CN106469312A - Weld seam visual identity method based on region growing label - Google Patents
Weld seam visual identity method based on region growing label Download PDFInfo
- Publication number
- CN106469312A CN106469312A CN201610867648.8A CN201610867648A CN106469312A CN 106469312 A CN106469312 A CN 106469312A CN 201610867648 A CN201610867648 A CN 201610867648A CN 106469312 A CN106469312 A CN 106469312A
- Authority
- CN
- China
- Prior art keywords
- image
- weld seam
- weld
- label
- region
- 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
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- 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
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
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)
- Image Processing (AREA)
Abstract
The present invention provides a kind of weld seam visual identity method based on region growing label, obtains the weld image of object to be detected by image capture device;Using median filter method, image denoising is carried out to the weld image being obtained, the noise that image enhaucament removes image is carried out using histogram of gradients, and improves the contrast of weld seam and ambient background;Weld image after Image semantic classification is carried out weld seam separate with ambient background, the algorithm using region growing enters line label to connected region, regions most for pixel count is carried out binaryzation, weld seam is split from image;Extract axis of a weld, obtain weld seam position in the picture;The weld seam visual identity method based on region growing label for this kind, and only need to common light source, the speed being capable of 10 frames/second carries out image acquisition, realizes real-time weld seam detection, ensure that the weld seam of extraction is more accurate.
Description
Technical field
The present invention relates to a kind of weld seam visual identity method based on region growing label.
Background technology
Image procossing is with computer, image to be analyzed, to reach the technology of results needed.Image procossing refers generally to
Digital Image Processing, digital picture refers to equipment such as industrial camera, video camera, scanneies through shooting the two-dimemsional number obtaining
Group, the element of this array is referred to as pixel, and its value is referred to as gray value.Image processing techniquess generally comprise image filtering and strengthen, scheme
As segmentation, 3 parts of feature extraction and identification.
The image processing algorithm flow process of weld seam detection:Image acquisition, Image semantic classification, image segmentation, feature extraction.At present
Weld seam detection algorithm be weld seam detection for welding process, and weld seam detection algorithm exists or algorithm is complicated, or accurately
The undesirable problem of property.Additionally, existing weld seam detection generally adopts laser as structure light source, need accurate optical device,
High cost, and for the setting angle of laser, there is higher requirement.
The problems referred to above are the problems that should pay attention to during weld seam detection and solve.
Content of the invention
It is an object of the invention to provide a kind of weld seam visual identity method based on region growing label, for existing weld seam
Detected, solved the above-mentioned problems in the prior art.
The technical solution of the present invention is:
A kind of weld seam visual identity method based on region growing label, comprises the following steps:
Image acquisition, obtains the weld image of object to be detected by image capture device;
Image semantic classification:Using median filter method, image denoising is carried out to the weld image being obtained, straight using gradient
Square figure carries out the noise that image enhaucament removes image, and improves the contrast of weld seam and ambient background;
Image segmentation:Weld image after Image semantic classification is carried out weld seam separate with ambient background, using region growing
Algorithm obtain connected region line label is entered to connected region, connected regions most for pixel count is carried out binaryzation, general
Weld seam splits from image;
Feature extraction:Extract axis of a weld, obtain weld seam position in the picture.
Further, in image segmentation, first using maximum variance between clusters, image is carried out binaryzation, then adopt region
The method of growth obtains connected region, and enters line label to each connected region, specially:In the horizontal direction and perpendicular respectively first
Nogata chooses a seed point to every several points, and the pixel value of each seed point puts different labels, the kind to each label
Son point investigates the eight neighborhood pixel of surrounding, if a certain neighborhood territory pixel point meets growth criterion, the picture of this neighborhood territory pixel point
Plain value and seed point put same label, are then investigated this neighborhood territory pixel point as new seed point, until not having pixel
Point can merge, and finally obtains the connected region of some labels.
