CN110490192A - A kind of commodity production date tag detection method and system - Google Patents

A kind of commodity production date tag detection method and system Download PDF

Info

Publication number
CN110490192A
CN110490192A CN201910642410.9A CN201910642410A CN110490192A CN 110490192 A CN110490192 A CN 110490192A CN 201910642410 A CN201910642410 A CN 201910642410A CN 110490192 A CN110490192 A CN 110490192A
Authority
CN
China
Prior art keywords
image
connected domain
character
industrial camera
commodity
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
Application number
CN201910642410.9A
Other languages
Chinese (zh)
Inventor
苏成悦
刘信宏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong University of Technology
Original Assignee
Guangdong University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guangdong University of Technology filed Critical Guangdong University of Technology
Priority to CN201910642410.9A priority Critical patent/CN110490192A/en
Publication of CN110490192A publication Critical patent/CN110490192A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
    • G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
    • G06V20/63—Scene text, e.g. street names
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/10—Character recognition
    • G06V30/14—Image acquisition
    • G06V30/148—Segmentation of character regions
    • G06V30/153—Segmentation of character regions using recognition of characters or words

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Image Analysis (AREA)

Abstract

The present invention relates to a kind of commodity production date tag detection method and systems, and the system comprises processor, bracket, industrial camera, annular light source, conveyer belt and racks；Bracket and conveyer belt are set on rack, industrial camera is set up in above conveyer belt by branch, when the communicated band of commodity to be detected is by below industrial camera lens, processor control industrial camera capture image simultaneously starts software systems recognizer, Threshold segmentation is carried out to target image using HSV space and gray space first, filters out suspicious character connected domain according to gray feature and morphological feature；The date of manufacture label Fast Segmentation detection scheme for proposing a kind of band rotation, detects the date of manufacture label of commodity upper surface.The present invention can effectively exclude, form comparable interference characteristic suitable with target character picture size, can quick Ground Split and detection character targets.

