Apparatus and method for identifying defects within the volume of a transparent sheet and use of the apparatus
09546967 ยท 2017-01-17
Assignee
Inventors
- Bruno Schrader (Gruenenplan, DE)
- Frank Macherey (Wiesbaden, DE)
- Holger Wegener (Alfeld, DE)
- Michael Stelzl (Mainz, DE)
Cpc classification
G01N2021/8967
PHYSICS
G01N21/896
PHYSICS
G01N21/958
PHYSICS
International classification
G01N21/958
PHYSICS
G01N21/896
PHYSICS
Abstract
An apparatus for identifying defects within the volume of a transparent sheet, such as a glass sheet, is provided. The apparatus includes an illumination device that directs incident light onto at least a portion of a surface of the sheet so as to illuminate the sheet, and an image detector onto which the light backscattered from the sheet is directed to image the sheet. The apparatus generates at least two interference images under different capturing conditions in order to perform identification of defects by evaluating the at least two interference images.
Claims
1. An apparatus for identifying defects within the volume of a transparent sheet, comprising: an illumination device that directs incident light onto at least a portion of a surface of the sheet to illuminate the sheet; and an image detector onto which the incident light that is backscattered from the sheet is directed, wherein the apparatus is adapted to generate interference patterns by superposition of components of the light backscattered from the sheet and, based on interference patterns, to generate at least two interference images under different capturing conditions in order to identify the defects by evaluating the at least two interference images.
2. The apparatus as claimed in claim 1, wherein the image detector generates the at least two interference images in order to identify the defects by comparing the at least two interference images.
3. The apparatus as claimed in claim 1, wherein the image detector generates the at least two interference images in order to identify the defects by determining differences between the at least two interference images.
4. The apparatus as claimed in claim 1, wherein the image detector generates the at least two interference images sequentially in time.
5. The apparatus as claimed in claim 1, wherein the image detector generates the at least two interference images at the same time.
6. The apparatus as claimed in claim 1, wherein the different capturing conditions are selected from the group consisting of incident light of at least two different wavelengths, incident light of least two different waveforms, incident light directed onto the sheet at least two different angles of illumination, capture of the backscattered light at least two different detection angles, incident light of at least two different phases, and any combination thereof.
7. The apparatus as claimed in claim 1, wherein the illumination device comprises at least one light source in form of a sodium vapor lamp or a laser.
8. The apparatus as claimed in claim 7, wherein the at least one light source has a coherence length greater than twice a thickness of the sheet.
9. The apparatus as claimed in claim 7, wherein the at least one light source has a coherence length greater than 3 mm.
10. The apparatus as claimed in claim 1, wherein the image detector comprises a screen onto which the backscattered light from the sheet is directed for displaying the interference images.
11. The apparatus as claimed in claim 1, wherein the image detector comprises at least one image sensor for capturing the backscattered light from the sheet.
12. The apparatus as claimed in claim 11, wherein the at least one image sensor is selected from the group consisting of a matrix camera, a line scan camera, a line scan camera with Time Delayed Integration (TDI) sensor, which is operated as a matrix camera, and combinations thereof.
13. The apparatus as claimed in claim 11, wherein the image detector further comprises a screen onto which the interference images captured by the image sensor are displayed.
14. The apparatus as claimed in claim 1, further comprising a computer connected to the image detector and/or to the illumination device.
15. The apparatus as claimed in claim 1, further comprising a conveyor for moving the sheet relative to the illumination device and/or the image detector.
16. The apparatus as claimed in claim 15, wherein the conveyor comprises a roller over which the sheet is guided.
17. The apparatus as claimed in claim 1, further comprising a pattern recognition system for evaluating the at least two interference images.
18. The apparatus as claimed in claim 1, wherein the illumination device comprises a device selected from the group consisting of a tunable light source for generating the incident light with a plurality of different wavelengths sequentially in time, a light source having a defined spectral width and a tunable filter for generating the incident light with a plurality of different wavelengths sequentially in time, a plurality of spaced apart light sources for simultaneously generating the incident light of a plurality of different wavelengths, and combinations thereof.
19. A method for identifying defects within the volume of a transparent sheet, comprising: generating and directing incident light onto at least a portion of a surface of the sheet; directing light backscattered from the sheet onto an image detector, so that components of the light backscattered from the sheet are superposed to generate interference patterns; capturing light backscattered from the sheet with the image detector; generating by means of the interference patterns at least two interference images at the image detector under different capturing conditions; and identifying the defects by evaluating the at least two interference images.
20. The method as claimed in claim 19, wherein identifying the defects comprises comparing the at least two interference images with one another.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
DETAILED DESCRIPTION
(11)
(12) Apparatus 11 is adapted to generate at least two interference images at different capturing conditions. By evaluating the at least two interference images the apparatus 11 enables to identify the defects.
(13) The defects are identified by comparing the interference images. According to the present embodiment, the comparison comprises determining the differences between the interference images. The evaluation of the interference images is performed by an operator.
(14) The different capturing conditions are provided by generating each interference image using light of a wavelength that is different from the wavelength of the light used for generating a further interference image. In the present case, the interference images are generated simultaneously.
(15) According to
(16) Beam splitter 26 exhibits selective light transmission in that the incident light 20 is reflected towards sheet 10, and the light 22 coming from sheet 10 is transmitted through the beam splitter 26.
(17) The interference pattern is captured by a sensor line 16 having the same width as the test object. The several interference patterns are created by illumination with wavelengths .sub.ii=1, 2, . . . which are observed separated in time, or spectrally split.
(18) The defects are identified by a comparison of the interference images, and apparatus 11 is configured to perform such a comparison. The comparison of the interference images comprises to determine differences between the interference images.
(19) One interference image of a series of a plurality of interference images generated with different wavelengths can be used to determine an expectation interference pattern which is used as a reference for the other interference patterns. The reference pattern or expectation value is compared with the interference patterns of each of the other images by subtracting the expectation value from the respective interference pattern of the other images.
(20) Apparatus 11 comprises a conveyor means 24 which moves the sheet 10 relative to the illumination means 12 and image detection means 16.
(21)
(22) A surface area of the test object is illuminated by a coherent light source 12. The interference pattern is observed on a screen 14 and a camera/lens system 16. The image rate of the camera 16 is selected so that each point is detected a plurality of times, under different illumination angles, i.e. with different interference patterns.
(23) A point of sheet 10, which is at a position x.sub.0 at a time t (
(24) Sheet 10 is moved by conveyor means 24 in a forward direction. Consequently, at a later time t+t said point of sheet 10 will be located at a position x.sub.0+x and will scatter the incident light 20 onto screen 14 at a detection angle or angle of reflection of +. In this way, said point can be imaged on screen 14 at a plurality of illumination angles , +, . . . , and for each angle of illumination an interference image will be generated.
(25) Apparatus 11 comprises a computing unit 18 connected to image detection means 14, 16 and to illumination means 12. The evaluation of the interference images including the detection of a disturbance and identification of a defect is performed by means of computing unit 18.
(26) Advantageously, by using screen 14 onto which the backscattered light 22 is directed for visualizing an image, the image sensor 16 does not need to directly capture the backscattered light 22. Thus, a smaller and cheaper image sensor 16 can be used.
(27) An interference image is created due to the fact that incident light 20 is reflected from the upper and lower surfaces of sheet 10. The reflected light 22 therefore has two components 22a, 22b which superimpose each other. The superposition of reflected light components 22a, 22b results in an interference image at the location of image sensor 16 or screen 14. The interference pattern is generated by a variation in thickness D(x,y) of sheet 10.
I.sub.s,p(,l,D)=R.sub.s,pI.sub.0(1(1R.sub.s,p)e.sup.i2/.sup.
(28)
(29) An observation in reflection in a reflected-light setup requires a defined position of the test object 10 (distance to image sensor 16). In case of the glass application, local warpage of the glass ribbon 10 will occur due to internal stresses, which makes a defined position even more difficult.
(30) In order to solve this problem for observing reflection as in the interference inspection described, the glass sheet in form of a glass ribbon 10 is guided over roller 25, so that due to bending an external tension is imposed on the glass, which compensates for the internal stresses. In this manner, the position of is unambiguously defined.
(31)
(32) Each of
(33) The interference effect caused by the defect is greater by many times (about 10 to 14 times greater) than the defect itself.
(34) The defect is detected by comparing the interference images, the comparison comprising to determine differences between the interference images. In the present exemplary embodiment, interference images are used which have been produced at the same time.
(35)
(36) 1) metallic inclusion having a core size of about 0.150 mm (
(37) 2) gas bubble having a size of 0.06 mm (
(38) Both samples 10 were inspected by an apparatus 11 according to
(39) A comparison of
(40) In case the defect is a glass bubble (
(41) In the manufacturing of thin glass (thickness from about 0.02 to 1 mm), high feeding rates are partly necessary because of the process, partly economically desirable. At the same time, high glass defect sensitivities are required, especially for very thin glasses. The invention permits to achieve high defect sensitivity at high feeding rates under manufacturing conditions (tolerance to variations in height, glass deflection) with reasonable effort.
(42) For image analysis, a plurality of images ii>2 are captured for each position. Interference patterns caused by defects and impurities are small compared to the image field. The interference pattern of the undisturbed sheet material is created by a gradual change of the sheet thickness, in contrast to the interference pattern caused by defects. Therefore, an optical thickness can be determined from each image by smoothing, which is identical for all images ii.
(43) Therefore, by forward calculation taking into account the changed capturing situation, an expectation interference pattern can be determined for the respective other interference images, which is subtracted from the actual image for all images ii. A defect will then be distinguished by the fact that it deviates from the expectation value in different images. This deviation between all images is added up, so that small defects can be identified and larger defects can be identified in more detail. Dirt and stain can be discriminated due to their consistent behavior in all images.
(44) Defects in the glass will cause a local disturbance in the glass that depends on the defect type (bubble, inclusion, etc.) and the defect size relative to the glass thickness. The disturbance locally changes the optical wavelength of the light beam that passes through this region as compared to a light beam which passes through an undisturbed region. In an arrangement as shown in
(45) When using a light source 12 with a sufficient coherence length, the light beams 22a, 22b reflected at the upper and lower surfaces will interfere. Due to the high homogeneity of the glass (in terms of variations in thickness and variations in refractive index), the path difference of the interfering partial beams changes only slowly when compared to a region in the vicinity of a defect. Here, the striped pattern will exhibit a disturbance which can be observed by an image sensor 16 focused on screen 14.
(46) The invention has a variety of advantages.
(47) The local disturbance is greater by many times than the actual defect, usually by 10 to 12 times. Thus, a lower optical resolution is necessary, with a positive impact on the cost and the achievable depth of field. However, the resolution should be sufficient to resolve a light/dark/light or complementary pattern.
(48) The disturbance is caused by glass defects. Dust/dirt on the surface do not produce any interference effect, so that glass defects can be distinguished from dust.
(49) The method employs reflected light instead of transmitted light, so that no interruption of the conveyor is necessary in case of continuous material, which in turn simplifies the configuration.
(50) By guiding a flexible glass ribbon 10 over a roller 25, the position relative to illumination means 12 and camera 16 is fixed in a simple manner. Moreover, the tension imposed by the bending compensates for warpage caused by local stresses in the glass ribbon.
(51) The method is insensitive to variations in height. The angle tolerance required can be accommodated by beam expansion or by employing area scan cameras. Apparatus 11 initially uses only one channel, and in conjunction with the lower resolution it is suitable for high feeding rates.
(52) By using a further wavelength which inverts the pattern, a further information is resulting for each glass defect, so that smaller defects would be found or a lower resolution is needed for the same defect sensitivity.
(53) In summary, the apparatus 11 of the invention provides for a simple and cost efficient configuration with a very high performance in terms of detection of defects, in particular small defects.
REFERENCE NUMERALS
(54) 10 Transparent sheet, glass ribbon, test object, sample 11 Apparatus for identifying a defect within the volume of a transparent sheet 12 Illumination means, light source 14 Screen 16 Image sensor, sensor line, camera 18 Computing unit 20 Incident light 22 Light backscattered or reflected from the sheet 22a Light reflected from the upper surface of the sheet 22b Light reflected from the lower surface of the sheet 24 Conveyor means 25 Roller 26 Beam splitter