Partial coherence mitigation by illumination apodization in video measurement systems

Abstract

A video metrology system for measuring a test object, comprising an imaging system and an illumination system. The imaging system comprises an imager having an imaging pupil arranged to view at least a portion of the silhouette of the test object by receiving light transmitted past the test object over a first angular range. The illumination system includes (i) an illumination source, (ii) an output having a second angular range in object space that is greater than the first angular range received by the imaging pupil, and (iii) a substrate arranged to diffuse light from the illumination source, the substrate having an axial centerline and an optical obscuration element. The optical obscuration element is at least approximately coaxial with the axial centerline of the substrate, and the illumination system and the imaging system pupil are at least approximately conjugate image planes.

Description

[Technical Field]

[0001] The present invention relates to an illumination system for a video measurement system, and more particularly to an apodizing illumination system. [Background technology]

[0002] Video metrology machines collect metrology data from a test object. U.S. Patent Nos. 10,701,259 and 9,784,564 teach various aspects of such video metrology machines and are incorporated herein by reference in their entireties. One-way video metrology machines collect metrology data using backlighting. In this machine, the test object is illuminated from one direction and imaged from the opposite direction. When the test object is backlit, the test object itself appears dark to the imaging system, while the remaining background appears bright. This causes the test object to appear in silhouette. The object's profile is identified by transition points between light and dark, where light surrounding or passing through the test object is contrasted with adjacent portions where light is blocked. The imaging system images the test object's silhouette. The object's profile can be identified by transition points, where light surrounding or passing through the test object is contrasted with adjacent portions where light is blocked. Video metrology backlights are typically designed to produce an angularly uniform illumination distribution. However, this can cause an apparent shift of the edges of backlit test objects observed with the video metrology machine. The apparent edge shift progresses from "dark" to "light" regardless of the edge orientation, causing objects illuminated with an opaque backlight to have larger than expected dimensions and smaller than expected inner diameter rings. This is largely due to an optical phenomenon called partial coherence.

[0003] Several techniques are used to solve the problem of shifted edges. Typically, the aperture stop of the illuminator is nearly aligned with the aperture stop of the imaging system, limiting the range of angles at which the object is illuminated. An overfilled imaging system aperture allows higher angles of light to enter the imaging system aperture through specular reflection or diffusion, which can obscure the object's silhouette boundary. Therefore, a range of angles is collected by the imaging system to image the silhouette boundary, while limiting the range of illumination angles to avoid unnecessarily illuminating the test object from different directions. Current configurations of video metrology machines can provide apparent edge shift magnitudes of less than 10 μm for backlit test objects. However, this edge shift is observable and still significant in many applications. Incoherent illumination, or significantly increasing the angular range of illumination, has also been considered as a solution to edge shift. For an F / 100 imager, increasing illumination from F / 50 to F / 5 is known to significantly reduce the magnitude of apparent edge shift. Additionally, placing a diffuser immediately after the backlight or widening the backlight pupil can help measure objects closer to nominal "zero thickness" chrome, such as on a glass reticle. However, these techniques still create problems when measuring certain test parts. For example, the wraparound effect can be observed when measuring test objects with curved or sloped surfaces, such as gauge pins. The wraparound effect is the result of a wide range of illumination angles reflecting off curved or sloped surfaces in the test object profile and entering the imaging system aperture. In other words, edges formed by curved or sloped walls can specularly reflect light back into the imager, introducing another source of error at the edge. Another possible technique for resolving the shifted edge issue is to use software to correct values ​​after the edge of interest is located. In this case, the true edge location is determined by special post-processing of the edge profile, both in the system and the object.Typically, algorithms are used to find the true edge location, requiring input of the illumination angle range and edge depth of the object. However, it would be more desirable to acquire images without such corrections and without requiring prior knowledge of the object. Summary of the Invention

[0004] The present invention contemplates a video metrology system for measuring test objects in which partial coherence induced edge shifts are mitigated through illumination apodization. a video measurement system for measuring a test object, the video measurement system comprising an imager having an imaging pupil, the imager positioned to view at least a portion of a silhouette of the test object by receiving light transmitted past the test object over a first angular range; and an illumination system, the illumination system including (i) an illumination source; and (ii) a second illumination source in object space that is greater than the first angular range received by the imaging pupil. an output having a range of angles; and (iii) a substrate having an axial centerline and a light obscuration element positioned to diffuse light from said illumination source. (iv) an illumination pupil; the optical obscuration element is at least approximately coaxial with the axial centerline of the substrate; The illumination pupil and the imaging pupil are At least approximately in a conjugate image plane.

[0005] In one configuration, the substrate of the video metrology system includes a front surface illuminated by the illumination source and a back surface, and the optical obscuration element is disposed on the front surface and is at least approximately coaxial with the axial centerline of the substrate. In one configuration, the substrate is a volume diffuser. In one configuration, the illumination system of the video metrology system has an object space numerical aperture greater than the object space numerical aperture of the imager. Typically, the axial centerline of the illumination pupil is at least approximately aligned with the optical obscuration element. In one configuration, the lighting systemThe second angular range is twice as large as the first angular range received by the imaging pupil. the illumination pupil includes an axial centerline; The substrate is front and a ball disposed within the hole, the hole being approximately coaxial with the axial centerline of the illumination pupil. In one configuration, the ball is an opaque spherical ball and the hole is substantially cylindrical.