Description

A kind of commodity production date tag detection method and system
Technical field
The present invention relates to non-contact intelligent observation and control technology fields, examine more particularly, to a kind of commodity production date tag Survey method and system.
Background technique
Date of manufacture label defect detection technology belongs to computer vision technique, is one and shoots object by camera, obtains Image information is taken, then carries out image analysis, to obtain the engineering discipline of the information such as pose and the form of object.Compared to tradition Manually check detection, computer vision is not by artifact, with the advantages such as quick, accurate, non-contact and low in cost, pole The earth reduces cost of labor, improves production efficiency.
In computer vision detect target often need to be split image processing, its object is to by object to be measured with carry on the back Other signals such as scape, noise are separated.It can be used in cutting procedure between fixed threshold split plot design, edge split plot design or maximum kind The automatic threshold segmentations method such as variance.Regardless of using which kind of dividing method, all it cannot be guaranteed that completely target and interference signal Separated, this requires detection scheme must be directed to have the case where interference signal.
In addition, direction of rotation has randomness when date of manufacture label is imprinted on product surface, meanwhile, there is ruler in image The presence of very little and gray scale and the close interference characteristic of character targets, these to the segmentation and detection of date of manufacture label propose compared with High requirement.
The present invention provides the commodity production date tag detection system and its method of a kind of band rotation, and effective exclusive PCR is special Sign is realized and is detected to the Fast Segmentation of the commodity production date tag with rotation.
Summary of the invention
The present invention is the defect for overcoming the detection of label described in the above-mentioned prior art to can not rule out interference characteristic, provides one kind Commodity production date tag detection method and system.
The described method includes:
S1: obtaining commodity surface RGB image, and pre-process to RGB image, obtains bianry image；Wherein, RGB is The color in three channels of red, green, blue is represented, it is mesh that this standard, which almost includes all colours that human eyesight can perceive, It is preceding to use most wide one of color system.
S2: connected domain screening is carried out to the bianry image that S1 is obtained；The connected domain collection of suspicious object after being screened It closes.
S3: the date of manufacture label Fast Segmentation with rotation is carried out to the connected domain set of suspicious object and is handled, is given birth to Produce the image T of date character combinationout。
S4: character comparison database is established；The image of the character on date of manufacture label is obtained by experiment, forms image collection CH={ CHi, i=1,2,3 ..., n；The character include 0~9, a~z, A~Z, " ", "/", " " and "-".
S5: by image ToutIn each connected domain be cut into, successively carry out knn Forecasting recognition；And obtain prediction result.
Preferably, S1 the following steps are included:
S1.1: commodity surface RGB image O and its gray level image G are obtained；By image O convert to HSV space (Hue, Saturation, Value), obtain image H；Threshold segmentation is fixed to image G, since target object and detection environment are solid Fixed, segmentation threshold can be obtained by many experiments, obtain bianry image B, and object pixel is set as 255, and background pixel is set as 0；
Wherein HSV (Hue, Saturation, Value) was by A.R.Smith according to the intuitive nature of color in 1978 years A kind of color space of creation, also referred to as hexagonal pyramid model (Hexcone Model), the parameter difference of color in this model It is: tone (H), saturation degree (S), lightness (V).
S1.2: any pixel of image H is set as Hi, HSV component is { h, s, v }；If in image B with HiLocation of pixels phase Same pixel is Bi；If HiAnd BiMeet condition C on, then by Bi0 is set, wherein condition C on are as follows:
Wherein Rh、Rs、RvIndicate the range of object pixel H, S, V component numerical value；Since target object and detection environment are solid It is fixed, Rh、Rs、RvInterval range is obtained by many experiments；Image H and image B is traversed, new acquisition bianry image B is obtained.
Preferably, S2 the following steps are included:
S2.1: whole connected domains in bianry image B after the completion of detection pretreatment obtain connected domain set Re= {Rei, i=1,2,3 ..., n；
