Multi-channel image alignment method and device based on hierarchical calibration pattern, and medium

Abstract

The invention discloses a multi-channel image alignment method based on graded calibration patterns, according to the method, a first-level calibration object and a second-level calibration object are arranged on a measurement jig, the first-level calibration object is a dotted pattern on a preset high-precision photoetching calibration plate, and the second-level calibration object is a preset low-precision calibration body of which the material is similar to and matched with a measured object. The first-level calibration object and the second-level calibration object can be scanned at the same time in a pre-scanning mode, then the position of the second-level calibration object is calculated according to the position of the first-level calibration object, and therefore the positions of the first-level calibration object and the second-level calibration object can be known. As the material of the secondary calibration object is close to that of the measured object, the secondary calibration object and the product can be imaged clearly at the same time without overexposure or over-darkness, so that the problem that the product and the calibration object cannot be shot clearly at the same time can be solved. After the position relation of the first-level calibration object and the second-level calibration object is calibrated, high-precision calibration points of the first-level calibration object at four corners are reserved, and when multi-channel alignment is carried out, images of all channels can obtain better projection transformation precision.

Description

Technical Field

[0001] The present application relates to the field of precision measurement technology, and more specifically, to a multi-channel image alignment method, device, and medium based on a graded calibration pattern. Background Art

[0002] When using a line scan camera for high-precision measurement of large objects (such as semiconductor wafers), the camera's narrow field of view requires dozens of reciprocating scans to obtain a complete image of the object. Specifically, the camera typically scans a line in the X direction, then slightly shifts the camera in the Y direction by the width of the field of view, and then scans the next line in the X direction. This reciprocating process produces multiple strip-shaped partial images of the object, which are then precisely stitched together (even without stitching, the precise positional relationship between the images must be determined for subsequent processing) to create a complete image.

[0003] Due to the start-stop triggering of the line scan camera, the starting position in the X direction will have a large offset, which causes the local image of the strip to have a large random offset in the X direction. At the same time, the accuracy of the field of view movement in the Y direction depends on the accuracy of the mechanism (generally, an accuracy below 100nm can be ignored). However, with the current level of precision measurement technology, without the use of air-floating or magnetic levitation guide rails, the superposition of systematic errors and random errors will result in a mechanism accuracy of more than 500nm. It can be seen that due to the influence of the camera's electronic control and mechanism accuracy, there will be errors in the X and Y directions of the line scan camera during the reciprocating scanning process, and these errors cannot be predicted before scanning, and therefore cannot be compensated in advance. At present, optical, software, and algorithmic compensation is generally performed post-process based on the scanning results.

[0004] Furthermore, the object under test often requires scanning and imaging using multiple optical channels, such as brightfield, darkfield, and Raman systems. Each optical system is referred to as an optical channel. If different optical channels share a single area scan camera and lens, and the product remains unchanged, simply switching the light source and capturing the next channel image, the images from the different channels are naturally aligned, and the same defect will appear in the same location across different channel images. However, in reality, these different optical channels are physically distinct optical systems and are often located at different workstations. Furthermore, different optical systems exhibit varying magnification and optical distortion. For example, if a brightfield image is taken at workstation A, followed by a darkfield image at workstation B, and then a Raman image at workstation C, the magnification and optical distortion of the brightfield, darkfield, and Raman images will differ. Sometimes, a single defect requires joint identification using different optical channels. For example, a "triangle" defect may appear as a triangle in brightfield images but a black shadow in Raman images. Both of these conditions must be met for it to be identified as a triangle. However, relying solely on traditional AI technology to locate the same defect in images from different channels is error-prone and unstable. Summary of the Invention

[0005] In response to at least one defect or improvement need in the prior art, the present application provides a multi-channel image alignment method, device and medium based on a graded calibration pattern, which is used to overcome the defects of the prior art and achieve accuracy compensation based on real-time scanning results through a new technical means, thereby improving the measurement accuracy of tiny objects to be tested and realizing accurate identification and positioning of tiny defects.

[0006] To achieve the above objectives, in a first aspect, the present application provides a multi-channel image alignment method based on a hierarchical calibration pattern, comprising:

[0007] The position mapping relationship between the primary and secondary calibration objects is obtained through pre-scanning. The primary and secondary calibration objects are set on the measurement fixture. The primary calibration object is a dot pattern on a preset high-precision photolithography calibration plate, and the secondary calibration object is a preset low-precision calibration object with a material that closely matches the object being measured.

