JP4792314B2 - Substrate inspection apparatus and substrate inspection method - Google Patents

Description

  The present invention relates to a substrate inspection apparatus and method. For example, a substrate inspection in which a two-dimensional scanning image is obtained by moving a linear imaging region relative to a plate-like object such as a semiconductor wafer or a liquid crystal substrate, and defect extraction is performed based on the scanning image. The present invention relates to an apparatus and a method.

For example, in a manufacturing process of a substrate such as a semiconductor wafer or a liquid crystal substrate, a so-called macro inspection is performed in which an inspector visually inspects defects such as film unevenness, scratches, dust, and exposure pattern defects.
In recent years, from the viewpoint of inspection efficiency and inspection accuracy, substrate inspection apparatuses that automatically inspect this macro inspection based on image data of a substrate have come to be used. For example, there is known a substrate inspection apparatus that acquires a two-dimensional scan image by scanning a substrate with a line image imaging unit such as a line sensor camera and performs defect inspection based on the scan image.
In such a substrate inspection apparatus, since various image noises caused by substrate conditions and imaging conditions are included in the scanned image, it is necessary to accurately remove these image noises to prevent detection of pseudo defects.
On the other hand, in such an automatic inspection, an inspector may confirm the acquired scanned image in order to finally check whether defect extraction is performed correctly. At that time, if there is an image that is not judged as a defect because the difference in brightness with respect to the background brightness is very small, for example, an image such as a streak is detected sharply by human vision. It may be judged as an image.
For this reason, in the substrate inspection apparatus, it is necessary to remove image noise different from the defect with high accuracy before the defect determination process.
Conventionally, in order to remove such image noise, shading correction has been performed on the basis of a scanned image of a non-defective substrate.
In addition, as a technique related to such image noise removal, Patent Document 1 discloses an image picked up at a first position and a pixel of an image pickup means in order to remove the influence of a moire of a test object having a repetitive pattern. And a defect inspection apparatus that detects a defect by comparing an image picked up at a second position where the image pickup position is shifted so that the positional relationship between the pattern and the repeated pattern is substantially the same.
In Patent Document 2, as a technique for removing streak noise related to an image reading apparatus such as a digital copying machine, a facsimile machine, and a scanner and an image reading method thereof, two reading units are arranged apart from each other in the sub-scanning direction. An image reading apparatus and an image reading method for detecting and removing streak noise by comparing the images are described.
However, the streak noise in this case is generated when the amount of received light changes drastically due to dust adhering to the reading means, and is not a subtle luminance unevenness that is removed by the above shading correction or the like. . The streak noise in this case becomes a clear black streak on the image and can be eliminated by removing dust.
JP-A-11-194098 (FIGS. 1 and 2) JP 2000-295416 A (FIGS. 1 and 2)

However, the conventional substrate inspection apparatus as described above has the following problems.
In the inspection of bright field illumination of a substrate inspection apparatus that performs macro inspection, the imaging means is arranged in the reflection angle direction of the illumination light to capture an interference image for detecting defects such as film unevenness on the substrate, and illumination In some cases, an imaging means is arranged in the direction of the diffraction angle of the light to capture a diffraction image for detecting poor exposure of a repetitive pattern on the substrate or adhesion of large dust. Conventional shading correction is an interference image. It was done using. That is, as a non-defective substrate as a reference for correction, an image is taken in the reflection angle direction using raw glass as a substrate material or a uniform film portion of a substrate to be inspected to create shading correction data.
However, a certain kind of vertical stripe noise appears slightly differently between the interference image and the diffraction image, and there is a problem that the vertical stripe noise cannot be removed even if the shading correction data is applied to the diffraction image.
Although it is conceivable to use a substrate to be inspected having a diffraction pattern as a non-defective substrate, for example, a liquid crystal substrate or the like is remarkably large, and it is difficult to manufacture a non-defective substrate that is small enough to be used as a reference. . Even if a non-defective substrate can be manufactured, the intensity of the vertical streak noise in the diffraction pattern varies depending on the substrate type and process, so the operator must create a large amount of shading correction data according to the combination of the type and process. is there. For this reason, there is a problem that it takes a lot of time to create, and mistakes at the time of data creation tend to occur.
On the other hand, the techniques described in Patent Documents 1 and 2 can remove image noise without using a non-defective substrate, but the former depends on the positional relationship between the pixel array pitch and the repetitive pattern. Moire is removed, and the latter removes only vertical stripe noise caused by dust on the reading means. Therefore, there is a problem that vertical stripe noise due to other factors cannot be removed.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide a substrate inspection apparatus and method capable of removing vertical stripe noise by simple means.