S2.2: the connected domain subset for setting suspicious object is combined into Rc, detects connected domain ReiMinimum circumscribed rectangle, long side Length is a, bond length b, if a ∈ RaAnd b ∈ Rb, then by ReiIt is stored in Rc, wherein RaAnd RbRespectively indicate connected domain minimum The interval range of the long side length of boundary rectangle and the interval range of bond length, RaAnd RbIt can be obtained by many experiments；
S2.3: traversal connected domain set Re, the connected domain subclass Rc of the suspicious object after being screened, after screening Connected domain contain date of manufacture character and connected domain with similar in date of manufacture character form is interfered in part；
Preferably, S3 the following steps are included:
S3.1: the centre coordinate of the minimum circumscribed rectangle of each connected domain in connected domain subclass Rc is calculated, is obtained Heart coordinate set C={ Ci, i=1,2,3 ..., n, and with the position of corresponding element in coordinate set characterization Rc；
S3.2: definition set N={ Ni, i=1,2,3 ..., n；Enable i=0, j=0；
S3.3: with Elements CiCentered on, selection is one wide, and (length in pixels of date of manufacture character string, passes through with Gao Douwei w Many experiments obtain) rectangular area Rect_w, and count the region Rect_w include set C in element number n, if n > t (wherein t is preset threshold, can be obtained by many experiments), then enable Nj=Ci, j=j+1, otherwise starting in next step；
S3.4: i=i+1 is executed；And judge CiWhether ∈ C is true, if so, step S3.3 is then returned to, under otherwise starting One step；
S3.5: the scheme pond that j element is chosen from set N is established according to permutation and combination:
P={ Pk, k=1,2,3 ..., n
That is the whole circumstances that j element is chosen from set N, each element P in set P are contained in set PkIt represents One kind choosing the case where j element, and each element inequality of set P from set N；Enable k=0；
S3.6: scheme P is usedkJ element is chosen from set N, this j element is constituted into an entirety T, and calculate T Minimum circumscribed rectangle Rect_T short side and long side length ratio f；
S3.7: setting ratio interval threshold range as ratio, i.e. the ratio between short side of connected domain minimum circumscribed rectangle and long side Range, ratio section is fixed according to character size, and since detection target is fixed, value can be obtained by many experiments, if f ∈ ratio, Then obtain the rotation angle of Rect_TThen otherwise starting step S3.10 enables k=k+1, restart in next step；
S3.8: if Pk∈ P then returns to step S3.6, otherwise enables j=j-1, restarts in next step；
S3.9: if j > nt, then step S3.5 is returned to, wherein ntIt is determined by target character number, otherwise output segmentation is unsuccessfully believed Breath, and terminate dividing processing；
S3.10: the element general image in connected domain set Rc in the region Rect_T is cut into, and according to rotation angle DegreeIt is rotated to image long side and is overlapped with horizontal direction, obtain the image T of date of manufacture character combinationout。
Preferably, S5 specifically: by image ToutIn each connected domain be cut into, successively carry out KNN (k- NearestNeighbor, K nearest neighbor algorithm) prediction, then prediction result and probability threshold value pn are compared, probability threshold value pn by Many experiments obtain, if its prediction probability is greater than pn, otherwise character value of the prediction result as the connected domain determines the connection Domain character has defect.
Wherein, KNN prediction calculates the distance between test data and library data, adjusts the distance and sorts from small to large, chooses Apart from the smallest k library data, as prediction result；
The present invention also provides a kind of detection systems using commodity production date tag detection method, and the system comprises platforms Frame, conveyer belt, bracket, industrial camera, light source, processor；
Conveyer belt is set on rack；For transmitting commodity to be detected；
Rack side is fixed in one end of bracket, and the other end is used to fix industrial camera, so that industrial camera is located at transmission Band top, industrial camera lens downward, for capture the commodity to be detected transmitted on conveyer belt date of manufacture label image, And captured image information is transferred to processor；
Light source is set to below industrial camera, is used to supplement light source for industrial camera；
Processor controls industrial camera and carries out image capture, and carries out data processing to industrial camera captured image, Including carrying out Threshold segmentation to target image using HSV space and gray space, screened according to gray feature and morphological feature Suspicious character connected domain out.
Preferably, the light source is annular light source.
Compared with prior art, the beneficial effect of technical solution of the present invention is: the automatic detection of commodity production date tag Compared to traditional artificial detection, there are the advantages such as quick, accurate, non-contact and low in cost, greatly reduce cost of labor, Improve production efficiency.
