CN109583535A - A kind of logistics bar code detection method, the computer program of view-based access control model - Google Patents
A kind of logistics bar code detection method, the computer program of view-based access control model Download PDFInfo
- Publication number
- CN109583535A CN109583535A CN201811443403.8A CN201811443403A CN109583535A CN 109583535 A CN109583535 A CN 109583535A CN 201811443403 A CN201811443403 A CN 201811443403A CN 109583535 A CN109583535 A CN 109583535A
- Authority
- CN
- China
- Prior art keywords
- bar code
- cargo
- picture
- logistics
- view
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06K—GRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
- G06K17/00—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations
- G06K17/0022—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations arrangements or provisious for transferring data to distant stations, e.g. from a sensing device
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Warehouses Or Storage Devices (AREA)
- Image Processing (AREA)
Abstract
The invention belongs to logistics detection technique fields, disclose logistics bar code detection method, the computer program of a kind of view-based access control model;Including picture capture;Identification and positioning cargo location;Identification and positioning barcode position;Read bar code;Error correction alarm.The present invention uses computer vision library, and the one-dimensional code position and its rotation angle oriented under complex background except means of making an uproar is added, intercepts out horizontal one-dimension code picture;Ensure that each cargo only corresponds to a label, and discovery mechanism of the cargo without label alarm is added;The operation of irrelevant data in picture is avoided, arithmetic speed is improved.While the present invention realizes that assembly line transports package, high definition camera is installed in fixed position, identifies logistics parcel location within the scope of camera fields of view, obtains the one-dimension code information pasted on cargo and be decoded, and then replace traditional assembly line either scans completely.
Description
Technical field
The invention belongs to the logistics bar code detection method of logistics detection technique field more particularly to a kind of view-based access control model,
Computer program.
Background technique
Currently, the prior art commonly used in the trade is such that
A large amount of with internet popularize, and logistic industry also rapidly develops therewith to emerge, and pursue the promotion of logistics speed, mention
High information-based, the intelligent specific gravity in logistics, and replace an inevitable development trend of a part of manpower as logistic industry.
In the pipelining of logistic industry, the general of industry is had become to logistics package number, label using one-dimensional bar code
Way, and logistics company big absolutely still needs to man-hour manually hand-held barcode scanning gun and is scanned to package bar code.This way not only increases manpower
Cost, and since scanner scanning direction needs horizontal aligument one-dimensional bar code and scanning distance is limited, seek operator must
It looks for and artificially finds and be aligned barcode position and angle, when logistics capacity increases, this mode will be difficult to meet speed requirement.
Currently, mobile camera installation is widely available, obtains digital picture and just become increasingly easy.Meanwhile artificial intelligence
Energy technology this year grows rapidly, and vision application technology is constantly researched and developed, and has had sufficient basic condition exploitation view-based access control model to know
The industrialization one-dimensional bar code information of other technology reads algorithm.
Existing one-dimensional bar code recognizer in the prior art, some cell phone application applications can also be in alignment levels one
One-dimensional bar code is identified under the premise of dimension code.But one-dimension code pastes angle random and non-horizontal in logistics goods, but also needs
Meet the condition that single cargo only identifies unique barcode, it is corresponding merely to identify that bar code does not play the role of examining.
At present still lack judge single cargo correspond to unique barcode number method, existing visual identity bar code function it is complete
Property, practicability do not meet logistic industry requirement.
In conclusion problem of the existing technology is:
(1) existing logistics also needs to meet the condition that single cargo only identifies unique barcode, if there is cargo is omitted
Put up bar coded sticker, i.e., there is a situation where on assembly line cargo discovery can not correspond to a unique barcode, and system does not have
It checks that this kind of situation gives a warning, eventually causes one to enter sorting line without information encapsulation.If can carry out
Anticipation and alarm can and alarm omission the case where putting up label and remind operator to handle.
(2) upper bar code pastes angle random and non-horizontal in existing logistics goods, and recognizer such as ZBar at present, needs
Bar code to be identified, which is in a horizontal position, just can be carried out correct identification.When bar code has biggish level angle in the picture
When deviation, the correct coding that can not read bar code will lead to.And can occur multiple bar codes in the visual field of video camera simultaneously,
When there is the case where multiple bar codes exist jointly, it is difficult to find a fixed direction of rotation for all bar code return water
Straight angle degree is read out.So must accomplish to rotate all bar codes in the visual field in this technology to angle can be read and carry out
It reads, to guarantee correctly to read bar code number and not omit.
Solve the difficulty and meaning of above-mentioned technical problem:
(1) how to judge how many cargo and bar code in camera fields of view, and could judge goods under what circumstances
Bar coded sticker is not puted up on object.Solve the problems, such as the case where this can be found that on cargo without pasting bar coded sticker.
(2) number represented by the bar coded sticker pasted on cargo how is decoded, and the bar coded sticker decodes
Digit length have to comply with actual bar code number format.Solve the problems, such as that this can prevent from misreading not meeting correct coding side
The bar code of formula.
(3) certain fault tolerant mechanism need to be designed, identification case occurs in a certain frame or a few frame images as noise jamming will lead to
Son reads label in generation quantity and precision mistake in short-term, but not issues wrong report.Due to image taking noise jamming etc.