In order to solve the above problems, a substrate inspection apparatus according to the present invention includes an illumination unit that illuminates a subject, an imaging unit that captures a line image on the subject, the illumination unit and the imaging unit, a moving mechanism for moving relative to the subject, by controlling the moving operation and an imaging operation of said imaging portion of said moving mechanism, an imaging control unit which allowed to obtain a 2-dimensional scanned image of the subject, the imaging The vertical streak noise detecting means for detecting the vertical streak noise by calculating the image data of the scanned image of the subject imaged by the unit, and the vertical streak noise detected by the vertical streak noise detecting means is removed and corrected An image noise processing unit having a vertical streak noise removing unit that generates an image; and a defect extraction processing unit that extracts a defect from the corrected image generated by the image noise processing unit , wherein the vertical streak noise detecting unit includes: The scan A luminance comparison unit that reads a luminance profile for each line image from an image, and detects a luminance change region in the high luminance region among the high luminance region and the low luminance region of the luminance profile as vertical stripe noise information for each line image; A vertical streak noise determination unit that determines vertical streak noise from all the vertical streak noise information detected for each line image .

The substrate inspection method of the present invention obtains a two-dimensional scanning image by phase TaiUtsuri dynamic to isosamples the imaging unit to the subject to acquire line image on the object, based on the scanned image A substrate inspection method for performing defect extraction on the subject, wherein a luminance profile for each line image is read from the scanned image, and a luminance change in a high luminance region among a high luminance region and a low luminance region of the luminance profile A step of detecting a region as vertical stripe noise information for each line image , a vertical stripe noise determination step of determining vertical stripe noise from all the vertical stripe noise information detected for each line image, and the vertical stripe noise determination and having the steps of the process therefore the determined vertical streak noise removal processing to generate a corrected image, and a step of performing defect extraction from the corrected image.

  According to the substrate inspection apparatus and method of the present invention, the vertical stripe noise is detected by calculating the image data of the captured scanned image and the vertical stripe noise is removed. For example, a reference object such as a non-defective substrate is used. There is an effect that vertical stripe noise can be removed by simple means without using it.

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.
A substrate inspection apparatus according to an embodiment of the present invention will be described.
FIG. 1 is a front view showing a schematic configuration of a substrate inspection apparatus according to an embodiment of the present invention. 2A and 2B are a plan view and a side view showing a schematic configuration of the substrate inspection apparatus according to the embodiment of the present invention. FIG. 3 is a functional block diagram of a control unit of the substrate inspection apparatus according to the embodiment of the present invention. FIG. 4 is a functional block diagram of the image noise processing unit of the board inspection apparatus according to the embodiment of the present invention.

The substrate inspection apparatus 50 according to this embodiment is a so-called automatic macro inspection apparatus that performs image inspection for defects of the substrate 1 (subject), such as film unevenness, scratches, dust, and circuit pattern defects.
Examples of the substrate 1 include a semiconductor wafer and a liquid crystal substrate on which a repetitive pattern composed of a circuit pattern or the like is formed on the surface by a photolithography process or the like. Hereinafter, an example in which the substrate 1 is a liquid crystal substrate having a rectangular shape in a plan view will be described.
As shown in FIGS. 1, 2, and 3, the schematic configuration of the substrate inspection apparatus 50 includes a floating holder 2, a substrate holding / moving unit 4 (moving mechanism), a line illumination unit 3 (illuminating unit), and a line sensor camera 6 (imaging unit). ), And the control unit 8.

The levitating holder 2 is a mechanism for levitating the substrate 1 on a substantially horizontally arranged base in order to move the substrate 1 smoothly, and is arranged so that the longitudinal direction is the horizontal direction in the figure. As the levitation holder 2, for example, a gas levitation mechanism that discharges a gas such as air from a large number of air discharge ports 2 a provided on the substrate can be suitably employed.
Although not particularly illustrated, a position detection sensor that can detect the timing at which the tip of the substrate 1 reaches the imaging region P is provided in the vicinity of the imaging region P described later.

