CN110956619B - Curved glass defect detection method - Google Patents
Curved glass defect detection method Download PDFInfo
- Publication number
- CN110956619B CN110956619B CN201911167579.XA CN201911167579A CN110956619B CN 110956619 B CN110956619 B CN 110956619B CN 201911167579 A CN201911167579 A CN 201911167579A CN 110956619 B CN110956619 B CN 110956619B
- Authority
- CN
- China
- Prior art keywords
- glass
- image
- gray value
- area
- pixel
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
A method for detecting defects of curved glass is disclosed, which comprises collecting images formed after stripe light penetrates through glass; dividing the image into a plurality of first areas according to a gray value threshold value division method and determining a skeleton of each first area; each first area is divided into a plurality of strip-shaped second areas with pixel columns perpendicular to the framework along the framework, and if the gray value is changed violently along the pixel column direction, the position of the violent change has a defect. The method solves the problem that the image defect cannot be detected by comparing the image with a defect-free image due to the limitation of the processing precision of the curved glass.
Description
Technical Field
The invention relates to the field of glass detection, in particular to a curved glass defect detection method.
Background
In the prior art, the defects of the glass are generally detected by analyzing an image formed by reflecting light of curved glass by using a machine vision analysis technology, however, the defects are detected by using the principle, and the highlight formed by reflection can interfere with the development of the defects of the glass.
Chinese patent application CN103344651A published in 2013, 10, 9 and discloses a method for acquiring an image formed by a strip light source penetrating through glass, comparing the acquired image with an image of defect-free glass, and acquiring a glass defect according to a phase difference. The intrinsic law of defect visualization in this method is unknown, but it has been confirmed by experiments that it is really feasible. However, in practice, this method cannot be applied to defect detection of curved glass because the two pieces of curved glass without defects have different image phases after being modulated by the fringe light due to the limitation of the processing precision of the curved glass. Therefore, in practice, since a defect-free image of the curved glass to be inspected cannot be obtained, an image defect cannot be detected by a phase difference obtained by comparing the image with the defect-free image.
Disclosure of Invention
The invention aims to overcome the defects or problems in the background art and provide a curved glass defect detection method, which solves the problem that highlight formed by reflection interferes with the development of glass defects and also solves the problem that image defects cannot be detected by comparing with defect-free images due to the limitation of the processing precision of curved glass.
In order to achieve the purpose, the following technical scheme is adopted:
the first technical scheme defines a curved glass defect detection method, which comprises the following steps: collecting an image formed by a light and shade alternative stripe light source after penetrating through glass to be detected; at least dividing a part of the image corresponding to the curved surface area of the glass to be detected into a plurality of connected first areas according to a gray value threshold segmentation method; determining a skeleton of each first region; dividing each first area into a plurality of second areas which are connected or intersected with each other along the framework, wherein all the second areas cover the first areas; the second area is in a strip shape which is vertical to the framework, two ends of the second area coincide with the boundary of the first area, pixel rows of the second area are defined to be vertical to the framework, and if the second area is found to have drastic gray value change along the direction of the pixel rows, the corresponding position of the glass to be detected corresponding to the position of the drastic change in the image has defects.
Based on the first technical scheme, a second technical scheme is also disclosed, wherein a specific method for finding that the gray value of the second area is changed violently along the pixel column direction is as follows: defining pixels in a direction perpendicular to the pixel columns to form pixel rows, the pixel rows including one or more pixels; averaging the gray values of the pixels of each pixel row to obtain a row gray value; deriving the gray value of the row pixel along the direction of the pixel column to obtain a derivative value of the row pixel; dividing a row pixel gray value curve into one or more sections of a positive derivative value section, a near zero section, a negative derivative value section and a near zero section which are connected with each other by a threshold segmentation method for row pixel derivative values along the direction of pixel columns; and if a negative derivative value appears in the positive derivative value section or a positive derivative value appears in the negative derivative value section or the absolute value of the difference between the gray value of a certain row of pixels and the average value of the gray value of the pixels of the row of the near-zero section appears in the near-zero section is larger than a specific multiple of the image noise value, determining that the gray value of the second area is changed violently along the pixel column direction.