Reason, program cannot be guaranteed that the accuracy and resolution ratio, the reading accuracy of bar code of object in every frame picture reach 100%,
Solve the problems, such as that this can prevent sensibility and report an error.
Summary of the invention
In view of the problems of the existing technology, the present invention provides a kind of logistics bar code detection method of view-based access control model,
Computer program.
The invention is realized in this way a kind of logistics bar code detection method of view-based access control model, specifically includes the following steps:
Step 1: picture capture: high definition high speed camera is fixed on a certain height and position right above conveyer belt, it is ensured that the visual field
The interior conveyer belt comprising certain length and clearly cargo profile and label picture can be taken under the premise of cargo movement.If
After having set photographing mode, camera starts to shoot high definition picture, while picture is converted into from cv::Mat format by program;
Step 2: identification and positioning cargo location: what a YOLO target detection model of pre-arranged, to the figure in step 1
Piece is calculated, and obtains all kind of object (cargo, label) and their coordinate position (object in captured picture
Xy coordinate value in picture), bounding box size (length of shared picture pixels);In this step, what be will test is all kinds of
Object is saved into different data types, wherein save each object coordinate position, bounding box size, and existed with bounding box
The picture intercepted out in original image;
Step 3: identification and positioning barcode position: after saving all kinds of object identifications in receiving step two, program is first
First " label " type objects are handled.The interception picture that " label " type objects save carries out region by bar code processing routine
Screening with return just, the bar code picture in available horizontal direction.
Step 4: it reads bar code: identifying bar code using Zbar method, handle, obtain horizontal by above step
The barcode size or text field picture, to the picture using being identified in Zbar program, the number of you can get it bar code;
Step 5: error correction alarm: " cargo " type objects identified in camera fields of view and " label " class picture are decoded
Number corresponded to.If when discovery has " cargo class " object to lack corresponding label number, issuing report in continuous multiple frames picture
It warns and operator is notified to handle.
Further, in step 1, the screening-mode of high definition high speed camera is set, light can be installed before camera fields of view position
The settable triggering mode of screening-mode is triggered grating, order camera is continuously clapped when there is cargo that will enter coverage by grid
According to a period of time;It can also set screening-mode to continuously to take pictures, no matter whether there is or not cargos on conveyer belt, all in accordance between certain time
It takes pictures every continuously.
Further, in step 2, the kind of object that need to detect identification is divided into " cargo class " (hexahedron chest or packed packet
Wrap up in), " tag class " (bar coded sticker), orient place center (coordinate) and bounding box size of the object in photo
(width, height under Picture Coordinate system);Used method is the YOLO algorithm of target detection in deep learning, arithmetic speed energy
Enough achieve the effect that real-time detection, meets target call;Specific step is as follows:
(1) what a reliable object detection model of the invention by pre-training, and model algorithm is deployed in program, it should
Model is a large amount of actual picture data of acquisition, and is trained the maturity model formed after operation;When camera has shot a figure
After piece, which will carry out target detection to the cargo in picture;
(2) target detection model will handle shooting picture one by one, obtain kind of object present in captured picture, a
Several and its bounding box information;It will appear multiple objects due to estimated in the same visual field, detection model need to detect own " cargo "
Class and " label " type objects, and format is designed to them, kind of object (chest, packed package, mark are saved wherein
Label), coordinate position (xy coordinate value of the object in picture), the information such as bounding box size (length of shared picture pixels), together
When the rectangle frame occupied by them is saved into data format from being cut into subgraph in original image, be sent to one-dimensional in next step
It is calculated in the program of barcode detection;As there is not cargo in the visual field, then this is cut and continued to execute without image
Step.Design data format is as follows.
Cargo class:
Name variable record be box or package, center_position variable records object coordinates, bounding
Variable records rectangle frame range, and arean matrix stores the rectangle frame subgraph intercepted from original image, and link_code variable indicates should
Whether fritter is corresponding with a bar code number.
Tag class:
Center_position variable records tag coordinate, and arean stores the rectangle frame subgraph intercepted from original image,
Num_type represents its bar code type decoded, and num_code represents digital content.Link_packegebox_num is indicated
" cargo " the type objects number that the label is pasted.
Further, in step 3, in the horizontal direction and when pixel is more visible, irrelevant is less just due to one-dimension code picture
It is easy to be read by reading code program, the identification bar code algorithm in the present invention is utilized that one-dimensional bar code image profile is obvious, image
The consistent feature of gradient is found the consistent region in change of gradient direction in the interception picture of " label " type objects identified and is made
For one-dimension code region that may be present, process is implemented are as follows:
(1) the picture P0 of " label " type objects intercepted out in step 2 (intercepting sub-pictures from original image) is subjected to gray processing
Processing is converted into gray level image, and by the noise in Gaussian Blur removal image, obtains gray level image P1;
(2) gradient value of each pixel both direction in image P1, i.e. two image arrays are calculated using Cany operator
Gx and Gy;Image array Gx is added with Gy, the available one visually apparent image array P2 of contour feature;In P2
The profile of bar code can be recognized clearly, and the main gradient of each profile point can be calculated according to the gradient value in XY both direction
Value and direction, calculation formula are as follows:
D=sqrt (x (I) ^2+y (I) ^2)
(3) image P2 will be divided into M fritter, the data structure of each fritter with the rectangle of m*m size are as follows:
The data structure class can be reserved for fritter pixel coverage in the picture, fall into the fritter in all contour images P2
The quantity of the point in region and the main gradient direction of point.