The substrate holding / moving unit 4 is a mechanism that moves while fixing and holding the end portion of the substrate 1 levitated on the levitating holder 2. In the present embodiment, it is provided on both ends of the levitation holder 2 and can be moved along guides 4 a that extend along the longitudinal direction of the levitation holder 2.
On the upper surface of each substrate holding and moving unit 4, a holding surface 4 b is provided for detachably fixing and holding the end portion of the substrate 1 at the same height as the lower surface of the substrate 1 levitated on the levitation holder 2. Yes. As a method for fixing the holding surface 4b, for example, a vacuum suction method or the like can be employed.
The substrate 1 can be accurately positioned manually or automatically with respect to the substrate holding and moving unit 4 by using an appropriate transfer mechanism. A positioning mechanism for mounting and fixing the substrate 1 at a predetermined reference position and a position detection sensor for detecting the mounting position may be provided.

The line illumination unit 3 illuminates a line-shaped imaging region P set with the direction perpendicular to the moving direction of the substrate holding and moving unit 4 as the longitudinal direction on the horizontal plane on which the upper surface of the substrate 1 moves. Although not shown in FIGS. 1 and 2, the imaging region P is set within a plane orthogonal to the longitudinal direction of the imaging region P by an imaging direction setting mechanism 9 (see FIG. 3) having an appropriate rotation mechanism. The rotation angle is held as a center, and the incident angle θ ₁ of the illumination light with respect to the substrate 1 in the imaging region P can be varied.

The line sensor camera 6 is for capturing a line image of the imaging region P, and includes an imaging lens 6a and a line imaging element 6b disposed at the imaging position of the imaging lens 6a. Then, the imaging direction setting mechanism 9 (see FIG. 3) is held rotatably like the line illumination unit 3 so that the imaging direction angle θ ₂ with respect to the normal of the substrate 1 in the imaging area P can be varied. .
As the line image sensor 6b, an appropriate sensor can be adopted depending on, for example, required sensitivity, the number of pixels, etc., for example, a one-dimensional CCD can be adopted.
On the optical path between the imaging region P and the imaging lens 6a, there is an interference filter 5 composed of an appropriate bandpass filter or the like that limits the wavelength band of light from the imaging region P in order to obtain a clear interference image. It is provided so that it can move forward and backward. Although this advance / retreat mechanism may be built in the line sensor camera 6, in the present embodiment, it is provided in the imaging direction setting mechanism 9, and similarly to the imaging direction setting mechanism 9, a device control unit 13 described later. The advancing / retreating operation is controlled by.

  The control unit 8 performs overall apparatus control of the board inspection apparatus 50 and image processing control, and may be configured by dedicated hardware, but in this embodiment, a CPU, a memory, and various input / output interfaces And other hardware that performs various operations by communicating control signals and data through various input / output interfaces (all not shown). Then, various functions and operations corresponding to the respective programs and hardware are realized by executing the programs appropriately loaded in the memory.

Below, the schematic structure of the control unit 8 is demonstrated for every functional block shown in FIG.
The functional blocks of the control unit 8 are a camera control unit 11, an illumination control unit 12, an apparatus control unit 13, an image storage unit 14, a display control unit 18, an image noise processing unit 15, a defect extraction processing unit 16, and a defect information registration unit. 17.
In addition, although not specifically described or illustrated, an operation unit that performs operation input to a computer or hardware is naturally provided.

  The camera control unit 11 controls the imaging operation of the line sensor camera 6 in accordance with the control signal from the device control unit 13, and performs signal processing on the imaging signal acquired by the line sensor camera 6, so that each scanning line Line image data is acquired and stored in the image storage unit 14.

  The illumination control unit 12 controls the illumination conditions of the line illumination unit 3, for example, lighting timing, light amount, etc., according to a control signal from the device control unit 13.

  The device control unit 13 sends a control signal to the camera control unit 11, the illumination control unit 12, the imaging direction setting mechanism 9, the substrate holding / moving unit 4, and the flying holder 2, so that each control unit and each mechanism By controlling, the imaging operation of the board inspection apparatus 50 is mainly controlled.

The apparatus control unit 13 sends the arrangement angle setting information of the line illumination unit 3 and the line sensor camera 6 to the imaging direction setting mechanism 9 to control the rotation angle position of the line illumination unit 3 and the line sensor camera 6. Then, the incident angle θ ₁ of the illumination light and the imaging direction angle θ ₂ are set.
For example, when an interference image is captured, θ ₁ = θ ₂ is set, and the interference filter 5 is advanced on the optical path.
When capturing a diffraction image, the relative relationship between the angles θ ₁ and θ ₂ is set in accordance with the diffraction direction of a specific order of diffracted light, and the interference filter 5 is retracted from the optical path.

