CN104820827A - Method for recognizing punctiform characters on surfaces of cables - Google Patents
Method for recognizing punctiform characters on surfaces of cables Download PDFInfo
- Publication number
- CN104820827A CN104820827A CN201510206069.4A CN201510206069A CN104820827A CN 104820827 A CN104820827 A CN 104820827A CN 201510206069 A CN201510206069 A CN 201510206069A CN 104820827 A CN104820827 A CN 104820827A
- Authority
- CN
- China
- Prior art keywords
- character
- image
- cable
- point
- characters
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01B—CABLES; CONDUCTORS; INSULATORS; SELECTION OF MATERIALS FOR THEIR CONDUCTIVE, INSULATING OR DIELECTRIC PROPERTIES
- H01B7/00—Insulated conductors or cables characterised by their form
- H01B7/36—Insulated conductors or cables characterised by their form with distinguishing or length marks
- H01B7/365—Insulated conductors or cables characterised by their form with distinguishing or length marks being indicia imposed on the insulation or conductor
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Character Input (AREA)
Abstract
The invention discloses a method for recognizing punctiform characters on the surfaces of cables and belongs to the image processing field, in particular, belongs to the automated optical character recognition field. The method mainly comprises the steps of Otsu method threshold extraction and binarization, image preprocessing, cable regional coarse extraction, cable tilt correction, character segmentation, character tilt correction, template matching and identification logic design. The processing method of the punctiform characters on the surfaces of cables is put forward for the first time. With the method adopted, punctiform digits, English characters and few symbols on the surfaces of various kinds of cables can be accurately recognized. In a cable production process, character tilt appears in character jet printing inevitably, while, with the method of the invention adopted, accurate reorganization can be still performed when characters tilt slightly, and therefore, the method has a certain anti-interference ability.
Description
Technical field
The invention belongs to image processing field, particularly automated optical character recognition field.
Background technology
Electric wire face character is for marking the product important informations such as cable product batch number, production code member, date of manufacture, production length, therefore the apparent character machining of high-speed cable is for improving product image, improve the cable quality of production, enhance productivity, reduce production cost and be significant.Electric wire face character major part is laser printing or inkjet printing, and character style is mainly divided into continuation character and point-like character.Point-like character, due to the noncontinuity of its font, needs through special treatment step, then identifies.Current most domestic cable manufacturer adopts the mode of hand inspection, with high costs, inefficiency, loss are high, work multiplicity is high.
Summary of the invention
The deficiency technical matters to be solved that the present invention is directed to background technology designs a kind of efficiency high, simple to operate, electric wire superficial punctate character identifying method with low cost.
The point-like character identifying method on a kind of electric wire surface of the present invention, the method comprises the following steps:
Step 1: extract the cable gray level image comprising point-like character;
Step 2: image binaryzation; Adopt Otsu algorithm to extract the binary-state threshold of gray level image, then utilize binary-state threshold to carry out binaryzation to gray level image;
Step 3: Image semantic classification; First medium filtering is carried out to the image after binaryzation, remove less noise, then use morphological method to carry out opening operation to image, make the round dot belonging to same character in image close to bonding;
Step 4: cable region is slightly extracted; The cable image obtained after pre-service is projected, calculates the cut point at two ends up and down extracting image, according to cut point position, original image is cut, thus obtain comprising whole cable region and the little as far as possible image of area;
Step 5: cable regional dip in image is corrected; Due in process of production, cable is in high-speed motion state, and the cable image major part therefore obtained exists the situation of small angle inclination, needs to correct it;
Step 6: character skewness correction and segmentation; Because character is in jet printing process, for high-speed motion state, there is the situation that spray printing character is not exclusively vertical with cable in cable unavoidably, therefore needs to tilt to correct to character, obtain the character picture vertical with cable border;
Step 7: character is separated; Character each in image is separated, obtains the independent image of each character;
Step 8: template matches; The independent image of each character is mated with masterplate character, identifies character picture.
The concrete steps of described step 5 are:
Step 5.1: be extracted cable edge pixels, carries out least square fitting to edge pixels, thus obtains cable angle of inclination;
Step 5.2: utilize matrix rotation algorithm to carry out rotation correction to image.
The concrete steps of described step 7 are: obtain the image after character correction, because point-like character picture exists uncontinuity, first image are carried out lateral expansion, longitudinal projection, then carry out longitudinal cut point searching, so just can obtain each character picture after separating;
The concrete steps of described step 8 are:
Step 8.1: by image consistent sized by the independent image conversion of each character;
Step 8.2: character is slightly identified according to the size of character in image;
Step 8.3: adopt character masterplate to mate character in image, identify character.