[0006] In another configuration, the substrate of the video measurement system includes a front surface illuminated by the illumination source and a back surface, and the light obscuration element includes a pair of linear polarizing filters, the pair including (i) a first linear polarizing filter disposed between the illumination source and the front surface of the substrate, and (ii) a second linear polarizing filter overlapping the first linear polarizing filter, one of the first and second linear polarizing filters being rotated relative to the other of the first and second linear polarizing filters. the illumination pupil includes an axial centerline; The second linear polarizing filter has a diameter smaller than the diameter of the illumination pupil and is at least approximately coaxial with the axial centerline of the illumination pupil, and the second linear polarizing filter is larger than the diameter of a conjugate image of the imaging pupil at the back surface of the substrate. In one configuration, the second linear polarizing filter is glued to the back surface of the substrate. In another configuration, the second linear polarizing filter is glued to the front surface of the substrate. In yet another configuration, the second linear polarizing filter is positioned between the illumination source and the first linear polarizing filter. Rotating the substrate can rotate the second linear polarizing filter relative to the first linear polarizing filter. In one configuration, a mechanism rotates the first linear polarizing filter relative to the second linear polarizing filter.

[0007] According to another approach, a video measurement system for measuring a test object includes an imaging system, an illumination system, a substrate, and an opaque ball. Imager withand wherein the imager is positioned to view at least a portion of the silhouette of the test object by receiving light transmitted past the test object over a first angular range, the illumination system including an illumination source and configured to receive light over a second angular range greater than the first angular range received by the imaging pupil, and an output and an illumination pupil, wherein the illumination pupil and the imaging pupil are The substrate is positioned to diffuse light from the illumination source, the substrate having a front surface and a back surface, the front surface being illuminated by the illumination source. The opaque ball is positioned within the hole in the front surface of the substrate and is sized to block a portion of the light from the illumination source. In one configuration, the front surface of the substrate is diffuse and the back surface of the substrate is diffuse. In one configuration, the ball is a spherical ball bearing and the hole is substantially cylindrical. The imaging pupil and the ball each have an axial centerline, the imaging pupil being at least approximately coaxial with the axial centerline of the ball.

[0008] In yet another approach, a video measurement system for measuring a test object includes an imaging system, an illumination system, a substrate, a first linear polarizing filter, and a second linear polarizing filter. The imaging system includes an imager having an imaging pupil positioned to view at least a portion of the silhouette of the test object by receiving light transmitted past the test object over a first angular range. The illumination system includes an illumination source, a second angular range greater than the first angular range received by the imaging pupil, an output, and With illuminated eyes, It has. the illumination pupil and the imaging pupil are at least approximately conjugate image planes. The substrate is positioned to diffuse light from the illumination source and has a front surface and a back surface. The first linear filter is positioned between the illumination source and the front surface of the substrate. The second linear polarizing filter is positioned to diffuse light from the first linear filter. Polarizing filterand overlap. One of the first and second linear polarizing filters is rotated relative to the other of the first and second linear polarizing filters. In one configuration, the illumination system further includes an illumination pupil, the illumination pupil having an axial centerline, the second linear polarizing filter being smaller than the illumination pupil and at least approximately coaxial with the axial centerline of the illumination pupil, and the first linear polarizing filter being larger than the imaging pupil. The second linear polarizing filter is disposed on the back surface of the substrate, disposed between the first polarizing filter and the substrate, or disposed between the illumination source and the first linear polarizing filter. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 illustrates a video measurement system for backlighting a test object, the video measurement system including an illumination system and an imaging system for capturing a silhouette of the backlit test object, where the angular range of the backlight is greater than the angular range experienced by the imaging pupil.

[0010] [Figure 2] FIG. 1 shows a lighting system for a video measurement system that backlights a test object.

[0011] [Figure 3] 1 is a diagram of a video measurement system for backlighting a test object, showing the test object on a mounting table.

[0012] [Figure 4A] FIG. 1 is an enlarged top view of the front side of a volume diffuser with obscuration elements within the holes.

[0013] [Figure 4B] 4B is a cross-sectional view of a volume diffuser with holes taken along line LL in FIG. 4A.

[0014] [Figure 5] 4B is an image of the illuminance distribution using an obscuration element in the hole of FIG. 4A.

[0015] [Figure 6] FIG. 1 is a close-up view of a light obscuration element using first and second linear polarizers, where the first linear polarizing filter is positioned close to at least the front surface of the diffuser and the second linear polarizing filter is positioned on the back surface of the diffuser.

[0016] [Figure 7] FIG. 1 is a close-up view of a light obscuration element using first and second linear polarizers, with the second linear polarizing filter placed on the front surface of the diffuser.

[0017] [Figure 8] FIG. 1 is a close-up view of a light blocking element using first and second linear polarizers, where the second linear polarizing filter is positioned between the illumination source and the first linear polarizing filter.

[0018] [Figure 9] 7 is a set of images of the distribution, shown in grayscale, obtained by using the first and second linear polarizers shown in FIG. 6 and rotating the second linear polarizer relative to the first linear polarizer.