In addition, the apparatus control unit 13 sets a conveyance speed V such that the substrate 1 is conveyed by a distance corresponding to the pixel size at each imaging timing of one line of the line imaging element 6b with respect to the substrate holding and moving unit 4. Then, it is driven in the sub-scanning direction.
Moreover, on the levitation holder 2, on / off control of levitation and discharge pressure control for stabilizing the levitation height are performed.

It should be noted that various information that can be changed by the control of the device control unit 13 such as the line illumination unit 3, the angles θ ₁ and θ _{2 of the} line sensor camera 6, the substrate holding / moving unit 4, the position coordinates of the substrate 1, etc. The information is communicated to the device control unit 13 and sent to the display control unit 18 (described later) as necessary to be displayed on the display unit 10.

The image storage unit 14 can communicate with the camera control unit 11, the image noise processing unit 15, the defect extraction processing unit 16, and the display control unit 18 to read and write various image data.
The display control unit 18 displays an image signal to be displayed on the display unit 10 based on information from the device control unit 13, the image noise processing unit 15, the defect extraction processing unit 16, and the image data stored in the image storage unit 14. Is generated.
The display unit 10 includes a monitor and a dedicated display device.

The image noise processing unit 15 reads the image of the substrate 1 stored in the image storage unit 14 through the camera control unit 11, removes vertical streak noise, and stores the image in the image storage unit 14.
As shown in FIG. 4, the schematic configuration includes vertical stripe noise detection means 20 and vertical stripe noise removal means 21.

The vertical stripe noise detection means 20 includes a main scanning direction luminance acquisition unit 22, a luminance comparison unit 23, a vertical stripe noise information storage unit 24, and a vertical stripe noise determination unit 25.
The main scanning direction luminance acquisition unit 22 reads line image data in the main scanning direction from the scanned image stored in the image storage unit 14 and sequentially acquires luminance distribution profiles.
The luminance comparison unit 23 compares the luminance of adjacent pixels in the main scanning direction in the luminance profile sequentially acquired by the main scanning direction luminance acquisition unit 22, and if there is a band-like range in which a luminance change of a predetermined range or more occurs, Record as a vertical streak noise candidate. For example, information such as the center position of the band, the band width, the full width of the luminance change, and the luminance of the band boundary line is stored in the vertical stripe noise information storage unit 24 as vertical stripe noise information.
The vertical streak noise information storage unit 24 stores the vertical streak noise information detected by the luminance comparison unit 23 together with the position information of each line image data in the sub-scanning direction. For example, a storage area is secured in the memory. Constitute.
The vertical streak noise determination unit 25 uses the vertical streak noise information stored in the vertical streak noise information storage unit 24 over the entire length in the sub-scanning direction, for the number of pixels corresponding to the preset vertical streak noise length determination condition. A continuous line in the sub-scanning direction is searched and determined as vertical stripe noise.

The vertical streak noise removing unit 21 includes a shading correction data creation unit 26 and a shading correction unit 27.
The shading correction data creation unit 26 creates shading correction data for removing the vertical stripe noise determined by the vertical stripe noise determination unit 25.
In this embodiment, shading correction data that replaces the luminance of the vertical stripe noise portion with the correction luminance generated from the luminance in the vicinity of the vertical stripe noise is used.
For example, the vertical stripe noise information storage unit 24 is referred to as the corrected luminance, the band boundary line luminance is acquired, and the luminance data of the vertical stripe noise area is replaced with the band boundary line luminance.
The shading correction unit 27 performs shading correction of the scanned image using the shading correction data created by the shading correction data creation unit 26 and stores the corrected image data in the image storage unit 14.
The shading correction unit 27 notifies the defect extraction processing unit 16 when the vertical stripe noise removal is completed. Further, if necessary, a control signal is sent to the display control unit 18 so that the image data after shading correction is output to the display unit 10.

The defect extraction processing unit 16 (see FIG. 3) is for performing defect extraction from the image data from which vertical stripe noise has been removed by the image noise processing unit 15. As the defect extraction algorithm, any known algorithm may be used according to the type of defect to be extracted. However, in the case of including operations such as relaxation and removal of vertical stripe noise and determination of a vertical stripe noise portion, these operation portions can be omitted.
The defect information registration unit 17 registers the defect information extracted by the defect extraction processing unit 16, for example, the type of defect and the position coordinates of the defect, in the external storage unit 19.