Further, in image segmentation, connected regions most for pixel count is carried out binaryzation and is specially:Pixel count is most
The pixel value of connected region be set to 0, the pixel value of other points is set to 255.
Further, in feature extraction, extract axis of a weld and be specially:By obtaining weld seam upper half area and bottom half
The center-of-mass coordinate in domain, 2 points of lines obtain axis of a weld.
Further, image acquisition is to be realized by building image capture device, and image capture device includes camera, mirror
Head and light source, to obtain weld image.
Further, image acquisition carries out the collection of image by DirectShow Streaming Media kit, and speed is 8-15
Frame/second.
The invention has the beneficial effects as follows:The weld seam visual identity method based on region growing label for this kind, is for existing
The detection of weld seam, obtains connected region using algorithm of region growing, and connected region is entered with line label, extracts pixel count most
Connected region carries out binaryzation, and by weld seam from the background segment of surrounding out, the speed being capable of 10 frames/second carries out image and adopts
Collection, realizes real-time weld seam detection.On the basis of the method carries out binaryzation in maximum variance between clusters, using the side of region growing
Method obtains the connected region of some labels, noise connected region is separated with weld seam connected region, further noise reduction, extracts pixel
The satisfactory connected region of number, the typically connected region of maximum, the weld seam of extraction is more accurate.This kind is based on region growing
The weld seam visual identity method of label, it is only necessary to common light source, low cost, is installed simple, is easy to debug.
Brief description
Fig. 1 is the schematic flow sheet based on the weld seam visual identity method of region growing label for the embodiment of the present invention.
Fig. 2 is the schematic flow sheet adopting the method for region growing to obtain connected region in embodiment.
Fig. 3 is the weld seam original image of embodiment image acquisition.
Fig. 4 is the weld image in embodiment after Image semantic classification.
Fig. 5 is the weld image in embodiment after image segmentation.
Specific embodiment
Describe the preferred embodiments of the present invention below in conjunction with the accompanying drawings in detail.
The weld seam visual identity method based on region growing label of embodiment, for the weld seam detection of large-size spherical tank, leads to
Cross the positional information that image processing techniquess obtain weld seam.
Embodiment
A kind of weld seam visual identity method based on region growing label, such as Fig. 1, comprise the following steps:
Image acquisition, obtains the weld image of object to be detected, such as Fig. 3 by image capture device;
Image semantic classification:Using median filter method, image denoising is carried out to the weld image being obtained, straight using gradient
Square figure carries out image enhaucament, improves the contrast of weld seam and ambient background, such as Fig. 4;
Image segmentation:Weld image after Image semantic classification is carried out binaryzation using maximum variance between clusters, using area
The algorithm of domain growth enters line label to connected region, extracted region most for pixel count is weld seam, by weld seam from image
Split, such as Fig. 5;
Feature extraction:Extract axis of a weld, obtain weld seam position in the picture.
In embodiment, image capture device, including camera, camera lens and light source, by DirectShow Streaming Media kit
Carry out the collection of image, speed is 8-15 frame/second, preferably 10 frames/second.
In embodiment, image segmentation is specially:First using maximum variance between clusters, image is carried out binaryzation, then adopt
The method of region growing obtains connected region, and enters line label to each connected region, such as Fig. 2, in the horizontal direction and perpendicular first
Nogata chooses a seed point to every 5 points, and the pixel value of each seed point puts different labels, the seed point to each label
Eight neighborhood pixel around investigating, if a certain neighborhood territory pixel point meets growth criterion, its pixel value and seed point put with
Then this neighborhood territory pixel point is investigated as new seed point, until not having pixel can merge, is finally obtained by one label
Obtain the connected region of some labels, the pixel value of the wherein most connected region of pixel count is set to 0, the pixel value of other points is set to
255.
In feature extraction, extract axis of a weld and be specially:By obtaining the barycenter of weld seam upper half area and lower half region
Coordinate, 2 points of lines obtain axis of a weld.
The weld seam visual identity method based on region growing label for this kind of embodiment, is the detection for existing weld seam,
Using the algorithm of region growing, connected region is entered with line label, extract the most connected region of pixel count and carry out binaryzation, will weld
Seam from the background segment of surrounding out, extracts the most connective region of pixel count and is weld seam, be capable of the speed of 10 frames/second
Degree carries out image acquisition, realizes real-time weld seam detection.
Image acquisition is to be realized by building image capture device, and image capture device includes camera, camera lens and light source,
To obtain weld image.The weld seam visual identity method based on region growing label for this kind it is not necessary to the structure light source of laser class,
Only need to common light source, low cost, install simple, be easy to debug.
The weld seam visual identity method based on region growing label for this kind, carries out the base of binaryzation in maximum variance between clusters
On plinth, obtain the connected region of some labels using the method for region growing, noise connected region and weld seam connected region are divided
From, further noise reduction, extracts the satisfactory connected region of pixel count, the typically connected region of maximum, the weld seam of extraction is more
Plus accurately.
Claims (6)
1. a kind of weld seam visual identity method based on region growing label is it is characterised in that comprise the following steps:
Image acquisition, obtains the weld image of object to be detected by image capture device;
Image semantic classification:Using median filter method, image denoising is carried out to the weld image being obtained, using histogram of gradients
Carry out the noise that image enhaucament removes image, and improve the contrast of weld seam and ambient background;
Image segmentation:Weld image after Image semantic classification is carried out weld seam separate with ambient background, using the calculation of region growing
Method obtains connected region and enters line label to connected region, connected regions most for pixel count is carried out binaryzation, by weld seam
Split from image;
Feature extraction:Extract axis of a weld, obtain weld seam position in the picture.
2. the weld seam visual identity method based on region growing label as claimed in claim 1 it is characterised in that:Image segmentation
In, first using maximum variance between clusters, image is carried out binaryzation, then adopt the method for region growing to obtain connected region, and
Line label is entered to each connected region, specially:Choose one with vertical direction every several points in the horizontal direction respectively first
Individual seed point, the pixel value of each seed point puts different labels, to the eight neighborhood pixel around the seed point investigation of each label
Point, if a certain neighborhood territory pixel point meets growth criterion, the pixel value of this neighborhood territory pixel point and seed point put same label, so
Afterwards this neighborhood territory pixel point is investigated as new seed point, until not having pixel can merge, finally obtained some marks
Number connected region.
3. the weld seam visual identity method based on region growing label as claimed in claim 1 it is characterised in that:Image segmentation
In, connected regions most for pixel count is carried out binaryzation and is specially:The pixel value of the most connected region of pixel count is set to 0,
The pixel value of other points is set to 255.
4. the weld seam visual identity method based on region growing label as described in any one of claim 1-3 it is characterised in that:
In feature extraction, extract axis of a weld and be specially:By obtaining the center-of-mass coordinate of weld seam upper half area and lower half region, 2 points
Line obtains axis of a weld.
5. the weld seam visual identity method based on region growing label as described in any one of claim 1-3 it is characterised in that:
Image acquisition is to be realized by building image capture device, and image capture device includes camera, camera lens and light source, to obtain weldering
Seam image.
6. the weld seam visual identity method based on region growing label as claimed in claim 5 it is characterised in that:Image acquisition
Carry out the collection of image by DirectShow Streaming Media kit, speed is 8-15 frame/second.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610867648.8A CN106469312A (en) | 2016-09-29 | 2016-09-29 | Weld seam visual identity method based on region growing label |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610867648.8A CN106469312A (en) | 2016-09-29 | 2016-09-29 | Weld seam visual identity method based on region growing label |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106469312A true CN106469312A (en) | 2017-03-01 |
Family
ID=58230758
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610867648.8A Pending CN106469312A (en) | 2016-09-29 | 2016-09-29 | Weld seam visual identity method based on region growing label |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106469312A (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107154042A (en) * | 2017-05-18 | 2017-09-12 | 湖南大学 | A kind of seed-coating machine visible detection method and device |
CN107274420A (en) * | 2017-06-15 | 2017-10-20 | 中国水产科学研究院东海水产研究所 | The oceanic front extracting method split based on image |
CN107909567A (en) * | 2017-10-31 | 2018-04-13 | 华南理工大学 | The slender type connected region extracting method of digital picture |
CN109146866A (en) * | 2018-08-23 | 2019-01-04 | 深圳市神视检验有限公司 | The method and device that robot handles weld seam |
CN109514043A (en) * | 2018-12-13 | 2019-03-26 | 刘堂斌 | A kind of effective welding robot welding control system |
CN110021017A (en) * | 2019-04-02 | 2019-07-16 | 南通大学 | A method of extracting axis of a weld |
CN111126392A (en) * | 2019-12-25 | 2020-05-08 | 江苏恒创软件有限公司 | Digital image rapid identification method and device for ladle label |
CN112053368A (en) * | 2019-09-23 | 2020-12-08 | 华北电力大学(保定) | Welding seam center identification method and system for sheet welding |
CN112132807A (en) * | 2020-09-23 | 2020-12-25 | 泉州装备制造研究所 | Weld joint region extraction method and device based on color similarity segmentation |
CN112581398A (en) * | 2020-12-22 | 2021-03-30 | 上海电机学院 | Image noise reduction method based on region growing labels |
CN114820629A (en) * | 2022-07-01 | 2022-07-29 | 山东意吉希精密制造有限公司 | Welding identification method for automobile parts |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105205821A (en) * | 2015-09-21 | 2015-12-30 | 江苏科技大学 | Weld image segmentation method |
CN105588845A (en) * | 2016-01-04 | 2016-05-18 | 江苏科技大学 | Weld defect characteristic parameter extraction method |
-
2016
- 2016-09-29 CN CN201610867648.8A patent/CN106469312A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105205821A (en) * | 2015-09-21 | 2015-12-30 | 江苏科技大学 | Weld image segmentation method |
CN105588845A (en) * | 2016-01-04 | 2016-05-18 | 江苏科技大学 | Weld defect characteristic parameter extraction method |
Non-Patent Citations (7)
Title |
---|
JINTAO YU ET AL.: "Weld Seam Recognition Based On CCD Image Processing", 《INTERNATIONAL JOURNAL OF SIGNAL PROCESSING, IMAGE PROCESSING AND PATTERN RECOGNITION》 * |
孙太生 等: "基于改进区域生长算法的焊缝图像分割", 《现代焊接》 * |
李秀忠 等: "基于二值形态学的焊缝图像处理", 《试验与研究》 * |
林立峰: "基于机器视觉的胶囊检测系统", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
段大高 等: "《图像处理与应用》", 31 January 2013 * |
王璇: "焊接机器人焊缝信息视觉检测系统研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
聂欢欢 等: "基于区域生长法提取二值图像中的连通区域", 《计算机时代》 * |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107154042A (en) * | 2017-05-18 | 2017-09-12 | 湖南大学 | A kind of seed-coating machine visible detection method and device |
CN107154042B (en) * | 2017-05-18 | 2020-02-21 | 湖南大学 | Visual detection method and device for coating machine |
CN107274420B (en) * | 2017-06-15 | 2020-08-14 | 中国水产科学研究院东海水产研究所 | Ocean frontal surface extraction method based on image segmentation |
CN107274420A (en) * | 2017-06-15 | 2017-10-20 | 中国水产科学研究院东海水产研究所 | The oceanic front extracting method split based on image |
CN107909567A (en) * | 2017-10-31 | 2018-04-13 | 华南理工大学 | The slender type connected region extracting method of digital picture |
CN107909567B (en) * | 2017-10-31 | 2022-02-15 | 华南理工大学 | Method for extracting slender connected region of digital image |
CN109146866A (en) * | 2018-08-23 | 2019-01-04 | 深圳市神视检验有限公司 | The method and device that robot handles weld seam |
CN109514043A (en) * | 2018-12-13 | 2019-03-26 | 刘堂斌 | A kind of effective welding robot welding control system |
CN109514043B (en) * | 2018-12-13 | 2020-11-24 | 南昌市龙诚电器设备有限公司 | Effective welding control system of welding robot |
CN110021017A (en) * | 2019-04-02 | 2019-07-16 | 南通大学 | A method of extracting axis of a weld |
CN112053368A (en) * | 2019-09-23 | 2020-12-08 | 华北电力大学(保定) | Welding seam center identification method and system for sheet welding |
CN112053368B (en) * | 2019-09-23 | 2023-07-21 | 华北电力大学(保定) | Weld joint center identification method and system for sheet welding |
CN111126392A (en) * | 2019-12-25 | 2020-05-08 | 江苏恒创软件有限公司 | Digital image rapid identification method and device for ladle label |
CN112132807A (en) * | 2020-09-23 | 2020-12-25 | 泉州装备制造研究所 | Weld joint region extraction method and device based on color similarity segmentation |
CN112132807B (en) * | 2020-09-23 | 2024-02-23 | 泉州装备制造研究所 | Weld joint region extraction method and device based on color similarity segmentation |
CN112581398A (en) * | 2020-12-22 | 2021-03-30 | 上海电机学院 | Image noise reduction method based on region growing labels |
CN114820629A (en) * | 2022-07-01 | 2022-07-29 | 山东意吉希精密制造有限公司 | Welding identification method for automobile parts |
CN114820629B (en) * | 2022-07-01 | 2022-09-02 | 山东意吉希精密制造有限公司 | Welding identification method for automobile parts |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106469312A (en) | Weld seam visual identity method based on region growing label | |
CN106709436B (en) | Track traffic panoramic monitoring-oriented cross-camera suspicious pedestrian target tracking system | |
KR101403876B1 (en) | Method and Apparatus for Vehicle License Plate Recognition | |
CN102663760B (en) | Location and segmentation method for windshield area of vehicle in images | |
CN104616275A (en) | Defect detecting method and defect detecting device | |
CN108982512A (en) | A kind of circuit board detecting system and method based on machine vision | |
CN109993099A (en) | A kind of lane line drawing recognition methods based on machine vision | |
US10586321B2 (en) | Automatic detection, counting, and measurement of lumber boards using a handheld device | |
CN105262991B (en) | A kind of substation equipment object identifying method based on Quick Response Code | |
JP2013065304A (en) | High-speed obstacle detection | |
CN105044114A (en) | Electrolytic capacitor appearance package defect image detection system and electrolytic capacitor appearance package defect image detection method | |
CN109741314A (en) | A kind of visible detection method and system of part | |
CN104268853A (en) | Infrared image and visible image registering method | |
CN109886935A (en) | A kind of road face foreign matter detecting method based on deep learning | |
CN110648330B (en) | Defect detection method for camera glass | |
CN104282018A (en) | Method for on-line detection of overall diameter of industrial product based on machine vision | |
CN110414308A (en) | A kind of target identification method for dynamic foreign matter on transmission line of electricity | |
CN111257329A (en) | Smartphone camera defect detection method and detection system | |
WO2017120796A1 (en) | Pavement distress detection method and apparatus, and electronic device | |
CN111368797A (en) | Target real-time ranging method based on road end monocular camera | |
CN105957300B (en) | A kind of wisdom gold eyeball identification is suspicious to put up masking alarm method and device | |
CN105718964A (en) | Transmission line vibration damper visual detection method | |
CN108961262B (en) | Bar code positioning method in complex scene | |
US20200193561A1 (en) | Automatic Detection, Counting, and Measurement of Lumber Boards Using a Handheld Device | |
CN103390259A (en) | Ground image processing method in visual guidance AGV |
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 |
Application publication date: 20170301 |
|
RJ01 | Rejection of invention patent application after publication |