(4) Screening Treatment is done, rejects and does not include profile point and the image fritter very few comprising profile point, only retain profile point
Fritter (anchor_block) of the number in m/2 or more;
(5) the main gradient direction of remaining image fritter is calculated.Every 20 degree of increase is divided into one kind since 0 degree, has
1-9 totally 9 angular ranges, and according to the θ value of each profile point gradient direction determine belonging to number class number, i.e. anchor_
Belong_angle value in block data structure class.When the quantity for having profile point in a certain angular range accounts for all profiles
When putting 60% or more of quantity, it is determined that the angular range is the main angle direction of this fritter.If not provided, indicating this
Gradient is more chaotic in fritter, which is deleted.
(6) connection processing is done to the identical fritter of each gradient direction.Pixel distance is within d and main gradient direction one
The fritter Cluster merging of cause is at a region, and using the gradient direction as the gradient direction in " region " after merging;The step
It is middle can weed out distance farther out, the fritter that can not cluster of number rareness.
(7) judge that all merging rear region areas whether more than threshold value S, if it is, retaining, otherwise reject the block area
Domain;
(8) it can show that remaining region is bar code position by above step, and the region is obtained in (6)
Gradient direction θ is bar code level angle deviation, is rotated clockwise θ degree and obtains identifiable one-dimensional bar code picture.
Further, in step 4, which is open source program, is easy the information of reading horizontal one-dimensional bar code,
Output result is a certain length numeric string.Bar code be generally UPCUPC code, EANEAN code, 39 yards of Code, in 128 yards of Code
One kind the type of target bar can be exported together if identified successfully.
Further, in step 5, the step logical design are as follows:, can be by just by each after step 1 to step 4
" tag class " object encoded really is distinguished out, if respective coordinates shaft position is located at the rectangle frame model of " cargo " type objects
In enclosing, then judge that it can be somebody's turn to do " cargo " type objects and be labeled coding.When there are " cargo " objects not to be labeled for discovery
When, illustrate that corresponding " tag class " object is not detected in detection model, program record does not correspond to number once.It is such as continuous more
The generation of frame picture does not correspond to situation, then not corresponding to number can continue to increase, and can then issue alarm higher than certain certain numerical value, remind behaviour
The case where not putting up bar code on cargo has occurred as personnel.If all " cargo " class objects detected in period frame picture
Body can correspond to a number, then do not correspond to number clearing.Simultaneously, it is contemplated that can occur to have recognized when object has just enter into the visual field
Not the case where " cargo " and label not can enter the visual field out, when " cargo " type objects are in visual field boundary position and corresponding coding
When, it not will increase and do not correspond to number.Above procedure can interfere the hair for leading to sensitive alarm situation to avoid individual uncertain factors
Raw, only determining has object that can just alarm in the case where can not all reading label number in multiframe picture.
The bar code of logistics selection recognition is generally UPCUPC code, EANEAN code, 39 yards of Code, one in Code128 code
Kind, if the numeric type that may detect bar code in step 4 is not Selective type, cannot be gone with this number pair
Answer " cargo " object.
In conclusion advantages of the present invention and good effect are as follows:
The detection network structure source code of freight detection of the present invention is write by C/C++ language, and prototype is YOLO target detection net
Network structure changes its input layer, output layer structure in the present invention, and voluntarily acquisition data carry out the training of network model,
Stable algorithm structure is formed;Source code, network configuration and network parameter in the protection present invention.By testing 543 pictures
In include 853 cargos and label, can correctly identify wherein 827 object types, Classification and Identification accuracy reaches
96.96%, the accuracy of Area computing reaches 90.42%;
The method that the present invention finds one-dimensional bar code using gradient, source code are independently write using C Plus Plus, and principle is
Detect the gradient direction of profile point and profile point, thus filter out gradient reversely most consistent region as one-dimension code location
Domain;Used computer vision library (OPENCV), be added oriented except making an uproar means one-dimensional code position under complex background and its
Angle is rotated, and horizontal one-dimension code picture can be intercepted out.
The control test method and source code of cargo and label of the present invention can ensure that each cargo only corresponds to a label, and
It joined discovery mechanism of the cargo without label alarm, while in view of when cargo is not completely into the visual field, possible bar shaped
Code does not enter the visual field and reads wrong situation, avoids false alarm.It is tested in 319 cargos on assembly line and is contaminated with 38
The cargo of non-adhesive label removes, except 2 cargos are not by addition to not alarming, remaining, which is able to detect, pastes bar code information and alarm.
The method that deep learning of the present invention is combined with several picture processing, is extracted each using the method for deep learning
Then the segmented image of cargo handles segmented image using geometry method, avoids the operation of irrelevant data in picture, mention
High arithmetic speed, can be realized real-time detection.Test single picture target identification time-consuming be averaged 0.04s, and bar code recognition is put down
Time-consuming 0.05, every frame picture calculate average used time 0.09s, FPS 11.
