Inspection device and inspection method
The inspection apparatus and method address misalignment issues by selecting a suitable registration pattern for alignment, ensuring accurate defect detection and optimizing data capacity through composite image synthesis.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- TORAY ENG CO LTD
- Filing Date
- 2024-10-24
- Publication Date
- 2026-05-12
AI Technical Summary
Existing inspection methods face misalignment issues due to poor positioning accuracy of the stage and imaging means, leading to false defect detection when comparing inspection images with reference images.
An inspection apparatus and method that selects a suitable registration pattern for alignment by detecting patterns on the object, aligning inspection images with reference images based on pattern positions, and synthesizing composite images to determine variation, allowing for the selection of an optimal registration pattern.
Enables accurate alignment and detection of defects by selecting the most suitable registration pattern, reducing false defect detection and optimizing data capacity through efficient image synthesis.
Smart Images
Figure 2026076823000001_ABST
Abstract
Description
[Technical Field]
[0001] This disclosure relates to an inspection device and an inspection method. [Background technology]
[0002] As shown in Patent Document 1, an inspection method is known in which an object to be inspected is inspected by comparing an inspection image obtained by imaging the object with a reference image. [Prior art documents] [Patent Documents]
[0003] [Patent Document 1] Japanese Patent Application Publication No. 7-325046 [Overview of the project] [Problems that the invention aims to solve]
[0004] In this type of inspection method, the object to be inspected is divided into multiple fields of view, and the imaging device is moved relative to the object to be inspected while sequentially imaging the object in each field of view with the imaging device, thereby obtaining an image for each field of view.
[0005] The captured images are cropped for each field of view to obtain inspection images for each field of view. The inspection images obtained for each field of view are compared with reference images that have been pre-generated for each field of view. This allows the object under inspection to be inspected, and for example, defects occurring in the object under inspection can be detected.
[0006] If the positioning accuracy of the stage supporting the object under inspection or the imaging means is poor, a positional misalignment will occur between the inspection image being compared and the reference image.
[0007] To prevent such misalignment, a registration pattern for alignment is created for each field of view in the reference image. The acquired image (field of view) is then cropped based on the position of the registered pattern in the reference image to obtain the inspection image. The obtained inspection image is then aligned with the reference image and compared with it.
[0008] The registration pattern for alignment should preferably be as unique as possible from among multiple patterns formed on the object under inspection, in order to avoid misidentification with other patterns. Furthermore, the registration pattern for alignment should preferably be placed at locations on the object under inspection where the manufacturing error for each product is large.
[0009] The purpose of this disclosure is to suitably select a registration pattern for alignment used when comparing an inspection image with a reference image in an inspection method that inspects an object to be inspected by comparing an inspection image obtained by imaging the object to be inspected with a reference image. [Means for solving the problem]
[0010] The inspection apparatus according to this disclosure is an inspection apparatus that inspects an object to be inspected by comparing an inspection image obtained by imaging the object to be inspected with a reference image, and comprises an imaging unit for imaging the object to be inspected and a processing unit, wherein the processing unit registers a pattern for alignment from among a plurality of patterns formed on the object to be inspected as a registered pattern in the reference image, detects a pattern similar to the registered pattern from among a plurality of patterns formed on the object to be inspected as a detection pattern in the inspection image obtained by imaging the object to be inspected, aligns the inspection image to the reference image based on the position of the detected pattern in the inspection image and the position of the registered pattern in the reference image, and synthesizes a composite image by superimposing a plurality of captured images obtained by imaging the object to be inspected for each product, based on a provisionally registered pattern provisionally registered from among a plurality of patterns formed on the object to be inspected, and determines the variation among the plurality of captured images constituting the composite image in order to determine whether the provisionally registered pattern is suitable as the registered pattern.
[0011] To determine whether a provisional registration pattern is suitable as a registered pattern, the variation among the multiple captured images that make up the composite image is calculated. This allows for the selection of a suitable registration pattern for alignment when comparing the inspection image with a reference image.