[0019] [Figure 10] 10 is a graph of the radial profile of illumination corresponding to an approximation of the distribution extracted by digital image analysis.

[0020] [Figure 11A] 9 is a grayscale representation of a color heat map of edge shift obtained from a video measurement system rotated with the first and second linear polarizers shown in FIG. 6. The grayscale representation shows a median "dark width minus light width" difference of -0.986px when the second linear polarizer is rotated relative to the first linear polarizer to provide an image of the backlight distribution labeled "0" in FIG.

[0021] [Figure 11B]9 is a grayscale representation of a color histogram of values ​​obtained from a video measurement system rotated with the first and second linear polarizers shown in FIG. 6. It shows that the median "dark width minus light width" difference is -0.986px when the second linear polarizer is rotated relative to the first linear polarizer to provide the image of the backlight distribution labeled "0" in FIG.

[0022] [Figure 12A] 9 is a grayscale representation of a color heat map of edge shift obtained from a video measurement system rotated with the first and second linear polarizers shown in FIG. 6. The grayscale representation shows a median "dark width minus light width" difference of -0.423px when the second linear polarizer is rotated relative to the first linear polarizer to provide an image of the backlight distribution labeled "1" in FIG.

[0023] [Figure 12B] 9 is a grayscale representation of a color histogram of values ​​obtained from a video measurement system rotated with the first and second linear polarizers shown in FIG. 6. It shows that the median "dark width minus light width" difference is -0.423px when the second linear polarizer is rotated relative to the first linear polarizer to provide an image of the backlight distribution labeled "1" in FIG.

[0024] [Figure 13A] 9 is a grayscale representation of a color heat map of edge shift obtained from a video measurement system rotated with the first and second linear polarizers shown in FIG. 6. The grayscale representation shows a median "dark width minus light width" difference of -0.008px when the second linear polarizer is rotated relative to the first linear polarizer to provide an image of the backlight distribution labeled "2" in FIG.

[0025] [Figure 13B]9 is a grayscale representation of a color histogram of values ​​obtained from a video measurement system rotated with the first and second linear polarizers shown in FIG. 6. It shows that the median "dark width minus light width" difference is -0.008px when the second linear polarizer is rotated relative to the first linear polarizer to provide an image of the backlight distribution labeled "2" in FIG.

[0026] [Figure 14A] 9 is a grayscale representation of a color heat map of edge shifts obtained from a video measurement system rotated with the first and second linear polarizers shown in FIG. 6. The grayscale representation shows a median "dark width minus light width" difference of +0.196px when the second linear polarizer is rotated relative to the first linear polarizer to provide an image of the backlight distribution labeled "3" in FIG.

[0027] [Figure 14B] 9 is a grayscale representation of a color histogram of values ​​obtained from a video measurement system rotated with the first and second linear polarizers shown in FIG. 6. It shows that the median "dark width minus light width" difference is +0.196px when the second linear polarizer is rotated relative to the first linear polarizer to provide an image of the backlight distribution labeled "3" in FIG.

[0028] [Figure 15A] Grayscale representation of the color heat map of edge shift obtained from a video measurement system without a light obscuration device, showing a median "dark width - light width" difference of +0.377px.

[0029] [Figure 15B] Grayscale representation of the color histogram of values ​​obtained from a video measurement system without a light obscuration device, showing a median "dark width - light width" difference of +0.377px.

[0030] [Figure 15C]15B is an image of the backlight distribution shown in grayscale that produces the plot shown in FIG.

[0031] [Figure 16A] Grayscale representation of a color heat map of edge shifts obtained from a video measurement system without a light obscuration element and with an improperly collimated illumination source, showing a median "dark width minus light width" difference of +1.032px.

[0032] [Figure 16B] Grayscale representation of the color histogram of values ​​obtained from a video measurement system without a light obscuration element and with an improperly collimated illumination source, showing a median "dark width - light width" difference of +1.032px. Detailed Description of the Invention

[0033] At the outset, it should be understood that like reference numerals are intended to identify the same structural elements, portions, or surfaces consistently throughout the drawings. These elements, portions, and surfaces are further described throughout the specification, of which the detailed description of the invention is a part. Unless otherwise noted, the drawings are intended to be read together with the specification and should be considered part of the overall written description of the invention.

[0034] The video measurement system 10 shown in Figures 1-3 presents a layout featuring an illumination system 12 for backlighting a test object 100 and an imaging system 14 configured to measure the silhouette of the test object 100. Measuring the silhouette of a backlit test object is known as shadow imaging or silhouette imaging. The video measurement system 10 for opaque parts, one of many possible configurations, includes a test object 100 positioned on a mounting stage 16. The mounting stage 16 may be constructed of a transparent optical material or may provide optical transmission to transmit light passing through the test object 100 to the imaging system 14. The test object 100 is backlit by an illumination system 12 having an illumination source 20 and an illumination lens 22. While one illumination lens is shown, it should be understood that additional illumination lenses may be included in the configuration. Typically, the illumination lens 22 is a collimating lens that provides a collimated light beam, as shown in Figure 2, for example. The illumination system 12 further includes an illumination pupil 26. The illumination system 12 in one configuration is aligned with the optical axis or centerline 28 of the video measurement system 10. The imaging system 14 is positioned to detect transmitted light from the illumination device on the opposite side of the test object 100 (known as diascopic illumination).