Next, the operation of the substrate inspection apparatus 50 will be described focusing on the operation of the image noise processing step.
FIG. 5 is a schematic diagram illustrating an example of a scanned image. FIG. 6 is a flowchart showing an operation flow of the image noise processing step of the substrate inspection apparatus according to the embodiment of the present invention. FIG. 7 is a schematic graph showing a luminance profile along the line AA in FIG. The horizontal axis represents the pixel position, and the vertical axis represents the luminance. FIG. 8 is a schematic diagram showing an example of a corrected image corresponding to the scanned image shown in FIG.

In order to inspect the substrate 1 by the substrate inspection apparatus 50, first, the floating holder 2 is turned on, the substrate 1 is floated on the floating holder 2, and both ends of the substrate 1 are sucked and fixed on the substrate holding and moving unit 4. .
Then, the imaging direction setting mechanism 9 is driven according to the type of inspection, for example, whether to capture an interference image or a diffraction image, and the positional relationship between the line illumination unit 3 and the line sensor camera 6 and the interference filter 5. Set the advance / retreat position.
Since the image noise processing step of this embodiment can be applied to both interference images and diffraction images in principle, description of operations unique to each inspection type is omitted.

Next, the line illumination unit 3 is turned on, the substrate holding / moving unit 4 is driven at the conveyance speed V, and the movement of the substrate 1 is started. And the line image of the board | substrate 1 which reached | attained the imaging area P with the line sensor camera 6 is imaged continuously.
The line image is converted into line image data by the camera control unit 11 and sequentially stored in the image storage unit 14. When the scanning is completed, the image data of the two-dimensional scanned image of the substrate 1 is stored in the image storage unit 14.

For example, assume that the scanned image 100 shown in FIG. 5 is stored. However, the scanned image 100 is simplified and schematically shown for convenience of explanation.
The scanned image 100 is a portion in which four rectangular high-brightness portions 100a having substantially constant high-brightness are formed in a lattice pattern on the low-brightness portion 100b and have a low-brightness that is different from a normal image of the high-brightness portion 100a. As shown, a defect 101, a streak defect 102, a streak defect 103, and vertical streak noise 104a, 104b, 105a, 105b, 106a, 106b are recognized.

The defect 101 is a low-luminance portion that extends in a substantially elliptical shape in a slightly narrow region relative to the region of the high-luminance portion 100a.
The streak defect 102 is a streaky low-intensity portion distributed in an oblique direction with respect to the sub-scanning direction, and has a length shorter than the outer dimension of the high-luminance portion 100a.
The streak defect 103 is a streak-like low-luminance portion that substantially extends in the sub-scanning direction, and has a length shorter than the outer dimension of the high-luminance portion 100a, like the streak defect 102.
The vertical streak noises 104a and 104b are vertical streak-like low brightness parts that penetrate the high brightness parts 100a and 100a adjacent in the sub scanning direction along the sub scanning direction at the same position in the main scanning direction.
The vertical streak noises 105a and 105b and the vertical streak noises 106a and 106b are vertical streak noises 104a and 104b, respectively, and similar vertical streak-like low-intensity portions that appear at different positions in the main scanning direction.

The amount of decrease in luminance such as vertical streak noise 104a varies depending on the cause of occurrence, but even when the amount of decrease in luminance is low, it is continuous due to continuous vertical streaks in the same direction. When an inspector checks an inspection result, it is easy to be cited as an abnormal image.
In this case, the automatic determination result based on the luminance information and shape conflicts with the inspector's visual check result, and it takes time and effort to perform re-inspection and re-determination.

Therefore, in the present embodiment, an image noise processing step is provided to remove such vertical streak noise that should not be determined as a defect before performing defect extraction.
The operation flow of the image noise processing step will be described with reference to FIG.