The present invention realizes that fixed height installs high definition camera above assembly line, knows while assembly line is transported and wrapped up
Logistics parcel location within the scope of other camera fields of view positions the one-dimension code picture pasted on cargo and is decoded reading, and then complete
Replace traditional assembly line either scans entirely;It has the advantage that
(1) recognition speed precision can satisfy delivery industry requirement, not omit detection substantially within the scope of conveyor belt speed
Bar code;
(2) within the scope of camera fields of view, can identify and be only capable of identification be pasted onto the one-dimensional bar code label that logistics superscribes
(being usually located at express packaging bag, express delivery box surface), can initiative alarming when finding that logistics package does not paste bar code;
(3) one-dimension code label position can be accurately positioned, if one-dimension code has horizontal departure angle, can be asked by algorithm
Its angular dimension is solved, and is rotated automatically to just;
(4) it can be read, decode one-dimension code encoded information and be converted to digital number, and can check number format whether just
Really, and the nonsignificant digit read out is abandoned.
Detailed description of the invention
Fig. 1 is the logistics bar code detection method flow chart of view-based access control model provided in an embodiment of the present invention.
Fig. 2 is the logistics bar code detection method Technology Roadmap of view-based access control model provided in an embodiment of the present invention.
Fig. 3 is freight detection flow chart provided in an embodiment of the present invention.
Fig. 4 is that barcode detection provided in an embodiment of the present invention reads flow chart.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to embodiments, to the present invention
It is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not used to
Limit the present invention.
Application principle of the invention is further elaborated with reference to the accompanying drawing;
As shown in Figure 1, the logistics bar code detection method of view-based access control model provided in an embodiment of the present invention, specifically includes following
Step:
S101: picture capture: high definition high speed camera is fixed on a certain height and position right above conveyer belt, it is ensured that in the visual field
Conveyer belt comprising certain length and clearly cargo profile and label picture can be taken under the premise of cargo movement;
S102: using the YOLO algorithm of target detection in deep learning, judge all types of merchandize in captured picture,
Label position, bounding box size;Save the data type for forming freight all kinds and label;
S103: the interception picture saved in tag class data in processing step S102 through noise reduction and is rotated to level, is obtained
It is easy to the one-dimensional bar code picture read, and make cargo class data spatially to correspond to a tag class data;
S104: it reads bar code: identifying bar code using Zbar method, handled by above step, obtain one-dimension code figure
The horizontal picture of piece obtains the data of bar code using identifying in Zbar program;
S105: meeting reminding alarm, when necessary artificial error correction when error situation error correction alarm: occurs.
In step S101, high definition high speed camera provided in an embodiment of the present invention can install grating before camera fields of view position,
By the settable triggering mode of screening-mode, when there is cargo that will enter coverage, grating is triggered, order camera continuously takes pictures one
The section time;It can also set screening-mode to continuously to take pictures, no matter whether there is or not cargos on conveyer belt, all in accordance with certain time interval company
It is continuous to take pictures.
In step S102, logistics goods type provided in an embodiment of the present invention is divided into hexahedron chest and packed package,
Orient place center (coordinate) and bounding box size (width, height Picture Coordinate system under) of the cargo in photo;Simultaneously
Center and bounding box size where also identifying that label, the subgraph for intercepting outgoing label and cargo are analyzed for lower step;Institute
For the method used for the YOLO algorithm of target detection in deep learning, arithmetic speed can achieve the effect that real-time detection, symbol
Close target call;Specific step is as follows:
(1) what a reliable object detection model of the invention by pre-training, and model algorithm is deployed in program;When
After camera has shot a picture, program will be to the cargo and label progress target detection in picture;
(2) target detection model will handle shooting picture one by one, obtain the center of cargo and label in shooting picture
And its bounding box information;Will appear multiple objects due to estimated in the same visual field, detection model need to detect all objects and according to
It is secondary to be cut into subgraph, it is sent in the program of next step one-dimensional bar code detection and is calculated;As do not occurred in the visual field
Label or cargo then cut without image and continue to execute this step.
In step S103, just it is easy when one-dimension code picture provided in an embodiment of the present invention is in the horizontal direction and pixel is more visible
It being read by reading code program, the feature that one-dimensional bar code image profile is obvious, image gradient is consistent is utilized in identification bar code algorithm,
The characteristic will be judged to meet in image and pixel region that region is concentrated is for one-dimension code, and by the gradient direction in these regions work
For the rotation angle of one-dimension code;Specific implementation process are as follows:
(1) by each label sub-image gray processing, and Gaussian Blur, noise reduction process are carried out;
(2) gradient direction of the profile point and profile point in image is calculated using Cany operator;
(3) cargo subgraph will be drawn with the rectangle of m*m (numerical value of m is adjusted according to actual effect, generally takes 5-10) size
It is divided into several fritters, is recorded in the gradient direction of the profile point and profile point that in each fritter include;
(4) Screening Treatment is done, rejects and does not include profile point and the image fritter very few comprising profile point;
(5) the main gradient direction for calculating remaining image fritter, i.e., most concentrated with the gradient direction of profile point in fritter
Direction be main gradient direction;If the more dispersed confusion of gradient direction, is rejected in fritter;
(6) connection processing, the fritter Cluster merging that closely located, density is concentrated are done to the identical fritter of each gradient direction
At a region, and using the gradient direction as the gradient direction for merging rear region;Reject distance farther out, number rareness can not
The fritter of cluster;
(7) judge to merge rear region area whether more than a certain threshold value, if it is, retaining, otherwise reject;
(8) it can show that remaining region is one-dimension code position by above step, and can be obtained according to gradient direction
The level angle deviation of the one-dimensional bar code obtains identifiable one-dimensional bar code through rotation.