[0012] In an inspection device that inspects an object to be inspected by comparing an inspection image obtained by imaging the object with a reference image, it is possible to suitably select a registration pattern for alignment used when comparing the inspection image and the reference image.
[0013] In one embodiment, the plurality of captured images are stored in a memory, the provisional registration patterns include a first provisional registration pattern and a second provisional registration pattern, and the composite image includes a first composite image synthesized by overlapping the plurality of captured images with each other based on the first provisional registration pattern, and a second composite image synthesized by overlapping the plurality of captured images with each other based on the second provisional registration pattern. The first composite image and the second composite image are synthesized using the common plurality of captured images stored in the memory.
[0014] Since both the first composite image and the second composite image are synthesized using the common plurality of captured images stored in the memory, the data capacity can be saved as compared with the case where the first composite image and the second composite image are synthesized using separate captured images.
[0015] In one embodiment, the inspection device includes a display unit, the variation includes a first variation among the plurality of captured images constituting the first composite image and a second variation among the plurality of captured images constituting the second composite image, and the display unit displays the first provisional registration pattern, the second provisional registration pattern, the first variation of the first composite image, and the second variation of the second composite image.
[0016] The provisional registration pattern and the variation of the corresponding composite image can be visualized by the display unit.
[0017] In one embodiment, the provisional registration pattern selected by the user from the first provisional registration pattern and the second provisional registration pattern displayed on the display unit is registered as the registration pattern.
[0018] The user can select the most suitable provisional registration pattern as the registration pattern from among the plurality of provisional registration patterns while looking at the display unit.
[0019] In one embodiment, the variation includes a first variation among the plurality of captured images constituting the first composite image and a second variation among the plurality of captured images constituting the second composite image. When the first variation is smaller than the second variation, the first provisional registration pattern is applied as the registration pattern.
[0020] Based on the magnitude of the variation, it is possible to easily determine which provisional registration pattern should be used as the registration pattern.
[0021] In one embodiment, the variation in brightness for each pixel among the plurality of captured images constituting the composite image is obtained.
[0022] The variation can be easily obtained.
[0023] In one embodiment, the variation includes at least any one of an average value, a standard deviation, a variance value, and a coefficient of variation.
[0024] The variation can be easily obtained.
[0025] The inspection method relating to this disclosure is an inspection method for inspecting an object to be inspected by comparing an inspection image obtained by imaging the object to be inspected with a reference image, wherein in the reference image, a pattern for alignment is registered as a registered pattern from among a plurality of patterns formed on the object to be inspected, and in the inspection image obtained by imaging the object to be inspected, a pattern similar to the registered pattern is detected as a detection pattern from among a plurality of patterns formed on the object to be inspected, and the inspection image is aligned with the reference image based on the position of the detected pattern in the inspection image and the position of the registered pattern in the reference image, and in the inspection method, a composite image is synthesized by superimposing a plurality of captured images obtained by imaging the object to be inspected for each product, based on a provisionally registered pattern provisionally registered from among a plurality of patterns formed on the object to be inspected, and in order to determine whether the provisionally registered pattern is suitable as the registered pattern, the variation among the plurality of captured images constituting the composite image is determined, and the provisionally registered pattern is registered as the registered pattern based on the variation.
[0026] In the inspection method described above, which involves inspecting an object to be inspected by comparing an inspection image obtained by imaging the object with a reference image, a registration pattern for alignment used when comparing the inspection image and the reference image can be suitably selected. [Effects of the Invention]
[0027] According to this disclosure, in an inspection method that inspects an object to be inspected by comparing an inspection image obtained by imaging the object to be inspected with a reference image, the registration pattern for alignment used when comparing the inspection image and the reference image can be suitably selected. [Brief explanation of the drawing]
[0028] [Figure 1] Figure 1 shows an inspection device that inspects an object to be inspected by comparing the inspection image obtained from imaging the object with a reference image. [Figure 2]Figure 2 shows the alignment of the examination image and the reference image. [Figure 3] Figure 3 shows the selection of registration patterns. [Figure 4] Figure 4 shows the display unit. [Figure 5] Figure 5 shows the method for determining the variability. [Figure 6] Figure 6 shows a flowchart for determining the registration pattern. [Modes for carrying out the invention]
[0029] Hereinafter, an embodiment of the present disclosure will be described in detail with reference to the drawings. The following description of preferred embodiments is essentially illustrative and is not intended in any way to limit the present disclosure, its applications or uses.
[0030] (Inspection device) Figure 1 shows the inspection device 1. Figure 2 shows the alignment of the inspection image J and the reference image K. The inspection device 1 inspects the object to be inspected A by comparing the inspection image J, obtained by imaging the object to be inspected A, with the reference image K.
[0031] The object under inspection A is, for example, the surface of a chip on a substrate W. Multiple chips are provided on the surface of the substrate W. The substrate W is, for example, a semiconductor wafer.
[0032] Inspection device 1 detects defects occurring in the object A under inspection. Examples of defects include foreign matter, scratches, poor edge rinsing, through-hole abnormalities, uneven coating, peeling, scratches, bonding pad abnormalities, color unevenness, pattern abnormalities, stains, discoloration, deformation, etc.
[0033] The inspection device 1 comprises an imaging unit 10, a processing unit 20, and a display unit 30.
[0034] The imaging unit 10 is, for example, a camera. The imaging unit 10 takes an image of the object A to be inspected.
[0035] The processing unit 20 is, for example, a computer. The processing unit 20 includes, for example, a processor mounted on a circuit board and memory for storing software to operate the processor.
[0036] In the inspection device 1, the object to be inspected A is divided into multiple fields Fj, and the imaging unit 10 is moved relative to the object to be inspected A, while the object to be inspected A is sequentially imaged by the imaging unit 10 for each field Fj, thereby obtaining an image for each field Fj.
[0037] The processing unit 20 of the inspection device 1 trims the outer edge of the captured image for each field of view Fj to obtain an inspection image J for each field of view Fj. The processing unit 20 of the inspection device 1 compares the inspection image J obtained by imaging the object to be inspected A with the reference image K.
[0038] Reference image K is also called a fitting image or statistical good image. Reference image K is obtained, for example, by averaging images of the inspected object A that has been judged as good. Reference image K is set in advance. Reference image K is created in advance for each field of view Fk of the imaging unit 10 in the same manner as inspection image J.
[0039] The visual field Fj used to obtain the examination image J (hereinafter referred to as "examination visual field Fj") and the visual field Fk used to obtain the reference image K (hereinafter referred to as "reference visual field Fk") are the same size and have the same shape (a rectangle in this example).
[0040] The object under inspection A is inspected by comparing the inspection image J with the reference image K. For example, defects occurring in the object under inspection A are detected.
[0041] However, if the positioning accuracy of the stage 2 supporting the object under inspection A or the imaging unit 10 is poor, a misalignment will occur between the inspection field Fj and the reference field Fk. As a result, a misalignment will also occur between the inspection image J and the reference image K. In such misaligned areas, false defects (excess defects) may be detected.
[0042] To prevent such misalignment, the processing unit 20 pre-registers, as registered pattern Pk, a pattern P for alignment purposes from among multiple patterns P formed on the object A under inspection in the reference image K for each reference field of view Fk.
[0043] Then, for each field of view Fj, the processing unit 20 detects a pattern P similar to the registered pattern Pk from among multiple patterns P formed on the object A under inspection in the inspection image J, and uses this as the detected pattern Pj.
[0044] In Figure 2, in the inspection field Fj and reference field Fk, the gray areas represent the base material of the object under inspection A (plain, for example, the surface of a semiconductor substrate chip), while the white or black areas represent areas that may contain some kind of pattern P.
[0045] Pattern P can be various things, such as a wiring pattern, an integrated circuit shape pattern, or a substrate shape pattern W.
[0046] The detection of detection pattern Pj in subject A is performed within a pattern matching search range D, which is narrower than the inspection field Fj and inspection image J, but wider than the detection pattern Pj. In other words, in subject A, detection pattern Pj is detected within the pattern matching search range D.
[0047] The processing unit 20 aligns the inspection image J with the reference image K (reference image Fk) based on the position Lj of the detected pattern Pj in the inspection image J (inspection field Fj) and the position Lk of the registered pattern Pk in the reference image K (reference field Fk) (the amount of positional shift). In other words, the inspection image J is aligned with the reference image K based on the amount of positional shift of the detected pattern Pj relative to the registered pattern Pk. This type of alignment method is called template matching.
[0048] Specifically, the outer edge of the examination field (imaging image) Fj is trimmed to obtain examination image J, and the obtained examination image J is aligned with reference image K. Then, examination image J is compared with reference image K. This allows the subject A to be examined.
[0049] Figure 2 illustrates an example of the position Lj of the detected pattern Pj in the examination image J (examination field of view Fj), showing the distance between the lower right corner of the examination image J (examination field of view Fj) and the lower right corner of the detected pattern Pj. Similarly, an example of the position Lk of the registered pattern Pk in the reference image K (reference field of view Fk) is illustrated, showing the distance between the lower right corner of the reference image K (reference field of view Fk) and the lower right corner of the registered pattern Pk.
[0050] The registration pattern Pk for alignment should preferably be a unique pattern P selected from among multiple patterns P formed on the object A under inspection, in order to avoid misidentification with other patterns P. Furthermore, it is preferable that the registration pattern Pk for alignment be placed at locations on the object A under inspection where the manufacturing error for each product is large.
[0051] As shown in Figure 1, the display unit 30 is connected to the processing unit 20. The display unit 30 is, for example, a well-known display. The display unit 30 is displayed to the user U (human).
[0052] (Selection of registration pattern) Figure 3 shows the selection of the registered pattern Pk. The processing unit 20 synthesizes the composite image G. The reference image K may be the composite image G as is, or it may be used after performing predetermined processing (such as changing the size or color) on the composite image G. The composite image G may be discarded after the registered pattern Pk has been selected. Subsequently, the reference image K may be generated based on the selected registered pattern Pk (as a reference).
[0053] The processing unit 20 synthesizes a composite image G by superimposing multiple captured images g onto each other. Multiple captured images g are obtained by imaging the object under inspection A for each product. Each captured image g is basically almost the same, but differs slightly from product to product due to manufacturing errors, etc. The more captured images g there are, the better. The number of captured images g is, for example, several tens to several hundred.
[0054] The processing unit 20 has a memory 21. The memory 21 is, for example, a semiconductor memory device. Multiple captured images g are stored (remembered) in the memory 21.
[0055] The processing unit 20 synthesizes a composite image G by superimposing multiple captured images g on each other, based on a provisional registration pattern Q. The provisional registration pattern Q is provisionally registered from among multiple patterns P formed on the object A under inspection. The provisional registration pattern Q may be selected by a human or by machine.
[0056] The provisional registration pattern Q includes the first provisional registration pattern Qa, the second provisional registration pattern Qb, and the third provisional registration pattern Qc. The composite image G includes the first composite image Ga, the second composite image Gb, and the third composite image Gc.
[0057] The first composite image Ga, the second composite image Gb, and the third composite image Gc are synthesized using a common set of acquired images g stored in memory 21.
[0058] The first provisional registration pattern Qa is a mark placed near the center of the object under inspection A. The second provisional registration pattern Qb is a corner of the object under inspection A. The third provisional registration pattern Qc is a stepped step on the object under inspection A. The first provisional registration pattern Qa is unique and is placed in a location with a large manufacturing error for each product. The second provisional registration pattern Qb is unique but is placed in a location with a small manufacturing error for each product. The third provisional registration pattern Qc is not unique (stepped steps are common).
[0059] The first composite image Ga is created by superimposing multiple captured images g on each other, using the first provisional registration pattern Qa as a reference. The second composite image Gb is created by superimposing multiple captured images g on each other, using the second provisional registration pattern Qb as a reference. The third composite image Gc is created by superimposing multiple captured images g on each other, using the third provisional registration pattern Qc as a reference.
[0060] The processing unit 20 determines the variation E among the multiple captured images g that make up the composite image G in order to determine whether the provisional registration pattern Q is suitable as the registration pattern Pk.
[0061] The variation E includes a first variation Ea among the multiple acquired images g constituting the first composite image Ga, a second variation Eb among the multiple acquired images g constituting the second composite image Gb, and a third variation Ec among the multiple acquired images g constituting the third composite image Gc.
[0062] The processing unit 20 may register the provisional registration pattern Q as the registered pattern Pk based on the variation E.
[0063] For example, among multiple provisional registration patterns Q (first provisional registration pattern Qa, second provisional registration pattern Qb, and third provisional registration pattern Qc), the provisional registration pattern Q with the smallest variation E may be applied as the registered pattern Pk. As an example, when the first variation Ea is smaller than the second variation Eb and the third variation Ec, the first provisional registration pattern Qa may be applied as the registered pattern Pk.
[0064] The variation E includes at least one of the mean, standard deviation, variance, and coefficient of variation among multiple acquired images g.
[0065] In the composite image G of Figure 3, the darker the color, the greater the variation E, and the lighter the color, the smaller the variation E.
[0066] Figure 4 shows the display unit 30. The display unit 30 displays the first provisional registration pattern Qa, the second provisional registration pattern Qb, and the third provisional registration pattern Qc, as well as the first variation Ea of the first composite image Ga, the second variation Eb of the second composite image Gb, and the third variation Ec of the third composite image Gc.
[0067] Here, the temporary registration pattern Q selected by user U from among the multiple temporary registration patterns Q displayed on the display unit 30 may be registered as the registration pattern Pk. In this example, the temporary registration pattern Q selected by user U from among the first temporary registration pattern Qa, the second temporary registration pattern Qb, and the third temporary registration pattern Qc displayed on the display unit 30 may be registered as the registration pattern Pk.
[0068] User U selects the optimal temporary registration pattern Q from among multiple temporary registration patterns Q while looking at the display unit 30. In this example, User U selects the optimal temporary registration pattern Q from the first temporary registration pattern Qa, the second temporary registration pattern Qb, and the third temporary registration pattern Qc while looking at the display unit 30.
[0069] For example, user U selects the optimal temporary registration pattern Q from among the multiple temporary registration patterns Q displayed on the display unit 30 and clicks the cursor 31. A corresponding selection button 32 may be placed near each temporary registration pattern Q. The user may also click the selection button 32 with the cursor 31. Alternatively, if the display unit 30 is a touch panel (equipped with a touch sensor), user U touches the optimal temporary registration pattern Q selected from among the multiple temporary registration patterns Q displayed on the display unit 30 with their finger.
[0070] Instead of simply considering the magnitude of the variation E, the user U may decide which provisional registration pattern Q to register as the registered pattern Pk based on the overall distribution of variation E and their empirical rules. Alternatively, the user U may decide which provisional registration pattern Q to register as the registered pattern Pk based on the variation E in areas where manufacturing errors tend to be large.
[0071] FIG. 5 shows a method for obtaining the variation E. For example, the variation E of the brightness x for each pixel M among a plurality of captured images g constituting the composite image G is obtained. The index of the brightness x may be, for example, hue, lightness, darkness, chroma, luminosity, illuminance, etc.
[0072] As an example, the captured image g includes a first captured image g1, a second captured image g2, and a third captured image g3. In each captured image g, the numerical value entered in each pixel M indicates the magnitude of the brightness x. In the composite image G, the numerical value entered in each pixel M indicates the variation E of the brightness x for each pixel M among the plurality of captured images g. As the variation E, the standard deviation s is exemplified. In obtaining the standard deviation s, first, the variance s 2 is obtained. The variance s 2 is obtained by [Equation 1].
[0073]
Equation
[0074] n is the number of captured images g. In this example, n is 3. x i is the numerical value of the brightness x in each captured image g.
[0075] X ave indicates the average value of the brightness x in the plurality of captured images g. Taking the pixel M in the upper left corner as an example, the brightness x i is x1 = 90 in the first captured image g1 (n = 1), x2 = 90 in the second captured image g2 (n = 2), and x3 = 10 in the third captured image g3 (n = 3).
[0076] The average value X ave is X ave = (x1 + x2 + x3) / n = (90 + 90 + 10) / 3 = 63.33…. (x1 - X ave ) 2 = (90 - 63.33…) 2 = (26.66…) 2 = 711.1…. (x2 - X ave ) 2 = (90 - 63.33…) 2=(26.66…) 2 =711.1... (x³-X ave ) 2 =(10-63.33…) 2 =(-53.33…) 2 = 2844.1...
[0077] dispersion s 2 is, s 2 =(1 / 3)(711.1...+711.1...+2844.1...)=1422.1.
[0078] The standard deviation s is s = 1422.1 (1 / 2) =37.71...≈38.
[0079] Note that the coefficient of variation C v This is obtained by dividing the standard deviation s by the mean Xave (C v =s / Xave).
[0080] (flowchart) Figure 6 shows a flowchart for determining the registered pattern Pk.
[0081] Starting from the beginning, in the first step T1, the subject A to be inspected is imaged for each product, and multiple captured images g are obtained.
[0082] In the second step T2, multiple captured images g are saved to memory 21.
[0083] In the third step T3, a provisional registration pattern Q is provisionally registered from among multiple patterns P formed on the object A under inspection.
[0084] In the fourth step T4, multiple captured images g are read from memory 21.
[0085] In the fifth step T5, multiple captured images g are superimposed on each other using a provisionally registered pattern Q as a reference to obtain a composite image G.
[0086] In the sixth step T6, the variation E among the multiple captured images g that make up the composite image G is determined.
[0087] In step 7, T7, the suitability of the provisional registration pattern Q as a registered pattern Pk is determined based on the variation E. If the provisional registration pattern Q is determined to be suitable as a registered pattern Pk, the process proceeds to step 8, T8. If the provisional registration pattern Q is determined to be unsuitable as a registered pattern Pk, the process proceeds to step 9, T9. This suitability determination may be performed by human eyes or by machine based on predetermined criteria.
[0088] In step 8, T8, the provisional registration pattern Q is registered as the registered pattern Pk. Then, the process ends.
[0089] In step 9, T9, a different provisional registration pattern Q is provisionally registered from among the multiple patterns P formed on the object A under inspection. Then, the process returns to step 4, T4.
[0090] (Effects and Benefits) To determine whether a provisional registration pattern Q is suitable as a registered pattern Pk, the variation E among the multiple captured images g that make up the composite image G is calculated. This allows for the appropriate selection of a registration pattern Pk for alignment used when comparing the inspection image J with the reference image K.
[0091] In the inspection device 1 that inspects an object A by imaging the object A and comparing the inspection image J obtained with a reference image K, a registration pattern Pk for alignment used when comparing the inspection image J and the reference image K can be suitably selected.
[0092] Since both the first composite image Ga and the second composite image Gb (and the third composite image Gc) are synthesized using a common set of acquired images g stored in memory 21, data capacity can be saved compared to when the first composite image Ga and the second composite image Gb (and the third composite image Gc) are synthesized using separate acquired images g.
[0093] The variation E between the provisional registration pattern Q and the corresponding composite image G can be visualized by the display unit 30.
[0094] User U can select the most suitable temporary registration pattern Q as the registration pattern Pk from among multiple temporary registration patterns Q while looking at the display unit 30.
[0095] Which provisional registration pattern Q should be designated as the registered pattern Pk can be easily determined based on the magnitude of the variation E.
[0096] The variation E of brightness x for each pixel M among the multiple captured images g that make up the composite image G can be determined. The variation E can be easily calculated.
[0097] Variability E includes at least one of the mean, standard deviation, variance, and coefficient of variation. Variability E can be easily calculated.
[0098] (Other embodiments) Although this disclosure has been described above with reference to preferred embodiments, this description is not limiting, and various modifications, substitutions, or combinations are, of course, possible.
[0099] The composite image G may be discarded after selecting (registering) the registered pattern Pk. In this case, a reference image K may be generated by superimposing multiple (new) captured images based on (using as a reference) the registered pattern Pk. Alternatively, the reference image K may be generated by other methods. The composite image G may be used solely for the purpose of selecting the registered pattern Pk. The captured image g used to synthesize the composite image G and the other captured image used to generate the reference image K may be different from each other.
[0100] There may be just one provisional registration pattern Q. In this case, the suitability of the single provisional registration pattern Q as a registration pattern Pk is determined.
[0101] In Figure 1, the position Lj of the detected pattern Pj in the examination image J (examination field of view Fj) and the position Lk of the registered pattern Pk in the reference image K (reference field of view Fk) are illustrated by the distance between their lower right edges. However, the explanation is not limited to this, and the positions may also be represented by the distance between other edges, such as the upper left edges, or the distance between the centers.
[0102] The examination field of view Fj (examination image J) and the reference field of view Fk (reference image K) are not limited to quadrilaterals; for example, they may be polygons other than quadrilaterals, or circles other than polygons.
[0103] The provisional registration pattern Q is not limited to those described above, and various forms are possible (however, it is preferable that it be unique).
[0104] The object under inspection A is not limited to chips provided on the surface of substrate W. Substrate W is not limited to a semiconductor wafer; for example, it may be glass, metal, resin, etc. The object under inspection A may not be related to the chips on substrate W.
[0105] The inspection method according to this disclosure inspects an object to be inspected A by comparing an inspection image J obtained by imaging the object to be inspected A with a reference image K. In this inspection method, in the reference image K, a alignment pattern P is registered as a registered pattern Pk from among a plurality of patterns P formed on the object to be inspected A. In the inspection image J obtained by imaging the object to be inspected A, a pattern P similar to the registered pattern Pk is detected as a detection pattern Pj from among a plurality of patterns P formed on the object to be inspected A. Based on the position Lj of the detected pattern Pj in the inspection image J and the position Lk of the registered pattern Pk in the reference image K, the inspection image J is aligned with the reference image K. In this inspection method, a provisionally registered pattern Q is provisionally registered from among a plurality of patterns P formed on the object to be inspected A as a reference. A composite image G is synthesized by superimposing a plurality of captured images g obtained by imaging the object to be inspected A for each product. In order to determine whether the provisionally registered pattern Q is suitable as a registered pattern Pk, the variation E among the plurality of captured images g constituting the composite image G is determined, and the provisionally registered pattern Q is registered as a registered pattern Pk based on the variation E.
[0106] In an inspection method for inspecting an object A by imaging the object A and comparing the inspection image J obtained with a reference image K, the registration pattern Pk used for alignment when comparing the inspection image J and the reference image K can be suitably selected. [Industrial applicability]
[0107] This disclosure is extremely useful and highly industrially applicable, as it can be applied to inspection devices and inspection methods. [Explanation of Symbols]
[0108] 1. Inspection device 2 stages 10 Imaging Unit 20 Processing Units 30 Display section 31 Cursor 32 Select button 21 memory Fj Examination Field Fk Reference Field J Examination image K Reference image Lj position Lk position D Pattern matching search range A Subject to Inspection W board P pattern Pk registration pattern Pj detection pattern Q Temporary registration pattern Qa First provisional registration pattern Qb 2nd Provisional Registration Pattern Qc 3rd Provisional Registration Pattern G Composite image Ga 1st composite image Gb 2nd composite image Gc 3rd composite image g Captured image g1 First captured image g2 Second captured image g3 Third captured image E Variation Ea First Variation Eb Second Variation Ec Third Variation M pixels x Brightness U User
Claims
1. An inspection device that inspects an object to be inspected by comparing an inspection image obtained by imaging the object to be inspected with a reference image, The imaging unit captures an image of the object to be inspected, A processing unit, and The aforementioned processing unit, In the aforementioned reference image, the alignment pattern is registered as a registered pattern from among the multiple patterns formed on the object to be inspected. In the inspection image obtained by imaging the object to be inspected, a pattern similar to the registered pattern is detected as a detection pattern from among a plurality of patterns formed on the object to be inspected. Based on the position of the detection pattern in the inspection image and the position of the registration pattern in the reference image, the inspection image is aligned with the reference image. The aforementioned processing unit, Based on a provisionally registered pattern selected from among the multiple patterns formed on the object to be inspected, a composite image is synthesized by superimposing multiple captured images obtained by imaging the object to be inspected for each product, thereby combining them. An inspection device that determines the variability among a plurality of captured images constituting the composite image in order to determine whether the provisional registration pattern is suitable as the registration pattern.
2. Multiple of the aforementioned captured images are stored in memory. The aforementioned provisional registration pattern includes a first provisional registration pattern and a second provisional registration pattern. The composite image includes a first composite image formed by superimposing a plurality of the captured images on each other based on the first provisional registration pattern, and a second composite image formed by superimposing a plurality of the captured images on each other based on the second provisional registration pattern. The inspection apparatus according to claim 1, wherein the first composite image and the second composite image are synthesized using a common plurality of the acquired images stored in the memory.
3. Equipped with a display unit, The aforementioned variation includes a first variation among the plurality of acquired images constituting the first composite image, and a second variation among the plurality of acquired images constituting the second composite image. The inspection apparatus according to claim 2, wherein the display unit displays the first provisional registration pattern and the second provisional registration pattern, as well as the first variation of the first composite image and the second variation of the second composite image.
4. The inspection apparatus according to claim 3, wherein the provisional registration pattern selected by the user from among the first provisional registration pattern and the second provisional registration pattern displayed on the display unit is registered as the registered pattern.
5. The aforementioned variation includes a first variation among the plurality of acquired images constituting the first composite image, and a second variation among the plurality of acquired images constituting the second composite image. When the first variation is smaller than the second variation, the first provisional registration pattern is applied as the registration pattern. The inspection apparatus according to claim 2.
6. An inspection apparatus according to any one of claims 1 to 5, wherein the variation in brightness of each pixel among the plurality of captured images constituting the composite image can be determined.
7. The inspection apparatus according to any one of claims 1 to 5, wherein the variation includes at least one of the mean, standard deviation, variance, and coefficient of variation.
8. An inspection method for inspecting an object to be inspected by comparing an inspection image obtained by imaging the object to be inspected with a reference image, In the aforementioned reference image, the alignment pattern is registered as a registered pattern from among the multiple patterns formed on the object to be inspected. In the inspection image obtained by imaging the object to be inspected, a pattern similar to the registered pattern is detected as a detection pattern from among a plurality of patterns formed on the object to be inspected. Based on the position of the detection pattern in the inspection image and the position of the registration pattern in the reference image, the inspection image is aligned with the reference image. In the aforementioned inspection method, Based on a provisionally registered pattern selected from among the multiple patterns formed on the object to be inspected, a composite image is synthesized by superimposing multiple captured images obtained by imaging the object to be inspected for each product, thereby combining them. In order to determine whether the provisional registration pattern is suitable as the registration pattern, the variation among the multiple captured images that constitute the composite image is determined. An inspection method for registering the provisional registration pattern as the registered pattern based on the aforementioned variation.