[0035] The imaging system 14 includes at least an imager, such as an array image sensor 48, aligned along the common optical axis 28 of the video measurement system 10. The illumination system 12 includes a relatively high angular range compared to the imaging system 14. The imaging system 14 further includes at least one imaging front-end lens 42 and an imaging pupil 44. The imaging system 14 also includes a rear lens 46 and an image sensor 48 arrayed at an image plane. The front-end lens 42 cooperates with the collimating lens 22 to image the illumination pupil 26 of the illumination system 12 onto the imaging pupil 44 of the imaging system 14. A silhouette of the test object 100 is collected from a wider range of off-axis angles. However, the imaging system 14 has an object space numerical aperture smaller than that of the illumination system, thereby limiting the opportunity for stray light from the test object 100 to enter the imaging pupil. With this video metrology system 10, the angular extent 40a of the illumination output 24 in the object space 34 is greater than the angular extent 40b experienced by the imaging pupil 44. In one configuration, the angular extent 40a of the illumination source 20 in the object space 34 is approximately twice the angular extent 40b experienced by the imaging pupil 44. It should be understood that the illumination pupil 26 and the imaging pupil 44 are at least approximately conjugate image planes, and typically are conjugate image planes. The illumination pupil 26 and the imaging pupil 44 each include axial centerlines 30, 32 that are aligned with the optical axis 28 of the video metrology system 10.

[0036] The illumination source 20 may include a standard backlight with a high angular range. In one configuration, the angular range of the illumination source 20 in the target space is approximately twice the angular range experienced by the imaging pupil 44. The backlight target space f-number (F / #) may be approximately half that of the associated imager. For example, the illumination system 12 may have a target space F / 50 and the imaging system 14 may have an F / 100. In one configuration, the illumination source 20 may be a single light-emitting diode (LED) or multiple LEDs. In another configuration, the illumination source 20 is an incandescent bulb, a high-intensity discharge (HID) lamp, or a superluminescent diode (SLD or SLED). The illumination system 12 further includes a substrate 56 configured to diffuse light from the illumination source 20. The substrate 56 is preferably at least approximately aligned with the optical axis 28. More preferably, the substrate 56 is a diffuser 50 aligned with the optical axis 28. The illumination system 12 also includes a light obscuration device 18. The light obscuration device 18 comprises a substrate 56 and a light obscura element 52 configured to reduce a portion of the illumination distribution directly captured by the imaging system 14. In one configuration, the light obscuration element 52 is at least approximately coaxial with the axial centerline 30 of the illumination pupil 26, and more preferably is coaxial with the axial centerline 30 of the illumination pupil 26. More preferably, the imaging pupil 44, the light obscuration element 52, and the illumination pupil are coaxial with the optical axis 28. As described below, most preferably, the imaging pupil 44 is centered on a dark spot formed by the light obscuration element 52. It should be understood that the substrate 56 and the light obscura element 52 can be positioned within the region of the axial position R between the illumination source 20 and the illumination pupil 26.

[0037] The light obscuration device 18, including the light obscuration element 52 and the substrate 56, can take many different configurations. Many examples of light obscuration elements 52 and substrates 56 are described below. It should be understood that each of the described illumination distribution structures emits light that is not directly captured by the imaging optical system. Furthermore, it should be understood that the illumination system can optionally include additional optical elements between the illumination source 20 and the substrate 56.

[0038] As shown in FIGS. 1-4B, the substrate 56 can be a volumetric diffuser 50. Here, "diffuse" and "diffusing" refer to scattering of light, for example, by reflection, refraction, or diffraction. A volumetric diffuser 50 is generally a substrate that scatters light throughout the volume of a bulk material. One example of a volumetric diffuser 50 is a white acrylic sheet, such as Acrylite 020-4, Plexiglas 2447, or similar materials. However, it should be understood that other suitable materials are possible and are intended to be within the spirit and scope of the present invention. In one configuration, the volumetric diffuser 50 includes a diffusing front surface and / or a diffusing back surface. In another configuration, the volumetric diffuser 50 has a polished or buffed surface.

[0039] In another configuration, the illumination system includes a substrate 56 that is a non-volumetric double-sided diffuser or a surface diffuser. A surface diffuser refers to a substrate in which scattering points are limited to a surface, typically by a roughened exterior finish. Examples of surface diffusers include, but are not limited to, a clear glass piece with a roughened front and back surface, or a single-sided diffuser.

[0040] The optical obscuration element 52 is positioned to apodize the illumination pupil 26 by providing a dark center and a bright annular region near the pupil edge. The optical obscuration element 52 and substrate 56 control the angular distribution of light emitted from the illumination system 12. It should be appreciated that rotational symmetry within the imaging pupil 44 is important to avoid anisotropic measurement errors, especially when the object is away from best focus.

[0041] One effective way to achieve this distribution is to use a thick volume diffuser 50 that is uniformly illuminated from the front side 54. As shown in FIGS. 4A and 4B, a hole 58 is drilled in the center of the diffuser member, and a ball 60 is placed within the hole 58. The volume diffuser 50 has a thickness ranging from 2 mm to 6 mm and a diameter ranging from 3 mm to 7 mm. In one embodiment, the volume diffuser 50 is 2.9 mm thick and 5.65 mm in diameter. The ball 60 is sized to fit snugly within the hole 58, typically approximately the same size as the hole 58, with a tolerance that results in a minimum gap size. In one configuration, the diameter of the hole 58 is 1 mm to 5 mm. More preferably, the hole 58 has a cylindrical shape with a diameter of 1.5 mm to 3 mm. In some configurations, the hole 58 further includes a tapered edge or a conical base. In one configuration, the base is a right cone with an apex angle of approximately 118°. In some configurations, the depth of the hole 58 is less than the diameter of the ball 60. Thus, in some configurations, the ball 60 can protrude from the surface of the diffuser 50 when positioned within the hole 58. For example, in one configuration, the ball 60 can protrude from the front surface 62 of the diffuser 50 by 0.1 mm to 0.4 mm, and more preferably 0.2 mm to 0.3 mm. While a spherical steel ball bearing is shown as the light obscuration element 52, it should be understood that other materials can be used and other shapes that fit within the diffuser hole 58 are possible. The geometry of the light obscuration element 52 must vary depending on the properties of the particular diffuser material to produce the appropriate distribution. The light obscuration element 52 blocks some or all of the light near or at the center of the illumination pupil 26 and can be any material or shape. In one configuration, the ball is approximately spherical. When diffuser 50 is a volume diffuser and ball 60 is not directly visible on the output side of diffuser 50, the resulting distribution typically does not provide abrupt changes in light intensity. Adjusting the depth of hole 58 and the diameter of hole 58 provides control parameters for the resulting illuminance distribution. In one configuration, hole 58 is coaxial with the axial centerline of diffuser 50.Furthermore, ball 60 and hole 58 are at least approximately coaxial with the axial centerlines of illumination pupil 26 and imaging pupil 44, respectively, and more preferably are coaxial with the axial centerlines of illumination pupil 26 and imaging pupil 44.

[0042] It should be understood that other single, static, central obscuration components that provide less than 100% transmission can be used instead of or in addition to the ball. Additionally, other optical obscuration elements can be used, including, but not limited to, rod stock, deposited metal dots (e.g., chrome on glass), shim stock disks, black paint fills in the holes of a diffuser, metal foil disks, and the like.

[0043] In one configuration, the substrate 56 is part of an assembly of an illumination system, with the illumination source 20 on one side of the substrate 56 and the illumination lens element 22 on the other side.

[0044] Referring to FIG. 5, this figure shows an image of the illumination distribution using the light obscuration element in the hole of FIG. 4A. While the imaging system 14 can be aligned with the illumination source 20 along the optical axis 28 so that the circular imager receiving area is centered in the illumination distribution, it should be understood that the imaging system 14 does not need to be aligned with the illumination source 20. The light obscuration element 52 provides a brighter annular portion 66 near the edge of the illumination pupil 26 and a dark spot 68 near the center. The light obscuration element 52 can be approximately centered on a sharp outer edge. Alternatively, the light obscuration element 52 may not have a sharp outer edge. However, the imaging pupil 44 is preferably centered on the dark spot 68 formed by the light obscuration element 52.

[0045] In another approach, the illumination distribution is obtained by a light obscuration device 18 that apodizes the illumination pupil 26 using a pair of linear polarizing filters, as shown in Figures 6-8. The intended effect of adjustable edge shift is achieved by rotating one polarizing filter relative to the other.

[0046] In one configuration, the light obscuration element 52 is a pair of linear polarizing filters 70, 72. As shown in FIG. 6 , the linear polarizing filter 70 is positioned between the illumination source 20 and the front surface 62 of the substrate 56, with the linear polarizing filter 72 overlapping the linear polarizing filter 70. In this configuration, the illumination system 12 includes the illumination source 20, the substrate 56, and the first and second polarizing filters 70, 72, as described above. One of the polarizing filters 70, 72 is rotated relative to the other linear polarizing filter 70, 72 at different, adjustable orientations to provide a darker center and a brighter annular portion near the edge of the illumination pupil 26. Typically, the linear polarizing filter 72 has a smaller diameter than the illumination pupil 26, attenuates the corresponding central portion of the distribution by an adjustable amount, and is at least approximately coaxial with the axial centerline of the illumination pupil 26. Typically, the diameter of the linear polarizing filter 72 is larger than the diameter of the imaging pupil 44. In one configuration, the substrate 56 is the diffuser 50. In one configuration, the linear polarizing filter 72 is disposed on the back surface 74 of the substrate 56, as shown in FIG. 6. In this configuration, the linear polarizing filter 72 is disposed on the substrate 56 itself. The linear polarizing filter 72 is affixed or glued to the back surface 74 of the substrate 56 in a manner that does not completely block light transmission. Methods for securing the linear polarizing filter 72 to the diffuser include, but are not limited to, optically clear adhesives, clear tape, or press-fitting into holes in the substrate 56, e.g., the diffuser 50. While the diffuser 50 makes a convenient substrate, the linear polarizing filter 72 does not actually need to be physically attached to the diffuser 50. Rotating the diffuser 50 adjusts the transmission of light through the linear polarizing filter 72. For example, the linear polarizing filter 72 can be rotated in approximately 10-degree increments.

[0047] In another configuration shown in Figure 7, a linear polarizing filter 72 is disposed between the polarizing filter 70 and the substrate 56. In yet another configuration shown in Figure 8, the linear polarizing filter 72 is disposed between the illumination source 20 and the linear polarizing filter 70.

[0048] Reference is now made to Figures 9 and 10. Figure 9 shows a set of images of the distribution, shown in grayscale. These images were obtained by rotating the linear polarizing filter 72 relative to the linear polarizer 70, exemplarily using the linear polarizing filters 70 and 72 shown in Figure 6. Figure 10 shows a graph of the approximate corresponding radial illumination distribution profile extracted by digital image analysis. In this graph, the x-axis represents the distance from the center of the distribution, and the y-axis represents the relative intensity at each radial position. In this example, the linear polarizing filter 72 has a diameter slightly larger than the imaging pupil 44 and is affixed to the diffuser 50 and rotated in approximately 10-degree increments (see labels 0, 1, 2, and 3), starting from a rotation position that is at least transparent, until the position of the edge of the image of the test object 100 appears unshifted. In this configuration, the linear polarizing filter 72 is aligned with the center of the bright portion of the distribution generated by the illumination source 20 by translating the diffuser 50 by eye. The image of the distribution itself is shown below in grayscale, labeled 0, 1, 2, and 3, with the central portion 80 appearing brighter as the diffuser 50 is rotated. The edge shift map, as shown in the grayscale representation of the color heatmap in Figures 11A, 12A, 13A, and 14A, provides the median value of the "dark width minus light width" difference (error), which is displayed in the corresponding plot in the text. The error map plots displayed in Figures 11A, 12A, 13A, and 14A show a two-dimensional data visualization using a selected gradient. Here, the difference value between the "dark stripe width and the adjacent light stripe width" is plotted along a grayscale based on the maximum and minimum values ​​calculated in the "dark width minus light width" calculation. In the color heatmap versions of the images in Figures 11A, 12A, 13A, and 14A, different colors along the RGB scale correspond to different levels of grayscale brightness. More specifically, in one configuration of a color heatmap version of the images of Figures 11A, 12A, 13A, and 14A, the maximum "dark minus light" pixel value is represented in red, the median pixel value is represented in green, and the minimum "dark minus light" pixel value is represented in purple.

[0049] Further, histograms of each value are shown in FIGS. 11B, 12B, 13B, and 14B. FIGS. 11B, 12B, 13B, and 14B show histograms in grayscale using different luminance levels to represent colors along the RGB scale. In color versions of the grayscale histograms shown in FIGS. 11B, 12B, 13B, and 14B, different colors along the RGB scale correspond to luminance levels in the grayscale. More specifically, in the color versions of the grayscale histograms described herein, the histograms categorize the colors in the color error map and show the spread of values ​​along the RGB scale, from the highest "dark minus light" value (which may be represented by red at the top of the "dark minus light" pixel histogram), through the median "dark minus light" value (which may be represented by green and located approximately near the horizontal line of the histogram), to the lowest "dark minus light" value (which may be represented by purple at the bottom). The method for measuring the magnitude of edge shift in the existing video measurement system 10 is as follows:

[0050] First, a Ronchi ruling, chrome-deposited on glass with a 50% duty cycle, is placed at the focal point of the imaging system 14 and an image is captured. The digital image is then analyzed to locate various edges, preferably using the same edge-finding method used to measure the actual part. Edge locations are found across the entire image (i.e., the system's field of view (FOV)), and the difference between the dark fringe width and the widths of adjacent and light fringes is determined. If the difference is non-zero, there is an edge shift, which changes direction with the edge orientation. The dark fringe width minus the width of the adjacent light fringe is then mapped across the entire field of view of the imaging optical system. This is used to generate a color map of the difference magnitude. The median of all "dark minus light" values ​​is considered to summarize this into a single value. In the analysis output diagram, Figure 15A, for example, the median of all "dark minus light" values ​​is 0.377px. Additionally, Figure 15B shows a histogram of values, with the highest "dark minus light" value being 0.435px and shown in light gray in Figure 15B, the median being 0.377px and shown in medium gray located in the horizontal line area, and the lowest "dark minus light" value being 0.330px and shown in dark gray at the bottom of the histogram. In one configuration of a color version of the histogram shown in Figure 15B, the highest "dark minus light" value of 0.435px can be shown in red at the top, the median of all "dark minus light" values ​​being 0.377px can be shown in green located in the horizontal line area, and the lowest "dark minus light" value of 0.330px can be shown in purple at the bottom of the color version of the histogram. This shows the spread of values ​​in the plot, which can help indicate issues, such as a range of edge shift values ​​significantly higher than other machines of the same design.

[0051] Taking into account partial coherence effects, conventional lighting systems tend to shift edges from dark to light, and dark fringes tend to appear wider than adjacent light fringes. This causes dark-light (dark minus light) values ​​to tend to be positive. As an example of this tendency, Figure 15A is a grayscale display of a color error map, such as that described above, obtained from a multifunctional video measurement system with a backlight. An image of the backlight distribution that produced this plot (Figure 15C) is included in the lower left corner of the analysis output.

[0052] Having edge shifts measured across the entire FOV is useful for diagnosing inconsistent illumination angle distributions or misalignment between the illumination system 12 and the imaging system 14. For example, if the illumination source 20 (also called a backlight) is not properly collimated, a field-dependent error map is possible. Similarly, if the array image sensor 48 is not telecentric and the backlight is properly collimated, a pattern may be visible in the error map in the presence of edge shift effects of partial coherence. A field-dependent pattern indicates a system problem but does not identify the root cause. Figures 16A and 16B show an example of a field-dependent error map. In the analysis output diagram of Figure 16A, the median of all "dark minus bright" values ​​is +1.032px. Additionally, a histogram of the values ​​is shown in Figure 16B, where the highest "dark minus bright" value is 1.100px, shown in light gray, the median is 1.032px, shown in medium gray, and the lowest is 0.925px, shown in dark gray. In one configuration of the color histogram version of Figure 16B, the highest "dark minus light" value of 1.100px can be shown in red at the top, the median of all "dark minus light" values ​​of 1.032px can be shown in green, located in the horizontal line region of the histogram, and the lowest "dark minus light" value of 0.925px can be shown in purple at the bottom.

[0053] Other edge orientations can be tested by rotating the Ronchi ruling before taking images and performing analysis. A well-aligned system is expected to have edge shifts of isotropic partial coherence, so a single orientation can be used cautiously as a measure of this machine performance.

[0054] This method has the advantage of not requiring an exact (or any) physical pixel size, so at a minimum, there is no need to calibrate the magnification of the video measurement machine 10. This is because a found edge location is referenced to other found edge locations in the same image. The only thing required is an artifact with a verified 50% duty cycle, such as a Ronchi ruling in chrome on glass. To convert an edge from pixel units to a physical distance, the object space pixel size is required, but a good approximation of this parameter is usually adequate for this purpose.

[0055] Because the measurement is "dark width minus light width," isotropic edge shifts are counted several times (two widths with two edges per width). Assuming the edge shift is constant, the shift of any individual edge can be easily calculated by simply dividing the "dark minus light" value by four. Another convenient calculation is that objects whose widths are measured will typically have an error of half the "dark minus light" value, since the width comprises two measured edges.

[0056] An existing video measurement system was fitted with a rotated illumination system 12 using an optical obscuration device 18 to remove the edge shift of partial coherence, and the magnitude of the edge shift was measured using the method described above. The grayscale representation of the edge shift in the color heat maps of Figures 11A, 12A, 13A, and 14A shows that configuration "2" shown in Figure 13A has the closest measured edge shift to zero (0.008px) and therefore the most desirable correction factor.

[0057] If polarized illumination causes problems in measuring certain artifacts, a waveplate can be used after the linear polarizing filters 70, 72 to mitigate the effect of producing a non-linear output polarization.

[0058] It should be understood that alternative configurations of the pair of linear polarizing filters can be used. For example, an oversized linear polarizing filter 72 having a diameter significantly larger than the imaging pupil 44 can facilitate alignment sensitivity while maintaining sufficient adjustment to correct the desired amount of edge shift. Alternatively, the linear polarizing filter 72 can be placed on the side of the diffuser 50 closer to the illumination source 20, thus mitigating any features of the output distribution via the action of the diffusive substrate.

[0059] Finally, instead of rotating the diffuser 50 with the linear polarizing filter 72, a mechanism (not shown) for rotating the linear polarizing filter 70 may be used, since rotating the diffuser may cause shifts in the alignment of the system, including, but not limited to, the alignment of the linear polarizing filter 72 with the illumination source 20, the alignment of the linear filter 72 with the imaging system 12, or both.

[0060] Although particular configurations have been described in this disclosure, it is not intended to be limited to these configurations, and all alternative configurations, embodiments, modifications, and equivalents that occur to those skilled in the art are within the spirit and scope of the present invention as defined by the appended claims.

Claims

1. 1. A video measurement system for measuring a test object, comprising: (a) an imaging system comprising an imager having an imaging pupil, the imager positioned to view at least a portion of a silhouette of the test object by receiving light transmitted past the test object over a first angular range; (b) an illumination system comprising: (i) an illumination source; and (ii) a light source, in object space, at the imaging pupil. an output having a second angular range greater than the first angular range received by the illumination source; (iii) a substrate arranged to diffuse light from the illumination source, the substrate having an axial centerline and a light obscuration element; and (iv) an illumination pupil. wherein the optical obscuration element is at least approximately coaxial with the axial centerline of the substrate, and the illumination pupil and the imaging pupil are at least approximately conjugate image planes.

2. 2. The video measurement system of claim 1, wherein the substrate includes a front surface illuminated by the illumination source and a back surface, and the optical obscuration element is disposed on the front surface and is at least approximately coaxial with the axial centerline of the substrate.

3. The video measurement system of claim 1 , wherein the substrate is a volume diffuser.

4. The video measurement system of claim 1 , wherein the illumination system has an object space numerical aperture greater than an object space numerical aperture of the imager.