In step S104, the identification method provided in an embodiment of the present invention is open source program, is easy the one-dimensional item of reading horizontal
The information of shape code, output result are a certain length numeric string.
In step S105, when generation following error situation provided in an embodiment of the present invention can reminding alarm, when necessary manually
Error correction:
(1) feelings that can not be read when bar code is not pasted, is pasted onto camera fields of view dead range, bar code badly broken
Condition;When the item that discovery cargo has completely passed into the visual field, and can be read to step 4 processing without discovery by step 1
Code issues alarm;
(2) due to label printing issues, two pieces or more the barcode size or text field may be detected in step 4;If it find that
It is inconsistent that two pieces of bar codes interpret number, or interprets number and have length violation typical problem, issues alarm;
(3) since photo opporunity is shorter, in fact it could happen that identical goods appear in continuous multiple pictures, and the present invention is by basis
Conveyor belt speed estimates that the tolerable time span for repeating serial number of the same race permits after reading a new bar code
Perhaps the bar code of this same sequence number repeats to be read within certain a period of time later;If exceeded the time, conveyer belt is indicated
Cargo advance is not transmitted, issues alarm.
As shown in Fig. 2, the logistics bar code detection method Technology Roadmap of view-based access control model provided in an embodiment of the present invention.
As shown in figure 3, freight detection flow chart provided in an embodiment of the present invention.
As shown in figure 4, barcode detection provided in an embodiment of the present invention reads flow chart.
Application principle of the invention is further elaborated combined with specific embodiments below;
Interception is belonged to the bounding box of cargo by the present invention within sweep of the eye first, forms the subgraph of each cargo, then
The processing and merger that profile and gradient direction are done in these subgraphs, finally extract the pixel coverage for belonging to one-dimensional barcode.
After the one-dimension code picture rotation to level finally extracted, one-dimension code sequence is read out using decoding program.
Detailed process of the present invention is as follows:
(1) start process flow, photographing mode is set, it is ensured that each cargo and its label image can clearly be shot;
(2) position and the rectangle size of each cargo and label are identified, and intercepts out all cargos and label sub-image;
(3) the one-dimensional code position in each label sub-image is found, one-dimension code picture is extracted;
(4) digital coding of identification one-dimension code is read;
(5) if mistake occurs, alarm;Otherwise next picture is continued with since (2).
In the above-described embodiments, can come wholly or partly by software, hardware, firmware or any combination thereof real
It is existing.When using entirely or partly realizing in the form of a computer program product, the computer program product include one or
Multiple computer instructions.When loading on computers or executing the computer program instructions, entirely or partly generate according to
Process described in the embodiment of the present invention or function.The computer can be general purpose computer, special purpose computer, computer network
Network or other programmable devices.The computer instruction may be stored in a computer readable storage medium, or from one
Computer readable storage medium is transmitted to another computer readable storage medium, for example, the computer instruction can be from one
A web-site, computer, server or data center pass through wired (such as coaxial cable, optical fiber, Digital Subscriber Line (DSL)
Or wireless (such as infrared, wireless, microwave etc.) mode is carried out to another web-site, computer, server or data center
Transmission).The computer-readable storage medium can be any usable medium or include one that computer can access
The data storage devices such as a or multiple usable mediums integrated server, data center.The usable medium can be magnetic Jie
Matter, (for example, floppy disk, hard disk, tape), optical medium (for example, DVD) or semiconductor medium (such as solid state hard disk Solid
State Disk (SSD)) etc..
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Made any modifications, equivalent replacements, and improvements etc., should all be included in the protection scope of the present invention within mind and principle.
Claims (9)
1. a kind of logistics bar code detection method of view-based access control model, which is characterized in that the logistics bar code of the view-based access control model
Detection method, specifically includes the following steps:
Step 1: high definition high speed camera is fixed on a certain height and position right above conveyer belt, it is ensured that include a fixed length in the visual field
The conveyer belt of degree and clearly cargo profile and label picture can be taken under the premise of cargo movement;
Step 2: using the YOLO algorithm of target detection in deep learning, judge all types of merchandize, mark in captured picture
Sign position, bounding box size;Save the data type for forming freight all kinds and label;
Step 3: the interception picture saved in tag class data in processing step two through noise reduction and rotates to level, obtains and be easy to
The one-dimensional bar code picture of reading, and make cargo class data spatially and can correspond to a tag class data;
Step 4: identifying bar code using Zbar method, handle by above step, obtain the horizontal picture of one-dimension code picture,
Using identifying in Zbar program, the data of bar code are obtained;
Step 5: meeting reminding alarm, when necessary artificial error correction when error situation occurs.
2. the logistics bar code detection method of view-based access control model as described in claim 1, which is characterized in that in the step 1,
High definition high speed camera can install grating before camera fields of view position, by the settable triggering mode of screening-mode, when there is the cargo will
Into coverage, grating is triggered, order camera is continuously taken pictures a period of time;Screening-mode can also be set to continuously to take pictures,
No matter whether there is or not cargos on conveyer belt, continuously take pictures all in accordance with certain time interval.