The date of manufacture label Fast Segmentation Algorithm of band rotation proposed by the present invention, efficiently solves date of manufacture label When being imprinted on product surface, direction of rotation has detection difficult brought by randomness.
The date of manufacture label Fast Segmentation Algorithm of band rotation proposed by the present invention, can effectively exclude size and gray scale and word The close interference characteristic of target is accorded with, it can quick Ground Split and detection character targets.
Detailed description of the invention
Fig. 1 is a kind of commodity production date tag detection method flow chart.
Fig. 2 is the image of the connected domain set of the suspicious object after morphology screening.
Fig. 3 is the centre coordinate schematic diagram that connected domain is corresponded in Fig. 2；
Fig. 4 is Fast Segmentation Algorithm flow chart；
Fig. 5 is postrotational target character integrated images；
Fig. 6 is a kind of commodity production date tag detection system structure.
In figure, 1- processor, 2- bracket, 3 industrial cameras, 4- camera lens, 5- light source, 6- conveyer belt, 7- commodity to be detected, 8- Rack.
Specific embodiment
The attached figures are only used for illustrative purposes and cannot be understood as limitating the patent；
In order to better illustrate this embodiment, the certain components of attached drawing have omission, zoom in or out, and do not represent actual product Size；
To those skilled in the art, it is to be understood that certain known features and its explanation, which may be omitted, in attached drawing 's.
The following further describes the technical solution of the present invention with reference to the accompanying drawings and examples.
Embodiment 1:
The present embodiment provides a kind of commodity production date tag detection methods, as shown in Figure 1, which comprises
S1: obtaining commodity surface RGB image, and pre-process to RGB image, obtains bianry image；
S2: connected domain screening is carried out to the bianry image that S1 is obtained；The connected domain collection of suspicious object after being screened It closes；
S3: the date of manufacture label Fast Segmentation with rotation is carried out to the connected domain set of suspicious object and is handled, is given birth to Produce the image T of date character combinationout；
S4: character comparison database is established；By experiment obtain date of manufacture label on 0~9, a~z, A~Z, " ", "/", " " and the spcial characters such as "-" character picture, form image collection CH={ CHi, i=1,2,3 ..., n；
S5: by image ToutIn each connected domain be cut into, successively carry out KNN Forecasting recognition；And obtain prediction result.
S1 the following steps are included:
S1.1: commodity surface RGB image O and its gray level image G are obtained；By image O convert to HSV space (Hue, Saturation, Value), obtain image H；Threshold segmentation is fixed (since target object and detection environment are solid to image G Fixed, segmentation threshold can be obtained by many experiments), bianry image B is obtained, object pixel is set as 255, and background pixel is set as 0；
S1.2: any pixel of image H is set as Hi, HSV component is { h, s, v }；If in image B with HiLocation of pixels phase Same pixel is Bi；If HiAnd BiMeet condition C on, then by Bi0 is set, wherein condition C on are as follows:
Wherein Rh、Rs、RvIndicate the range of object pixel H, S, V component numerical value；Since target object and detection environment are solid It is fixed, Rh、Rs、RvInterval range is obtained by many experiments；Image H and image B is traversed, new acquisition bianry image B is obtained.
S2 the following steps are included:
S2.1: whole connected domains in bianry image B after the completion of detection pretreatment obtain connected domain set Re= {Rei, i=1,2,3 ..., n；
S2.2: the connected domain subset for setting suspicious object is combined into Rc, detects connected domain ReiMinimum circumscribed rectangle, long side Length is a, bond length b, if a ∈ RaAnd b ∈ Rb, then by ReiIt is stored in connected domain set Rc, wherein RaAnd RbIt respectively indicates The interval range of the long side length of connected domain minimum circumscribed rectangle and the interval range of bond length, RaAnd RbIt can be by multiple Experiment obtains；
S2.3: traversal connected domain set Re, the connected domain subclass Rc of the suspicious object after being screened, such as Fig. 2 institute Show, the connected domain after screening contains date of manufacture character and part is connected to interfering similar in date of manufacture character form Domain；
As shown in figure 4, S3 the following steps are included:
S3.1: the centre coordinate of the minimum circumscribed rectangle of each connected domain in connected domain subclass Rc is calculated, is obtained Heart coordinate set C={ Ci, i=1,2,3 ..., n, display is as shown in figure 3, and with coordinate set characterization in the picture The position of corresponding element in Rc；
S3.2: definition set N={ Ni, i=1,2,3 ..., n；Enable i=0, j=0；
S3.3: with Elements CiCentered on, selection is one wide, and (length in pixels of date of manufacture character string, passes through with Gao Douwei w Many experiments obtain) rectangular area Rect_w, and count the region Rect_w include set C in element number n, if n > t (wherein t is preset threshold, can be obtained by many experiments), then enable Nj=Ci, j=j+1, otherwise starting in next step；
S3.4: i=i+1 is executed；And judge CiWhether ∈ C is true, if so, step S3.3 is then returned to, under otherwise starting One step；
S3.5: the scheme pond that j element is chosen from set N is established according to permutation and combination:
P={ Pk, k=1,2,3 ..., n
That is the whole circumstances that j element is chosen from set N, each element P in set P are contained in set PkIt represents One kind choosing the case where j element, and each element inequality of set P from set N；Enable k=0；
S3.6: scheme P is usedkJ element is chosen from set N, this j element is constituted into an entirety T, and calculate T Minimum circumscribed rectangle Rect_T short side and long side length ratio f；
S3.7: setting ratio interval threshold range as ratio, i.e. the ratio between short side of connected domain minimum circumscribed rectangle and long side Range, ratio section is fixed according to character size, and since detection target is fixed, value can be obtained by many experiments, if f ∈ ratio, Then obtain the rotation angle of Rect_TThen otherwise starting step S3.10 enables k=k+1, restart in next step；
S3.8: if Pk∈ P then returns to step S3.6, otherwise enables j=j-1, restarts in next step；
S3.9: if j > nt, then step S3.5 is returned to, wherein ntIt is determined by target character number, otherwise output segmentation is unsuccessfully believed Breath, and terminate dividing processing；
S3.10: the element general image in connected domain set Rc in the region Rect_T is cut into, and according to rotation angle DegreeIt is rotated to image long side and is overlapped with horizontal direction, as shown in figure 5, obtaining the image T of date of manufacture character combinationout。
S5 specifically: by image ToutIn each connected domain be cut into, successively carry out KNN prediction again by prediction result with Probability threshold value pn is compared, and probability threshold value pn is obtained by many experiments, if its prediction probability is greater than pn, prediction result conduct Otherwise the character value of the connected domain determines that the connected domain character has defect.
Wherein, KNN prediction calculates the distance between test data and library data, adjusts the distance and sorts from small to large, chooses Apart from the smallest k library data, the present embodiment takes k=1, as prediction result；
Embodiment 2:
The detection system of commodity production date tag detection method described in a kind of Application Example 1 of the present embodiment, such as Fig. 6 institute Show, the system comprises rack 8, conveyer belt 6, bracket 2, industrial camera 3, light source 5, processors 1；
Conveyer belt is set on rack 8；For transmitting commodity 7 to be detected；
8 side of rack is fixed in one end of bracket 2, and the other end is used to fix industrial camera 3, pass so that industrial camera is located at Send the top of band 6,3 camera lens 4 of industrial camera downward, for capturing the date of manufacture label for transmitting commodity 7 to be detected on conveyer belt Image, and captured image information is transferred to processor；
Light source 5 is set to 3 lower section of industrial camera, is used to supplement light source for industrial camera 3；
Processor 1 controls industrial camera 3 and carries out image capture, and carries out at data to 3 captured image of industrial camera Reason, including Threshold segmentation is carried out to target image using HSV space and gray space, it is sieved according to gray feature and morphological feature Select suspicious character connected domain.
The light source 5 is annular light source.
The same or similar label correspond to the same or similar components；
The terms describing the positional relationship in the drawings are only for illustration, should not be understood as the limitation to this patent；
Obviously, the above embodiment of the present invention be only to clearly illustrate example of the present invention, and not be pair The restriction of embodiments of the present invention.For those of ordinary skill in the art, may be used also on the basis of the above description To make other variations or changes in different ways.There is no necessity and possibility to exhaust all the enbodiments.It is all this Made any modifications, equivalent replacements, and improvements etc., should be included in the claims in the present invention within the spirit and principle of invention Protection scope within.