5. A video measurement system as described in claim 1, wherein the second angular range of the illumination system is twice as large as the first angular range received by the imaging pupil.

6. A video measurement system as described in claim 3, wherein the illumination pupil includes an axial centerline, and the base further includes a hole in the front surface and a ball positioned within the hole, the hole being approximately coaxial with the axial centerline of the illumination pupil.

7. 7. The video measurement system of claim 6, wherein the ball is an opaque ball and the substrate is a volume diffuser.

8. 7. The video measurement system of claim 6, wherein the ball is spherical and the hole is substantially cylindrical.

9. The video measurement system of claim 1 , wherein the imaging pupil is at least approximately aligned with the optical obscuration element.

10. the substrate includes a front surface illuminated by the illumination source and a back surface; 2. The video measurement system of claim 1, wherein the light obscuration element comprises a pair of linear polarizing filters, the pair of linear polarizing filters including: (i) a first linear polarizing filter disposed between the illumination source and the front surface of the substrate; and (ii) a second linear polarizing filter overlapping the first linear polarizing filter, wherein one of the first and second linear polarizing filters is rotated relative to the other of the first and second linear polarizing filters.

11. the illumination pupil includes an axial centerline; the second linear polarizing filter is circular, has a diameter smaller than a diameter of the illumination pupil, and is at least approximately coaxial with an axial centerline of the illumination pupil; 11. The video measurement system of claim 10, wherein the second linear polarizing filter is larger than the diameter of a conjugate image of the imaging pupil at the back surface of the substrate.

12. 12. The video measurement system of claim 11, wherein the second linear polarizing filter is glued to the back surface of the substrate.

13. 12. The video measurement system of claim 11, wherein the second linear polarizing filter is adhered to the front surface of the substrate.

14. 12. The video measurement system of claim 11, wherein the second linear polarizing filter is positioned between the illumination source and the first linear polarizing filter.

15. 13. The video measurement system of claim 12, wherein rotating the substrate rotates the second linear polarizing filter relative to the first linear polarizing filter.

16. 13. The video measurement system of claim 12, further comprising a mechanism for rotating the first linear polarizing filter relative to the second linear polarizing filter.

17. 1. A video measurement system for measuring a test object, comprising: (a) an imaging system comprising an imager having an imaging pupil, the imager positioned to view at least a portion of a silhouette of the test object by receiving light transmitted past the test object over a first angular range; (b) an illumination system including an illumination source, the illumination system having a second angular range greater than the first angular range received by the imaging pupil, an output, and an illumination pupil, the illumination pupil and the imaging pupil being at least approximately conjugate image planes; (c) a substrate positioned to diffuse light from the illumination source, the substrate having a front surface and a back surface, the front surface having a hole formed therein, the front surface being illuminated by the illumination source; (d) an opaque ball disposed within the hole in the front surface of the substrate and sized to block a portion of the light from the illumination source; A video measurement system comprising:

18. 20. The video measurement system of claim 17, wherein the front surface of the substrate is diffuse and the back surface of the substrate is diffuse.

19. 18. The video measurement system of claim 17, wherein the substrate is a volume diffuser.

20. 18. The video measurement system of claim 17, wherein the substrate further includes an axial centerline, and the hole is generally coaxial with the axial centerline of the substrate.

21. 20. The video measurement system of claim 18, wherein the ball is a spherical ball bearing and the hole is substantially cylindrical.

22. 18. The video measurement system of claim 17, wherein the imaging pupil and the ball each have an axial centerline, and the imaging pupil is at least approximately coaxial with the axial centerline of the ball.

23. 1. A video measurement system for measuring a test object, comprising: (a) an imaging system comprising an imager having an imaging pupil, the imager positioned to view at least a portion of a silhouette of the test object by receiving light transmitted past the test object over a first angular range; (b) an illumination system having an illumination source, a second angular range received by the imaging pupil that is greater than the first angular range, an output, and an illumination pupil, wherein the illumination pupil and the imaging pupil are at image planes that are at least approximately conjugate; (c) a substrate positioned to diffuse light from the illumination source, the substrate having a front surface and a back surface; (d) a first linear polarizing filter disposed between the illumination source and the front surface of the substrate; (e) a second linear polarizing filter overlapping the first linear polarizing filter; wherein one of the first and second linear polarizing filters is rotated relative to the other of the first and second linear polarizing filters.

24. 24. The video measurement system of claim 23, wherein the front surface of the substrate is diffuse and the back surface of the substrate is diffuse.

25. 24. The video measurement system of claim 23, wherein the substrate is a volume diffuser.

26. the illumination system further includes an illumination pupil, the illumination pupil having an axial centerline; the second linear polarizing filter is smaller than the illumination pupil and is at least approximately coaxial with the axial centerline of the illumination pupil; 24. The video measurement system of claim 23, wherein the first linear polarizing filter is larger than the imaging pupil.

27. 24. The video measurement system of claim 23, wherein the second linear polarizing filter is disposed on the back surface of the substrate, disposed between the first linear polarizing filter and the substrate, or disposed between the illumination source and the first linear polarizing filter.

Images