The present invention proposes the disposal route of cable surface point-like character first, and the method accurately can identify point-like numeral, English character and a few symbols of all kinds of cable wire surface; In cable production run, occur the situation that character tilts during character spray printing unavoidably, this method still can accurately identify when character exists slight inclination, has certain antijamming capability.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of a kind of electric wire superficial punctate of the present invention character identifying method;
Fig. 2 is the picture that first this law is extracted;
Fig. 3 is the picture in the cable region that the present invention extracts after step 4;
Fig. 4 is the character picture of the present invention after step 6;
Fig. 5 is the independent character picture that the present invention finally obtains.
Embodiment
Below in conjunction with accompanying drawing, the present invention is further detailed.
Step 1: read in the cable gray level image comprising point-like character.
Step 2: image binaryzation; Adopt Otsu algorithm to extract the binary-state threshold of gray level image, then utilize binary-state threshold to carry out binaryzation to gray level image.
Step 3: Image semantic classification; First medium filtering is carried out to the image after binaryzation, remove less noise, then use morphological method to choose suitable morphology template and opening operation is carried out to image, make the initial point belonging to same character in image close to bonding;
Step 4: cable region is slightly extracted; The image obtained after pre-service is projected, chooses suitable threshold value, calculate the cut point at upper and lower two ends, cable region, according to cut point position, original image is cut, thus downscaled images size, reduce operand, improve arithmetic speed;
Step 5: cable regional dip corrects; Due in process of production, cable is for high-speed motion state, and the cable image major part therefore obtained exists the situation of small angle inclination.In order to correct inclination cable, the present invention is extracted cable edge pixels, carries out least square fitting to edge pixels, thus obtains cable angle of inclination, and recycling matrix rotation algorithm carries out rotation correction to image.The edge pixels vector calculated is designated as, because cable edge is commonly straight line, adopts a curve to carry out matching here.Suppose to treat fit curve equation, bring edge pixels vector into curvilinear equation successively, meter mean square deviation is then have
Obtain thus about quadratic equation, try to achieve respectively, make two partial derivatives equal zero, variance can be tried to achieve respectively and minimum time, bring into curvilinear equation can obtain matching after curve, obtain curve slope thus.Utilize
Postrotational image can be obtained, thus complete the slant correction of cable image.
Step 6: character skewness correction and segmentation; Because character is in jet printing process, for high-speed motion state, there is the situation that spray printing character is not exclusively vertical with cable in cable unavoidably, therefore needs to tilt to correct to character, obtain the character picture vertical with cable border, be convenient to later algorithm process.Within experimentally finding that the angle of inclination on all characters and cable border is all positioned at a scope, therefore the present invention adopts priori value, carries out translation, carry out correction character image by the control of translation pixel count to the character pixels of diverse location.
Step 7: Character segmentation; After obtaining the character picture after correcting, because point-like character picture exists uncontinuity, first image is carried out lateral expansion, longitudinal projection, then carry out longitudinal cut point searching, so just can obtain the character picture after splitting; Matrix template for carrying out expanding is complete 1 vector of one 22 row, obtains now expansion effect best through test; By judging the position at zero point and the distance between zero point in projection vector, just can obtain suitable Character segmentation point, thus obtain the single character picture after splitting.
Step 8: template matches; Cause because the image obtained after segmentation is not of uniform size, be unfavorable for the carrying out of template matches, therefore, use Image Zooming algorithm, picture size is normalized to 75 × 95.Then the image after normalization is mated with Character mother plate, find the template that related coefficient is the highest, as recognition result.The computing formula of related coefficient is as follows:
Step 9: character recognition logic designs; The size that different characters is corresponding may be inconsistent, and such as character " 1 ", the width of " 0 " and the width of other numerical characters exist bigger difference, before therefore can carrying out template matches again, utilize dimension information first classify to character or slightly identify.
