CN105000411B - A kind of obbbin excess detection method based on machine vision technique - Google Patents
A kind of obbbin excess detection method based on machine vision technique Download PDFInfo
- Publication number
- CN105000411B CN105000411B CN201510418902.1A CN201510418902A CN105000411B CN 105000411 B CN105000411 B CN 105000411B CN 201510418902 A CN201510418902 A CN 201510418902A CN 105000411 B CN105000411 B CN 105000411B
- Authority
- CN
- China
- Prior art keywords
- obbbin
- image
- width
- interest region
- imaging sensor
- 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
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B65—CONVEYING; PACKING; STORING; HANDLING THIN OR FILAMENTARY MATERIAL
- B65H—HANDLING THIN OR FILAMENTARY MATERIAL, e.g. SHEETS, WEBS, CABLES
- B65H26/00—Warning or safety devices, e.g. automatic fault detectors, stop-motions, for web-advancing mechanisms
- B65H26/06—Warning or safety devices, e.g. automatic fault detectors, stop-motions, for web-advancing mechanisms responsive to predetermined lengths of webs
-
- A—HUMAN NECESSITIES
- A24—TOBACCO; CIGARS; CIGARETTES; SIMULATED SMOKING DEVICES; SMOKERS' REQUISITES
- A24C—MACHINES FOR MAKING CIGARS OR CIGARETTES
- A24C5/00—Making cigarettes; Making tipping materials for, or attaching filters or mouthpieces to, cigars or cigarettes
Landscapes
- Length Measuring Devices By Optical Means (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of novel obbbin excess detection method based on machine vision technique passes through imaging sensor and acquires remaining obbbin side(That is thickness)Image and send computer to, handled by computer by image of the image processing techniques to side and judge surplus, when reaching setting value to cigarette making and tipping machine send switching signal.The present invention obbbin surplus can be controlled it is very low, avoid previous excessive surplus and caused by waste, and be adapted to the spool of different-diameter and the obbbin of different-thickness, have a wide range of applications space.
Description
Technical field
The present invention relates to cigarette rolling up and connecting machine obbbin detection method field, specifically a kind of obbbins based on machine vision technique
Excess detection method.
Background technique
Cigarette cigarette making and tipping machine is there are two the fixed obbbin of pallet, and during the production of cigarette cigarette, unit is in current obbbin
Another disk is switched to before being finished to keep the continuity of production.The principle of detection obbbin surplus is as follows at present:Obbbin pallet side
Edge has equidistant detent projection, and by detecting detent projection close to switch, obbbin every revolution can generate fixation close to switch
Generally this pulse sum (namely obbbin rotation number of total coils) is arranged in machine main controller in the pulse of quantity, and system is utilized and connect
The number of total coils of nearly switch detection obbbin rotation, carries out obbbin switching after having arrived pre- fixing turn.Since auxiliary material provider is to obbbin
Overall length does not have strict control, and in addition also difference is larger for the diameter of spool, therefore the number of total coils of every bobbin paper is also different.In order to keep away
Exempt to set the not timed out obbbin of fixing turn just to have run out and cause to shut down, operator is generally arranged obbbin surplus in tens circles.
Remaining bobbin length often has more than ten meters to tens meters every time in this way, accumulates over a long period and will cause very big waste.
At home and abroad in tobacco business, machine vision technique using more and more extensive, be mainly used in packet appearance matter
Foreign body detecting etc. in amount detection, barrel appearance quality detection and Primary Processing.It, can with the continuous maturation of machine vision technique
The surplus of obbbin is detected using machine vision technique.
Summary of the invention
The object of the present invention is to provide a kind of obbbin excess detection method based on machine vision technique, to detect cigarette volume
Pick obbbin surplus.
In order to achieve the above object, the technical scheme adopted by the invention is as follows:
A kind of obbbin excess detection method based on machine vision technique, it is characterised in that:Cooperated using imaging sensor
Strip source acquires obbbin side image in cigarette rolling up and connecting machine, and described image sensor face obbbin can show the side of thickness,
For the optical lens of imaging sensor perpendicular to obbbin side, strip source is symmetricly set on imaging sensor two sides, and strip light
Source is emitted light direction and tilts 45 degree of angles respectively relative to imaging sensor optical lens direction, and imaging sensor is by collected disk
Paper side image is sent to computer, successively filters to the collected obbbin side image of imaging sensor in a computer
After wave, gray processing, dividing processing, then upright projection is carried out to the image after segmentation, then using calculus of finite differences in upright projection
It searches search obbbin both sides of the edge and calculates obbbin width pixel number, the obbbin width pixel that finally calculus of finite differences is calculated
Number is averaged again after carrying out quadratic fit, more accurate obbbin width pixel number is obtained, to obtain obbbin surplus.