3. the logistics bar code detection method of view-based access control model as described in claim 1, which is characterized in that in the step 2,
Logistics goods type is divided into hexahedron chest and packed package, orients place center and boundary of the cargo in photo
The place center and bounding box size of frame size, label in photo;Since label exists only in cargo surfaces, therefore can cut
The subgraph of label is taken out, and is saved into tag class data and is analyzed for lower step;YOLO algorithm of target detection in deep learning;
Specific step is as follows:
(1) what a reliable object detection model by pre-training, and model algorithm is deployed in program;When camera has been shot
After one picture, program will carry out target detection to the cargo in picture;
(2) target detection model will handle shooting picture one by one, obtain shooting picture in the number of cargo and label, position and its
Bounding box information;It will appear multiple objects due to estimated in the same visual field, detection model need to detect all objects and successively cut
It is cut into subgraph, saves into corresponding class data and be sent into the program of next step one-dimensional bar code detection and calculated;The visual field
In as there is not cargo or label, then cut without image and continue to execute this step.
4. the logistics bar code detection method of view-based access control model as described in claim 1, which is characterized in that in the step 3,
It is just easy to be read by reading code program when one-dimension code picture is in the horizontal direction and pixel is more visible, identification bar code algorithm is utilized one
The dimension feature that bar code image profile is obvious, image gradient is consistent will judge the picture for meeting the characteristic in image and region is concentrated
Plain region is one-dimension code, and using the gradient direction in region as the rotation angle of one-dimension code;Specific implementation process are as follows:
(1) by the subgraph gray processing of tag class data, and Gaussian Blur, noise reduction process are carried out;
(2) gradient direction of the profile point and profile point in image is calculated using Cany operator;
(3) cargo subgraph will be divided into several fritters with the rectangle of m*m size, is recorded in the profile in each fritter included
The gradient direction of point and profile point;
(4) Screening Treatment is done, rejects and does not include profile point and the image fritter very few comprising profile point;
(5) the main gradient direction of remaining image fritter, i.e., the side most concentrated with the gradient direction of profile point in fritter are calculated
Xiang Weizhu gradient direction;If the more dispersed confusion of gradient direction, is rejected in fritter;
(6) connection processing is done to the identical fritter of each gradient direction, the fritter Cluster merging that closely located, density is concentrated is at one
A region, and using the gradient direction as the gradient direction for merging rear region;Reject distance farther out, number rareness can not cluster
Fritter;
(7) judge to merge rear region area whether more than a certain threshold value, if it is, retaining, otherwise reject;
(8) can show that remaining region is one-dimension code position by above step, and can be obtained according to gradient direction this one
The level angle deviation for tieing up bar code obtains identifiable one-dimensional bar code through rotation.
5. the logistics bar code detection method of view-based access control model as described in claim 1, which is characterized in that in the step 4,
The identification method is open source program, is easy the information of reading horizontal one-dimensional bar code, and output result is a certain length numeric string.
6. the logistics bar code detection method of view-based access control model as described in claim 1, which is characterized in that in the step 5,
Meeting reminding alarm when following error situation occurs, artificial error correction when necessary:
(1) the case where can not being read when bar code is not pasted, is pasted onto camera fields of view dead range, bar code badly broken;
When discovery cargo has completely passed into the visual field, and by step 1 to step 4 processing without the bar code that can be read of discovery,
Issue alarm;
(2) due to label printing issues, two pieces or more the barcode size or text field may be detected in step 4;If it find that two pieces
It is inconsistent that bar code interprets number, or interprets number and have a length violation typical problem, issues alarm;
(3) since photo opporunity is shorter, in fact it could happen that identical goods appear in continuous multiple pictures, and the present invention will be according to transmission
Tape speed estimates that the tolerable time span for repeating serial number of the same race allows this after reading a new bar code
The bar code of a same sequence number repeats to be read within certain a period of time later;If exceeded the time, indicate that conveyer belt does not pass
Delivery object advances, and issues alarm.
7. a kind of computer journey for the logistics bar code detection method for realizing view-based access control model described in claim 1~6 any one
Sequence.
8. a kind of information data for the logistics bar code detection method for realizing view-based access control model described in claim 1~6 any one
Processing terminal.