Claims (4)
1. the point-like character identifying method on electric wire surface, the method comprises the following steps:
Step 1: extract the cable gray level image comprising point-like character;
Step 2: image binaryzation; Adopt Otsu algorithm to extract the binary-state threshold of gray level image, then utilize binary-state threshold to carry out binaryzation to gray level image;
Step 3: Image semantic classification; First medium filtering is carried out to the image after binaryzation, remove less noise, then use morphological method to carry out opening operation to image, make the round dot belonging to same character in image close to bonding;
Step 4: cable region is slightly extracted; The cable image obtained after pre-service is projected, calculates the cut point at two ends up and down extracting image, according to cut point position, original image is cut, thus obtain comprising whole cable region and the little as far as possible image of area;
Step 5: cable regional dip in image is corrected; Due in process of production, cable is in high-speed motion state, and the cable image major part therefore obtained exists the situation of small angle inclination, needs to correct it;
Step 6: character skewness correction and segmentation; Because character is in jet printing process, for high-speed motion state, there is the situation that spray printing character is not exclusively vertical with cable in cable unavoidably, therefore needs to tilt to correct to character, obtain the character picture vertical with cable border;
Step 7: character is separated; Character each in image is separated, obtains the independent image of each character;
Step 8: template matches; The independent image of each character is mated with masterplate character, identifies character picture.
2. the point-like character identifying method on a kind of electric wire surface as claimed in claim 1, is characterized in that the concrete steps of described step 5 are:
Step 5.1: be extracted cable edge pixels, carries out least square fitting to edge pixels, thus obtains cable angle of inclination;
Step 5.2: utilize matrix rotation algorithm to carry out rotation correction to image.
3. the point-like character identifying method on a kind of electric wire surface as claimed in claim 1, it is characterized in that the concrete steps of described step 7 are: obtain the image after character correction, because point-like character picture exists uncontinuity, first image is carried out lateral expansion, longitudinal projection, carry out longitudinal cut point searching again, so just can obtain each character picture after separating.
4. the point-like character identifying method on a kind of electric wire surface as claimed in claim 1, is characterized in that the concrete steps of described step 8 are:
Step 8.1: by image consistent sized by the independent image conversion of each character;
Step 8.2: character is slightly identified according to the size of character in image;
Step 8.3: adopt character masterplate to mate character in image, identify character.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510206069.4A CN104820827A (en) | 2015-04-28 | 2015-04-28 | Method for recognizing punctiform characters on surfaces of cables |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510206069.4A CN104820827A (en) | 2015-04-28 | 2015-04-28 | Method for recognizing punctiform characters on surfaces of cables |
Publications (1)
Publication Number | Publication Date |
---|---|
CN104820827A true CN104820827A (en) | 2015-08-05 |
Family
ID=53731117
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510206069.4A Pending CN104820827A (en) | 2015-04-28 | 2015-04-28 | Method for recognizing punctiform characters on surfaces of cables |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104820827A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108596173A (en) * | 2018-04-19 | 2018-09-28 | 长春理工大学 | One camera full view wire size real-time distinguishing apparatus and its detection method |
CN108805128A (en) * | 2017-05-05 | 2018-11-13 | 北京京东金融科技控股有限公司 | A kind of character segmentation method and device |
EP3460719A1 (en) * | 2017-04-21 | 2019-03-27 | DTP spólka z ograniczona odpowiedzialnoscia | The device for identifying wire markings and the method for identifying wire markings |
CN110119648A (en) * | 2018-02-05 | 2019-08-13 | 国家计算机网络与信息安全管理中心 | A kind of facsimile signal classification method based on optical character identification |
CN112150397A (en) * | 2020-10-15 | 2020-12-29 | 江西威力固智能设备有限公司 | Character processing method and jet printing equipment |
CN117037185A (en) * | 2023-10-10 | 2023-11-10 | 天津普传控制设备有限公司 | Wire number recognition method and device based on optical character recognition technology |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2004252810A (en) * | 2003-02-21 | 2004-09-09 | Fujitsu Ltd | Method, device and program for matching image and document |
CN101739556A (en) * | 2010-01-21 | 2010-06-16 | 上海交通大学 | Method for automatically identifying number of steel billet |
CN102024144A (en) * | 2010-11-23 | 2011-04-20 | 上海海事大学 | Container number identification method |
CN102945368A (en) * | 2012-10-17 | 2013-02-27 | 西安理工大学 | Method for positioning and identifying laser character of beer bottle cap |