A kind of obbbin excess detection method based on machine vision technique, it is characterised in that:When dividing processing, adopt
With the method for local sampling, by the image segmentation after gray processing at many fritters, the crooked radian of obbbin is neglected in each fritter
Slightly disregard, can regard straight as, then respectively to each fritter upright projection, is then looked into upright projection using calculus of finite differences
It looks for search obbbin both sides of the edge and calculates obbbin width pixel number.
The beneficial effects of the invention are as follows:
Compared with current detection method, detection accuracy greatly improves obbbin excess detection method proposed by the present invention, remains
The length of remaining obbbin can be accurate to each circle, avoid the waste of previous tens circle of residue.And using the method for vision-based detection
It is adapted to the spool of different-diameter and the obbbin of different-thickness, has a wide range of applications space.
Detailed description of the invention
Fig. 1 is the image sensor of that present invention and light source setting angle figure.
Fig. 2 is the original image of imaging sensor acquisition.
Fig. 3 is filtered image.
Fig. 4 is the image after gray scale.
Fig. 5 is original image compared with filtered image upright projection, wherein:
Fig. 5 a is original image, and Fig. 5 b is filtered image.
Fig. 6 is interest region vertical projection diagram, wherein:
Fig. 6 a is the 8th interest region vertical projection diagram, and Fig. 6 b is the 14th interest region vertical projection diagram.
Fig. 7 is interest region upright projection difference curves figure, wherein:
Fig. 7 a is the 8th interest region upright projection difference curves figure,
Fig. 7 b is the 14th interest region upright projection difference curves figure.
Specific embodiment
A kind of obbbin excess detection method based on machine vision technique, using imaging sensor cooperation strip source acquisition
Obbbin side image in cigarette rolling up and connecting machine, imaging sensor face obbbin can show the side of thickness, the optics of imaging sensor
Camera lens is perpendicular to obbbin side, and strip source is symmetricly set on imaging sensor two sides, and strip source outgoing light direction is opposite
45 degree of angles are tilted respectively in imaging sensor optical lens direction, and collected obbbin side image is sent to by imaging sensor
Computer is in a computer successively filtered the collected obbbin side image of imaging sensor, gray processing, dividing processing
Afterwards, then to the image after segmentation upright projection is carried out, search obbbin two sides is then searched in upright projection using calculus of finite differences
Edge simultaneously calculates obbbin width pixel number, after the obbbin width pixel number progress quadratic fit that finally calculus of finite differences is calculated again
It averages, obtains more accurate obbbin width pixel number, to obtain obbbin surplus.
When dividing processing, using the method for local sampling, by the image segmentation after gray processing at many fritters, each small
The crooked radian of obbbin is ignored in block, can regard straight as, is then then used to each fritter upright projection respectively
Calculus of finite differences searches search obbbin both sides of the edge in upright projection and calculates obbbin width pixel number.
The setting angle of imaging sensor and light source is as shown in Figure 1, the side of imaging sensor face obbbin is installed, optics
Camera lens is vertical with obbbin side, and what imaging sensor took is the direct picture of obbbin side, intermediate white obbbin and two
The dark-background of side forms sharp contrast.Two strip sources are loaded on imaging sensor two sides, tilt about 45 degree, the photograph of light source
Firing angle degree is fine-tuning, which had not only guaranteed to illuminate obbbin but also obbbin reflection light is avoided to enter camera lens to influence into image quality
Amount.
Fig. 2 is the collected original image of imaging sensor, on the image for causing acquisition due to factors such as external contexts
There is noise, understands some small burrs after upright projection on image, as shown in Fig. 5 (a), although these small burrs are unobvious, handle
When image segmentation is at many fritters, these small burrs can interfere query search obbbin edge after projection, so using equal
Value filtering is first filtered original image, removes these small noises.Fig. 3 is schemed after the filtering obtained using mean filter method
Picture.Fig. 5 (b) is the upright projection of filtered image, it can thus be seen that after filtering curve it is obvious it is flat very much, the influence of small burr becomes
It is small.Fig. 4 is to result after filtered image gray processing.Gray processing is the color image in order to take imaging sensor
Gray level image is converted to, does not need to carry out gray processing processing if using black white image sensor.
Because the image of obbbin has certain radian, the method for directlying adopt projection searches the obbbin that the edge of obbbin obtains
Width error can be larger, and the present invention is by the way of local sampling, by the image segmentation after gray processing at many fritters, each
The crooked radian of obbbin is ignored in fritter, can be regarded as vertical.Then to each fritter upright projection, pass through calculus of finite differences
It searches obbbin edge and obtains obbbin width, quadratic fit then is carried out to obbbin width data, averages, obtains accurate disk
Paper width.Specific step is as follows:
1. choosing k interest region in the image after gray processing, the width in each interest region is the width of image,
Height is n pixel.The mode of selection is from top to bottom that the height every n pixel takes an interest region.N value is smaller, obtains
The result arrived is more accurate, while computational processing is also larger, can select suitable n value according to demand.