[0008] Perform multi-line scanning imaging, alternating between two optical parameter configurations; wherein the scanning lines in the first and last regions use the first optical parameter to obtain a clear image of the primary calibration object, and the scanning lines in the middle region use the second optical parameter to obtain a clear image of the object under test and the secondary calibration object;

[0009] Building a dynamic coordinate transformation model based on the continuous scanned images of the secondary calibration object, compensating for mechanical errors in the X / Y directions, and generating a spliced ​​reference image;

[0010] Extract the corner features of the primary calibration object in the scanning lines of the first and last areas of each channel, and map each channel image to the coordinate system of the reference image through a projection transformation matrix;

[0011] The reference image is geometrically calibrated and rotated, and the transformation parameters are synchronized to all channel images to complete multi-channel image alignment.

[0012] Furthermore, obtaining the position mapping relationship between the primary calibration object and the secondary calibration object includes:

[0013] The theoretical position (x) of the secondary calibration object is obtained according to the coordinates of the marking points of the primary calibration object. i ′,y i ′), and compared with the actual scanning position (x i ,y i )Establish the error compensation equation:

[0014]

[0015] Among them, (Δx i , Δy i ) represents the actual position deviation of the secondary calibration object in the X / Y direction of the i-th calibration point; a0 and b0 represent the system translation error components; a1 and b1 represent the scaling and coupling error coefficients in the X direction; a2 and b2 represent the coupling and rotation error coefficients in the Y direction.

[0016] Furthermore, the dynamic coordinate transformation model includes:

[0017]

[0018] Among them, x n 、y n represents the coordinates of the reference coordinate system of the nth scanning line; θ represents the mechanical rotation error angle of adjacent lines; Δx n , Δy n Indicates the actual position deviation of the secondary calibration object in the X / Y direction of the nth calibration point.

[0019] Furthermore, mapping each channel image to the coordinate system of the reference image through a projection transformation matrix includes:

[0020] Extract the coordinates of the four corner points of the first-level calibration object in the scan line of the first and last area of ​​each channel, construct the homography matrix with the corresponding corner points in the reference image, and solve it through singular value decomposition:

[0021]

[0022] Where H represents a 3×3 homography matrix; (u k , v k ) represents the coordinates of the corner points in the image to be aligned; (xk ,y k ) represents the coordinates of the corresponding corner points in the reference image.

[0023] Furthermore, the difference between the first optical parameter and the second optical parameter includes:

[0024] Reduce the light source intensity by 30%-50% to avoid overexposure of the primary calibration object;

[0025] The camera exposure time is shortened to 1 / 2-1 / 3 to match the dynamic range of the object being measured;

[0026] The optical filter is switched to narrowband mode to improve the contrast of the calibration pattern.

[0027] Furthermore, performing geometric center calibration and rotation correction on the reference image includes:

[0028] Obtaining the centroid coordinates of the reference image;

[0029] Generate a translation vector according to the center of mass offset compensation;

[0030] The overall rotation error angle of the image is obtained through Radon transform;

[0031] Apply a uniform affine transformation to all channel images:

[0032]

[0033] Among them, α represents the overall rotation error angle of the image; (x c ,y c ) represents the centroid coordinates of the reference image; (δ x , δ y ) represents the center of mass offset compensation amount.

[0034] Furthermore, the design standards of the secondary calibration object meet the following requirements:

[0035] The refractive index of the material differs from that of the object being measured by less than 5%;

[0036] The surface roughness Ra value is in the range of 50-200nm;

[0037] The geometry includes at least three asymmetric raised structures to provide rotationally invariant features.

[0038] Furthermore, the production standards of the first-level calibration object meet the following requirements:

[0039] The substrate is fused quartz glass with a thermal expansion coefficient of less than 5×10-7 / ℃;

[0040] The marking line width is 2μm±0.1μm, and the edge roughness is less than 50nm;

[0041] Contains periodically arranged cross marks and pseudo-random coded dot matrix for absolute position calibration.

[0042] In a second aspect, the present application provides an electronic device comprising at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit is enabled to perform the steps of any of the multi-channel image alignment methods described above.

[0043] In a third aspect, the present application provides a storage medium storing a computer program executable by an access authentication device. When the computer program runs on the access authentication device, the access authentication device is enabled to perform the steps of any of the multi-channel image alignment methods described above.

[0044] In general, the above technical solutions conceived by this application can achieve the following beneficial effects compared with the existing technology:

[0045] The present application discloses a multi-channel image alignment method based on a graded calibration pattern, in which a primary calibration object and a secondary calibration object are set on a measuring fixture, wherein the primary calibration object is a dot pattern on a preset high-precision photolithography calibration plate, and the secondary calibration object is a preset low-precision calibration body whose material closely matches the object being measured. The primary calibration object and the secondary calibration object can be scanned at the same time by pre-scanning, and then the position of the primary calibration object is used to calculate the position of the secondary calibration object, so that the positions of the primary calibration object and the secondary calibration object are also known. Because the secondary calibration object is close to the material of the object being measured, it can be clearly imaged at the same time as the product without being overexposed or too dark, thus solving the problem of not being able to clearly photograph the product and the calibration object at the same time. After the positional relationship between the primary calibration object and the secondary calibration object is calibrated, the calibration points of the high-precision primary calibration object at the four corners are retained, so that the images of each channel can obtain better projection transformation accuracy when performing multi-channel alignment. Through such multi-channel image alignment operations, the errors in the XY directions of the scan are also compensated. A new technical means is used to achieve precision compensation based on real-time scanning results, improve the measurement accuracy of tiny objects to be tested, and achieve accurate identification and positioning of tiny defects. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0047] Figure 1A core flow chart of a multi-channel image alignment method based on hierarchical calibration patterns provided in an embodiment of the present application;

[0048] Figure 2 Schematic diagram of the primary and secondary calibration objects provided in the embodiments of the present application;

[0049] Figure 3 This is the image before bright field stitching provided in the embodiment of the present application;

[0050] Figure 4 The image before dark field stitching provided in the embodiment of the present application;

[0051] Figure 5 The bright field stitching image provided in the embodiment of the present application;

[0052] Figure 6 The dark field stitching image provided in the embodiment of the present application;

[0053] Figure 7 The image provided in the embodiment of the present application after bright field stitching and alignment with other channels;

[0054] Figure 8 The image after dark field stitching and alignment with other channels provided in the embodiment of the present application;

[0055] Figure 9 This is a graph after all four channels, namely bright field, dark field, PL NIR (photoluminescence, infrared band), and PL VIS (photoluminescence, visible light band), are aligned, as provided in the embodiments of the present application.

[0056] Figure 10 A detailed diagram of the alignment of a "triangle defect" provided in an embodiment of the present application;

[0057] Figure 11 A block diagram of an electronic device suitable for implementing the multi-channel image alignment method described above is provided in an embodiment of the present application. DETAILED DESCRIPTION

[0058] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely for the purpose of explaining this application and are not intended to limit this application. In addition, the technical features involved in the various embodiments of this application described below may be combined with each other as long as they do not conflict with each other.

[0059] The terms "first," "second," or "nth" in the specification, claims, or drawings of this application may be used to distinguish different objects or to describe a specific order, depending on the specific scenario. In addition, the terms "including" or "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products, or devices.

[0060] In response to at least one of the technical problems existing in the prior art described in the background technology section, the present application provides a multi-channel image alignment method, device and medium based on graded calibration patterns, which are used to overcome the defects of the prior art and achieve accuracy compensation based on real-time scanning results through a new technical means, thereby improving the measurement accuracy of tiny objects to be tested and realizing accurate identification and positioning of tiny defects.

[0061] Traditional AI technology, used to locate the same defect in images from different channels, is error-prone and unstable. To effectively address this issue, processing can be performed at the source of the optical image. Specifically, images acquired from different channels can be aligned. This function is called multi-channel image alignment. Multi-channel image alignment is closely related to compensating for errors in the XY direction of the scan.

[0062] This application uses calibration plates and calibration objects to solve this problem. Among them, these calibration objects are produced by photolithography on a high-precision glass plate, and the accuracy can reach about 10nm. For practical applications, this level of accuracy is excessive and can be considered as a gold standard and its error can be ignored. The calibration plate is a glass plate containing the calibration object, which is fixedly mounted on a measuring fixture. During measurement, the product to be measured is placed on the fixture, and the product to be measured and the calibration plate are scanned at the same time. Due to reasons such as product placement error, the position of the product is not fixed, but since the position of the calibration plate is fixed, the precise position of the product can be obtained based on the image of the calibration plate.

[0063] Furthermore, in practical applications, high-precision calibration plates must be produced on glass plates using a photolithography process. If the material of the product being tested differs significantly from that of the calibration plate, it can be difficult to obtain clear images of both simultaneously. This can result in the product being tested being clearly imaged while the calibration plate is overexposed. Furthermore, when the calibration plate is clearly imaged, defects in the product being tested may be unclear. To address this issue, this application proposes the following solution.

[0064] refer to Figures 1-10One embodiment of the present application proposes a multi-channel image alignment method based on a hierarchical calibration pattern. The multi-channel image alignment method may include the following steps.

[0065] Step 1: Pre-scan to obtain the position mapping relationship between the primary and secondary calibration objects; the primary and secondary calibration objects are set on the measurement fixture. Among them, the primary calibration object is a dot pattern on a preset high-precision photolithography calibration plate, and the secondary calibration object is a preset low-precision calibration object with a material that closely matches the object being measured.

[0066] Specifically, select an object with a material similar to that of the product to be tested (such as a cylindrical silicon support column or calibration column), and use ordinary technology to make a low-precision secondary calibration object (a low-precision calibration column, i.e. Figure 2 The left edge is a piano key pattern), which is fixed on the periphery of the measuring fixture of the product to be tested. After installation, it is fixed and the position accuracy is low. However, it can scan the primary calibration object (the MARK pattern on the calibration plate made by photolithography) at the same time through pre-scanning. Figure 2 The position of the primary calibration object can be used to calculate the position of the secondary calibration object. At this point, the positions of both the primary and secondary calibration objects are known. Because the secondary calibration object is made of a similar material to the product under test, it can be clearly imaged simultaneously with the product under test without overexposure or darkening. This solves the problem of not being able to clearly capture both the product under test and the calibration object.

[0067] Figure 2 The densely arranged small dots in the middle are primary calibration objects, which are made by photolithography on glass with an accuracy of less than 100nm. Figure 2 The piano-key-shaped object on the left is a secondary calibration object, machined to an accuracy of approximately 10μm. Because the dots on the primary calibration object are produced using a photolithography process, their precision is very high, while the accuracy of the secondary calibration object is relatively low.

[0068] The piano-key-shaped secondary calibration object is made of a similar or identical material to the circular wafer in the middle (i.e., the product or object to be tested). Therefore, under different lighting conditions, as long as the wafer in the middle is visible, the piano-key-shaped secondary calibration object at the edge will also be visible. Moreover, since the secondary calibration object does not require high precision, it can be directly machined using the same or similar material as the product to be tested, resulting in a wide range of material options. However, due to the limited precision requirements, the primary calibration object in the form of a small dot in the middle cannot be arbitrarily changed in material or rough-machined. More careful material selection and photolithography accuracy must be ensured.

[0069] The secondary calibration object is made of a material that is the same or similar to the material of the object being measured. For example, if the object being measured is a silicon wafer, the column of the secondary calibration object is made of silicon; if the object being measured is 304 stainless steel, the column of the secondary calibration object is made of 304 stainless steel; if the object being measured is a silicon carbide wafer, considering that the preparation of high-purity silicon carbide is very difficult and the cost is very high, the secondary calibration object can be replaced by ion-doped yellow-green calcium sodium glass. In short, the imaging characteristics of the column of the secondary calibration object under each optical system must be the same or similar to those of the object being measured. More specifically, the design standards of the secondary calibration object meet the following requirements:

[0070] The refractive index of the secondary calibration material should differ by less than 5% from that of the object being measured. The refractive index of the silicon nitride ceramic secondary calibration material used for glass panel testing is set to 1.98, which is approximately 3% different from the glass refractive index of 1.92.

[0071] The surface roughness Ra value is in the range of 50-200 nm, preferably, the surface roughness Ra=85 nm, and sandblasting is performed.

[0072] The geometric shape comprises at least three asymmetric raised structures to provide rotationally invariant features. Preferably, the side lengths of the triangular raised structures are 1 mm / 2 mm / 3 mm.

[0073] The secondary calibration object is made by machining, with an accuracy usually above 10μm, and is relatively easy to manufacture.

[0074] The selection of the material and structure of the secondary calibration object can reduce optical distortion by 72% and the rotation recognition accuracy can reach 99.5%.

[0075] The production standards of first-level calibration objects meet the following requirements:

[0076] The substrate is fused quartz glass with a thermal expansion coefficient of less than 5×10-7 / ℃;

[0077] The marking line width is 2μm±0.1μm, and the edge roughness is less than 50nm;

[0078] Contains periodically arranged cross marks and pseudo-random coded dot matrix for absolute position calibration.

[0079] More specifically, the primary calibration object is a dot pattern on a calibration plate used by a photolithography machine. The specific technical parameters of the calibration plate include:

[0080] Substrate: Corning 7980 fused quartz (thermal expansion coefficient 0.05×10-6 / ℃)

[0081] Marking: Chrome crosshairs (line width 2μm±0.05μm)

[0082] Contains 1024 pseudo-random points (position accuracy ±0.1μm)

[0083] The first-level calibration object provides 0.05μm absolute positioning reference, and the temperature fluctuation effect is <0.008μm / ℃

[0084] When installing the primary calibration plate (the calibration plate on which the primary calibration object is located), affix a fused silica calibration plate with photolithographic markings (primary calibration plate) to the scanning system reference plane. The plate must be sized to cover the maximum displacement range of the object under test and, optionally, extend 20% beyond the object boundary. The primary calibration plate should contain a high-contrast pseudo-random dot matrix with, optionally, 1024 marking points and a line width of 2.05μm ± 0.08μm.

[0085] The secondary calibration object is constructed from alumina ceramic with a refractive index difference of less than 5% from the object being measured, and a surface roughness Ra of 50-200nm. Three asymmetrical raised calibration objects are symmetrically mounted on the edge of the measurement fixture. Their geometry incorporates a helical structure to provide rotational invariance.

[0086] Obtain the position mapping relationship between the primary calibration object and the secondary calibration object. After obtaining the position mapping relationship, the position of the primary calibration object can be used to calculate the position of the secondary calibration object. At this time, the positions of the primary calibration object and the secondary calibration object are also known. Obtaining the position mapping relationship between the primary calibration object and the secondary calibration object includes:

[0087] The theoretical position (x) of the secondary calibration object is obtained according to the coordinates of the marking points of the primary calibration object. i ′,y i ′), and compared with the actual scanning position (x i ,y i )Establish the error compensation equation:

[0088]

[0089] Among them, (Δx i , Δy i ) represents the actual position deviation of the secondary calibration object in the X / Y direction of the i-th calibration point; a0 and b0 represent the system translation error components; a1 and b1 represent the scaling and coupling error coefficients in the X direction; a2 and b2 represent the coupling and rotation error coefficients in the Y direction.

[0090] parameter Physical meaning Typical values Measurement method <![CDATA[a0]]> X-axis reference deviation 0.15μm Laser interferometer measurement <![CDATA[a1]]> Thermal expansion coefficient in X direction <![CDATA[4.3e -6 / ℃]]> Constant temperature box gradient test <![CDATA[a2]]> XY axis coupling coefficient 0.002 Orthogonal grid calibration plate scanning

[0091] In a certain PCB inspection system, after 5 pre-scans, it was measured that: a0 = 0.12 μm (reflecting the installation deviation of the carrier), a1 = 3.8e -6 / ℃ (corresponding to the material properties of the aluminum alloy stage), the position error can be reduced from ±2.1μm to ±0.3μm after compensation.

[0092] After calibrating the positional relationship between the primary and secondary calibration objects, it is best not to cancel the primary calibration object. This is because the high-precision primary calibration points at the four corners can help achieve better projection transformation accuracy for each channel image during multi-channel alignment. For more details, please refer to the other steps below.

[0093] Step 2: Perform multi-line line scanning imaging, alternating between two optical parameter configurations; wherein, the scanning lines in the first and last areas use the first optical parameter to obtain a clear image of the primary calibration object, and the scanning lines in the middle area use the second optical parameter to obtain a clear image of the object being measured and the secondary calibration object.

[0094] Specifically, the difference between the first optical parameter and the second optical parameter includes:

[0095] Reduce the light source intensity by 30%-50% to avoid overexposure of the primary calibration object.

[0096] The camera exposure time is shortened to 1 / 2-1 / 3 to match the dynamic range of the object being measured.

[0097] The optical filter is switched to narrowband mode to improve the contrast of the calibration pattern.

[0098] More specifically, in a certain semiconductor wafer inspection system, a 20cm×30cm fused quartz primary calibration plate (line width 2.1μm) and an alumina ceramic secondary calibration object (refractive index difference 3.5%) are used. The scanning process uses a line array camera to perform multispectral imaging at a speed of 500 lines / second:

[0099] When scanning from the beginning to the end, use a 450nm narrowband filter, adjust the light source intensity to 700 lux, and use an exposure time of 3ms. The high-precision calibration plate (primary calibration plate) should be larger than the product to be tested, and ensure that the first and last 5% of the scan lines are scanned (preferably the first and last lines, as long as the beginning and end are clearly scanned). The product to be tested should not be visible during scanning. When scanning the high-precision calibration plate, use the primary optical parameters that can clearly capture the high-precision calibration plate to ensure the image quality of the primary calibration object.

[0100] When scanning in the middle, switch to full spectrum mode, increase the light intensity to 1200 lux, and extend the exposure time to 10ms. In addition to scanning the first and last rows mentioned above, all other rows are scanned using the secondary optical parameters that are most conducive to clearly capturing the product under test, while also obtaining a clear image of the secondary calibration object.

[0101] Step 3: Build a dynamic coordinate transformation model based on the continuous scanned images of the secondary calibration object to compensate for the mechanical errors in the X / Y directions and generate a spliced ​​reference image. The dynamic coordinate transformation model includes:

[0102]

[0103] Among them, x n 、y n represents the coordinates of the reference coordinate system of the nth scanning line; θ represents the mechanical rotation error angle of adjacent lines; Δx n , Δy n Indicates the actual position deviation of the secondary calibration object in the X / Y direction of the nth calibration point.

[0104] parameter Physical meaning Measurement methods Typical values θ Rotational error caused by screw backlash Encoder differential measurement 0.0025rad <![CDATA[Δx n ]]> Displacement deviation caused by guide rail wear Laser displacement meter online monitoring ±0.18μm

[0105] In the OLED screen scanning device, after scanning every 10 lines, a cumulative error of θ = 0.0032 rad (about 0.18°) is detected. After dynamic compensation, the splicing misalignment can be reduced from 3 pixels to 0.2 pixels.

[0106] Step 4: Extract the corner features of the primary calibration object in the scan lines of the first and last regions of each channel, and map each channel image to the coordinate system of the reference image through a projection transformation matrix, specifically including:

[0107] Extract the coordinates of the four corner points of the first-level calibration object in the scan line of the first and last area of ​​each channel, construct the homography matrix with the corresponding corner points in the reference image, and solve it through singular value decomposition:

[0108]

[0109] Where H represents a 3×3 homography matrix; (u k , v k ) represents the coordinates of the corner points in the image to be aligned; (x k ,y k ) represents the coordinates of the corresponding corner points in the reference image.

[0110] parameter Physical meaning How to obtain Accuracy requirements uk UV channel corner coordinates SIFT feature extraction ±0.1 pixel xk Visible light reference coordinates Template matching ±0.05 pixels

[0111] In a certain material analysis microscope, four-channel (bright field / dark field / fluorescence / Raman) images are aligned using this homography matrix. The homography matrix calculation takes less than 8ms, and the cross-channel registration error is less than 0.15 pixels.

[0112] Step 5: Perform geometric center calibration and rotation correction on the reference image, and synchronize the transformation parameters to all channel images to complete multi-channel image alignment. Performing geometric center calibration and rotation correction on the reference image includes:

[0113] Get the centroid coordinates of the reference image.

[0114] Generates a translation vector based on the center of mass offset compensation.

[0115] The overall image rotation error angle is obtained through Radon transform.

[0116] Apply a uniform affine transformation to all channel images:

[0117]

[0118] Among them, α represents the overall rotation error angle of the image; (x c ,y c ) represents the centroid coordinates of the reference image; (δ x , δ y ) represents the center of mass offset compensation amount.

[0119] parameter Physical meaning Measurement method Typical values α Rotation angle caused by stage tilt Dual-frequency laser interferometer 0.0017rad δx Camera installation eccentricity Cross-hair calibration 1.2 pixels

[0120] In a specific embodiment, through this correction, a wafer inspection machine can achieve a center of mass positioning accuracy of 0.03 pixels, and a full-width rotation error of a 300mm wafer is less than 0.003°.

[0121] refer to Figure 10 , a phenomenon known as the "triangle defect." Figure 10 The image on the left shows the appearance of a triangular defect under a bright field microscope. Figure 10 The middle image shows its morphology in the photoluminescence infrared band, and Figure 10 The image on the right shows its morphology in the photoluminescence visible light band.

[0122] exist Figure 9 In the four-channel image, from left to right: the first channel is bright field, the second channel is dark field, the third channel is photoluminescence (infrared band), and the fourth channel is photoluminescence (visible light band). Figure 10 Detailed diagram of triangle defect alignment, Figure 10 The first channel (sorted from left to right, the same as the attached figure) corresponds to Figure 9 The first channel; Figure 10 The second channel corresponds to Figure 9 The third channel; Figure 10 The third channel corresponds to Figure 9 The fourth channel.

[0123] This "triangular defect" appears differently in different channel images, but physically it is the same defect. Initially, its position is different in the different channel images. After using the alignment method of the embodiment of the application, it can be aligned in the corrected different channel images.

[0124] The present application discloses a multi-channel image alignment method based on a graded calibration pattern, in which a primary calibration object and a secondary calibration object are set on a measuring fixture, wherein the primary calibration object is a dot pattern on a preset high-precision photolithography calibration plate, and the secondary calibration object is a preset low-precision calibration body whose material closely matches the object being measured. The primary calibration object and the secondary calibration object can be scanned at the same time by pre-scanning, and then the position of the primary calibration object is used to calculate the position of the secondary calibration object, so that the positions of the primary calibration object and the secondary calibration object are also known. Because the secondary calibration object is close to the material of the object being measured, it can be clearly imaged at the same time as the product without being overexposed or too dark, thus solving the problem of not being able to clearly photograph the product and the calibration object at the same time. After the positional relationship between the primary calibration object and the secondary calibration object is calibrated, the calibration points of the high-precision primary calibration object at the four corners are retained, so that the images of each channel can obtain better projection transformation accuracy when performing multi-channel alignment. Through such multi-channel image alignment operations, the errors in the XY directions of the scan are also compensated. A new technical means is used to achieve precision compensation based on real-time scanning results, improving the measurement accuracy of tiny objects to be tested and achieving accurate identification and positioning of tiny defects.

[0125] Figure 11 The block diagram of an electronic device suitable for implementing the multi-channel image alignment method described above according to an embodiment of the present application is schematically shown. Figure 11 The electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present application.

[0126] like Figure 11 As shown, the electronic device 1000 described in this embodiment includes: a processor 1001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage part 1008 into a random access memory (RAM) 1003. The processor 1001 may, for example, include a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (for example, an application-specific integrated circuit (ASIC)), etc. The processor 1001 may also include an on-board memory for caching purposes. The processor 1001 may include a single processing unit or multiple processing units for executing different actions of the multi-channel image alignment method process according to an embodiment of the present application.

[0127] In RAM 1003, various programs and data required for the operation of electronic device 1000 are stored. Processor 1001, ROM 1002 and RAM 1003 are connected to each other via bus 1004. Processor 1001 performs various operations of the multi-channel image alignment method according to the embodiment of the present application by executing the programs in ROM 1002 and / or RAM 1003. It should be noted that the program can also be stored in one or more memories other than ROM 1002 and RAM 1003. Processor 1001 can also perform various operations of the multi-channel image alignment method according to the embodiment of the present application by executing the programs stored in the one or more memories.

[0128] According to an embodiment of the present application, electronic device 1000 may further include an input / output (I / O) interface 1005, which is also connected to bus 1004. Electronic device 1000 may further include one or more of the following components connected to I / O interface 1005: an input portion 1006 including a keyboard, mouse, etc.; an output portion 1007 including devices such as a cathode ray tube (CRT), liquid crystal display (LCD), and speakers; a storage portion 1008 including a hard disk; and a communication portion 1009 including a network interface card such as a LAN card or modem. Communication portion 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to I / O interface 1005 as needed. Removable media 1011, such as a magnetic disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed in drive 1010 as needed, so that computer programs read from the removable media can be installed into storage portion 1008 as needed.

[0129] The multi-channel image alignment method process according to an embodiment of the present application can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program contains program code for executing the multi-channel image alignment method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1009, and / or installed from the removable medium 1011. When the computer program is executed by the processor 1001, the above-mentioned functions defined in the system of the embodiment of the present application are executed. According to an embodiment of the present application, the systems, devices, means, modules and / or units described above can be implemented by computer program modules.

[0130] Embodiments of the present application also provide a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments, or may exist independently and not be incorporated into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the steps of the multi-channel image alignment method according to the embodiments of the present application.

[0131] According to an embodiment of the present application, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, it may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In an embodiment of the present application, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, apparatus, or device. For example, according to an embodiment of the present application, the computer-readable storage medium may include one or more memories other than the ROM 1002 and / or RAM 1003 described above.

[0132] It should be noted that the functional modules in the various embodiments of the present application can be integrated into a processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above-mentioned integrated modules can be implemented in the form of hardware or in the form of software functional modules. If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product.

[0133] The flowcharts and / or block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flowchart and / or block diagram can represent a module, a program segment or a part of code, and the part of the above-mentioned module, program segment or code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. It should also be noted that each box in the block diagram or flowchart, and the combination of the boxes in the block diagram or flowchart, can be implemented by a special hardware-based system that performs the specified function or operation, or can be implemented by a combination of special hardware and computer instructions.

[0134] Those skilled in the art will appreciate that the features described in the various embodiments and / or claims of this application may be combined and / or coupled in various ways, even if such combinations and / or couplings are not explicitly described in this application. In particular, without departing from the spirit and teachings of this application, the technical features described in the various embodiments and / or claims of this application may be combined and / or coupled in various ways, and all such combinations and / or couplings fall within the scope of this application.

[0135] Although the present application has been shown and described with reference to certain exemplary embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made to the present application without departing from the spirit and scope of the present application as defined by the appended claims and their equivalents. Therefore, the scope of the present application should not be limited to the above-described embodiments, but should be determined not only by the appended claims but also by the equivalents of the appended claims.

Claims

1. A multi-channel image alignment method based on hierarchical calibration patterns, characterized in that: include: The position mapping relationship between the primary calibration object and the secondary calibration object is obtained through pre-scanning; The primary and secondary calibration objects are set on the measuring fixture. The primary calibration object is a dot pattern on a preset high-precision photolithography calibration plate, and the secondary calibration object is a preset low-precision calibration object whose material closely matches the object being measured. Perform multi-line scanning imaging, alternating between two optical parameter configurations; wherein the scanning lines in the first and last regions use the first optical parameter to obtain a clear image of the primary calibration object, and the scanning lines in the middle region use the second optical parameter to obtain a clear image of the object under test and the secondary calibration object; Building a dynamic coordinate transformation model based on the continuous scanned images of the secondary calibration object, compensating for mechanical errors in the X / Y directions, and generating a spliced ​​reference image; Extract the corner features of the primary calibration object in the scanning lines of the first and last areas of each channel, and map each channel image to the coordinate system of the reference image through a projection transformation matrix; The reference image is geometrically calibrated and rotated, and the transformation parameters are synchronized to all channel images to complete multi-channel image alignment.

2. The multi-channel image alignment method according to claim 1, wherein: Obtaining the position mapping relationship between the primary calibration object and the secondary calibration object includes: The theoretical position (x) of the secondary calibration object is obtained according to the coordinates of the marking points of the primary calibration object. i ′,y i ′), and compared with the actual scanning position (x i ,y i )Establish the error compensation equation: Among them, (Δx i , Δy i ) represents the actual position deviation of the secondary calibration object in the X / Y direction of the i-th calibration point; a0 and b0 represent the system translation error components; a1 and b1 represent the scaling and coupling error coefficients in the X direction; a2 and b2 represent the coupling and rotation error coefficients in the Y direction.

3. The multi-channel image alignment method according to claim 1, wherein: The dynamic coordinate transformation model includes: Among them, x n 、y n represents the coordinates of the reference coordinate system of the nth scanning line; θ represents the mechanical rotation error angle of adjacent lines; Δx n , Δy n Indicates the actual position deviation of the secondary calibration object in the X / Y direction of the nth calibration point.

4. The multi-channel image alignment method according to claim 1, wherein: Mapping each channel image to the coordinate system of the reference image through the projection transformation matrix includes: Extract the coordinates of the four corner points of the first-level calibration object in the scan lines of the first and last areas of each channel, construct a homography matrix with the corresponding corner points in the reference image, and solve it through singular value decomposition: Where H represents the 3×3 homography matrix; (u k , v k ) represents the coordinates of the corner points in the image to be aligned; (x k ,y k ) represents the coordinates of the corresponding corner points in the reference image.

5. The multi-channel image alignment method according to claim 1, wherein: The difference between the first optical parameter and the second optical parameter includes: Reduce the light source intensity by 30%-50% to avoid overexposure of the primary calibration object; The camera exposure time is shortened to 1 / 2-1 / 3 to match the dynamic range of the object being measured; The optical filter is switched to narrowband mode to enhance the contrast of the calibration pattern.

6. The multi-channel image alignment method according to claim 1, wherein: Performing geometric center calibration and rotation correction on the reference image includes: Obtaining the centroid coordinates of the reference image; Generate a translation vector according to the center of mass offset compensation; The overall rotation error angle of the image is obtained through Radon transform; Apply a uniform affine transformation to all channel images: Among them, α represents the overall rotation error angle of the image; (x c ,y c ) represents the centroid coordinates of the reference image; (δ x , δ y ) represents the center of mass offset compensation amount.

7. The multi-channel image alignment method according to claim 1, wherein: The design standards of the secondary calibration object meet the following requirements: The refractive index of the material differs from that of the object being measured by less than 5%; The surface roughness Ra value is in the range of 50-200nm; The geometry includes at least three asymmetric raised structures to provide rotationally invariant features.

8. The multi-channel image alignment method according to claim 1, wherein: The production standards of the first-level calibration objects meet the following requirements: The substrate is fused quartz glass with a thermal expansion coefficient of less than 5×10-7 / ℃; The marking line width is 2μm±0.1μm, and the edge roughness is less than 50nm; Contains periodically arranged cross marks and pseudo-random coded dot matrix for absolute position calibration.

9. An electronic device, characterized in that: The method comprises at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit is enabled to perform the steps of the multi-channel image alignment method according to any one of claims 1 to 8.

10. A storage medium, characterized in that: It stores a computer program executable by an access authentication device. When the computer program runs on the access authentication device, the access authentication device is enabled to perform the steps of the multi-channel image alignment method according to any one of claims 1 to 8.

Images