The image noise processing step is started from the time when the imaging step is completed and the scanned image 100 is acquired.
Step S <b> 1 is a vertical streak noise detection step of detecting vertical streak noise by performing arithmetic processing on the scanned image 100.
First, the luminance profile 110 of the line image of the scanned image 100 is called by the main scanning direction luminance acquisition unit 22. As an example, Figure 7 shows the intensity profile along the line A-A of FIG.
The luminance profile 110 corresponds to the two high luminance portions 100a of the scanned image 100 and corresponds to the high luminance regions 110A and 110B having the luminance b1, and the luminance b ₀ corresponding to the low luminance portion 100b on each side (however, , B ₀ > b ₁ ) corresponding to the low-luminance region 110C corresponding to a substantially rectangular wave luminance distribution.
In addition, luminance reduction portions 110a, 110b, and 110c appear at pixel positions L _a , L _b , and L _c corresponding to the images of the vertical stripe noises 104a, 105a, and 106a.
For example, the luminance lowering unit 110a changes the luminance b ₁ to the minimum luminance b ₂ (where b ₁ > b ₂ ) between the pixel positions L ₁ and L ₂ (where L ₁ <L _a <L ₂ ). Thus, a concave change reaching luminance b ₁ is shown. Here, the luminance difference (b ₁ −b ₂ ) is a minute value for both the luminance b ₁ and b ₂ . In many cases, the brightness difference is smaller than that of the defect 101 and the streak defects 102 and 103.

Next, the luminance comparison unit 23 performs arithmetic processing on the luminance profile 110 to detect vertical stripe noise 104a and the like. In the present embodiment, by comparing the brightness of the adjacent images, for example, in the high luminance region 110A, detects the pixel position L ₁ to turn on the luminance decreased from the brightness b _1, followed by continuing the comparison, detects the continuity of the luminance reduction Then, the minimum luminance value b ₂ is detected, and further, the increase in luminance is detected, and the pixel position L ₂ that returns to the luminance b ₁ is detected.
In this case, since a luminance change area having a width is detected, the vertical stripe noise information is stored in the vertical stripe noise information storage unit 24 by using the band-like area between the pixel positions L ₁ and L ₂ as a vertical stripe noise candidate. . As vertical stripe noise information, for example, pixel position information for specifying the sub-scanning direction position of the line image, pixel positions L ₁ and L _{2 in the} main scanning direction for giving the band boundary position, luminance b _{1 of} the band boundary line, minimum value _{b 2} of the luminance, the luminance difference _(b 1 _-b 2), the total width of the band _(L 2 _-L 1), include information such as the center of the band width _{L a = (L 1 + L} 2) / 2.

After these processes are performed on all line image data of the scanned image 100, the vertical streak noise determination unit 25 selects the vertical streak noise from the vertical streak noise candidates based on the information stored in the vertical streak noise information storage unit 24. Determine streak noise.
In this embodiment, at each position where the vertical streak noise information is stored, the number of pixels continuous in the sub-scanning direction is counted, and when the number of pixels corresponding to a preset length determination condition is counted, Judge as noise.
As the length determination condition, an appropriate range longer than the length assumed as the defect can be set so as not to mistakenly exclude the defect.
In addition, it is preferable that the determination of the identity of the pixel position in the main scanning direction has a certain width in consideration of the instability of the line width of the vertical stripe noise.

For example, in the example of the scanned image 100, in order to exclude a defect on the high luminance portion 100a, the length may be slightly shorter than the sub-scanning length of the high luminance portion 100a.
In this case, the streak defect 102 has a streak shape, but extends in an oblique direction. Therefore, the number of pixels continuous in the sub-scanning direction is extremely short and is not determined as vertical streak noise.
Further, the streak defect 103 is a streak continuous along the sub-scanning direction, but is not determined to be vertical streak noise because it is shorter than the length of the high luminance portion 100a in the sub-scanning direction.
In addition, when the luminance difference is small, it is considered that the vertical streak noise appears as a broken line at the same pixel position in the main scanning direction. In such a case, sub-scanning at a certain pixel position in the main scanning direction is performed. If the appearance rate of the vertical streak noise candidate is high to some extent in the direction, it may be determined as vertical streak noise even in a broken line shape.

If the vertical line noise determination unit 25 determines that there is no vertical line noise, the process proceeds to step S4.
If it is determined that there is vertical stripe noise, step S2 is executed.
This completes the vertical stripe noise detection process.

After step S2, the vertical line noise removal process starts.
In step S2, the shading correction data creating unit 26 creates shading correction data.
In the example of the scanned image 100, for example, in the case of the vertical streak noise 104a, data that replaces the luminance data of the pixel positions L _{1 to} L ₂ with the luminance b ₁ of the band boundary line is created.
Generally, vertical streak noise is easy to see due to the narrow line width, and the background luminance distribution often changes slowly in the range of the vertical streak noise line width. The brightness may not match. In this case, an average value may be taken. In addition, it is more preferable if the data is replaced with a luminance distribution obtained by linear approximation.

  In step S <b> 3, shading correction of the scanned image 100 is performed using the shading correction data created by the shading correction data creation unit 26 by the shading correction unit 27. Thereby, the vertical stripe noises 104a, 104b, 105a, 105b, 106a, 106b are removed, and for example, a corrected image 200 as shown in FIG. 8 is obtained.

In step S4, the corrected image 200 is stored in the image storage unit 14. Then, the defect extraction processing unit 16 is notified that the vertical stripe noise removal has been completed. Further, if necessary, a control signal is sent to the display control unit 18 to display the corrected image 200 on the display unit 10. In this case, even if the inspector looks at the corrected image 200, no vertical streak noise is recognized, so there is no possibility of making an erroneous determination as a pseudo defect.
This completes the image noise processing step.

The correction image 200 is subjected to a defect extraction processing step by the defect extraction processing unit 16. At this time, since the vertical streak noise is removed, even if the scanned image 100 has a luminance difference in which the vertical streak noise protrudes, it is not subject to defect extraction and may be automatically detected as a pseudo defect. There is nothing.
The defects extracted by the defect extraction processing unit 16 are registered in the external storage unit 19 by the defect information registration unit 17.
In this way, the inspection process based on the scanned image 100 acquired under certain imaging conditions is completed.
When further inspection is continued under other imaging conditions, the above operation is repeated from the beginning.

As described above, according to the substrate inspection apparatus 50 of the present embodiment, the vertical stripe noise is detected by calculating the image data of the captured scanned image 100, and the shading correction data is detected from the image data by the image noise processing unit 15. Because it automatically generates and eliminates vertical streak noise, it eliminates vertical streak noise by a simple means without using the operator to create shading correction data using a non-defective substrate for each manufacturing process. be able to.
The substrate inspection apparatus 50 is very convenient because it can remove the vertical streak noise in the same manner when capturing an interference image or a diffraction image.
In addition, according to the substrate inspection method of the present embodiment, since the defect extraction step is performed after the image noise processing step, defect extraction can be performed with high accuracy without erroneously detecting vertical streak noise as a pseudo defect. it can.

In the above description, when detecting as a vertical streak noise candidate, an example has been described in which all luminance change portions that can become vertical streak noise are stored. However, in advance, the specification of the apparatus, the operating environment of the apparatus, the inspector If the vertical streak noise conditions to be removed are specified to some extent from the level of the current level, etc., if the conditions for detecting the vertical streak noise candidates are specified based on those conditions, the computation processing time Can be shortened, which is preferable.
For example, even if the luminance difference is the same, it is thin to some extent that human eyes can easily recognize it as vertical stripe noise. Therefore, when the full width (L ₂ −L ₁ ) is larger than a predetermined value, it may not be set as a vertical stripe noise candidate.
Also, and the size of the luminance difference _(b 1 -b _2), by a combination of the luminance difference between _(b 1 -b ₂₎ and the total width _(L 2 -L _1), may be excluded from the candidates.

  In the above description, the shading correction data is described as an example in the case of replacing the luminance at the boundary line of the vertical stripe noise, but the vertical stripe noise portion can be removed to the extent that the inspector does not erroneously detect it, And if it is the range which does not interfere with the detection of a genuine defect, a suitable arithmetic processing can be employ | adopted. For example, the luminance may be averaged within an appropriate processing area.

In the above description, in the vertical streak noise detection step, the vertical line noise candidate is obtained by comparing the luminance change of the adjacent pixels on the luminance profile. However, the detection of the candidate is limited to such a luminance comparison. It is not something. For example, the luminance profile may be detected after appropriate arithmetic processing.
For example, depending on the noise level of the luminance profile, the luminance change may be determined based on the average of the luminance of continuous pixels on the luminance profile or smoothing processing. In this case, since the noise component of the luminance change is averaged, the detection accuracy can be improved.
Further, when there is little change in the brightness profile, the brightness change component of the brightness profile may be detected after being emphasized.

  In the above description, the image noise processing is described as an example in which the scan image is acquired and stored. However, the image noise processing step includes a number of lines corresponding to the length determination condition of the vertical stripe noise. Since it can be performed in units of image data, the line image data may be stored in a buffer memory that stores the line image data of the required number of lines, and the imaging process and the image noise processing process may be performed in parallel. .

  In the above description, the example in which the vertical stripe noise partially decreases in the high luminance region in the luminance profile has been described. However, the same applies to the case where the luminance increases partially in the low luminance region. And can be determined as vertical stripe noise.

  In the above description, the example in which the imaging region P is fixed and the substrate 1 moves has been described. However, the illumination unit, the imaging unit, and the subject only need to be relatively movable in the sub-scanning direction. For example, the illumination unit and the imaging unit may move with respect to the subject.

It is a front view which shows schematic structure of the board | substrate inspection apparatus which concerns on embodiment of this invention. It is the top view and side view which show schematic structure of the board | substrate inspection apparatus which concerns on embodiment of this invention. It is a functional block diagram of the control unit of the board | substrate inspection apparatus which concerns on embodiment of this invention. It is a functional block diagram of the image noise processing part of the board | substrate inspection apparatus which concerns on embodiment of this invention. It is a schematic diagram which shows an example of a scanning image. It is a flowchart which shows the operation | movement flow of the image noise processing process of the board | substrate inspection apparatus which concerns on embodiment of this invention. It is a schematic graph showing the brightness | luminance profile in alignment with the AA of FIG. It is a schematic diagram which shows an example of the correction | amendment image corresponding to the scanning image shown in FIG.

Explanation of symbols

1 Substrate (subject)
3 line lighting unit (lighting unit)
4 Substrate holding and moving unit (moving mechanism)
5 Interference filter 6 Line sensor camera (imaging part)
6b Line imaging device 8 Control unit 9 Imaging direction setting mechanism 10 Display unit 13 Device control unit (imaging control unit)
14 Image storage unit 15 Image noise processing unit 16 Defect extraction processing unit 20 Vertical stripe noise detection unit 21 Vertical stripe noise removal unit 22 Main scanning direction luminance acquisition unit 23 Luminance comparison unit 24 Vertical stripe noise information storage unit 25 Vertical stripe noise determination unit 26 Shading correction data creation unit 27 Shading correction unit 50 Substrate inspection apparatus 100 Scanned images 104a, 104b, 105a, 105b, 106a, 106b Vertical stripe noise 110 Luminance profile 200 Correction image P Imaging region

Claims (5)

An illumination unit for illuminating the subject;
An imaging unit for imaging a line image on the subject;
And the illuminating unit and the imaging unit, and the moving mechanism for relative movement between the object,
An imaging control unit that acquires a two-dimensional scanning image of the subject by controlling a moving operation of the moving mechanism and an imaging operation of the imaging unit;
Longitudinal streak noise detecting means for detecting vertical streak noise by calculating image data of the scanned image of the subject imaged by the imaging unit, and processing for removing the vertical streak noise detected by the vertical streak noise detecting means. An image noise processing unit having vertical streak noise removing means for generating a corrected image
A defect extraction processing unit that performs defect extraction from the corrected image generated by the image noise processing unit ;
Have
The vertical stripe noise detection means includes:
Luminance comparison for reading out a luminance profile for each line image from the scanned image and detecting, for each line image, a luminance change region in the high luminance region of the luminance profile as vertical stripe noise information. And
A vertical stripe noise determination unit for determining vertical stripe noise from all the vertical stripe noise information detected for each line image;
A board inspection apparatus comprising: The vertical streak noise determination unit counts the number of continuous pixels at each position where the vertical streak noise information for each line image is stored, and the number of pixels corresponding to a preset length determination condition is counted. The substrate inspection apparatus according to claim 1 , wherein the vertical line noise is determined in the case. The vertical streak noise removing means is
2. The vertical streak noise removal processing is performed by replacing the vertical streak noise detected by the vertical streak noise detection means with a corrected brightness generated from the brightness in the vicinity of the vertical streak noise. Or the board | substrate inspection apparatus of 2.   The luminance difference with respect to the background luminance of the vertical stripe noise detected by the image noise processing unit is smaller than the luminance difference with respect to the background luminance of the defect extracted by the defect extraction processing unit. The board inspection apparatus according to claim 1. Wherein the imaging unit that captures an image of a line image on an object to obtain a two-dimensional scan images in phase TaiUtsuri kinematic be isosamples to the subject, a defect extraction of the object based on the scanned image substrate An inspection method ,
Reading a luminance profile for each line image from the scanned image, and detecting, for each line image, a luminance change region in the high luminance region as vertical stripe noise information among the high luminance region and the low luminance region of the luminance profile; ,
A vertical streak noise determination step of determining the vertical streak noise from all of the vertical stripe noise information detected for each of the line image,
And as factories that generates a corrected image by removing handle vertical stripe noise is determined by said longitudinal muscle noise determination step,
And performing a defect extracted from the corrected image,
Substrate inspection method characterized by having a.

Images