CN104112132A (en) * | 2014-07-03 | 2014-10-22 | 中国人民解放军第二炮兵工程大学 | Automatic gun number identification method |
CN104331688A (en) * | 2014-11-05 | 2015-02-04 | 中北大学 | Detonator shell dot character identifying method |
CN104463124A (en) * | 2014-12-11 | 2015-03-25 | 天津普达软件技术有限公司 | Milk box spray-printed character recognition method |
-
2015
- 2015-04-28 CN CN201510206069.4A patent/CN104820827A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2004252810A (en) * | 2003-02-21 | 2004-09-09 | Fujitsu Ltd | Method, device and program for matching image and document |
CN101739556A (en) * | 2010-01-21 | 2010-06-16 | 上海交通大学 | Method for automatically identifying number of steel billet |
CN102024144A (en) * | 2010-11-23 | 2011-04-20 | 上海海事大学 | Container number identification method |
CN102945368A (en) * | 2012-10-17 | 2013-02-27 | 西安理工大学 | Method for positioning and identifying laser character of beer bottle cap |
CN104112132A (en) * | 2014-07-03 | 2014-10-22 | 中国人民解放军第二炮兵工程大学 | Automatic gun number identification method |
CN104331688A (en) * | 2014-11-05 | 2015-02-04 | 中北大学 | Detonator shell dot character identifying method |
CN104463124A (en) * | 2014-12-11 | 2015-03-25 | 天津普达软件技术有限公司 | Milk box spray-printed character recognition method |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP3460719A1 (en) * | 2017-04-21 | 2019-03-27 | DTP spólka z ograniczona odpowiedzialnoscia | The device for identifying wire markings and the method for identifying wire markings |
EP4102471A1 (en) * | 2017-04-21 | 2022-12-14 | DTP spólka z ograniczona odpowiedzialnoscia | Method for identifying wire markings |
CN108805128A (en) * | 2017-05-05 | 2018-11-13 | 北京京东金融科技控股有限公司 | A kind of character segmentation method and device |
CN108805128B (en) * | 2017-05-05 | 2023-11-07 | 京东科技控股股份有限公司 | Character segmentation method and device |
CN110119648A (en) * | 2018-02-05 | 2019-08-13 | 国家计算机网络与信息安全管理中心 | A kind of facsimile signal classification method based on optical character identification |
CN108596173A (en) * | 2018-04-19 | 2018-09-28 | 长春理工大学 | One camera full view wire size real-time distinguishing apparatus and its detection method |
CN112150397A (en) * | 2020-10-15 | 2020-12-29 | 江西威力固智能设备有限公司 | Character processing method and jet printing equipment |
CN112150397B (en) * | 2020-10-15 | 2024-02-20 | 江西威力固智能设备有限公司 | Character processing method and jet printing equipment |
CN117037185A (en) * | 2023-10-10 | 2023-11-10 | 天津普传控制设备有限公司 | Wire number recognition method and device based on optical character recognition technology |
CN117037185B (en) * | 2023-10-10 | 2024-01-12 | 天津普传控制设备有限公司 | Wire number recognition method and device based on optical character recognition technology |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104820827A (en) | Method for recognizing punctiform characters on surfaces of cables | |
Greenhalgh et al. | Recognizing text-based traffic signs | |
CN105229669B (en) | Image processing apparatus and image processing method | |
CN102663378B (en) | Method for indentifying joined-up handwritten characters | |
EP3309705B1 (en) | Qr code feature detection method and system | |
Lelore et al. | FAIR: a fast algorithm for document image restoration | |
US9235755B2 (en) | Removal of underlines and table lines in document images while preserving intersecting character strokes | |
LeBourgeois | Robust multifont OCR system from gray level images | |
CN104361336A (en) | Character recognition method for underwater video images | |
Kita et al. | Binarization of color characters in scene images using k-means clustering and support vector machines | |
CN108133216B (en) | Nixie tube reading identification method capable of realizing decimal point reading based on machine vision | |
CN105426890B (en) | A kind of graphical verification code recognition methods of character distortion adhesion | |
CN107122775A (en) | A kind of Android mobile phone identity card character identifying method of feature based matching | |
Cohen et al. | Using scale-space anisotropic smoothing for text line extraction in historical documents | |
Bijalwan et al. | Automatic text recognition in natural scene and its translation into user defined language | |
CN108205678B (en) | Nameplate character recognition processing method containing bright spot interference | |
Karanje et al. | Survey on text detection, segmentation and recognition from a natural scene images | |
CN106650719B (en) | Method and device for identifying picture characters | |
JP5439069B2 (en) | Character recognition device and character recognition method | |
Huang et al. | Scene character detection and recognition based on multiple hypotheses framework | |
JP6890849B2 (en) | Information processing system | |
Mohan et al. | Handwritten character recognition: a comprehensive review on geometrical analysis | |
Abu-Ain et al. | Text normalization framework for handwritten cursive languages by detection and straightness the writing baseline | |
CN108596182B (en) | Manchu component cutting method | |
CN110738268A (en) | intelligent stereoscopic warehouse goods automatic identification method based on SIFT and DDIS |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
EXSB | Decision made by sipo to initiate 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: 20150805 |