Based on the first technical scheme, a third technical scheme is also disclosed, which further comprises: for the part corresponding to the plane area of the glass to be detected in the image, dividing the part in the image into a plurality of connected third areas along the direction vertical to the light and shade change direction, wherein the third areas are strip-shaped, the length direction of the third areas is defined to be parallel to the light and shade change direction, and the width direction of the third areas is defined to be vertical to the light and shade change direction and comprises one or more pixels; averaging the gray values of all pixels in the width direction along the length direction to obtain the average gray value in the width direction; carrying out Fourier transformation and inverse transformation on the frequency information of the width direction average gray value curve along the length direction to obtain a width direction average gray value curve without stripe textures; if a drastic change of the gray value exists in the width-direction average gray value curve for removing the stripe texture, the corresponding position of the glass to be measured corresponding to the position in the image of the drastic change has a defect.
Based on the first, second or third technical scheme, a fourth technical scheme is also disclosed, the method further comprises the steps of rotating the stripe light source relative to the glass to be detected and stopping at least at two positions, wherein the changing directions of the light and shade phases of the stripe light source at each position are intersected with each other; the collecting of the image formed by the light and shade alternate stripe light source after penetrating through the glass to be detected specifically comprises collecting the image formed by the light emitted by the stripe light source after penetrating through the area to be detected of the glass to be detected when the stripe light source is positioned at each position; and each image acquired is subjected to subsequent analysis.
Compared with the prior art, the scheme has the following beneficial effects:
in the first technical scheme, because the images formed after the light and shade alternative stripe light sources along the specific direction penetrate through the glass to be measured are collected, the problem of highlight interference defect development caused by collecting reflected light can not be caused. Secondly, because the curved surface area in the image is divided into a plurality of first areas extending along the stripes (according to a threshold segmentation method, the first areas in the image all extend along the junctions of the bright stripes and the dark stripes), then the skeleton of the first areas is searched, and the first areas are divided into a plurality of second areas vertical to the skeleton along the skeleton, thus, even if the curved surface area stripes of the curved surface glass are possibly distorted, the image defect searching problem can be still decomposed into the defect searching problem of a plurality of strip-shaped areas vertical to the stripe direction (namely, the light and shade change direction).
The reason why the second region is set to be perpendicular to the stripe direction (skeleton direction) is that the applicant found that the gray value of the position of the glass defect is more greatly changed than that of the surrounding pixels by taking one pixel strip in the direction perpendicular to the stripe direction than taking one pixel strip in the stripe direction. This rule is not presently disclosed in any prior art.
In the first technical solution, it is further noted that since the defect position is determined by finding a drastic change in the gray level value in the pixel column direction, the comparison with the defect-free glass is not required any more, and therefore, the problem that the image defect cannot be detected by comparing with the defect-free image due to the limitation of the processing precision of the curved glass is solved.
In the second technical solution, a method for recognizing the drastic change is specifically given, and of course, many methods exist. In which the variation due to noise can be better controlled by taking the mean of the gray values of the pixel lines.
In the third technical means, since the portion corresponding to the flat area does not need to be divided into the plurality of first areas because the flat area does not generate distortion, it is possible to easily separate the gray value change due to noise and the gray value change due to a defect by eliminating the texture generated by the fringe light source by using the fourier transform and the inverse fourier transform.
In the fourth solution, by rotating the stripe light source, some directional defects (e.g. scratches) on the glass are displayed, which is still the principle that the stripe area perpendicular to the stripe direction is easier to display the defects. Specifically, the stripe light source irradiates the glass to be detected in different brightness change directions, so that even if some directional glass defects (such as scratches) are not easily developed in one brightness change direction, the directional glass defects are developed after the brightness change direction is changed, and the defect omission ratio is reduced.
Drawings
In order to more clearly illustrate the technical solution of the embodiments, the drawings needed to be used are briefly described as follows:
FIG. 1 is a schematic structural diagram of an image capturing device for capturing images;
FIG. 2 is a schematic diagram of the distortion of the fringes formed by the curved surface region in the image formed by the light of the fringe light source after passing through the curved glass;
FIG. 3 is a portion of an image with a defect at a first location in the image in a second region;
FIG. 4 is an enlarged view of FIG. 3 at the defect location;
FIG. 5 is a diagram of the gray scale values of the pixels in the second area along the row direction in FIG. 4;
FIG. 6 is a portion of an image with a defect at a second location in the image in a second region;
FIG. 7 is an enlarged view of FIG. 6 at the defective location;
FIG. 8 is a graph of the gray scale values of the pixels in the second area along the row direction in FIG. 7;
FIG. 9 is a schematic diagram showing the relationship between the position of the stripe light source and the position of the glass to be measured when the stripe light source rotates to the first stop position;
FIG. 10 is a schematic diagram illustrating a relationship between the position of the fringe light source and the position of the glass to be measured when the fringe light source rotates to the second stop position.
