CN107169961A - A kind of cigarette sorting detecting system and method based on CIS IMAQs - Google Patents
A kind of cigarette sorting detecting system and method based on CIS IMAQs Download PDFInfo
- Publication number
- CN107169961A CN107169961A CN201710338029.4A CN201710338029A CN107169961A CN 107169961 A CN107169961 A CN 107169961A CN 201710338029 A CN201710338029 A CN 201710338029A CN 107169961 A CN107169961 A CN 107169961A
- Authority
- CN
- China
- Prior art keywords
- image
- bar
- cigarette
- stick
- bar cigarette
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10052—Images from lightfield camera
-
- 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
- G06T2207/20016—Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform
-
- 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
- G06T2207/20036—Morphological image processing
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of cigarette sorting detecting system and method based on CIS IMAQs, detecting system includes：Conveyer, image collecting device, data processing module and display device；Image collector is setting in the surface of conveyer, and image collecting device is connected with data processing module；When bar cigarette is transported by conveyer and when bar cigarette is by below image collecting device, described image harvester is to bar cigarette progress IMAQ, and by image transmitting to data processing module, data processing module carries out analysis detection to it.Detection method includes：S1：IMAQ is carried out to bar cigarette by CIS image acquisition units；S2：The bar cigarette image of collection is handled, the feature of bar cigarette image is obtained；S3：The type and quantity of bar cigarette are determined according to the feature of the bar cigarette image and bar stick image data base；S4：Judge that whether sorting bar cigarette is consistent, and show result with original order according to the type and quantity of bar cigarette.
Description
Technical field
The present invention relates to field of industry detection, more particularly to one kind is based on CIS (Contact Image Sensor, contact
Formula imaging sensor) image acquisition units cigarette sorting detecting system and method.
Background technology
China's tobacco business carries out the system of " unified leadership, vertical management, franchise sale monopoly " always.The logistics network of tobacco
National tobacco is support from producing to being sold and the whole process consumed.During tobacco circulation, originate in each cigarette
Factory, various types of cigarette converges at Tobacco Distribution Center, wholesale to retailer afterwards, finally enters in consumer's hand.
Sorting cigarette is as link most complicated during whole tobacco circulation, and its function is the order point assigned by client
Sort out the cigarette of respective numbers and species.Sorting cigarette mode mainly includes hand-sorted, semi-automatic sorting and automatic sorting.Hand
It is to be based on retail customer order demand that work point, which is picked, is chosen by hand, a kind of sorting mode with goods.Because human cost is high,
Labor intensity is big, sort efficiency is low, and manual sorting only has application in some remote districts at present.Automatic sorting is super in order to meet
Large-scale delivery sorting requires, order data is carried out into network analysis research, is carried out using different cigarette sorting equipment high
Imitate a kind of sorting cigarette mode of processing.Because sorting system is complicated, cost is high, system maintenance difficulty is big, operating cost is high, from
Dynamic sorting is currently without popularization.Semi-automatic sorting refers to utilize Computerized Information Processing Tech, automatic control technology and automation
Equipment coordinates, and non-key process uses operating personnel, and order demand and bar cigarette are chosen to the sorting mode of processing.Due to sorting speed
Degree is fast, and human cost is low, and semi-automatic sorting mode is commonly used now.Manually various types of bar cigarette is placed in sorter first
(sorter only sorts same bar cigarette), then sorts bar cigarette according to customer order, is then passed by conveyer
It is sent to last encapsulation.Such as：Customer order is：10 Yellow Crane Towers, 5 China, 8 lotus kings, then correspond to these three cigarettes
Sorter sorts out the bar cigarettes of respective numbers to conveyer, carries out being sent to last encapsulation by conveyer.Due to
The transmission rate issues of sorter and transmitting device, may result in the problem of order is not inconsistent with sorting product.Therefore work out
The cigarette sorting detecting system and method for accurate stable are necessary.
The content of the invention
For the defect of prior art, it is an object of the invention to provide a kind of cigarette sorting based on CIS IMAQs
Detect detecting system and method, it is intended to which order is with sorting the problem of product is not inconsistent during solving cigarette sorting.
The invention provides a kind of cigarette sorting detecting system based on CIS IMAQs, including：Conveyer, image
Harvester, data processing module and display device；Image collector is setting in the surface of the conveyer, and image adopts
Acquisition means are connected with data processing module；When bar cigarette is transported by conveyer and when bar cigarette passes through under image collecting device
Fang Shi, image collecting device carries out IMAQ to bar cigarette, and by image transmitting to data processing module, data processing module pair
It carries out analysis detection.
Further, image collecting device includes：Three CIS image acquisition units and two strip sources, the first CIS
Image acquisition units are arranged on the surface of conveyer, the 2nd CIS image acquisition units and the 3rd CIS image acquisition units point
The both sides of conveyer are not arranged on, and the first CIS image acquisition units both sides are respectively provided with a strip source, for causing quilt
The brightness of image of the bar cigarette of collection is uniform, and the first CIS image acquisition units are used for the image of pick-up slip cigarette upper surface；2nd CIS
Image acquisition units and the 3rd CIS image acquisition units are respectively used to the image of two sides of pick-up slip cigarette.
Further, three CIS image acquisition units structures are identical, include：FPGA module, respectively with the FPGA
CIS image capture modules, analog-to-digital conversion module, memory module, picking rate matching module and the image output mould of module connection
Block, and for providing the power module of working power；CIS image capture modules are used to carry out IMAQ to bar cigarette and exported
A series of analogue datas；The analog-to-digital conversion module is used to analogue data being converted into data signal；The FPGA module is used for
Realize the SECO of whole system；The picking rate matching module is used to adjust CIS according to the transfer rate of conveyer
Image module gathers image rate；The memory module is used to cache the digital image data collected；Described image exports mould
Block is used to the bar cigarette image collected being sent to the data processing module.
Present invention also offers a kind of cigarette sorting detection method based on CIS IMAQs, comprise the steps：
S1：IMAQ is carried out to bar cigarette by CIS image acquisition units；
S2：The bar cigarette image of collection is handled, the feature of bar cigarette image is obtained；
S3：The type and quantity of bar cigarette are determined according to the feature of the bar cigarette image and bar stick image data base；
S4：Judge that whether sorting bar cigarette is consistent, and show result with original order according to the type and quantity of bar cigarette.
Further, the foundation of bar stick image data base is specifically included：
(1.1) using CCD camera pick-up slip stick and bar stick image is obtained；
(1.2) bar stick image is handled using morphological method, obtains bar stick contour images；
(1.3) bar stick contour images are handled using image pyramid method, obtains bar stick feature set；
The bar stick feature set includes the mark of different types of cigarette；
(1.4) bar stick feature set is marked.