9. a kind of computer readable storage medium, including instruction, when run on a computer, so that computer is executed as weighed
Benefit requires the logistics bar code detection method of view-based access control model described in 1-6 any one.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811443403.8A CN109583535B (en) | 2018-11-29 | 2018-11-29 | Vision-based logistics barcode detection method and readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811443403.8A CN109583535B (en) | 2018-11-29 | 2018-11-29 | Vision-based logistics barcode detection method and readable storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109583535A true CN109583535A (en) | 2019-04-05 |
CN109583535B CN109583535B (en) | 2023-04-18 |
Family
ID=65925299
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811443403.8A Active CN109583535B (en) | 2018-11-29 | 2018-11-29 | Vision-based logistics barcode detection method and readable storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109583535B (en) |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110322199A (en) * | 2019-06-28 | 2019-10-11 | 成都诚至诚商务物流有限责任公司 | The safe outbound device systems and method of image recognition |
CN110659706A (en) * | 2019-10-15 | 2020-01-07 | 广东燕塘乳业股份有限公司 | Two-dimensional code identification method based on stacking |
CN111553914A (en) * | 2020-05-08 | 2020-08-18 | 深圳前海微众银行股份有限公司 | Vision-based goods detection method and device, terminal and readable storage medium |
CN111770263A (en) * | 2020-05-22 | 2020-10-13 | 盟立自动化(昆山)有限公司 | Visual method for accurately capturing image of flowing product at specific visual angle |
CN111914579A (en) * | 2020-07-01 | 2020-11-10 | 上海视界纵横智能科技有限公司 | Industrial scanning device and method for self-learning beat recognition |
CN112507747A (en) * | 2020-12-10 | 2021-03-16 | 北京爱创科技股份有限公司 | Tracing code scanning device and method |
CN112800796A (en) * | 2019-11-14 | 2021-05-14 | 杭州海康机器人技术有限公司 | Code reading method and device and logistics system |
CN112818720A (en) * | 2021-01-26 | 2021-05-18 | 深圳市微目腾科技术有限公司 | Two-dimensional code detection method, detection device and storage medium |
CN113114877A (en) * | 2021-02-23 | 2021-07-13 | 广州弥特智能科技有限公司 | Multi-bottle rotary acquisition and identification method and equipment |
CN113449532A (en) * | 2020-03-25 | 2021-09-28 | 杭州海康机器人技术有限公司 | Method, device, computing equipment, logistics system and storage medium for detecting packages |
CN114330407A (en) * | 2021-12-30 | 2022-04-12 | 深圳创维-Rgb电子有限公司 | Method, device, equipment and storage medium for detecting and identifying bar code |
CN115081467A (en) * | 2022-07-22 | 2022-09-20 | 深圳市成为信息股份有限公司 | Method for collecting original image by handset, handset and storage medium |
WO2022227879A1 (en) * | 2021-04-30 | 2022-11-03 | 南方科技大学 | Logistics management method and system based on qr code, and server and storage medium |
CN116882432A (en) * | 2023-07-11 | 2023-10-13 | 深圳市裕源欣电子科技有限公司 | Method and system for scanning multiple materials, readable storage medium and computer equipment |
CN117764094A (en) * | 2024-02-21 | 2024-03-26 | 博诚经纬软件科技有限公司 | Intelligent warehouse management system and method for customs |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103810454A (en) * | 2012-11-14 | 2014-05-21 | 苏州工业园区高泰电子有限公司 | Automatic scanner |
CN104281931A (en) * | 2014-10-22 | 2015-01-14 | 天津慧博科技发展有限公司 | Goods checking and classifying system |
CN204399926U (en) * | 2015-01-01 | 2015-06-17 | 广州市嘉诚国际物流股份有限公司 | Intelligent cargo cabinet enters goods system |
CN107633192A (en) * | 2017-08-22 | 2018-01-26 | 电子科技大学 | Bar code segmentation and reading method under a kind of complex background based on machine vision |
CN108647553A (en) * | 2018-05-10 | 2018-10-12 | 上海扩博智能技术有限公司 | Rapid expansion method, system, equipment and the storage medium of model training image |
-
2018
- 2018-11-29 CN CN201811443403.8A patent/CN109583535B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103810454A (en) * | 2012-11-14 | 2014-05-21 | 苏州工业园区高泰电子有限公司 | Automatic scanner |
CN104281931A (en) * | 2014-10-22 | 2015-01-14 | 天津慧博科技发展有限公司 | Goods checking and classifying system |
CN204399926U (en) * | 2015-01-01 | 2015-06-17 | 广州市嘉诚国际物流股份有限公司 | Intelligent cargo cabinet enters goods system |
CN107633192A (en) * | 2017-08-22 | 2018-01-26 | 电子科技大学 | Bar code segmentation and reading method under a kind of complex background based on machine vision |
CN108647553A (en) * | 2018-05-10 | 2018-10-12 | 上海扩博智能技术有限公司 | Rapid expansion method, system, equipment and the storage medium of model training image |
Cited By (24)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110322199B (en) * | 2019-06-28 | 2024-05-14 | 四川省烟草公司成都市公司 | Safe ex-warehouse equipment system and method for image recognition |
CN110322199A (en) * | 2019-06-28 | 2019-10-11 | 成都诚至诚商务物流有限责任公司 | The safe outbound device systems and method of image recognition |
CN110659706A (en) * | 2019-10-15 | 2020-01-07 | 广东燕塘乳业股份有限公司 | Two-dimensional code identification method based on stacking |
CN112800796B (en) * | 2019-11-14 | 2023-05-26 | 杭州海康机器人股份有限公司 | Code reading method, code reading device and logistics system |
CN112800796A (en) * | 2019-11-14 | 2021-05-14 | 杭州海康机器人技术有限公司 | Code reading method and device and logistics system |
CN113449532A (en) * | 2020-03-25 | 2021-09-28 | 杭州海康机器人技术有限公司 | Method, device, computing equipment, logistics system and storage medium for detecting packages |
CN113449532B (en) * | 2020-03-25 | 2022-04-19 | 杭州海康机器人技术有限公司 | Method, device, computing equipment, logistics system and storage medium for detecting packages |
CN111553914A (en) * | 2020-05-08 | 2020-08-18 | 深圳前海微众银行股份有限公司 | Vision-based goods detection method and device, terminal and readable storage medium |
CN111553914B (en) * | 2020-05-08 | 2021-11-12 | 深圳前海微众银行股份有限公司 | Vision-based goods detection method and device, terminal and readable storage medium |
CN111770263A (en) * | 2020-05-22 | 2020-10-13 | 盟立自动化(昆山)有限公司 | Visual method for accurately capturing image of flowing product at specific visual angle |
CN111914579A (en) * | 2020-07-01 | 2020-11-10 | 上海视界纵横智能科技有限公司 | Industrial scanning device and method for self-learning beat recognition |
CN111914579B (en) * | 2020-07-01 | 2021-03-12 | 上海视界纵横智能科技有限公司 | Industrial scanning device and method for self-learning beat recognition |
CN112507747A (en) * | 2020-12-10 | 2021-03-16 | 北京爱创科技股份有限公司 | Tracing code scanning device and method |
CN112507747B (en) * | 2020-12-10 | 2024-04-19 | 北京爱创科技股份有限公司 | Device and method for scanning traceability codes |
CN112818720A (en) * | 2021-01-26 | 2021-05-18 | 深圳市微目腾科技术有限公司 | Two-dimensional code detection method, detection device and storage medium |
CN113114877A (en) * | 2021-02-23 | 2021-07-13 | 广州弥特智能科技有限公司 | Multi-bottle rotary acquisition and identification method and equipment |
WO2022227879A1 (en) * | 2021-04-30 | 2022-11-03 | 南方科技大学 | Logistics management method and system based on qr code, and server and storage medium |
US11631261B2 (en) | 2021-04-30 | 2023-04-18 | Southern University Of Science And Technology | Method, system, server, and storage medium for logistics management based on QR code |
CN114330407A (en) * | 2021-12-30 | 2022-04-12 | 深圳创维-Rgb电子有限公司 | Method, device, equipment and storage medium for detecting and identifying bar code |
CN115081467A (en) * | 2022-07-22 | 2022-09-20 | 深圳市成为信息股份有限公司 | Method for collecting original image by handset, handset and storage medium |
CN116882432B (en) * | 2023-07-11 | 2024-03-22 | 深圳市裕源欣电子科技有限公司 | Method and system for scanning multiple materials, readable storage medium and computer equipment |
CN116882432A (en) * | 2023-07-11 | 2023-10-13 | 深圳市裕源欣电子科技有限公司 | Method and system for scanning multiple materials, readable storage medium and computer equipment |
CN117764094A (en) * | 2024-02-21 | 2024-03-26 | 博诚经纬软件科技有限公司 | Intelligent warehouse management system and method for customs |
CN117764094B (en) * | 2024-02-21 | 2024-05-10 | 博诚经纬软件科技有限公司 | Intelligent warehouse management system and method for customs |
Also Published As
Publication number | Publication date |
---|---|
CN109583535B (en) | 2023-04-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109583535A (en) | A kind of logistics bar code detection method, the computer program of view-based access control model | |
Laroca et al. | An efficient and layout‐independent automatic license plate recognition system based on the YOLO detector | |
US9830704B1 (en) | Predicting performance metrics for algorithms | |
CN107617573B (en) | Logistics code identification and sorting method based on multitask deep learning | |
US11443133B2 (en) | Computer vision system for industrial equipment gauge digitization and alarms | |
EP3899508A1 (en) | Automated inspection system and associated method for assessing the condition of shipping containers | |
US20140153779A1 (en) | Object Segmentation at a Self-Checkout | |
CN109815863B (en) | Smoke and fire detection method and system based on deep learning and image recognition | |
CN109635797B (en) | Steel coil sequence accurate positioning method based on multi-carrier identification technology | |
CN106355367A (en) | Warehouse monitoring management device | |
CN111223260A (en) | Method and system for intelligently monitoring goods theft prevention in warehousing management | |
CN110348293B (en) | Commodity identification method and system | |
CN111597857B (en) | Logistics package detection method, device, equipment and readable storage medium | |
CN113962274A (en) | Abnormity identification method and device, electronic equipment and storage medium | |
CN115880260A (en) | Method, device and equipment for detecting base station construction and computer readable storage medium | |
Mi et al. | Research on a Fast Human‐Detection Algorithm for Unmanned Surveillance Area in Bulk Ports | |
CN111680680B (en) | Target code positioning method and device, electronic equipment and storage medium | |
CN107784625B (en) | Electronic device, virtual sample generation method and storage medium | |
CN106033526A (en) | Method of positioning bar code area based on gradient direction characteristic matching algorithm | |
Noceti et al. | A multi-camera system for damage and tampering detection in a postal security framework | |
CN114491648A (en) | Block chain data privacy protection method for video live broadcast social big data | |
CN110942008A (en) | Method and system for positioning waybill information based on deep learning | |
Wang | Recognition and Positioning of Container Lock Holes for Intelligent Handling Terminal Based on Convolutional Neural Network. | |
CN114596102B (en) | Block chain-based anti-counterfeiting traceability federated learning training method and device | |
US20230057340A1 (en) | Systems, methods, and devices for automated meter reading for smart field patrol |
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 | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20220704 Address after: 410073 Hunan province Changsha Kaifu District, Deya Road No. 109 Applicant after: National University of Defense Technology Address before: Room 1505, building 2, Xincheng Science Park, 588 Yuelu West Avenue, Changsha hi tech Development Zone, Hunan 410000 Applicant before: HUNAN SPEEDBOT ROBOT Co.,Ltd. |
|
TA01 | Transfer of patent application right | ||
GR01 | Patent grant | ||
GR01 | Patent grant |