Claims (8)

1. a kind of commodity production date tag detection method, which is characterized in that the described method includes:
S1: obtaining commodity surface RGB image, and pre-process to RGB image, obtains bianry image；
S2: connected domain screening is carried out to the bianry image that S1 is obtained；The connected domain set of suspicious object after being screened；
S3: the date of manufacture label Fast Segmentation with rotation is carried out to the connected domain set of suspicious object and is handled, production date is obtained The image T of phase character combinationout；
S4: character comparison database is established；The image of the character on date of manufacture label is obtained by experiment, forms image collection CH= {CHi, i=1,2,3 ..., n；
S5: by image ToutIn each connected domain be cut into, successively carry out KNN Forecasting recognition；And obtain prediction result.
2. commodity production date tag detection method according to claim 1, which is characterized in that S1 the following steps are included:
S1.1: commodity surface RGB image O and its gray level image G are obtained；Image O is converted to HSV space, image H is obtained；It is right Threshold segmentation is fixed in image G, obtains bianry image B, and object pixel is set as 255, and background pixel is set as 0；
S1.2: any pixel of image H is set as Hi, HSV component is { h, s, v }；If in image B with HiLocation of pixels is identical Pixel is Bi；If HiAnd BiMeet condition C on, then by Bi0 is set, wherein condition C on are as follows:
Wherein Rh、Rs、RvIndicate the range of object pixel H, S, V component numerical value；Since target object and detection environment are fixed, Rh、 Rs、RvInterval range is obtained by many experiments；Image H and image B is traversed, new acquisition bianry image B is obtained.
3. commodity production date tag detection method according to claim 2, which is characterized in that S2 the following steps are included:
S2.1: whole connected domains in bianry image B after the completion of detection pretreatment obtain connected domain set Re={ Rei, i= 1,2,3,......,n；
S2.2: the connected domain subset for setting suspicious object is combined into Rc, detects connected domain ReiMinimum circumscribed rectangle, long side length is A, bond length b, if a ∈ RaAnd b ∈ Rb, then by ReiIt is stored in Rc, wherein RaAnd RbRespectively indicate the minimum external square of connected domain The interval range of the long side length of shape and the interval range of bond length；
S2.3: connected domain set Re, the connected domain subclass Rc of the suspicious object after being screened, the company after screening are traversed Logical domain contains date of manufacture character and part and interferes connected domain with similar in date of manufacture character form.
4. commodity production date tag detection method according to claim 3, which is characterized in that S3 the following steps are included:
S3.1: calculating the centre coordinate of the minimum circumscribed rectangle of each connected domain in connected domain subclass Rc, obtains center seat Mark set C={ Ci, i=1,2,3 ..., n, and with the position of corresponding element in coordinate set characterization Rc；
S3.2: definition set N={ Ni, i=1,2,3 ..., n；Enable i=0, j=0；
S3.3: with Elements CiCentered on, one wide and Gao Douwei w rectangular area Rect_w is chosen, and count the region Rect_w packet The number n of element in C containing set, if n > t, wherein t is preset threshold, then enables Nj=Ci, j=j+1, otherwise starting in next step；
S3.4: i=i+1 is executed；And judge CiWhether ∈ C is true, if so, step S3.3 is then returned to, otherwise starting next step；
S3.5: the scheme pond that j element is chosen from set N is established according to permutation and combination:
P={ Pk, k=1,2,3 ..., n
That is the whole circumstances that j element is chosen from set N, each element P in set P are contained in set PkRepresent one kind The case where j element is chosen from set N, and each element inequality of set P；Enable k=0；
S3.6: scheme P is usedkJ element is chosen from set N, this j element is constituted into an entirety T, and calculate the minimum of T The short side of boundary rectangle Rect_T and the length ratio f of long side；
S3.7: setting ratio interval threshold range as ratio, i.e. the model of the ratio between short side of connected domain minimum circumscribed rectangle and long side It encloses, ratio section is fixed according to character size, and since detection target is fixed, value can be obtained by many experiments, if f ∈ ratio, Obtain the rotation angle of Rect_TThen otherwise starting step S3.10 enables k=k+1, restart in next step；
S3.8: if Pk∈ P then returns to step S3.6, otherwise enables j=j-1, restarts in next step；
S3.9: if j > nt, then step S3.5 is returned to, wherein ntIt is determined by target character number, otherwise output segmentation failure information, And terminate dividing processing；
S3.10: the element general image in connected domain set Rc in the region Rect_T is cut into, and according to rotation angle It is rotated to image long side and is overlapped with horizontal direction, obtain the image T of date of manufacture character combinationout。
5. commodity production date tag detection method according to claim 1, which is characterized in that character described in S4 includes 0~9, a~z, A~Z, " ", "/", " " and "-".
6. commodity production date tag detection method according to claim 1-5, which is characterized in that S5 is specific Are as follows: by image ToutIn each connected domain be cut into, successively carry out KNN prediction and again carry out prediction result and probability threshold value pn Comparison, if its prediction probability is greater than pn, otherwise character value of the prediction result as the connected domain determines that the connected domain character has Defect.
7. a kind of detection system using commodity production date tag detection method described in any one of claims 1-6, special Sign is that the system comprises rack, conveyer belt, bracket, industrial camera, light source, processors；
Conveyer belt is set on rack；For transmitting commodity to be detected；
Rack side is fixed in one end of bracket, and the other end is used to fix industrial camera, so that industrial camera is located on conveyer belt Side, industrial camera lens downward, for capture the commodity to be detected transmitted on conveyer belt date of manufacture label image, and will Captured image information is transferred to processor；
Light source is set to below industrial camera, is used to supplement light source for industrial camera；
Processor controls industrial camera and carries out image capture, and carries out data processing to industrial camera captured image, including Threshold segmentation is carried out to target image using HSV space and gray space, filtering out according to gray feature and morphological feature can Doubtful character connected domain.
8. commodity production date tag detection system according to claim 7, which is characterized in that the light source is annular Light source.
CN201910642410.9A 2019-07-16 2019-07-16 A kind of commodity production date tag detection method and system Pending CN110490192A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910642410.9A CN110490192A (en) 2019-07-16 2019-07-16 A kind of commodity production date tag detection method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910642410.9A CN110490192A (en) 2019-07-16 2019-07-16 A kind of commodity production date tag detection method and system

Publications (1)

Publication Number Publication Date
CN110490192A true CN110490192A (en) 2019-11-22

Family

ID=68547386

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910642410.9A Pending CN110490192A (en) 2019-07-16 2019-07-16 A kind of commodity production date tag detection method and system

Country Status (1)

Country Link
CN (1) CN110490192A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112183305A (en) * 2020-09-24 2021-01-05 江苏池丰科技有限公司 System for identifying production date and batch number of product

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08190610A (en) * 1995-01-12 1996-07-23 Oki Electric Ind Co Ltd Optical character reader
US20070201757A1 (en) * 2006-02-24 2007-08-30 Dariusz Madej System and method for optical character recognition in an image
CN103049743A (en) * 2012-12-27 2013-04-17 天津普达软件技术有限公司 Detection system and detection method for characters at bottom of bowl
CN105431866A (en) * 2013-07-16 2016-03-23 株式会社汤山制作所 Optical character recognition device
US20160371557A1 (en) * 2015-06-18 2016-12-22 Pfu Limited Image-processing apparatus, date identification apparatus, image-processing method, and computer program product
CN108364010A (en) * 2018-03-08 2018-08-03 广东工业大学 A kind of licence plate recognition method, device, equipment and computer readable storage medium

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08190610A (en) * 1995-01-12 1996-07-23 Oki Electric Ind Co Ltd Optical character reader
US20070201757A1 (en) * 2006-02-24 2007-08-30 Dariusz Madej System and method for optical character recognition in an image
CN103049743A (en) * 2012-12-27 2013-04-17 天津普达软件技术有限公司 Detection system and detection method for characters at bottom of bowl
CN105431866A (en) * 2013-07-16 2016-03-23 株式会社汤山制作所 Optical character recognition device
US20160371557A1 (en) * 2015-06-18 2016-12-22 Pfu Limited Image-processing apparatus, date identification apparatus, image-processing method, and computer program product
CN108364010A (en) * 2018-03-08 2018-08-03 广东工业大学 A kind of licence plate recognition method, device, equipment and computer readable storage medium

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
施成燕: "基于智能机器人的仪表示数识别技术与系统研制", 《中国优秀硕士学位论文全文数据库信息科技辑》 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112183305A (en) * 2020-09-24 2021-01-05 江苏池丰科技有限公司 System for identifying production date and batch number of product

Similar Documents

Publication Publication Date Title
CN110674746B (en) Method and device for realizing high-precision cross-mirror tracking by using video spatial relationship assistance, computer equipment and storage medium
CN111797653B (en) Image labeling method and device based on high-dimensional image
WO2022100470A1 (en) Systems and methods for target detection
CN109308447A (en) The method of equipment operating parameter and operating status is automatically extracted in remote monitoriong of electric power
EP3499414B1 (en) Lightweight 3d vision camera with intelligent segmentation engine for machine vision and auto identification
CN104598883B (en) Target knows method for distinguishing again in a kind of multiple-camera monitoring network
CN105493141B (en) Unstructured road border detection
CN109902592B (en) Blind person auxiliary walking method based on deep learning
CN108288289B (en) LED visual detection method and system for visible light positioning
CN107635099B (en) Human body induction double-optical network camera and security monitoring system
CN111814771A (en) Image processing method and device
CN108428220A (en) Satellite sequence remote sensing image sea island reef region automatic geometric correction method
CN111709994B (en) Autonomous unmanned aerial vehicle visual detection and guidance system and method
CN112818905B (en) Finite pixel vehicle target detection method based on attention and spatio-temporal information
Touati et al. A reliable mixed-norm-based multiresolution change detector in heterogeneous remote sensing images
Xiong et al. Early smoke detection of forest fires based on SVM image segmentation
CN110245592A (en) A method of for promoting pedestrian's weight discrimination of monitoring scene
CN111767826B (en) Timing and fixed-point scene anomaly detection method
CN109886336A (en) A kind of object detection method and system based on warship basic image
CN108022245B (en) Facial line primitive association model-based photovoltaic panel template automatic generation method
CN116994135A (en) Ship target detection method based on vision and radar fusion
CN107045630B (en) RGBD-based pedestrian detection and identity recognition method and system
CN112101260A (en) Method, device, equipment and storage medium for identifying safety belt of operator
CN117197789A (en) Curtain wall frame identification method and system based on multi-scale boundary feature fusion
CN110490192A (en) A kind of commodity production date tag detection method and system

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: 20191122