2. the image in pair this k interest region carries out upright projection respectively, calculation formula is as follows:
Wherein Pm[j] is the drop shadow curve in m-th of interest region, Qm(i, j) is on the gray level image in m-th of interest region
The gray value of i-th row, jth column pixel, width are the width of gray level image, and k One Dimensional Projection array can be obtained in this way.Fig. 6
It (a) and shown in (b) is two interest regions drop shadow curve therein.
3. in the drop shadow curve in interest region, there are obbbin and regional luminance without obbbin widely different, as shown in fig. 6,
Centre has obbbin regional luminance very high, can be searched by calculus of finite differences and distinguish them.By one-dimension array PmData be right in [j]
Subtracting each other after moving two with itself can be obtained difference function Sm[j], shown in institute Fig. 7, figure (a) and (b) correspond in Fig. 6 two it is emerging
The difference curves of interesting region projection curve, it can be seen that will form at obbbin left and right edges an obvious precipitous wave crest and
Trough.To difference function Sm[j] is searched, and S is recordedmJ value X when [j] maximum value and minimum valueR[m] and XL[m] is calculated
Formula is as follows:
The left edge position X of k group obbbin can be obtainedL(m) and right edge position XR(m).Then by each interest region
Right edge position, which subtracts its left edge position, can be obtained obbbin width Y [m].
4. pair k group obbbin width data Y [m] quadratic fit:Y [m] average value is sought, it is compared with Y [m], is removed
Difference is greater than the data of c, obtains new data M (x), then average to M (x), accurate obbbin width data can be obtained
Ave, calculation formula are as follows:
In this example, obbbin thickness is about 0.04mm, the imaging sensor used respectively rate for 1280*960.
The obbbin width data for taking k=20, n=18 to be calculated by above-mentioned steps is as shown in table 1.Obbbin after quadratic fit
Width is 131 pixels, and obbbin circle number is 42 circles at this time, then the pixel of every circle is about 3, and the length of remaining obbbin can be accurate to often
One circle, compared with current detection method, detection accuracy is greatly improved.It, can be using more if detecting the smaller obbbin of thickness
High-resolution imaging sensor.
Table 1 is the obbbin left and right sides marginal position that upright projection is searched after choosing 20 interest regions
Claims (1)
1. a kind of obbbin excess detection method based on machine vision technique, it is characterised in that:Using imaging sensor setting strip
Shape light source acquires obbbin side image in cigarette rolling up and connecting machine, and described image sensor face obbbin can show the side of thickness, schemes
As the optical lens of sensor is perpendicular to obbbin side, strip source is symmetricly set on imaging sensor two sides, and strip source
Outgoing light direction tilts 45 degree of angles relative to imaging sensor optical lens direction respectively, and imaging sensor is by collected obbbin
Side image is sent to computer, the collected obbbin side image of imaging sensor is successively filtered in a computer,
After gray processing, dividing processing, then upright projection is carried out to the image after segmentation, is then searched in upright projection using calculus of finite differences
Search obbbin both sides of the edge simultaneously calculate obbbin width pixel number, the obbbin width pixel number that finally calculus of finite differences is calculated into
It averages again after row quadratic fit, obtains more accurate obbbin width pixel number, to obtain obbbin surplus;
When dividing processing, using the method for local sampling, by the image segmentation after gray processing at many fritters, in each fritter
The crooked radian of obbbin is ignored, and can regard straight as, then respectively to each fritter upright projection, then uses difference
Method searches search obbbin both sides of the edge in upright projection and calculates obbbin width pixel number;Specific step is as follows:
(1)K interest region is chosen in image after gray processing, the width in each interest region is the width of image, high
Degree is n pixel, and the mode of selection is from top to bottom that the height every n pixel takes an interest region, and n value is smaller, obtains
Result it is more accurate, while computational processing is also larger, selects suitable n value according to demand;
(2)Upright projection is carried out respectively to the image in this k interest region, calculation formula is as follows:
WhereinFor the drop shadow curve in m-th of interest region,It is on the gray level image in m-th of interest region i-th
The gray value of row, jth column pixel,For the width of gray level image, k One Dimensional Projection array is obtained;
(3)In the drop shadow curve in interest region, there are obbbin and regional luminance without obbbin widely different, there is obbbin region in centre
Brightness is very high, is searched by calculus of finite differences and distinguishes them;By one-dimension arrayAfter middle data shift right two with phase itself
Subtract to obtain difference function;To difference functionIt is searched, is recordedJ value when maximum value and minimum valueWith, calculation formula is as follows:
Obtain the left edge position of k group obbbinAnd right edge position;Then by the right in each interest region
Edge position, which subtracts its left edge position, can be obtained obbbin width;
(4)To k group obbbin width dataQuadratic fit:It asksAverage value, by its withIt is compared, removes difference
Value is greater than the data of c, obtains new data, then it is rightIt averages, obtains accurate obbbin width data, meter
It is as follows to calculate formula:
。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510418902.1A CN105000411B (en) | 2015-07-14 | 2015-07-14 | A kind of obbbin excess detection method based on machine vision technique |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510418902.1A CN105000411B (en) | 2015-07-14 | 2015-07-14 | A kind of obbbin excess detection method based on machine vision technique |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105000411A CN105000411A (en) | 2015-10-28 |
CN105000411B true CN105000411B (en) | 2018-11-16 |
Family
ID=54373338
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510418902.1A Active CN105000411B (en) | 2015-07-14 | 2015-07-14 | A kind of obbbin excess detection method based on machine vision technique |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105000411B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106056617A (en) * | 2016-05-31 | 2016-10-26 | 中国电子科技集团公司第四十研究所 | Cigarette carton sealing machine hot melt adhesive detection method |
CN107218929B (en) * | 2017-04-18 | 2019-08-30 | 中国电子科技集团公司第四十一研究所 | A kind of cork paper flanging detection method based on machine vision technique |
CN107444712A (en) * | 2017-08-30 | 2017-12-08 | 红云红河烟草(集团)有限责任公司 | Device and method for detecting surplus of small box aluminum foil paper of packaging machine |
CN110834981A (en) * | 2019-11-21 | 2020-02-25 | 成都忠信机电技术有限公司 | Bobbin paper splicing method and system |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2042937A3 (en) * | 2007-09-27 | 2011-04-13 | Kabushiki Kaisha Toshiba | Sheet thickness measuring device and image forming apparatus |
CN101503123B (en) * | 2009-02-18 | 2011-01-05 | 王万年 | Bobbin detector |
CN203946659U (en) * | 2014-05-06 | 2014-11-19 | 南京文采科技有限责任公司 | Imaging type passim obbbin surplus Detection & Controling device |
CN105011358B (en) * | 2015-06-12 | 2017-01-25 | 中国电子科技集团公司第四十一研究所 | Dish paper remnant detection apparatus based on machine vision technology |
-
2015
- 2015-07-14 CN CN201510418902.1A patent/CN105000411B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN105000411A (en) | 2015-10-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105000411B (en) | A kind of obbbin excess detection method based on machine vision technique | |
KR102320061B1 (en) | Depth sensor calibration and per-pixel correction | |
US9998728B2 (en) | Information processing apparatus, control method for the same and storage medium | |
KR100656468B1 (en) | Automatic focus adjustment for projector | |
AU2007324081B2 (en) | Focus assist system and method | |
CN105976753B (en) | Abnormal lamp point in LED display correction detects compensation method | |
JP6727791B2 (en) | Tracking control device, tracking control method, and imaging device | |
TWI511081B (en) | Image capturing device and method for calibrating image deformation thereof | |
US9432563B2 (en) | Three-dimensional imaging through multi-image processing | |
JP2010176128A (en) | Two-dimensional polynomial model for depth estimation based on two-picture matching | |
US20110273610A1 (en) | Auto-focus image system | |
CN102572275B (en) | Imaging device, image processing device, image processing method and image processing software | |
GB2479229A (en) | Aberration correction in image processing apparatus | |
CN105701809B (en) | A kind of method for correcting flat field based on line-scan digital camera scanning | |
TW201837862A (en) | Depth information processing device | |
CN101494735A (en) | Imaging apparatus, imaging apparatus control method, and computer program | |
CN109084688A (en) | A kind of binocular distance measurement method based on varifocal camera | |
CN103176668B (en) | A kind of shooting method for correcting image for camera location touch system | |
JPH09116809A (en) | Picture steadiness correcting method in television film scanning of prescribed sequence changing method for performing picture object and device for performing the correcting method | |
CN106989681A (en) | The dimension measurement method and measuring apparatus of a kind of via | |
JP2011248085A (en) | Autofocus imaging system | |
US20150085078A1 (en) | Method and System for Use in Detecting Three-Dimensional Position Information of Input Device | |
CN102215332A (en) | Electronic camera | |
CN103776770A (en) | Multi-path light source self-adaptation system based on information entropy and working method of system | |
JP4656668B2 (en) | Light receiving unit and optical communication device |
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 |