Further, step (1.3) is specially：
(1.31) bar stick contour images are subjected to gray processing processing, obtain bar stick profile gray level image；
(1.32) Hessian conversion is carried out to bar stick gray level image, obtains bar stick changing image；
(1.33) by changing the scale size and repeat step (1.32) of operator in Hessian conversion, obtain a series of
Bar stick changing image, constitute image pyramid；
(1.34) bar stick characteristic point is found in image pyramid, bar cigarette marker characteristic is obtained；
(1.35) to mark repeat step (1.31)~(1.34) of different bar cigarettes, corresponding bar stick is obtained special
Levy, constituting bar stick feature set.
Further, step S2 is specially：
S21 carries out gray processing processing to bar cigarette image, obtains bar cigarette gray level image；
S22 carries out Hessian conversion to bar cigarette gray level image, obtains bar cigarette changing image；
S23 obtains a series of bar cigarette by changing the scale size and repeat step S22 of operator in Hessian conversion
Changing image, constitutes image pyramid；
S24 finds bar cigarette characteristic point in described image pyramid, and obtains the feature of bar cigarette image.
Further, step S3 is specially：
S31 by the Prototype drawing in the feature of bar cigarette image and the bar stick image data base by being compared realization
The preliminary matches of bar cigarette feature；
S32 realizes matching and correlation by rejecting erroneous matching in the preliminary matches；
S33 determines bar cigarette type and quantity according to the result after matching and correlation, and generates survey according to bar cigarette type and quantity
Trial order.
Further, determine that bar cigarette species is specially according to the result after matching and correlation：
Correct matching logarithm is obtained in step S32, logarithm will be correctly matched (special according to bar cigarette image with matching number threshold value
Number setting a little is levied, ten) usual value is compared for 6 the percent of bar cigarette image characteristic point number, if correct matching pair
Number is more than or equal to matching number threshold value, it is determined that the species of bar cigarette, if correctly matching logarithm is less than matching number threshold value, under
A kind of bar stick feature is matched with bar cigarette feature, until determining bar cigarette species.
Further, in step s 4, test is judged by comparing test order with the character string in original order
Whether order is consistent with original order, and will determine that result is shown.
By the contemplated above technical scheme of the present invention, compared with prior art, due to utilizing CIS image capturing systems
Figure is adopted in progress, it is possible to increase the speed of IMAQ, strengthens real-time, and reduce the beneficial effect of cost；Due to the present invention
The method for sorting of offer, it is ensured that the accuracy of order and ensure that sort efficiency, afterwards to increase the species of detector bar cigarette,
Bar stick image data base need to be only updated, versatility and practicality is improved, industry inspection now is met well
The demand in survey field.
Brief description of the drawings
Fig. 1 is the schematic diagram of the cigarette sorting system of the invention based on CIS IMAQs；
Fig. 2 is the control flow schematic diagram of the cigarette sorting method of the invention based on CIS IMAQs；
Fig. 3 is the theory diagram of CIS image acquisition units in the present invention；
The schematic flow sheet that Fig. 4 sets up for bar stick image data base in the present invention；
The schematic flow sheet that Fig. 5 is constituted for bar cigarette feature set in the present invention；
Fig. 6 is square box wave filter 9*9 template schematic diagrames in the present invention；Wherein, used in (a)Used in operator, (b)Operator (Operator), used in (c)Operator；
Fig. 7 is Harr small echo template schematic diagrames in the present invention；
Fig. 8 is the schematic flow sheet of acquisition bar cigarette feature in the present invention；
Fig. 9 is the schematic flow sheet of identification bar cigarette in the present invention；
Figure 10 is the schematic diagram of kd trees in the present invention；Wherein, (a) set of data samples, (b) finds the nearest of (2.1,3.1)
Adjoint point, (c) finds the nearest neighbor point of (2,4.5)；
Figure 11 is determines the schematic diagrames of multiple cigarette species in the present invention；Wherein, (a) detects multiple cigarette species, and (b) exists
Bar stick image data base determines starting matched position.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
A kind of cigarette sorting detecting system based on CIS image acquisition units that the present invention is provided, including：Conveyer,
Image collecting device, data processing module and display device；Different types of cigarette is transported by conveyer, when bar cigarette
During by below image collecting device, image collecting device is acquired to bar cigarette image and is transferred to data processing module, number
The processing that bar cigarette classification differentiates, counts and contrasted with original order is carried out to the image of collection according to processing module.
In embodiments of the present invention, image collecting device includes：Three CIS image acquisition units and two strip sources,
First CIS image acquisition units are arranged on the surface of conveyer, and the 2nd CIS image acquisition units and the 3rd CIS images are adopted
Collection unit is separately positioned on the both sides of conveyer, and the first CIS image acquisition units both sides are respectively provided with a strip source, use
In make it that the brightness of image of collected bar cigarette is uniform, the first CIS image acquisition units are used for the image of pick-up slip cigarette upper surface；
2nd CIS image acquisition units, the 3rd CIS image acquisition units are respectively used to the image of two sides of pick-up slip cigarette.
Because cigarette sorting detecting system is higher for system real time requirement in detection process, therefore the speed of IMAQ
Degree is fast, and CIS image acquisition units have the characteristics of picking rate is fast, simple in construction, easy to use, cost is low, so adopting
Bar cigarette image is acquired with CIS image acquisition units.
In embodiments of the present invention, CIS image acquisition units include：FPGA module, the CIS being connected respectively with FPGA module
Image capture module, modulus dress mold changing block, memory module, picking rate matching module and image output module, and for carrying
For the power module of operating voltage；The analog-to-digital conversion module is used to analogue data being converted to data signal；The FPGA moulds
Block is used for the SECO for realizing whole system；The picking rate matching module is used to be adjusted according to the transfer rate of conveyer
Whole CIS image modules gather image rate；The memory module is used to cache the digital image data collected；Described image is defeated
Going out module is used to the bar cigarette image collected being sent to the data processing module.
Strip source is used to provide illumination in image acquisition process, it is to avoid because the reason for external environment, causing illumination
Deficiency, so as to influence the quality of the image of collection.Rated voltage 24V, rated power 5W-10W strip light can such as be used
Source.
The present invention also provides a kind of cigarette sorting detection method based on CIS image acquisition units, comprises the steps：
(1) bar stick image data base is set up；
(2) IMAQ is carried out to bar cigarette by CIS image acquisition units；
(3) bar cigarette image is handled, obtains the feature of bar cigarette image；
(4) species and quantity of bar cigarette are determined；
(5) judge whether sorting bar cigarette is consistent with original order.
In embodiments of the present invention, step (1) is independently of step (2), (3).Step (1) can before step (2),
Can also be after step (3).
In embodiments of the present invention, step (1) is specially：
(1.1) IMAQ instrument (CCD camera, CIS image acquisition units etc.) pick-up slip stick is utilized, bar cigarette is obtained
Sign image.