Description of the main reference numerals:
a stripe light source 1; a carrier 2; an image collector 3; a lens 31; a camera 32; the glass 4 to be measured.
Detailed Description
In the claims and specification, unless otherwise specified the terms "first", "second" or "third", etc., are used to distinguish between different items and are not used to describe a particular order.
In the claims and specification, unless otherwise specified, the terms "central," "lateral," "longitudinal," "horizontal," "vertical," "top," "bottom," "inner," "outer," "upper," "lower," "front," "rear," "left," "right," "clockwise," "counterclockwise," and the like are used in the orientation and positional relationship indicated in the drawings and are used for ease of description only and do not imply that the referenced device or element must have a particular orientation or be constructed and operated in a particular orientation.
In the claims and the specification, unless otherwise defined, the terms "fixedly" or "fixedly connected" are to be understood in a broad sense as meaning any connection which is not in a relative rotational or translational relationship, i.e. including non-detachably fixed connection, integrally connected and fixedly connected by other means or elements.
In the claims and specification, unless otherwise defined, the terms "comprising", "having" and variations thereof mean "including but not limited to".
The technical solution in the embodiments will be clearly and completely described below with reference to the accompanying drawings.
The embodiment of the application relates to a curved glass defect detection method. The method is based on an image acquisition device. As shown in fig. 1, the specific image capturing device includes a stripe light source 1, a carrier 2, and an image capturing unit 3. Wherein, the carrying platform 2 is used for carrying the glass 4 to be measured. The stripe light source 1 can rotate relative to the carrier 2 or the glass 4 to be measured and stop at least at two positions, the image collector 3 comprises a lens 31 and a camera 32, and the camera 32 collects an image formed by light emitted by the stripe light source 1 after penetrating through the glass 4 to be measured and the lens 31.
The method comprises the following steps of detecting defects of the curved glass:
s1: the stripe light source 1 rotates to a first stop position (as shown in fig. 9) relative to the glass 4 to be measured to collect an image formed by the stripe light source with an interval of light and dark after transmitting through the glass to be measured. Because the image formed by the light and shade alternative stripe light source along the specific direction after penetrating through the glass to be measured is collected, the problem of highlight interference defect development caused by collecting reflected light can not be generated.
S2: dividing a part of the image corresponding to the curved surface area of the glass 4 to be detected into a plurality of connected first areas according to a gray value threshold segmentation method; as can be seen from fig. 2, the light and dark stripes appear distorted on the curved surface of the glass 4 to be measured. However, no matter what distortion occurs, the curved surface region can be partially divided into a plurality of first regions by a dynamic threshold segmentation method along a boundary (boundary point of light and dark stripes) where a gray value changes, and the first regions are formed by light stripes or dark stripes. The trend of the light stripe or the dark stripe can be a straight line, a broken line, a curve or even a ring formed by closed curves.
S3: determining a skeleton of each first region; in the field of machine vision, methods for determining the skeleton of a specific region using gray-scale values are known in the art. And will not be described in detail herein.
S4: dividing each first area into a plurality of second areas which are connected or intersected with each other along the framework, wherein all the second areas cover the first areas; the second region is a strip perpendicular to the frame, and both ends of the second region coincide with the boundaries of the first region, defining pixel rows (i.e., long edges of the strip) perpendicular to the frame.
If the second area is found to have drastic change of the gray value along the pixel column direction, the corresponding position of the glass to be measured corresponding to the position of the drastic change in the image has a defect. Specifically, it was found that the gradation value of the second region in the pixel column direction drastically changes by the following method:
s4.1: defining pixels in a direction perpendicular to the pixel columns to form pixel rows, the pixel rows including one or more pixels;
s4.2: averaging the gray values of the pixels of each pixel row to obtain a row gray value;
s4.3: deriving the gray value of the row pixel along the direction of the pixel column to obtain a derivative value of the row pixel;
s4.4: dividing a row pixel gray value curve into one or more sections of a positive derivative value section, a near zero section, a negative derivative value section and a near zero section which are connected with each other by a threshold segmentation method for row pixel derivative values along the direction of pixel columns; and if a negative derivative value appears in the positive derivative value section or a positive derivative value appears in the negative derivative value section or the absolute value of the difference between the gray value of a certain row of pixels and the average value of the gray value of the pixels of the row of the near-zero section appears in the near-zero section is larger than a specific multiple of the image noise value, determining that the gray value of the second area is changed violently along the pixel column direction.
