116453099A (en) * | 2023-06-14 | 2023-07-18 | 瑞幸咖啡烘焙(屏南)有限公司 | Automatic detection method, medium and equipment for coffee bean package |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5437702A (en) * | 1993-08-24 | 1995-08-01 | Coors Brewing Company | Hot bottle inspection apparatus and method |
JPH09178421A (en) * | 1995-12-21 | 1997-07-11 | Omron Corp | Center detection method of picture image data |
CN101403602B (en) * | 2008-11-17 | 2010-08-11 | 天津理工大学 | Embedded high-speed on-line machine vision detection method and device |
CN101799431B (en) * | 2010-04-08 | 2012-01-04 | 湖南大学 | Machine visual on-line detection method and equipment for tank cover quality in high-speed production line |
CN101858734B (en) * | 2010-05-19 | 2011-06-29 | 山东明佳包装检测科技有限公司 | Method for detecting PET bottleneck quality |
-
2011
- 2011-12-12 CN CN201110409840.XA patent/CN102519359B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN102519359A (en) | 2012-06-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102519359B (en) | Method for detecting label of polyethylene terephthalate (PET) bottle | |
CN108182413B (en) | Mine moving target detection and tracking identification method | |
CN105403147B (en) | One kind being based on Embedded bottle embryo detection system and detection method | |
Chen et al. | A new concentric circle detection method based on Hough transform | |
CN103439338A (en) | Classification method for film defects | |
CN103942808B (en) | Segmentation method for overlapping rock particle images | |
CN203376271U (en) | Machine vision-based non-continuous product online detection system | |
CN105160340A (en) | Vehicle brand identification system and method | |
CN102519972A (en) | Detection method of PET bottle cap and liquid level | |
CN104700423A (en) | Method and device for detecting bottle cap | |
CN110189363B (en) | Airport scene moving target low-visual-angle video speed measuring method | |
CN104182728A (en) | Vehicle logo automatic location and recognition method based on pattern recognition | |
IN2014MU03302A (en) | ||
CN104063877A (en) | Hybrid judgment identification method for candidate lane lines | |
CN106845482A (en) | A kind of license plate locating method | |
CN103020638B (en) | One is carried out welding bead based on chromatic information and is known method for distinguishing | |
CN103116743B (en) | A kind of railway obstacle detection method based on on-line study | |
CN109596631A (en) | A kind of cigarette packet seal mistake board detection device and method based on machine vision technique | |
CN103258218A (en) | Matte detection frame generation method and device and defect detection method and device | |
CN107945190A (en) | Bianry image high-speed communication region computational methods | |
CN101339663B (en) | Flotation video speed measurement method based on attribute matching | |
CN202676617U (en) | Embedded processor-based distributed paper defect detection system | |
Lei et al. | Railway obstacle detection using onboard forward-viewing camera | |
CN105979707B (en) | A kind of line layer is without copper area recognizing method and system | |
CN104408497A (en) | Product recycling system and online detection method based on full-angle two-dimensional code |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |