[0001] This application claims the benefit of priority under 35 U.S.C. § 119 of U.S. Provisional

Application Serial No. 62/384, 842, filed on September 8, 2016, the content of which is relied upon and incorporated herein by reference in its entirety.

FIELD

[0002] The present disclosure relates to optical inspection, and in particular to optical inspection systems and methods for detecting surface defects in a transparent sheet.

BACKGROUND

[0003] Transparent sheets are used in a variety of different device applications, including as covers for displays and panels such as LCD (liquid-crystal display) panels. Such devices and panels are being made increasing thinner and lightweight, requiring the transparent sheet that serves as the cover to be increasingly thinner and lightweight as well.

[0004] The typical transparent sheet for a display or panel is made using a glass substrate. An original or starting glass substrate can be thinned using a chemical process, e.g., etching and polishing, to achieve a desired thickness (e.g., 0.1 mm to 0.7 mm) for the final glass substrate. During the etching process, defects in form of dips (indentations) and bumps (protrusions) (also respectively referred to as "dimples" and "pimples") can be formed in or on the glass surfaces. The typical lateral extent of the defects can range from 10 microns to a few millimeters, and their typical vertical dimension (i.e., depth or height) can be large as a quarter of a micron. Such defects can be visible in LCD-based devices so that chemically thinned glass substrates needs to be inspected for surface defects so that they can be removed.

[0005] Presently, manual methods are used to inspect transparent sheets for defects. Unfortunately, such manual inspection is labor-intensive, inconsistent and extremely time consuming, e.g., it takes many hours to inspect a large production-sized sheet. There is also some difficulty in determining on which side of the substrate the defects reside and in distinguishing between the nature of the defect, i.e., whether they are dips or bumps. There is also some difficulty in distinguishing between surface defects and particles that reside on the surfaces of the transparent sheet.

SUMMARY

[0006] An aspect of the disclosure is an optical inspection system for inspecting a transparent sheet having upper and lower surfaces. The system includes: a defect camera system operably arranged relative to the transparent sheet and configured to capture either a dark-field or a bright-field defect- camera image of a region of the transparent sheet that includes at least one defect in the upper or lower surface and at least one particle on the upper or lower surface, wherein the defect-camera image includes image features defined by the at least one particle and the at least one defect; a particle camera system operably arranged relative to the transparent sheet and configured to capture a particle- camera image of the region, wherein the particle camera image includes an image of the at least one particle and does not include an image of the at least one defect; and a controller operably coupled to the defect camera system and the particle camera system, the controller being configured to receive and compare the defect-camera image to the particle-camera image to determine which of the image features of the defect-camera image are defined by the at least one defect and which are defined by the at least one particle.

[0007] Another aspect of the disclosure is the optical inspection system described above, wherein the defect camera system includes: first and second defect cameras having respective first and second camera axes with different angles relative to the upper and lower surfaces of the transparent sheet, with the first and second defect cameras being operably arranged relative to first and second collimated light sources, respectively; the first and second defect cameras having respective first and second image sensors that respectively capture first and second defect-camera images of the region and that are respectively electrically coupled to the controller; and wherein the controller is configured to process the first and second defect-camera images to determine whether the at least one defect is on the upper surface or the lower surface of the transparent sheet.

[0008] Another aspect of the disclosure is the optical inspection system described above, wherein the first and second defect cameras are arranged to capture the first and second defect-camera images of the region simultaneously. [0009] Another aspect of the disclosure is the optical inspection system described above, wherein the defect camera system and the particle camera system are movable relative to the transparent sheet and are arranged to sequentially view the region of the transparent sheet.

[0010] Another aspect of the disclosure is the optical inspection system described above, wherein the defect camera system and the particle camera system each use visible light.

[0011] Another aspect of the disclosure is the optical inspection system described above, and further including a schlieren camera system configured to capture a schlieren image of the region R, wherein the schlieren image has an image polarity that defines whether the at least one defect is a dimple or a pimple.

[0012] Another aspect of the disclosure is the optical inspection system described above, wherein the transparent sheet has a thickness in the range from 0.2 mm to 1 mm and the at least one defect has a dimension in the range from 0.01 mm to 1 mm.

[0013] Another aspect of the disclosure is the optical inspection system described above, wherein the defect camera system is configured to operate in a reflection mode.

[0014] Another aspect of the disclosure is the optical inspection system described above, wherein the defect camera system has a dark-field off-axis configuration.

[0015] Another aspect of the disclosure is a method of optically inspecting a transparent sheet that includes opposing upper and lower surfaces and a region that has at least one defect on at least one of the upper or lower surface and at least one particle on at least one of the upper or lower surface. The method includes: capturing at least one defect-camera image of the region, wherein the defect -camera image comprises either a dark-field image or a bright-field image and comprises image features that include a first particle image of at least one particle residing within the region and a first defect image of the at least one defect residing within the region, and wherein the first particle image and the first defect image of the dark-field image cannot be distinguished from one another; capturing a particle- camera image of the region, the particle-camera image including a second particle image of the at least one particle residing within the region and does not include a second defect image of the at least one defect residing within the region; and comparing the at least one defect-camera image and particle- camera image to distinguish which of the features in the defect-camera image is the first defect image and which is the first particle image. [0016] Another aspect of the disclosure is the method described above, wherein capturing the at least one defect-camera image includes capturing first and second dark-field defect-camera images of the region at different angles relative to the upper surface of the transparent sheet and measuring a shift between first and second defect images in the first and second dark-field defect-camera images to determine whether the at least one defect is on the upper surface or the lower surface of the transparent sheet.

[0017] Another aspect of the disclosure is the method described above, wherein capturing at least one defect-camera image includes capturing first and second dark-field defect-camera images of the region at different angles relative to the upper surface of the transparent sheet and performing triangulation to determine whether the at least one defect is on the upper surface or the lower surface of the transparent sheet.

[0018] Another aspect of the disclosure is the method described above, and further including capturing a schlieren image of the at least one defect to determine whether the at least one defect is a dimple or a pimple.

[0019] Another aspect of the disclosure is the method described above, and further including capturing the at least one defect-camera image and the particle-camera image of the region simultaneously.

[0020] Another aspect of the disclosure is the method described above, and further comprising capturing the at least one defect-camera image and the particle-camera image of the region sequentially.

[0021] Another aspect of the disclosure is the method described above, wherein the act of comparing is performed in a controller based on instructions embodied in a non-transitory computer-readable medium.

[0022] Another aspect of the disclosure is the method described above, wherein the at least one defect has a dimension in the range from 0.01 mm to 1 mm.

[0023] Another aspect of the disclosure is the method described above, wherein the transparent sheet is made of glass and has a thickness in the range from 0.2 mm to 1 mm. [0024] Another aspect of the disclosure is a method of optically inspecting a transparent sheet that includes opposite upper and lower surfaces and a region that has a defect on the upper or lower surface and a particle on the upper or lower surface. The method includes: capturing at least one dark-field image of the region, wherein the at least one dark-field image includes a first particle image of the particle and a first defect image of the defect, wherein the first particle image and the first defect image of the at least one dark-field image are not distinguishable from one another; capturing a conventional image of the region, wherein the conventional image includes only a second particle image and does not include a second defect image of the defect; and using the conventional image to distinguish between the first defect image and the first particle image of the at least one dark-field image.

[0025] Another aspect of the disclosure is the method described above, wherein the act of capturing the at least one dark-field image includes capturing first and second dark-field images of the region that respectively include first and second defect images, and further comprising using the first and second dark-field images to determine whether the defect resides on the upper surface or the lower surface of the transparent sheet by measuring a shift between the first and second defect images.

[0026] Another aspect of the disclosure is the method described above, wherein the act of using the first and second dark-field images to determine whether the defect resides on the upper surface or the lower surface of the transparent sheet includes performing a triangulation.

[0027] Another aspect of the disclosure is the method described above, and further including capturing a schlieren image of the defect to ascertain whether the defect is a dimple or a pimple.

[0028] Additional features and advantages are set forth in the Detailed Description that follows, and in part will be readily apparent to those skilled in the art from the description or recognized by practicing the embodiments as described in the written description and claims hereof, as well as the appended drawings. It is to be understood that both the foregoing general description and the following Detailed Description are merely exemplary, and are intended to provide an overview or framework to understand the nature and character of the claims.

BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The accompanying drawings are included to provide a further understanding, and are incorporated in and constitute a part of this specification. The drawings illustrate one or more embodiment(s), and together with the Detailed Description serve to explain principles and operation of the various embodiments. As such, the disclosure will become more fully understood from the following Detailed Description, taken in conjunction with the accompanying Figures, in which:

[0030] FIG. 1A is an elevated view of an example transparent sheet of the type that is inspected using the optical inspection systems and methods described herein;

[0031] FIG. IB is a close-up x-z cross-sectional view of a region of the transparent sheet of FIG. 1A, illustrating an example defect in the form of a dimple in the upper surface of the transparent sheet, and also showing a particle residing on the upper surface;

[0032] FIG. 1C is similar to FIG. IB and illustrates an example of a pimple defect in the upper surface of the transparent sheet;

[0033] FIGS. ID and IE are top-down close-up views of the transparent sheet illustrating different sizes and shapes of the defects that can be formed in the upper and lower the surfaces of the transparent sheet;

[0034] FIG. 2 is a schematic diagram of an example optical inspection system according to the disclosure and that includes a defect camera system and a particle camera system, and showing a transparent sheet operably disposed within the optical inspection system;

[0035] FIG. 3A is a schematic side view of an example configuration of a collimated light source that constitutes part of the defect camera system of the optical inspection system of FIG. 2;

[0036] FIG. 3B is a schematic side view of an example non-collimated light source that constitutes part of the particle camera system of FIG. 2, wherein the non-collimated light source is in the form of a planar backlight that emits non-collimated light over a wide range of angles;

[0037] FIG. 4 is a schematic side view of an example particle camera system;

[0038] FIG. 5A is a schematic side view of an example defect camera system that has a dark-field imaging configuration and that operates in a dark-field mode;

[0039] FIG. 5B is a schematic side view of an example defect camera system that has a bright-field imaging configuration and that operates in a dark-field mode; [0040] FIG. 5C is a schematic side view of an example defect camera system that has a reflection imaging configuration that operates in reflection mode;

[0041] FIG. 5D is a schematic side view of an example defect camera system similar to FIG. 5B and that shows an alternative off-axis dark-field configuration;

[0042] FIG. 6A is a representation of a dark-field defect-camera image as can be obtained by a dark- field defect camera system, wherein the dark-field defect-camera image shows image features that have yet to be identified as defects or particles;

[0043] FIG. 6B is a representation of a particle-camera image of the same region of the transparent sheet as shown in FIG. 6A, and that shows that some of the image features in the dark-field defect- camera image are particles;

[0044] FIG. 6C represents the result of an image-processing step that compares the representative particle-camera image of FIG. 6B to the representative dark-field defect-camera image of FIG. 6A, thereby allowing the features in the defect-camera image to be identified as surface defects in the form of either dimples or pimples;

[0045] FIG. 7A is a schematic diagram of an example defect camera system that includes two defect cameras and two collimated light sources configured relative to the transparent sheet in a manner that allows for determining whether a given defect resides on the upper surface or the lower surface of the transparent sheet;

[0046] FIG. 7B is similar to FIG. 7A and illustrates an embodiment of the defect camera system wherein the defect cameras and corresponding collimated light sources are configured to

simultaneously view and image the same region of the transparent sheet;

[0047] FIG. 7C is a close-up view of an example defect on the lower surface of the transparent substrate and shows the geometry of the projections of the defect onto the upper surface for the case where the camera axes cross at upper surface and the camera axis angles are equal, illustrating how the defect images will be spaced apart when viewed with the two defect cameras and then overlaid;

[0048] FIG. 8A is a view of overlapping representative defect-camera images from the two defect cameras of the defect camera system of FIG. 7A or 7B, illustrating an example of how two ring-type defect images can substantially overlap to indicate that the defect resides on the upper surface of the transparent sheet;

[0049] FIG. 8B is similar to FIG. 8A and illustrates an example of how the two ring-type defect images can be substantially spaced apart to indicate that the defect resides on the lower surface of the transparent sheet;

[0050] FIG. 9A is similar to FIG. 2 and illustrates an example configuration of the optical inspection system that includes a schlieren camera system that can be used to discern between dips and bumps (i.e., dimples and pimples);

[0051] FIG. 9B is a schematic side view of an example schlieren camera system as used in the optical inspection system of FIG. 9A;

[0052] FIG. 10A is a representative defect image and FIG. 10B is an actual defect image, illustrating an example orientation of the dark and bright regions of the defect image that indicate that the defect image is that for a dimple; and

[0053] FIG. 11A is a representative defect image and FIG. 11B is an actual defect image, illustrating an example orientation of the dark and bright regions of the defect image that indicate that the defect image is that for a pimple.

DETAILED DESCRIPTION

[0054] Reference is now made in detail to various embodiments of the disclosure, examples of which are illustrated in the accompanying drawings. Whenever possible, the same or like reference numbers and symbols are used throughout the drawings to refer to the same or like parts. The drawings are not necessarily to scale, and one skilled in the art will recognize where the drawings have been simplified to illustrate the key aspects of the disclosure.

[0055] The claims as set forth below a re incorporated into and constitute part of this Detailed Description.

[0056] Cartesian coordinates are shown in some of the Figures for the sake of reference and are not intended to be limiting as to direction or orientation. [0057] The terms "representative image" and "representation of an image" and like terms refer to images that are not actual acquired images and are used to depict in black and white the pertinent features of an actual image for ease of illustration and discussion.

[0058] The term "conventional imaging" as used in connection with the particle camera system and the particle camera means that the imaging process is that of a standard digital camera and does not include modifications that make the imaging process bright-field, dark-field, schlieren, etc.

[0059] The term "defect-camera image" is an image captured by the defect camera of the defect camera system.

[0060] The term "defect image" is a portion of the defect-camera image that includes an image of the defect, i.e., a feature in the detect-camera image that corresponds to a defect in the transparent sheet.

[0061] Transparent substrate and surface defects

[0062] FIG. 1A is an elevated view of an example transparent substrate 20 of the type that is inspected for surface features using the optical inspection systems and methods described herein. The transparent substrate 20 has a body 21 with a planar upper surface 22 and a planar lower surface 24, with the upper and lower surfaces being substantially parallel and defining a thickness TH. Thus, the transparent substrate 20 is in the form of a thin sheet, and so is referred to hereinafter as "transparent sheet."

[0063] The transparent sheet 20 also includes one or more edges 26, with four such edges shown for the example rectangular sheet. Also shown in FIG. 1A is a region of transparent sheet 20. In an example, optical inspections of transparent sheet 20 are made over multiple regions R, as discussed below.

[0064] In an example, transparent sheet is made of glass. In other example, transparent sheet is made of a material other than glass (e.g., plastic, acrylic, etc.). Exemplary glasses can be chemically strengthened glass or non-chemically strengthened glass. The glasses can be alkali containing or alkali- free glasses. In an example, transparent sheet 20 is made starting with a glass substrate that is subjected to chemical-based thinning process. In an example, the thickness TH is substantially constant and in the range from 0.1 mm to 5 mm, while in some cases the thickness TH is in the range from 0.2 mm to 1 mm. [0065] FIGS. IB and 1C are a close-up cross-sectional views in the x-z plane of an example region of transparent sheet 20, illustrating two different types of surface defects 30, namely an indentation or dimple 30D (FIG. IB) and a protrusion or pimple 30P (FIG. 1C). In an example, dimple 30D has a width dimension wl and a depth dimension dl while pimple 30P has a width dimension w2 and a height dimension h2. FIG. IB also shows an example of a particle 32 that resides on upper surface 22. As discussed below, an aspect of the disclosure includes distinguishing between particles 32 and surface defects 30, as well as determining whether a given defect resides on upper surface 22 or lower surface 24. It is noted that particles 32 can also reside on lower surface 22 (e.g., due to electrostatic effects) even though this surface faces downward.

[0066] FIGS. ID and IE are top-down views of a portion of transparent sheet 20 showing example surface defects 30 on upper surface 22, along with an example 1 mm reference scale. FIG. ID shows an example of an isolated round surface defect 30 (lower left corner), an isolated elongate surface defect (right side), and a cluster of surface defects (upper left corner). FIG. IE shows an example of a linear surface defect 30 (top portion) and an irregular shaped surface defect (lower portion). The surface defects 30 shown in FIGS. ID and IE can be dimples 30D, pimples 30P or a combination thereof.

[0067] In an example, the width dimensions wl and w2 and the depth and height dimensions dl and h2 are in the range from about 10 microns (0.01 mm) to about 1 mm. In an example, the width dimensions wl and w2 and the depth and height dimensions dl and h2 are maximum dimensions of the given defect. The 1 mm reference scale in FIGS. ID and IE is shown by way of non-limiting example.

[0068] Optical inspection system

[0069] FIG. 2 is a schematic side view of an example optical inspection system ("system") 100 for detecting surface defects 30 on transparent sheet 20. The transparent sheet 20 is operably supported within system 100 by a support device 106, which in an example is movable (e.g., a movable support stage). The system 100 includes an illumination system 110 arranged adjacent lower surface 24 and a detection system 210 arranged adjacent upper surface 22. System 100 also includes a controller 104 operably coupled to detection system 210 and optionally coupled to illumination system 110 and optionally coupled to substrate support device 106. The controller 104 can be a micro-controller, a computer, etc. that can perform image processing based on instructions (e.g., software, firmware, etc.) embodied in a non-transitory computer-readable medium. Thus, in an example, controller 104 is programmed to perform functions described herein to carry out the methods disclosed herein. As used herein, the term "controller" is not limited to just those integrated circuits referred to in the art as computers, but broadly refers to computers, processors, microcontrollers, microcomputers, programmable logic controllers, application specific integrated circuits, and other programmable circuits.

[0070] The illumination system 110 includes a collimated light source 120 that emits collimated light 122 along a first axis Al, and a non-collimated light source 130 that emits non-collimated light 132 generally along a second axis A2, wherein axes Al and A2 run generally in the z-direction. In an example, collimated light source 120 and non-collimated light source 130 are operably supported by a first movable support member 140, e.g., a movable stage.

[0071] FIG. 3A is a side view of an example collimated light source 120, which includes one or more light emitters 121, such as one or more light-emitting diodes (LEDs). In an example, the one or more light emitters 121 emit non-collimated light 122' at a visible wavelength, with an exemplary wavelength being in the blue range, e.g., 450 to 495 nm. Collimated light source 120 may also include a collimating optical system 125 that receives divergent light 122' emitted by the one or more light emitters 121 and forms collimated light 122. The collimated light source 120 may also include an aperture 126 disposed between the one or more light-emitters 121 and collimating optical system 135 to improve the degree of collimation of collimated light 122.

[0072] FIG. 3B is a side view of an example non-collimated light source 130 in the form of a planar backlight 131 that emits non-collimated light 132 over a wide range of angles. An example of such a backlight is an LED-based floodlight. I n an example, the non-collimated light 132 is visible light.

[0073] With reference again to FIG. 2, system 100 also includes a particle camera 330 arranged along second axis A2 and adjacent upper surface 22 and includes a defect camera 430 arranged along first axis Al and adjacent upper surface 22. The collimated light source 120 and defect camera 430 define a defect camera system 431, while the non-collimated light source 130 and the particle camera 330 define a particle camera system 331. The defect camera system 431 is configured to detect both defects 30 and particles 32, while the particle camera system 331 is configured to detect particles 32 only. In an example, defect camera system 431 operates using conventional imaging, such as used in standard digital cameras. [0074] In an example, defect camera 430 and particle camera 330 are operably supported by a second movable support member 240, e.g., a movable stage or movable gantry. In an example, first and second movable support members 140 and 240 are different parts of a single integrated movable support member or assembly. In an example, first and second movable support members 140 and 240 move in synchrony so that the defect camera 430 and collimated light source 120 maintain their relative and operable alignment along axis Al while particle camera 330 and non-collimated light source 130 maintain their relative and operable alignment along axis A2.

[0075] In an example, the first and second movable support members 140 and 240 are configured so that collimated light 122 passes through a first region Rl of transparent sheet 20 and is received by defect camera 430, while non-collimated light 232 passes through a second region R2 of the transparent sheet and is received by particle camera 330. The first and second regions Rl and R2 are spaced apart and include respective portions of the upper and lower surfaces 22 and 24 in which defects 30 can reside. In an example, defect camera 430 and particle 330 have substantially the same magnification and the first and second regions Rl and R2 have substantially the same dimensions.

[0076] The movement of the first and second movable stages 140 and 240 can be coordinated (e.g., by controller 104) to cover the entire (or substantially the entire) transparent sheet 20. This is accomplished in one example by having defect camera 430 and particle camera 330 capture respective defect-camera images and particle-camera images of all the regions R of the transparent sheet but at different times during the scanning process. Thus, in an example, the regions R of transparent sheet are imaged sequentially by defect camera 430 and particle camera 330.

[0077] The images of the same region R as captured by defect camera 430 and particle camera 330 are then aligned and compared, as discussed below. In addition, substrate support device 106 can also be movable to assist in scanning transparent sheet 20 relative to defect camera system 431 and particle camera system 331. In another example, substrate support device 106 is movable while the first and second support members 140 and 240 remain stationary.

[0078] Particle camera system

[0079] FIG. 4 is a schematic diagram of an example particle camera system 331 that includes a single particle camera 330. In an example, particle camera 330 is in the form of a digital camera that includes along axis A2 an imaging optical system 334 and an image sensor 340. The image sensor 340 is configured to capture an image ("particle-camera image") of second region 2 of transparent sheet 20 as formed by imaging optical system 334, which may comprise one or more lens elements. The image sensor 340 is defined by an array of pixels 342, as shown in the close-up inset. In an example, particle camera 330 is capable of resolving particles 32 that are 5 microns in size, i.e., the resolution is 5 microns per pixel 142. Other resolutions can also be used, depending on the minimum size of the particles 32 that need to be detected. Image sensor 340 receives the particle-camera image and forms an electronic particle-image signal SP1 that defines the electronic or digital particle-camera image.

[0080] In an example, image sensor 340 is electrically coupled to a digital image processor 350 that receives from the image sensor the electronic particle-image signal SP1 and performs one or more image-processing steps (e.g., filtering, formatting, etc.) to generate a processed electronic particle- image signal SP2, which is sent to controller 104.

[0081] The particle camera 330 can also include a microcontroller 356 electrically coupled to image sensor 340, digital image processor 350 and to a lens driver 360 operably arranged relative to imaging optical system 334. The lens driver 360 can be used to control the axial movement of imaging optical system 334, e.g., to adjust focus.

[0082] Defect camera system

[0083] FIG. 5A is a schematic diagram of an example defect camera system 431 that has a dark-field configuration and thus operates in a dark-field mode. The defect camera 430 of defect camera system 431 has a camera axis AC along which resides a collection lens 432, a dark-field aperture 434 arranged at an aperture plane AP, and an image sensor 440 arranged at a dark-field image plane DIP. In an example, image sensor 440 comprises an array of pixels 442, as shown in the close-up inset. In the example of FIG. 5A, the camera axis AC is co-axial with the first axis Al. The dark-field aperture 434 includes a central obscuration 436 surrounded by an optically transmissive portion 438, as shown in the close-up inset.

[0084] In an example, image sensor 440 of defect camera 430 is electrically coupled to a digital image processor 450 that receives from the image sensor the electronic defect-image signal SD1 and performs one or more image-processing steps (e.g., filtering, formatting, etc.) and generates a processed electronic defect-image signal SD2 that is sent to controller 104. In an example, digital image processor 450 is electrically coupled to a microcontroller 456. [0085] In the operation of defect camera 430, collimated light 122 from collimated light source 120 travels in the direction of axis Al through both the lower and upper surfaces 24 and 22 of transparent sheet 20, as well as through body 21. Some of collimated light 222 interacts with defect 30, thereby forming deflected light 122D. The collection lens 432 receives (collects) both collimated light 122 and deflected light 122D. The collection lens 432 directs the collimated light 122 to a focus position FP on the camera axis AC at aperture plane AP. The focused collimated light 122 is thus blocked by central obscuration 436 of dark-field aperture 434 and thus does not reach image sensor 440.

[0086] The deflected light 122D received by collection lens 432 is not focused at focus position FP by collection lens 432 since its trajectory has been altered by defect 30. The deflected light 122D passes around central obscuration 436 of dark-field aperture 434 and is transmitted through the transmissive portion 438. This light is then incident upon image sensor 440, where it forms a dark-field image ("defect-camera image") of defect 30. The defect-camera image is then converted to a digital defect- camera image embodied in electronic defect-image signals SDl and SD2. The defect 30 shown in FIG. 5A is a dimple defect 30D, but a similar result obtains in this dark-field embodiment for a pimple defect 30P.

[0087] FIG. 5B is similar to FIG. 5A and illustrates an example embodiment wherein defect camera 410 has a bright-field configuration. The bright-field configuration is defined by a bright-field aperture 436 arranged at aperture plane AP. The bright-field aperture 436 includes a central (i.e., on-axis) optically transmissive portion 437 and an outer non-transmissive (i.e., light-blocking) portion 438. In an example, the bright-field aperture 436 is in the form of an adjustable iris. The central transmissive portion 437 allows the focused light 122 to pass through and to be detected by image sensor 440 while the redirected light 122D is blocked by the outer non-transmissive portion 438. The defect 30 shown in FIG. 5B is a dimple defect 30D, but a similar result obtains in this bright-field embodiment for a pimple defect 30P. The image sensor 440 resides in a bright-field image plane BIP.

[0088] FIG. 5C is the schematic diagram of an example reflection configuration for defect camera system 431 so that it operates in a reflection mode. The collimated light source 120 is positioned on the same side of transparent sheet 10 as defect camera 430, with the collimated light 122 emitted orthogonal to the optical axis Al of defect camera 430. A beam splitter 127 is disposed so that it reflects collimated light 122 perpendicularly toward upper surface 22 of transparent sheet 20. The collimated light 122 is then reflected by upper surface 22 of transparent sheet 20 to form reflected collimated light 122R while also forming deflected collimated light 122D that travels towards defect camera 430. The reflected collimated light 122R and deflected collimated light 122D passes through the beam splitter 127 as it travels towards defect camera 430. The reflected collimated light 122R and deflected collimated light 122D is then received by defect camera 430 and the resulting electronic defect-image signals SD1 and SD2 are generated as described above. The defect camera system of FIG. 5C is capable of detecting smaller surface slope variations on the upper surface 22 of transparent 10 as compared to the transmission-mode configurations of FIGS. 5A and 5B.

[0089] FIG. 5D is the schematic diagram similar to FIG. 5A and shows an example of an alternative off- axis dark-field configuration for defect camera system 431. In some cases, collimated light 122 as well as deflected collimated light 122D can reflect multiple times between the surfaces of imaging lens 432, and cause stray light 461 to reach the defect-camera imaging plane DIP. If the stray light 461 reaches the imaging sensor 440, it will increase the background intensity reduce the contrast of the bright-field image of defect 30D.

[0090] To reduce the impact of stray light 461, the collimated light source 120 can be tilted relative to axis Al (i.e., away from perpendicular) by a small angle while also shifting the central obscuration 436 of dark-field aperture 434 to be off -axis. In this tilted or off-axis configuration, dark-field aperture 434 can still block the background collimated light 122 while letting most of deflected collimated light 122D form the image. Likewise, at least a substantial portion of stray light 461 is shifted and/or blocked by dark- field aperture 434 so that the stray light does not substantially alter the background intensity of the defect image. The close-up inset at the top of FIG. 5D shows an example of how some of the scattered light 461 can reach the defect-camera image plane DIP but fall outside of imaging sensor 440.

[0091] While the defect camera system 431 with either a dark-field or a bright-field imaging configuration, is good at detecting defects 30 and particles 32, it is typically difficult to distinguish between defect images and particle images within a given defect-camera image. In other words, the defect images and the particle images can be indistinguishable based on examining the defect-camera image alone. Thus, the particle camera system 331 is used to capture a second image (i.e., a particle- camera image) to assist in distinguishing between the defect images and the particle images captured in the defect-camera image.

[0092] Example method of operation [0093] With reference to FIGS. 6A and 6B, in an example method of operation of system 100, defect camera system 431 and particle camera system 331 are used to capture respective defect-camera images and particle-camera images I MD (FIG. 6A) and IMP (FIG. 6B) of select regions R of transparent substrate 20. The controller 104 can be used to track the defect-camera images I MD and particle- camera images I MP and match the images to the regions R of the transparent substrate. The bright-field or dark-field configuration of defect camera system 431 enables it to detect not only particles 32 but also dimples 30D and pimples 30P. On the other hand, the conventional imaging configuration of particle camera system 230 enables it to only detect particles 32. Thus, for a particle-camera image IMP of a given region of substrate 20 that includes both particles 32 and defects 30, there will be no images of the defects in the particle-camera image.

[0094] FIG. 6A is a representation of an example dark-field defect-camera image I MD of a select region R of transparent sheet 20 and that includes image features F that have yet to be identified as particle images 32' arising from particles 32 or defect images 30' arising from defects 30. FIG. 6B is a representative particle-camera image IMP of the same select region R and that includes image features F that are positively identified as particle images 32' because particle camera 330 detects only particles 30. Note that the particle images 32' are shown as dark spots in an otherwise bright field because the particles 32 scatter and/or absorb non-collimated light 132. In actual particle-camera images IMP, particle images 32' can appear as various shades (including black) of the imaging wavelength or wavelength band.

[0095] FIG. 6C is a final image I MF that is the result of an image-processing step (e.g., carried out in controller 104) wherein the representative particle-camera image I MP of FIG. 6B is compared to (e.g., subtracted from) the representative defect-camera image IMD of FIG. 6A. The image-processing operation can be described as IMF = I MP - I MD. In an example, the defect camera 430 and particle camera 330 have the same number of pixels 442 and 342 and the subtraction operation or comparison operation is done based on a pixel-by-pixel basis using the respective image intensities for each pixel. In cases where the pixel count and pixel sizes are different, controller 104 can be used to perform pixel averaging and image re-sizing in order to perform the appropriate image processing.

[0096] The comparison of the defect-camera image IMD and the particle-camera image IMP allows for the image features F in the particle-camera image IMD to be positively identified as defect images 30' arising from defects 30 or particle images 32' arising from particles 32. [0097] Optical inspection system with two defect cameras

[0098] FIG. 7A is a schematic diagram of an example configuration of defect camera system 431 that includes two (i.e., first and second) defect cameras 430a and 430b that respectively capture first and second defect-camera images I MDa and IM Db (see FIGS. 8A and 8B, introduced and discussed below) over respective imaging regions Rla and Rib of transparent sheet 20. The two defect cameras 430a and 430b have respective camera axes ACa and ACb that reside along respective system axes Ala and Alb. In an example, camera axes ACa and ACb are oriented at respective axes angles - 9a and + 9b relative to a surface normal N to upper surface 22, where a positive angle is measured clockwise from the surface normal. In an example, | 9a | = 19b | but 9a ≠ 9b (where " | x | " represents the absolute value of x).

[0099] The axes angles 9a and 9b can be measured from some other convenient reference line or surface other than the surface normal N, e.g., from planar upper surface 22. An example range on the magnitude of axes angles 9a and 9b is between 10 degrees and 60 degrees, with 30 degrees being an exemplary angle magnitude for both 9a and 9b. Controller 104 can be used to store the first and second defect-camera images I MDa and I MDb and their respective imaging regions Rla and Rib, and then align and compare the images as discussed below.

[00100] FIG. 7B is similar to FIG. 7A and shows an example defect camera system 431 where the two defect cameras 430a and 430b each view the same imaging region R at the same time, rather than viewing different imaging regions Rla and Rib at the same time, as shown in FIG. 7A. The configuration of the two defect cameras 430a and 430b of FIG. 7B obviates the need to account for an offset between the different imaging regions Rla and Rib. The first and second defect cameras 430a and 430b are associated with corresponding collimated light sources 120a and 120b that respectively direct collimated light 122a and 122b system axes Ala and Alb, respectively.

[00101] In an example, the configuration of the two defect cameras 420a and 420b is such that when the two defect-camera images IMDa and IMDb are aligned, the two images of a given defect 30 on upper surface 22 (denoted respectively as 30a' and 30b' and referred to as a "defect images") will substantially overlap. In each of the embodiments of FIG. 7A and 7B, the two defect cameras 430a and 430b view the same imaging region R at a different angle, with the configuration of FIG. 7B doing so simultaneously. This allows for optical inspection system 100 to determine where a given defect 30 resides in 3D space, and in particular, whether a given defect resides on upper surface 22 or lower surface 24. [00102] In an example, this determination includes performing a triangulation calculation (e.g., using controller 104) based on triangulation techniques known in the art. In an example, the two defect- camera images are used to create a stereoscopic image of the imaging region . In an example, the configuration of the two defect cameras 420a and 420b is selected based on the thickness TH of transparent sheet 20 and the anticipated size range of defects 30 so that the above-mentioned calculation can adequately resolve the surface location of a given defect 30, i.e., can determine whether the given defect resides on the upper surface 22 or lower surface 24 of transparent sheet 20.

[00103] FIG. 7C is a close-up view of an example defect 30 on lower surface 24 of transparent substrate 20 that illustrates an example of how the defect-camera images can show shifts in the defect images, depending on which surface the defect resides. FIG. 7C shows a defect 30 on lower surface 24 and also shows projections 30' and 30" of the defect onto the upper surface 22. Note that the camera axes ACa and ACb cross at lower surface 24. The total (center-to-center) separation S of the projected images 30' and 30" is given by S = 2-dx, where dx = TH-Tan (Asin(Sin(9)/n)), where n is the index of refraction of glass, for the case where 19a | = 19b | = Θ, and where "Asin" means "arcsine."

[00104] For defect 30 having a diameter d = 0.5 mm and for a transparent sheet thickness TH = 0.7 mm and for a index of refraction of glass n = 1.5 and an angle Θ = 30 degrees, the separation S = 2·(0.7 mm) · (Tan(Asin(Sin( 30)/1.5))) ~ 0.5 mm, which is similar to the defect diameter d of 0.5 mm. This allows for the separation S of the two images from defect cameras 430a and 430b to be easily seen when defect 30 resides upper surface 22. In an example, the configuration of two defect cameras 420a and 420b is selected such that the separation S between defect images 30' and 30" is at least half of the size (e.g., diameter d) of the defect image, i.e., S > d/2.

[00105] FIG. 8A is a representative depiction of camera images 30a' and 30b' respectively captured by defect cameras 420a and 430b of an example ring-type defect 30 that resides on upper surface 22. The two camera images 30a' and 30b' substantially overlap. FIG. 8B is a similar depiction of the two camera images 30a' and 30b' of an example defect that resides on lower surface 24 and shows how the two images are substantially spaced apart by a (center-to-center) separation S. This result obtains for the configuration of defect cameras 430a and 430b as shown in FIG. 7B, where the camera axis ACa and ACb intersect upper surface 22 at the same location with the given imaging region R (or regions Rla and Rib of FIG. 7A). [00106] Other configurations for the two defect cameras 420a and 420b can also be employed. For example, if the first and second camera axes Ala and Alb are made to cross at lower surface 24 such as shown in FIG. 7C, then the images of a given defect 30 on upper surface 22 will be substantially displaced while the images of a given defect on the lower surface 24 will substantially overlap. If the first and second camera axes Ala and Alb are made to cross at a location between upper and lower surface 22 and 24, then the images of a given defect 30 on upper surface 22 will be substantially displaced in one direction while the images of a given defect 30 on lower surface 24 will be displaced by the same amount in the opposite direction. Thus, in example, the two-camera embodiment of defect detection system 431 can be used to determine the surface location (i.e., upper or lower surface) of a given defect 30 by the amount of shift in the defect images 30' and 30" as determined from the two defect-camera images IM Da and IMDb.

[00107] Discriminating between dimples and pimples

[00108] While the above-described system and methods can be used to discriminate between defects 30 and particles 32 and to provide the locations, sizes, on which side of transparent sheet 10 the defects reside, the shape of the defects, etc., they generally cannot discriminate very well between a dimple 30D and a pimple 30P.

[00109] Thus, an aspect of the disclosure is directed to discriminating between a dimple 30D and a pimple 30P. FIG. 9A is similar to FIG. 2 and shows an example optical inspection system 100 that additionally includes a schlieren camera 530 and a second collimated light source 120 arranged along an axis A3. The schlieren camera 520 and second collimated light source 120 define a schlieren camera system 531.

[00110] The schlieren camera 530 is operably arranged adjacent upper surface 22 of transparent sheet 20 while second collimated light source 120 is operably arranged adjacent lower surface 24. In an example, schlieren camera 530 is operably supported by second support member 240, while second collimated light source 120 is operably supported by the first support member 140. This arrangement allows for controller 104 to move the first and second support members 140 and 240 in tandem so that schlieren camera system 531 can image a select region of transparent sheet 20. In another example, schlieren camera system 531 is independently supported and moved, e.g., it is not mechanically connected to first and second support members 140 and 240 and is supported by one or more separate support members, not shown. [00111] In an example, schlieren camera system 531 is moved to imaging regions of transparent sheet 20 that have been identified using the systems and methods described above as having at least one defect 30. As noted above, transparent substrate 20 can also be moved alone or in combination with the coordinated movement of first and second support members 140 and 240 to position schlieren camera system 531 to view one or more select imaging regions R of transparent sheet 20.

[00112] FIG. 9B is a more detailed view of an example schlieren camera system 531, shown along with a close-up of transparent sheet 20 having an imaging region R3 viewed by schlieren camera 530. The imaging region R3 includes an example defect 30 in the form of a dimple 30D by way of example. The schlieren camera 530 includes first and second lenses 532A and 532B. The first lens 532A defines a schlieren aperture plane SAP at which is operably disposed a schlieren aperture 534, which includes a semicircular knife-edge 535 and a semi-circular optically transmissive portion 536.

[00113] The close-up inset II shows different orientations of schleiren aperture 534. In an example, the close-up inset II represents different schlieren apertures 530 that can be inserted into schlieren aperture plane SAP (e.g., supported on a sliding or rotating member, not shown), or a single schlieren aperture that is rotatable (e.g., manually or via controller 104) within the schlieren aperture plane. The schlieren aperture 530 acts as a cut-off filter for intensity and knife edge 535 can be moved towards or away from axis A3 to adjust the image contrast.

[00114] Schlieren camera 530 also includes an image sensor 540 operably arranged at a schlieren image plane SIP, a digital image processor 550 electrically coupled to the image sensor, and a microcontroller 556 electrically coupled to the digital image processor. The digital image processor 550 can be electrically coupled to controller 104 so that schlieren images generated by image sensor 540 and digital image processor 550 can be stored, optionally processed, and further analyzed using the controller.

[00115] With continuing reference to FIG. 9B, in the operation of schlieren camera system 520, collimated light 122 from second collimated light source 120 travels through both lower and upper surfaces 24 and 22 of substrate 20, as well as through body 21. Some of the collimated light 122 interacts with defect 30, thereby forming deflected light 122D. The first lens 532A receives both collimated light 122 and deflected light 122D. The first lens 532A focuses the collimated light 122 to a focus position FP on axis A3 at schlieren aperture plane SAP. As noted above, schlieren aperture 534 is adjusted (e.g., rotated and translated) to pass a select amount of the focused light to adjust the image contrast.

[00116] Meanwhile, the deflected light 122D received by first lens 532A is not focused at focus position FP by first lens 532A since its trajectory has been altered by defect 30. A portion of deflected light 122D passes through transmissive region 536 of schlieren aperture 534 while another portion is blocked by knife edge 535. Thus, portions of deflected light 122D and focused light 122 are incident upon image sensor 540, which forms a digital schlieren image of defect 30. This digital image can be further processed by digital image processor 550 and then transmitted to controller 104.

[00117] The schlieren image obtained using schlieren camera system 531 can be used to discriminate between dimples 30D and pimples 30P by examining the image polarity of bright and dark regions of the defect image. FIG. 10A is a schematic representation of a schlieren image IMS that includes a defect image 30D' of a dimple 30D. The dimple image 30D' includes a bright region B on the right-side of the image and a dark (shadow) region D on the left side of the image. The orientation of bright region B and dark region D represents the image polarity and is defined by the configuration of schlieren camera 530. FIG. 11B is an actual schlieren image IMS that includes a defect mage 30P' of a more complex-shaped pimple 30P, wherein the defect image has the same image polarity as that of FIG. 11A.

[00118] FIGS. 11A and 11B are similar to FIGS. 10A and 10B, with FIG. 11A being a schematic representation of a schlieren image IMS and FIG. 11B being an actual schlieren image, with each showing a pimple image 30P' for a pimple 30P. The pimple image 30P' in FIG. 11B is for a more complex-shaped pimple 30P than the simple illustrative pimple that forms the pimple image of FIG. 11A.

[00119] For the pimple image 30P', the image polarization is reversed from that of a dimple image 30D', i.e., the bright region B is on the left-side of the image and the dark region D is on the left-side of the image.

[00120] In an example, schlieren camera 530 has a greater magnification and thus a greater resolution than defect camera 430 and particle camera 330. Thus, in an example, the defect camera system 431 and particle camera system 331 are used to perform a relatively low-resolution scan ("macro-scan") of transparent sheet 20 sufficient to discern defects 30 from particles 32 and to locate and perform a course size measurement (and optionally a coarse shape measurement) of any defects 30 that are discovered. Then, schlieren camera system 531 is used to revisit those measurement sites on transparent sheet 20 for which defects 30 were detected and make a finer measurement (i.e., perform a "micro-revisit"). For example, revisiting of the measurements sites and defects 30 therein by schlieren camera system 531 allows for a closer (i.e., higher-resolution) look at the defects 30 to characterize each defect as a dimple 30D or a pimple 30P, as discussed above. The higher-resolution revisit can also include determining (characterizing) the size and shape of each defect 30, as well as confirming or determining the side of transparent sheet 20 on which a given defect resides.

[00121] ft will be apparent to those skilled in the art that various modifications to the preferred embodiments of the disclosure as described herein can be made without departing from the spirit or scope of the disclosure as defined in the appended claims. Thus, the disclosure covers the modifications and variations provided they come within the scope of the appended claims and the equivalents thereto.