We can further understand the above step S4.4 by reflection on the image and gray value curves when the defect is located at the first and second positions in the image.
Fig. 3 shows a part of the image with a defect in a first position in the image, and fig. 4 is an enlarged view of fig. 3. From fig. 4 we can see that the defect is located at the intersection of the light and dark fringes. Since the first regions are divided into a plurality of first regions by the threshold segmentation method, it is known that each first region is normally composed of a gray value rising segment (positive derivative value segment), a near zero segment (small gray value change) and a gray value falling segment (negative derivative value segment), and a near zero segment (small gray value change). As can be seen from fig. 5, the defect is located in the gray value falling segment (i.e., the negative derivative value segment). At this point, we can find a sharp change from the gray value image that shows that there is a positive derivative value in the negative derivative value segment. Of course, similarly, if a negative derivative value is found in the positive derivative value segment, a drastic change may be assumed to exist.
Fig. 6 shows a part of the image with the defect at a second position in the image, and fig. 7 is an enlarged view of fig. 6. From fig. 7 we can see that the defect is located at the position of the bright stripe. Since the first regions have been segmented by the threshold segmentation method, we can see from fig. 8 that the defect is located in a near-zero segment (the gray value is very small). We can find a sharp change from the gray value image that appears as a difference between a certain gray value and the average gray value of the segment that is greater than a certain multiple of the noise value of the image. And the image noise value is determined by the signal-to-noise ratio and the average value of the gray value. In this embodiment, the specific multiple may be 2 times.
As can be seen from the above, the curved surface area in the image is divided into a plurality of first areas extending along the stripes (according to the threshold segmentation method, the first areas in the image all extend along the boundary between the bright stripes and the dark stripes), then the skeleton of the first areas is found, and the first areas are divided into a plurality of second areas perpendicular to the skeleton along the skeleton, so that even if the curved surface area stripes of the curved surface glass are possibly distorted, the image defect finding problem can be still decomposed into the defect finding problem of a plurality of strip-shaped areas perpendicular to the stripe direction (i.e. the light and shade change direction). The defect position is determined by finding the drastic change of the gray value in the pixel column direction, so that the comparison with the defect-free glass is not needed any more, and the problem that the image defect cannot be detected by comparing the image with the defect-free image due to the limitation of the processing precision of the curved glass is solved.
S5: for the part corresponding to the plane area of the glass to be detected in the image, manually cutting the part in the image into a plurality of connected third areas along the direction vertical to the light and shade change direction, wherein the third areas are long-strip-shaped, the length direction of the third areas is defined to be parallel to the light and shade change direction, and the width direction of the third areas is defined to be vertical to the light and shade change direction and comprises one or more pixels; because the stripe does not distort at the part corresponding to the plane area, the third area can be directly selected across the stripe without dividing the first area.
S6: and averaging the gray values of all pixels in the width direction along the length direction to obtain the average gray value in the width direction.
S7: the width-direction average gray value curve from which the stripe texture is removed is obtained by performing fourier transform and inverse transform on the frequency information of the width-direction average gray value curve along the length direction.
S8: if the sharp change of the gray value exists in the width direction average gray value curve without the stripe texture, the corresponding position of the glass to be measured corresponding to the position in the image of the sharp change has a defect. Here, the drastic change can also be detected by the difference between a certain gray value and the average gray value of the neighboring segments being greater than a certain multiple of the noise value of the image.
S9: the stripe light source 1 rotates to a second stop position (as shown in fig. 10) relative to the glass 4 to be measured to collect an image formed by the stripe light source with an interval of light and dark after transmitting through the glass to be measured.
Then, S2 to S8 are repeated.