(1.2) bar stick image is handled using morphological method, obtains bar stick contour images.
(1.3) bar stick contour images are handled using image pyramid, obtains bar stick feature set.
(1.4) bar stick feature set is marked.
In embodiments of the present invention, step (1.3) acquisition bar stick feature set is specially：
(1.31) bar stick contour images are subjected to gray processing processing, obtain bar stick profile gray level image.
(1.32) Hessian conversion is carried out to bar stick gray level image, obtains bar stick changing image.
(1.33) change the size of yardstick, obtain a series of bar stick changing image, obtain image pyramid.
(1.34) bar stick characteristic point is found in image pyramid, bar cigarette marker characteristic is obtained.
(1.35) to mark repeat step (1.31)-(1.34) of different bar cigarettes, corresponding bar stick feature is obtained,
Constituting bar stick feature set.
In embodiments of the present invention, the feature of acquisition bar cigarette image described in step (3) is specially：
(3.1) bar cigarette image is read, and gray processing processing is carried out to it, bar cigarette gray level image is obtained.
(3.2) Hessian conversion is carried out to bar cigarette gray level image, obtains bar cigarette changing image.
(3.3) change scale size, obtain a series of bar cigarette changing image, constitute image pyramid.
(3.4) bar cigarette characteristic point is found in image pyramid, bar cigarette feature is obtained.
In embodiments of the present invention, the type and quantity of step (4) the determination bar cigarette are specially：
(4.1) kd tree preliminary matches bar cigarette features and bar stick database are utilized.
(4.2) matching and correlation is carried out using random sampling unification algorism, rejects erroneous matching.
(4.3) acquisition in step (4.2) is correctly matched into logarithm to be compared with matching number threshold value, and counts bar cigarette
Quantity.
By the contemplated above technical scheme of the present invention, compared with prior art, due to utilizing CIS image capturing systems
Figure is adopted in progress, it is possible to increase the speed of IMAQ, strengthens real-time, and reduce the beneficial effect of cost；Due to the present invention
The method for sorting of offer, it is ensured that the accuracy of order and ensure that sort efficiency, versatility and practicality are higher, full well
The foot demand of present field of industry detection.
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
CIS (Contact Image Sensor, contact-type image sensor) IMAQ is based on the present invention relates to one kind
Cigarette sorting detecting system and method.As shown in figure 1, the cigarette sorting based on CIS image acquisition units that the present invention is provided
Detecting system includes：Conveyer 1, image collecting device 2, data processing module 3 and display device 4；Image collecting device 2
In the surface of conveyer 1, while image collecting device 2 is connected by data wire with data processing module 3.Bar cigarette is by passing
Device 1 is sent to be transported, when bar cigarette passes through 2 lower section of image collecting device, 2 pairs of bar cigarettes of image collecting device carry out image and adopted
Collection, then by image transmitting to data processing module 3, data processing module 3 carries out analysis detection to it.
Image collecting device 2 includes 3 CIS image acquisition units 21 and two uniform strip sources 22；Image collector
Put inner upper centre and the first CIS image acquisition units 21 are installed, both sides are respectively provided with a uniform bar inside device
Shape light source, the 2nd CIS image acquisition units 21 and the 3rd CIS image acquisition units 21 are separately positioned on the both sides of conveyer.
Such as Fig. 3, image collecting device 21 includes：CIS image capture modules, analog-to-digital conversion module, FPGA module (Field-
Programmable Gate Array, field programmable gate array), memory module, power module, picking rate matching module,
Image output module.When system works, the bar cigarette that CIS image capture modules pass through to lower section carries out IMAQ and output one is
Row analogue data, data signal is converted into by analog-to-digital conversion device, and FPGA is responsible for the SECO of whole system, picking rate
CIS image modules are adjusted according to the transfer rate of conveyer with module and gather image rate, FPGA is by the digitized map collected
As data buffer storage is into memory module, the bar cigarette image collected is then sent to by data processing mould by image output module
Block, power module is that whole system is powered, it is necessary to which outside 5V powers.Uniform strip source is a kind of each luminous point intensity of illumination one
The strip source of sample, it is therefore intended that make bar cigarette brightness of image uniform, it is to avoid because image each several part brightness is inconsistent, and influence it
Testing result afterwards.
The present invention carries out IMAQ by image collecting device, improves real-time and reduces cost, system and method
Sort bar cigarette, it is ensured that the accuracy of order and ensure that sort efficiency.
In the cigarette sorting system based on CIS image acquisition units that the present invention is provided, pass through CIS image acquisition units
Scanning, obtain bar cigarette coloured image, by coloured image carry out gray processing processing, then to bar cigarette gray level image carry out
Processing, obtains the feature of bar cigarette, is then matched with well-established bar stick image data base, determine the species of bar cigarette
And quantity, ultimately produce detection order and compared with original order, judge that whether sorting product is consistent, and pass through with original order
Display device is by result publicity.
Present invention also offers a kind of cigarette sorting detection method based on CIS image acquisition units, comprise the following steps that：
(1) bar stick image data base is set up
S11 utilizes IMAQ instrument (CCD camera, CIS image acquisition units etc.) pick-up slip stick, obtains bar cigarette-brand
Will image.
S12 is handled bar stick image using morphological method, obtains bar stick contour images.
S13 is handled bar stick contour images using image pyramid, obtains bar stick feature set.
Bar stick contour images are carried out gray processing processing by S131, obtain bar stick profile gray level image.
S132 carries out Hessian conversion to bar stick gray level image, obtains bar stick changing image.
S133 changes the size of yardstick, obtains a series of bar stick changing image, obtains image pyramid.
S134 finds bar stick characteristic point, obtains bar cigarette marker characteristic.
S135 obtains corresponding bar stick feature, constituting bar to the mark repeat step S131-S134 of different bar cigarettes
Stick feature set.
Bar stick feature set is marked S14.
Bar stick image data base includes different types of stick feature set, and it is special that feature set includes bar stick
Levy, wherein bar stick feature there are two classes：One class is used to distinguish bar cigarette species, a class for distinguishing under one species bar cigarette not
Same species.
(2) IMAQ is carried out to bar cigarette by CIS image capturing systems
Image collecting device is to carry out IMAQ, CIS image acquisition units run-downs by CIS image acquisition units
Some row image informations can be obtained, and (line number scanned each time is set according to the speed of conveyer and the resolution ratio of collection image
It is fixed), so as to obtain bar cigarette image.
(3) bar cigarette image is handled, obtains the feature of bar cigarette image
S31 reads bar cigarette image, and gray processing processing is carried out to it, obtains bar cigarette gray level image.
S32 carries out Hessian conversion to bar cigarette gray level image, obtains bar cigarette changing image.
S33 changes scale size, obtains a series of bar cigarette changing image, constitutes image pyramid.
S34 finds bar cigarette characteristic point, obtains bar cigarette feature.
(4) bar cigarette species and quantity are determined using bar stick image data base
S41 utilizes kd tree preliminary matches bar cigarette features and bar stick database.
S42 carries out matching and correlation using random sampling unification algorism, rejects erroneous matching.
S43 determines generation test order after bar cigarette species and quantity.
(5) generation detects order and is compared with original order, judges whether detection order is consistent with original order
The cigarette sorting system and method based on CIS image acquisition units that the present invention is provided, can at least bring following has
Beneficial effect：In the present invention, IMAQ is carried out using CIS image acquisition units, improves the speed of IMAQ, enhance
Real-time, and reduce cost.Meanwhile, the cigarette sorting method that the present invention is provided disclosure satisfy that point of variety classes bar cigarette
Pick, versatility is higher, and in the case where ensureing detection rates, the accuracy rate of sorting is higher, and present work is met well
The demand of industry detection field.
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below in conjunction with the accompanying drawings and implement
The present invention is specifically described example.Drawings in the following description are only some embodiments of the present invention.For this area
For those of ordinary skill, on the premise of not paying creative work, other accompanying drawings can also be obtained according to these accompanying drawings.
As shown in figure 1, the present invention provides a kind of cigarette sorting system based on CIS image acquisition units.Can be with from figure
Find out, the system includes：Conveyer 1, image collecting device 2, data processing module 3 and display device 4.Idiographic flow is such as
Under：Bar cigarette is transmitted by conveyer, during by below device case, and CIS IMAQs pipe carries out image to bar cigarette image
Collection, while image is transmitted into data processing module by USB3.0, data processing module by the bar cigarette image of acquisition with it is built
The bar stick image data base stood is matched, and obtains bar cigarette species, while obtaining the quantity of bar cigarette by counting, afterwards
Generation detects order and is compared with original order, judges whether sorting product is consistent with order, is filled finally by display
Put and will determine that result is shown.In image acquisition process, conveyer is passed in real time by cable to CIS image acquisition units
The speed of defeated current conveyer, the frequency acquisition of CIS IMAQ pipes is adjusted with this.
As shown in Fig. 2 the present invention provides a kind of cigarette sorting method based on CIS image acquisition units.Can be with from figure
Find out, the present invention includes 4 processes：Set up bar stick image data base, obtain the feature and bar stick figure of bar cigarette image
As database matching determines bar cigarette species and quantity, generation detection order and is compared with original order, judge that detection is ordered
It is single whether to be consistent with original order.
As shown in figure 4, setting up bar stick image data base specific steps includes：
It is the species for more quickly recognizing shaping cigarette in sort process in order to after to set up bar stick image data base, if
New bar tobacco kind is added afterwards, only need to be updated bar stick image data base, be substantially increased the reality of sorting system
With property and versatility, so the foundation of bar stick image data base is necessary.
Pick-up slip stick is (such as using IMAQ instrument (CCD camera or CIS image acquisition units) by S11：Chinese board is fragrant
" China " of cigarette, " Yellow Crane Tower " of Yellow Crane Tower board cigarette), obtain bar stick image.
CCD camera feature is that image pixel integrated level height of collection, small power consumption, sensitivity is high, performance stably, shock resistance,
CIS image acquisition units features are that picking rate is fast, simple in construction, easy to use, cost is low.CCD camera is relative to CIS images
Collecting unit, the image of collection is apparent, and picking rate is relatively slow.When setting up bar stick image data base, to figure
As quality requirement is higher, so carrying out IMAQ using CCD camera.
In selector bar stick, identification should be selected high, recognize distinct mark, such as：" China " of Chinese board cigarette.
Simultaneously it is noted that whether bar cigarette two sides mark is consistent, if inconsistent, it will choose, such as：" China " of Chinese board cigarette and
" chunghwa ", because during after, when bar cigarette is sent on conveyer by sorter, CIS image capturing systems are adopted
When collecting bar cigarette image, it is impossible to it is determined that any face collected.The difference of Soft Roll and Hard Roll is there may be additionally, due to same bar cigarette, because
The mark of this two sides will also be gathered, such as：The soft box and hard box of Chinese board cigarette, the mark of selection is " soft box ", " hard
Box ".For Chinese board cigarette, selection is masked as " China ", " chunghwa ", " soft box " and " hard box ".Wherein " China " and
" chunghwa " purpose is in order to distinguish bar cigarette species, and " soft box " and " hard box " purpose are in order to distinguish the species under Chinese board cigarette.
For other cigarettes, the selection of bar stick is also similar.
In pick-up slip stick image, we utilize CCD camera (Charge-coupled Device, Charged Couple member
Part) be in order to ensure the definition of bar stick image, while note using CCD camera gather sign image when, bar stick
Image should be located at the pinpointed focus of CCD camera, and positioned at image center, (bar stick is located at image other positions to bar stick
Also can), be so for convenience after bar cigarette identifying processing.
S12 is handled bar stick image using morphological method, obtains bar stick contour images.
Closing operation of mathematical morphology is to carry out dilation operation to image first, and erosion operation is then carried out again.Dilation operation is
All background dots contacted with object are merged into the process of the object.Erosion operation is the one of all boundary points for eliminating object
The process of kind, the result is that making remaining object along an inner edges pixel smaller than original object.
Black cap computing is the difference of design sketch and original image of the image after closed operation.Design sketch after black cap computing can be obtained
Obtain clearly bar stick profile.
As shown in figure 5, constituting bar stick feature set specific steps include
S13 is handled bar stick contour images using image pyramid, obtains bar stick feature set.
Bar stick contour images are carried out gray processing processing by S131, obtain bar stick profile gray level image.
S132 carries out Hessian conversion to bar stick gray level image, obtains bar stick changing image.
A point x=(x, y) ((x, y) is the transverse and longitudinal coordinate of pixel) in given image I, Hessian matrix Hs (x, σ)
Yardstick is that σ is defined as follows at x：
Here Lxx(x, σ), Lxy(x, σ), Lyx(x, σ), Lyy(x, σ) is difference Gauss second order derviation number Image I convolution is in x.
For Lxx(x, σ), Lxy(x, σ), Lyx(x, σ), Lyy(x, σ) is calculated,With image I's
Convolution algorithm is converted into the computing of square box wave filter, square box wave filter 9*9 templates as shown in Figure 6, figure (a), figure (b), figure (c) point
Do not representOperator,OperatorOperator withOperator is the same),The pixel in the region of mark 1 represents 1, mark in operator, figure
The pixel in 2 regions represents -1, and the pixel for there be not marked region represents 0.L can just be calculated by carrying out convolution by template and imagexx
(x, σ), Lxy(x, σ), Lyx(x, σ), Lyy(x, σ).
Now Det (H) ≈ LxxLyy-(wLxy)2(Det (H) represents Hessian determinants of a matrix；W represents wave filter
Associated weight, relevant with filter size, generally 0.9) we take, and by calculating Det (H), obtains certain point (x, y) in image
Response, then pixel all in traversing graph picture obtains bar stick changing image under a certain yardstick.
S133 changes the size of yardstick, obtains a series of bar stick changing image, obtains image pyramid.
Due to bar cigarette outer packing disunity, i.e., the height of bar cigarette is inconsistent, causes CIS image capturing systems to gather image
When can there are problems that the depth of field, therefore the bar cigarette classification after being ensured using the image pyramid of different yardstick image constructions is more
Plus it is accurate rapid.
Image pyramid is made up of O group images, and every group of image has S layers, and for the square box wave filter mould of every tomographic image
Board size size L=3 × (20+1(S+1)+1), now corresponding yardstickO=1, S=4 are generally selected, then square box
Filter size is 9*9,15*15,21*21,27*27.By changing the size of square box wave filter, repeat step S132 is obtained
A series of stick changing images, so as to form image pyramid.
S134 finds bar stick characteristic point, obtains bar cigarette marker characteristic.
Because image pyramid is made up of different scale images, in order to find characteristic point wherein, 3 × 3 are employed
The non-maximum of × 3 neighborhoods suppresses：Any pixel point in a certain tomographic image of image pyramid is used as detection characteristic point and same chi
Spend the pixel in layer in adjacent 8 field and above and under two scale layers, 9 pixels (totally 26 pixels)
It is compared, finds out the wherein maximum pixel of characteristic value (Hessian matrix determinants), then the point is characterized a little.By upper
Step is told, we are obtained with a series of bar stick characteristic point.
Because bar cigarette is arbitrarily put on conveyer, cause bar in the image that CIS image capturing systems collect
The position and direction of stick are also arbitrary.For bar cigarette of effectively classifying afterwards, so needing bar stick feature to have
Rotational invariance.
Harr wavelet characters in statistical nature point field, i.e., centered on characteristic point, it is that (σ is characterized 6 σ to calculate radius
Scale-value where point) circular field in point in x, the Haar small echos (the Haar small echo length of sides take 4 σ) in y directions are responded, Harr
(pixel in the region of mark 1 represents 1 in figure, and the pixel in the region of mark 2 represents -1, does not have the pixel of marked region by small echo template such as Fig. 7
Represent 0).Convolutional calculation is carried out by template and image to go out after response of the image on x the and y directions of Harr small echos, with
Centered on characteristic point, subtended angle is the fan-shaped sliding window of π/3, using step-length m (the usual values of m is 0.2rad) rotational slide window,
And image Harr small echo responses dx, dy in sliding window are added up, obtain a vector (mw, θw):
Then border circular areas is traveled through, the direction of most long vector is selected as the principal direction of this feature point.So, to characteristic point
Calculated one by one, obtain the principal direction of each characteristic point.
Centered on characteristic point, the σ (σ is characterized the scale-value at a place) of 20 σ × 20 is divided into 4 × 4 along principal direction
Block, each block carries out response calculating with the σ of size 2 Harr templates, and counts ∑ dx, Σ in each block | dx |, Σ
Dy, Σ | dy | characteristic vector is formed, so each characteristic point can just be represented with 64 dimensional feature vectors.
64 dimensional feature vector expressions are carried out to each characteristic point, it is special that these characteristic vector set are obtained into bar stick
Levy.
S135 obtains corresponding bar stick feature, constituting bar to the mark repeat step S131-S134 of different bar cigarettes
Stick feature set.
Bar stick feature set is marked S14.
Mark of the bar stick feature set comprising different types of cigarette, for convenience after bar cigarette classification processing, need
Bar stick feature is marked.Marking convention is：Such as really bar cigarette species is N, and the feature of every kind of cigarette is no more than M
Individual, then bar stick signature shape such as XY (1≤X≤N, 1≤Y≤M) is such as：The characteristic indication of Chinese board cigarette include " in
China ", " chunghwa ", " soft box ", " hard box "；For this four characteristic indications mark respectively correspond to " 11 ", " 12 ",
“13”、“14”。
As shown in figure 8, obtaining bar cigarette characteristics of image specific steps includes：
S31 reads bar cigarette image, and gray processing processing is carried out to it, obtains bar cigarette gray level image.
S32 carries out Hessian conversion to bar cigarette gray level image, obtains bar cigarette changing image.
It is similar with step S132, utilize matrixPass through square box wave filter 9*9 moulds simultaneously
Plate calculates Det (H) ≈ LxxLyy-(wLxy)2, the response of certain point (x, y) in image is obtained, is then owned in traversing graph picture
Pixel, obtain bar cigarette changing image under a certain yardstick.
S33 changes scale size, obtains a series of bar cigarette changing image, constitutes image pyramid.
Similar with step S133, selection square box filter size is 9*9,15*15,21*21,27*27.By changing square box
The size of wave filter, repeat step S132 obtains a series of cigarette changing images, so as to form 1 group 4 layers of image pyramid.
S34 finds bar cigarette characteristic point, obtains bar cigarette feature.
It is similar with step 134, employ the non-maximum of 3 × 3 × 3 neighborhoods and suppress：In a certain tomographic image of image pyramid
Any pixel point as the pixel in adjacent 8 field in detection characteristic point and same scale layer and above and under two
Individual 9 pixels of scale layer (totally 26 pixels) are compared, and find out wherein characteristic value (Hessian matrix determinants) maximum
Pixel, then the point be characterized a little.By appealing step, we are obtained with a series of bar stick characteristic point.System
The Harr wavelet characters in characteristic point field are counted, then selects the direction of most long vector as the principal direction of this feature point, passes through
Characteristic point is calculated one by one, the principal direction of each characteristic point is obtained.Centered on characteristic point, find 64 along principal direction and tie up
Characteristic vector.Finally, these characteristic vector set are obtained into bar cigarette feature.
As shown in figure 9, determining the type and quantity specific steps of bar cigarette includes：
S4 is matched the bar cigarette feature of acquisition and bar stick image data base, determines the species and number of bar cigarette
Amount.Specially：
S41 preliminary matches bar cigarette feature and bar stick database.
Image in bar cigarette marker image database is referred to as Prototype drawing, the bar cigarette that will be gathered with CIS image capturing systems
Image is referred to as real-time figure.It is R to make feature point description in Prototype drawingi=(ri1, ri2..., ri64), feature point description in real-time figure
Son is Si=(si1, s ..., rs64).Then the similarity measurement between any two descriptions can be expressed as：
A series of feature point description (R of matchings can be obtained according to above formulai, Si), they need to meet(SjRepresent in real-time figure apart from RiClosest approach, SpRepresent in real-time figure apart from RiSecondary near point,
Threshold represents distance threshold, what usual distance threshold took according to the actual requirements), whenDuring less than distance threshold, then this
Matching can take；Conversely, this matching can not be taken, it is necessary to remove.
During matching characteristic point, we carry out Feature Points Matching using kd trees.Its basic thought is：
(1) binary search tree first.By comparing the value and node to be checked of split vertexes division dimension, if learning to be checked
The value that node is less than or equal to split vertexes division dimension is put into left subtree, otherwise is put into right subtree branch until leaf knot
Point.
(2) along " searching route " with regard to the possibility point of potential arest neighbors can be found, at this approximate point and point to be checked
In same sub-spaces.
(3) trace back to again in searching route, and judge be in other child node spaces of the remaining node in searching route
It is no may have Distance query point closer to sample point.
(4) if it were possible, then need to jump to removal search in child node space (is added to search by other child nodes
In path).Repeat this process and know that searching route is sky.
In embodiments of the present invention, be as Figure 10 (a) assumes our sample set (2,3) (5,4) (9,6) (4,7) (8,
1) (7,2) }, our point to be checked is respectively (2.1,3.1) and (2,4.5).
Such as Figure 10 (b), we first search point (2.1,3.1) first, reach (5,4) in the test of (7,2) point, reach afterwards
(2,3), then the node of query path be { (7,2) (5,4) (2,3) }, from query path take out (2,3) as currently most
The nearest neighbor point of good node, distance is 0.141.
Then (5,4) are traced back to, with (2.1,3.1) for the center of circle, draws and justifies using 0.141 as radius, it is found that the circle is not peaceful
Face y=4 intersects, such as Figure 10 (b), so the right subspace removal search of (5,4) is not jumped to, because can not possibly have more in right subspace
Nearly sample point.
Then (7,2) are traced back to, are that radius draws circle discord plane x=with 0.141 with (2.1,3.1) for the center of circle similarly
7 intersect, so without being transferred to removal search in (7,2) right subspace.
Therefore, searching route is sky, terminates whole search, by the nearest neighbor point of (2,3) as (2.1,3.1), most low coverage
From for 0.141.
Carry out query point (2.4.5) below, such as Figure 10 (c) equally first finds searching route.(7,2) place test reach (5,
4), at (5,4) place, test reaches (4,7), and then the node in searching route is { (7,2) (5,4) (4,7) }, from searching route
Middle take out (4,7) are as the nearest neighbor point of current Best knots, and distance is 3.202.
Then (5,4) are dateed back, with (2,4.5) for the center of circle, is that radius picture circle intersects with plane y=4 with 3.202, such as schemes,
So needing to jump to the left subspace removal search of (5,4).So (2,3) are added in searching route, present searching route
In node be { (7,2) (2,3) }；The distance of (5,4) and (2,4.5) is 3.04 in addition<3.202, so (5,2) are set to most
Neighbor Points, and apart from for 3.04.
Trace back to (2,3), (2,3) be leaf node, directly judge (2,3) whether from (2,4.5) closer to, calculate obtain away from
From for 1.5, so nearest neighbor point is updated to (2,3), distance is 1.5
(7,2) are traced back to, are that radius picture circle discord plane x=7 intersects with 1.5 similarly with (2,4.5) for the center of circle, institute
With the right subspace removal search without jumping to node (7,2).
So far, searching route is sky, terminates whole search, and (2,3) are used as into the nearest neighbor point of (2,4.5), arest neighbors away from
From for 1.5.
S42 carries out matching and correlation, rejects erroneous matching.
Due to that there can be the matching of mistake in crucial Point matching, so to remove the matching of mistake, matching rate is improved.Profit
Erroneous matching is filtered out with random sampling unification algorism, its step is：
(1) it is random to extract 4 sample datas (be collinearly between this 4 samples) out from data set, calculate conversion square
Battle array H, is denoted as model M.
Transformation matrixDue to(wherein (x, y) represents special in Prototype drawing
A position is levied, (x ', y ') is characteristic point position in real-time figure, s is that scale parameter wherein generally makes h33=1), it is uncommon by 4
The sample data of line can just calculate h11、h12、h13、h21、h22、h23、h31、h32, so as to obtain transformation matrix H.While model
M is：
(2) projection error of all data and model M in data set is calculated, if error is less than projection threshold value, interior point is added
Collect I；
(3) if point set I element numbers are more than optimal interior point set best in current, best=I is updated, is changed while updating
Generation number k；
(3) if iterations is more than k (usual k takes 20), exit；Otherwise iterations adds 1, and repeats above-mentioned step
Suddenly；
By continuous iteration, erroneous matching can be rejected, correct matching is obtained, and count correct matching logarithm.
S43 determines generation test order after bar cigarette species and quantity.
Correct matching logarithm is compared with matching number threshold value, if correctly matching logarithm is more than or equal to matching number threshold
Value, it is determined that the species of bar cigarette, if correctly matching logarithm is less than matching number threshold value, under a kind of bar stick feature and bar
Cigarette feature is matched, until determining bar cigarette species.
When needing to detect the species of multiple cigarettes in real time, such as Figure 11 (a), first, when preceding article cigarette characteristics of image and bar cigarette
Flag sign collection is matched according to step S41-S43, from cigarette 1 to cigarette N, until determining bar cigarette species；Then, if when preceding article cigarette
Species is defined as after cigarette X, and the species for next rule cigarette is determined, we are using principle is closed on, shown in such as Figure 11 (b), from cigarette X
For starting matched position, matched according to cigarette X to cigarette X-1 order, until it is determined that bar cigarette species.Finally count variety classes
The quantity of bar cigarette.
It can be saved the time using above-mentioned matched rule, the real-time of detection be improved, because next rule cigarette is largely
May with as preceding article cigarette species, it is to avoid every time matching all since cigarette 1, cause waste of time.
Test order is compared with original order, is comprised the following steps that：
After bar cigarette type and quantity are determined in step S4, due to slug stick feature set described in step S14 before,
The test order then generated can exist in the form of a flag, such as：10 soft China, 5 hard China, then testing order is
" 1,310 145 ", wherein " 13 ", " 14 " represent soft Chinese, hard China respectively, " 10 ", " 5 " represent respectively soft Chinese quantity and
The quantity of hard China.Equally represent original order in this form, afterwards by compare character string judge test order with it is original
Whether order is consistent, finally will determine that result is shown by display device.
The specific embodiment of invention is described in detail above, but the present invention be not restricted to it is described above specific
Embodiment, it is intended only as example.To those skilled in the art, any equivalent modifications and replacement carried out to the system
Also all among scope of the invention.Therefore, the impartial conversion and modification made under the spirit and scope for not departing from invention, all
It should cover within the scope of the invention.
Claims (10)
1. a kind of cigarette sorting detecting system based on CIS IMAQs, it is characterised in that including：Conveyer (1), image
Harvester (2), data processing module (3) and display device (4)；
Described image harvester (2) is located at the surface of the conveyer (1), and described image harvester (2) and institute
State data processing module (3) connection；When bar cigarette is transported by conveyer (1) and when bar cigarette passes through image collecting device
(2) during lower section, described image harvester (2) carries out IMAQ to bar cigarette, and by image transmitting to data processing module
(3), data processing module (3) carries out analysis detection to it.
2. cigarette sorting detecting system as claimed in claim 1, it is characterised in that described image harvester (2) includes：Three
Individual CIS image acquisition units and two strip sources, the first CIS image acquisition units are arranged on the surface of conveyer, the
Two CIS image acquisition units and the 3rd CIS image acquisition units are separately positioned on the both sides of conveyer, and the first CIS images are adopted
Collection unit both sides are respectively provided with a strip source, for causing the brightness of image of collected bar cigarette uniform, the first CIS images
Collecting unit is used for the image of pick-up slip cigarette upper surface；2nd CIS image acquisition units and the 3rd CIS image acquisition units difference
Image for two sides of pick-up slip cigarette.
3. cigarette sorting detecting system as claimed in claim 2, it is characterised in that three CIS image acquisition units structure phases
Together, include：FPGA module, the CIS image capture modules being connected respectively with the FPGA module, analog-to-digital conversion module, storage
Module, picking rate matching module and image output module, and for providing the power module of working power；
The CIS image capture modules are used to carry out IMAQ to bar cigarette and export a series of analogue datas；The modulus turns
Mold changing block is used to analogue data being converted into data signal；The FPGA module is used for the SECO for realizing whole system；Institute
Stating picking rate matching module is used to adjust CIS image modules collection image rate according to the transfer rate of conveyer；It is described
Memory module is used to cache the digital image data collected；Described image output module is used to send out the bar cigarette image collected
Give the data processing module.
4. a kind of cigarette sorting detection method based on CIS IMAQs, it is characterised in that comprise the steps：
S1：IMAQ is carried out to bar cigarette by CIS image acquisition units；
S2：The bar cigarette image of collection is handled, the feature of bar cigarette image is obtained；
S3：The type and quantity of bar cigarette are determined according to the feature of the bar cigarette image and bar stick image data base；
S4：Judge that whether sorting bar cigarette is consistent, and show result with original order according to the type and quantity of bar cigarette.
5. cigarette sorting detection method as claimed in claim 4, it is characterised in that the bar stick image data base is built
It is vertical to specifically include：
(1.1) using CCD camera pick-up slip stick and bar stick image is obtained；
(1.2) bar stick image is handled using morphological method, obtains bar stick contour images；
(1.3) bar stick contour images are handled using image pyramid method, obtains bar stick feature set；It is described
Bar stick feature set includes the mark of different types of cigarette；
(1.4) bar stick feature set is marked.
6. cigarette sorting detection method as claimed in claim 4, it is characterised in that step (1.3) is specially：
(1.31) bar stick contour images are subjected to gray processing processing, obtain bar stick profile gray level image；
(1.32) Hessian conversion is carried out to bar stick gray level image, obtains bar stick changing image；
(1.33) by changing the scale size and repeat step (1.32) of operator in Hessian conversion, a series of bar is obtained
Stick changing image, constitutes image pyramid；
(1.34) bar stick characteristic point is found in image pyramid, bar cigarette marker characteristic is obtained；
(1.35) to mark repeat step (1.31)-(1.34) of different bar cigarettes, corresponding bar stick feature is obtained, is constituted
Bar stick feature set.
7. the cigarette sorting detection method as described in claim 4 or 5, it is characterised in that step S2 is specially：
S21 carries out gray processing processing to bar cigarette image, obtains bar cigarette gray level image；
S22 carries out Hessian conversion to bar cigarette gray level image, obtains bar cigarette changing image；
S23 obtains the conversion of a series of bar cigarette by changing the scale size and repeat step S22 of operator in Hessian conversion
Image, constitutes image pyramid；
S24 finds bar cigarette characteristic point in described image pyramid, and obtains the feature of bar cigarette image.
8. the cigarette sorting detection method as described in claim any one of 4-6, it is characterised in that step S3 is specially：
S31 realizes bar cigarette by the way that the Prototype drawing in the feature of bar cigarette image and the bar stick image data base is compared
The preliminary matches of feature；
S32 realizes matching and correlation by rejecting erroneous matching in the preliminary matches；
S33 determines bar cigarette type and quantity according to the result after matching and correlation, and is ordered according to the generation test of bar cigarette type and quantity
It is single.
9. cigarette sorting detection method as claimed in claim 8, it is characterised in that bar is determined according to the result after matching and correlation
Cigarette species is specially：
Correct matching logarithm is obtained in step S32, logarithm will be correctly matched with matching number threshold value (according to bar cigarette image characteristic point
Number setting, ten) usual value 6 percent is compared for bar cigarette image characteristic point number, if correctly matching logarithm is big
In equal to matching number threshold value, it is determined that the species of bar cigarette, if correctly matching logarithm is less than matching number threshold value, lower one kind
Bar stick feature is matched with bar cigarette feature, until determining bar cigarette species.
10. cigarette sorting detection method as claimed in claim 8, it is characterised in that in step s 4, is ordered by comparing test
Whether single character string with original order is consistent with original order to judge to test order, and will determine that result is shown.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710338029.4A CN107169961A (en) | 2017-05-15 | 2017-05-15 | A kind of cigarette sorting detecting system and method based on CIS IMAQs |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710338029.4A CN107169961A (en) | 2017-05-15 | 2017-05-15 | A kind of cigarette sorting detecting system and method based on CIS IMAQs |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107169961A true CN107169961A (en) | 2017-09-15 |
Family
ID=59816004
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710338029.4A Pending CN107169961A (en) | 2017-05-15 | 2017-05-15 | A kind of cigarette sorting detecting system and method based on CIS IMAQs |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107169961A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108787486A (en) * | 2018-06-05 | 2018-11-13 | 四川九哈科技股份有限公司 | Cigarette identification method for sorting and its cigarette based on deep learning identify sorting equipment |
| CN108896568A (en) * | 2018-05-21 | 2018-11-27 | 北京沃佳玛科技有限公司 | A kind of information detecting system |
| CN109493316A (en) * | 2018-09-21 | 2019-03-19 | 浙江中烟工业有限责任公司 | Cartoning sealing machine cigarette packet based on similarity analysis and feature incremental learning lacks an intelligent detecting method |
| CN109903109A (en) * | 2017-12-08 | 2019-06-18 | 北京京东尚科信息技术有限公司 | Test method and device |
| CN110569753A (en) * | 2019-08-26 | 2019-12-13 | 湖南航天天麓新材料检测有限责任公司 | machine vision-based cigarette carton online identification method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102275656A (en) * | 2010-06-12 | 2011-12-14 | 中国科学院沈阳自动化研究所 | On-line visual pack absence detection device and detection method for cigarette pack cartoning machine |
| CN102625056A (en) * | 2012-03-30 | 2012-08-01 | 广东正业科技股份有限公司 | FPGA (Field-Programmable Gate Array)-based CIS (Contact Image Sensor) image acquisition system and acquisition method thereof |
| CN102769711A (en) * | 2011-05-04 | 2012-11-07 | 广东工业大学 | Customizable hardware-based CIS (contact image sensor) sensing signals acquisition method by using FPGA (Field Programmable Gata Array) |
| CN103440702A (en) * | 2013-08-27 | 2013-12-11 | 华南理工大学 | Multi-spectrum bill image collecting device based on CIS and processing method |
| CN104048966A (en) * | 2014-03-14 | 2014-09-17 | 东华大学 | Big-law-based cloth cover defect detection and classification method |
| CN105783783A (en) * | 2014-12-23 | 2016-07-20 | 桂林电子科技大学 | Aggregate morphological characteristic acquisition device based on multi-view linear array camera |
-
2017
- 2017-05-15 CN CN201710338029.4A patent/CN107169961A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102275656A (en) * | 2010-06-12 | 2011-12-14 | 中国科学院沈阳自动化研究所 | On-line visual pack absence detection device and detection method for cigarette pack cartoning machine |
| CN102275656B (en) * | 2010-06-12 | 2013-03-27 | 中国科学院沈阳自动化研究所 | On-line visual pack absence detection device and detection method for cigarette pack cartoning machine |
| CN102769711A (en) * | 2011-05-04 | 2012-11-07 | 广东工业大学 | Customizable hardware-based CIS (contact image sensor) sensing signals acquisition method by using FPGA (Field Programmable Gata Array) |
| CN102625056A (en) * | 2012-03-30 | 2012-08-01 | 广东正业科技股份有限公司 | FPGA (Field-Programmable Gate Array)-based CIS (Contact Image Sensor) image acquisition system and acquisition method thereof |
| CN103440702A (en) * | 2013-08-27 | 2013-12-11 | 华南理工大学 | Multi-spectrum bill image collecting device based on CIS and processing method |
| CN104048966A (en) * | 2014-03-14 | 2014-09-17 | 东华大学 | Big-law-based cloth cover defect detection and classification method |
| CN105783783A (en) * | 2014-12-23 | 2016-07-20 | 桂林电子科技大学 | Aggregate morphological characteristic acquisition device based on multi-view linear array camera |
Non-Patent Citations (3)
| Title |
|---|
| 冯春等: ""半自动分拣线上运动条烟的在线识别"", 《烟草科技》 * |
| 汪洋: ""扫地机器人定位算法设计与嵌入式系统实现"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 沈镇等: ""一种适于应用程序员设计ＦＰＧＡ异构系统的框架"", 《单片机与嵌入式系统应用》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109903109A (en) * | 2017-12-08 | 2019-06-18 | 北京京东尚科信息技术有限公司 | Test method and device |
| CN108896568A (en) * | 2018-05-21 | 2018-11-27 | 北京沃佳玛科技有限公司 | A kind of information detecting system |
| CN108787486A (en) * | 2018-06-05 | 2018-11-13 | 四川九哈科技股份有限公司 | Cigarette identification method for sorting and its cigarette based on deep learning identify sorting equipment |
| CN109493316A (en) * | 2018-09-21 | 2019-03-19 | 浙江中烟工业有限责任公司 | Cartoning sealing machine cigarette packet based on similarity analysis and feature incremental learning lacks an intelligent detecting method |
| CN109493316B (en) * | 2018-09-21 | 2021-06-15 | 浙江中烟工业有限责任公司 | Intelligent detection method for cigarette packet missing strip of carton sealing machine based on similarity analysis and feature increment learning |
| CN110569753A (en) * | 2019-08-26 | 2019-12-13 | 湖南航天天麓新材料检测有限责任公司 | machine vision-based cigarette carton online identification method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107169961A (en) | A kind of cigarette sorting detecting system and method based on CIS IMAQs | |
| CN110533722A (en) | A kind of the robot fast relocation method and system of view-based access control model dictionary | |
| CN110378909B (en) | Single wood segmentation method for laser point cloud based on Faster R-CNN | |
| CN104303193B (en) | Target classification based on cluster | |
| CN108830188A (en) | Vehicle checking method based on deep learning | |
| CN109784204B (en) | Method for identifying and extracting main fruit stalks of stacked cluster fruits for parallel robot | |
| CN108710913A (en) | A kind of switchgear presentation switch state automatic identification method based on deep learning | |
| CN110263705A (en) | Towards two phase of remote sensing technology field high-resolution remote sensing image change detecting method | |
| CN103279765B (en) | Steel wire rope surface damage detection method based on images match | |
| CN109100741A (en) | A kind of object detection method based on 3D laser radar and image data | |
| CN104484681B (en) | Hyperspectral Remote Sensing Imagery Classification method based on spatial information and integrated study | |
| CN107492094A (en) | A kind of unmanned plane visible detection method of high voltage line insulator | |
| CN106846344A (en) | A kind of image segmentation optimal identification method based on the complete degree in edge | |
| CN106408011A (en) | Laser scanning three-dimensional point cloud tree automatic classifying method based on deep learning | |
| CN109544501A (en) | A kind of transmission facility defect inspection method based on unmanned plane multi-source image characteristic matching | |
| CN106951905A (en) | Apple identification and localization method on a kind of tree based on TOF camera | |
| CN112184746A (en) | Transmission line insulator defect analysis method | |
| Changyi et al. | Apple detection from apple tree image based on BP neural network and Hough transform | |
| CN109492665A (en) | Detection method, device and the electronic equipment of growth period duration of rice | |
| CN113313081B (en) | Road traffic rod object classification method integrating vehicle-mounted three-dimensional laser point cloud and image | |
| CN111160451A (en) | Flexible material detection method and storage medium thereof | |
| CN109949229A (en) | A kind of target cooperative detection method under multi-platform multi-angle of view | |
| CN112069985A (en) | High-resolution field image rice ear detection and counting method based on deep learning | |
| CN109101981A (en) | Winding detection method based on global image bar code under a kind of streetscape scene | |
| CN114862855A (en) | Textile defect detection method and system based on template matching |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170915 |