By rotating the stripe light source, some directional defects (such as scratches) on the glass can be displayed, which is the principle that the stripe area perpendicular to the stripe direction is easier to display the defects. Specifically, the stripe light source irradiates the glass to be detected in different brightness change directions, so that even if some directional glass defects (such as scratches) are not easy to be developed in one brightness change direction, the directional glass defects are developed after the brightness change direction is changed, and the defect omission ratio is reduced.
Of course, if no drastic changes are detected in all of the second and third regions in the image, the curved glass is free of defects.
The description of the above specification and examples is intended to be illustrative of the scope of the present invention and is not intended to be limiting.
Claims (4)
1. A curved glass defect detection method is characterized by comprising the following steps:
collecting an image formed by a light and shade alternative stripe light source after penetrating through glass to be detected;
at least dividing a part of the image corresponding to the curved surface area of the glass to be detected into a plurality of connected first areas according to a gray value threshold segmentation method;
determining a skeleton of each first region;
dividing each first area into a plurality of second areas which are connected or intersected with each other along the framework, wherein all the second areas cover the first areas; the second area is in a strip shape which is vertical to the framework, two ends of the second area coincide with the boundary of the first area, pixel rows of the second area are defined to be vertical to the framework, and if the second area is found to have drastic gray value change along the direction of the pixel rows, the corresponding position of the glass to be detected corresponding to the position of the drastic change in the image has defects.
2. The curved glass defect detection method according to claim 1, wherein the specific method for finding that the second region has a drastic change in gray level along the pixel column direction is as follows:
defining pixels in a direction perpendicular to the pixel columns to form pixel rows, the pixel rows including one or more pixels;
averaging the gray values of the pixels of each pixel row to obtain a row gray value;
deriving the gray value of the row pixel along the direction of the pixel column to obtain a derivative value of the row pixel;
dividing a row pixel gray value curve into one or more sections of a positive derivative value section, a near zero section, a negative derivative value section and a near zero section which are connected with each other by a threshold segmentation method for row pixel derivative values along the direction of pixel columns; and if a negative derivative value appears in the positive derivative value section or a positive derivative value appears in the negative derivative value section or a difference value between the gray value of a certain row of pixels and the average value of the gray values of the pixels of the row of the near-zero section appears in the near-zero section is larger than a specific multiple of the image noise value, determining that the gray value of the second area is changed violently along the direction of the pixel columns.
3. The method for detecting defects of curved glass according to claim 1, wherein the method comprises the following steps: further comprising:
for a part of the image corresponding to a planar area of the glass to be detected, dividing the part of the image into a plurality of connected third areas along a direction vertical to the light and shade change direction, wherein the third areas are strip-shaped, the length direction of the third areas is defined to be parallel to the light and shade change direction, and the width direction of the third areas is defined to be vertical to the light and shade change direction and comprises one or more pixels;
averaging the gray values of all pixels in the width direction along the length direction to obtain the average gray value in the width direction;
carrying out Fourier transformation and inverse transformation on the frequency information of the width direction average gray value curve along the length direction to obtain a width direction average gray value curve without stripe textures;
if the sharp change of the gray value exists in the width direction average gray value curve without the stripe texture, the corresponding position of the glass to be measured corresponding to the position in the image of the sharp change has a defect.
4. The curved glass defect detection method as claimed in claim 1, 2 or 3, further comprising rotating the streak light source relative to the glass to be detected and stopping at least at two positions, wherein the alternate light and dark changing directions of the streak light source intersect each other at each position;
the collecting of the image formed by the light and shade alternate stripe light source after penetrating through the glass to be detected specifically comprises collecting the image formed by the light emitted by the stripe light source after penetrating through the area to be detected of the glass to be detected when the stripe light source is positioned at each position; and each image acquired is subjected to subsequent analysis.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911167579.XA CN110956619B (en) | 2019-11-25 | 2019-11-25 | Curved glass defect detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911167579.XA CN110956619B (en) | 2019-11-25 | 2019-11-25 | Curved glass defect detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110956619A CN110956619A (en) | 2020-04-03 |
| CN110956619B true CN110956619B (en) | 2022-07-01 |
Family
ID=69976764
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911167579.XA Active CN110956619B (en) | 2019-11-25 | 2019-11-25 | Curved glass defect detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110956619B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112862800B (en) * | 2021-02-25 | 2023-01-24 | 歌尔科技有限公司 | Defect detection method and device and electronic equipment |
| CN113269731A (en) * | 2021-05-13 | 2021-08-17 | 苏州迪宏人工智能科技有限公司 | Defect detection method and system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011119303A1 (en) * | 2010-03-26 | 2011-09-29 | The Boeing Company | Detecting optical defects in transparencies |
| CN104101611A (en) * | 2014-06-06 | 2014-10-15 | 华南理工大学 | Mirror-like object surface optical imaging device and imaging method thereof |
| WO2015127950A1 (en) * | 2014-02-26 | 2015-09-03 | Heye International Gmbh | Method for detecting flaws in the walls of hollow glass items |
| CN106705897A (en) * | 2016-12-23 | 2017-05-24 | 电子科技大学 | Arc-shaped glass panel defect detecting method used for curved surface electronic display screen |
| CN108572181A (en) * | 2018-05-15 | 2018-09-25 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of mobile phone bend glass defect inspection method based on streak reflex |
-
2019
- 2019-11-25 CN CN201911167579.XA patent/CN110956619B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011119303A1 (en) * | 2010-03-26 | 2011-09-29 | The Boeing Company | Detecting optical defects in transparencies |
| WO2015127950A1 (en) * | 2014-02-26 | 2015-09-03 | Heye International Gmbh | Method for detecting flaws in the walls of hollow glass items |
| CN104101611A (en) * | 2014-06-06 | 2014-10-15 | 华南理工大学 | Mirror-like object surface optical imaging device and imaging method thereof |
| CN106705897A (en) * | 2016-12-23 | 2017-05-24 | 电子科技大学 | Arc-shaped glass panel defect detecting method used for curved surface electronic display screen |
| CN108572181A (en) * | 2018-05-15 | 2018-09-25 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of mobile phone bend glass defect inspection method based on streak reflex |
Non-Patent Citations (2)
| Title |
|---|
| Yang Y,et al..An Evaluation method of acceptable and failed spot welding products based on image classification with transfer learning technique.《Proceedings of the 2nd international conference on computer science and application engineering》.2018, * |
| 机器视觉技术在玻璃基板缺陷检测设备中的应用;李青 等;《玻璃与搪瓷》;20161231;第44卷(第3期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110956619A (en) | 2020-04-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP5443435B2 (en) | Tire defect detection method | |
| CN105784713B (en) | Sealing ring detection method of surface flaw based on machine vision | |
| US9171364B2 (en) | Wafer inspection using free-form care areas | |
| US8861831B2 (en) | Method for analyzing the quality of a glazing unit | |
| JP2018155690A (en) | Surface defect inspection method and surface defect inspection device | |
| BR112019016366A2 (en) | in-line apparatus for measuring small optical or obstructive defects in a selected area of a glass sheet and method for measuring small defects in a selected area of a glass sheet as a glass sheet is transported | |
| US9196034B2 (en) | Method of fast analysis of the relief elements featuring on the internal surface of a tyre | |
| CN110956619B (en) | Curved glass defect detection method | |
| CN104458764B (en) | Curved uneven surface defect identification method based on large-field-depth stripped image projection | |
| CN111046862B (en) | Character segmentation method, device and computer readable storage medium | |
| US9250067B2 (en) | System and method for evaluating the performance of a vehicle windshield/wiper combination | |
| KR20140116946A (en) | Segmentation for wafer inspection | |
| CN110849911B (en) | Glass defect image acquisition device, glass defect detection equipment and detection method | |
| CN110596116A (en) | Vehicle surface flaw detection method and system | |
| KR20220139292A (en) | Character segmentation method, apparatus and computer readable storage medium | |
| CN115330770B (en) | Cloth area type defect identification method | |
| Kaddah et al. | Automatic darkest filament detection (ADFD): a new algorithm for crack extraction on two-dimensional pavement images | |
| CN117252861A (en) | Method, device and system for detecting wafer surface defects | |
| JP3913517B2 (en) | Defect detection method | |
| CN112634252A (en) | Method for inspecting printed circuit | |
| CN117036259A (en) | Metal plate surface defect detection method based on deep learning | |
| CN113554688B (en) | O-shaped sealing ring size measurement method based on monocular vision | |
| CN116071304A (en) | Article edge line positioning method and device and electronic equipment | |
| CN110849912B (en) | Glass defect developing device and glass defect detection equipment | |
| KR100955736B1 (en) | Method for inspecting surfaces and apparatus for performing the method |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |