JP2010091360A - Method and device for inspecting image - Google Patents

Abstract

A highly accurate inspection is realized.
An image inspection apparatus includes a reference image creation unit that employs a normal image to be inspected that has been determined to be non-defective in advance as a reference image, and adds a preset upper threshold to the luminance value of each pixel of the reference image. Then, an upper threshold image is created, and a threshold image creating unit 404 that creates a lower threshold image by adding a preset lower threshold to the luminance value of each pixel of the reference image, and a CCD camera 3 that captures the inspection target And inspection means (positioning unit 406, expansion / contraction processing unit 407, for comparing the inspection image obtained by photographing the inspection object with the upper threshold image and the lower threshold image for each pixel to determine the quality of the inspection object. A defect location extraction unit 408 and a determination unit 409).
[Selection] Figure 1

Description

  The present invention relates to an image inspection method for inspecting a defect to be inspected such as a circuit pattern formed on the surface of a fine chip such as a sensor chip (such as a shape defect such as cracking, peeling, deformation, or the presence of foreign matter) with an image. And an image inspection apparatus.

  A fine chip such as a sensor chip created by applying a micromachining technique has a large number of chips formed on a single silicon wafer, for example, with a side of about 1 to 2 mm. Various circuit patterns are formed on each chip, and the patterns tend to be highly integrated and miniaturized. Detecting such pattern defects is an important factor in determining the quality of a chip. There are a wide variety of pattern defects such as foreign matter such as dust adhering to the chip, cracking, deformation, and peeling of the pattern, and since the chip to be inspected is very small and numerous, it is difficult to detect the defect visually. is there. Therefore, a technique for automatically detecting such a defect has been developed (for example, see Patent Document 1).

  In the conventional inspection apparatus disclosed in Patent Document 1, a circuit pattern on a chip surface to be inspected is imaged with a CCD camera or the like, and the captured image (inspection image) is compared with a reference image prepared in advance. Pattern defects were detected.

  12A and 12B are diagrams for explaining a method of creating a reference image in a conventional inspection apparatus. 12A shows the luminance value L for each pixel of a plurality of normal images, and FIG. 12B shows the luminance value Lref for each pixel of the reference image. The reference image is created by first imaging about 30 normal chips, preparing about 30 normal images, calculating luminance values for all pixels of these images, and for each normal image for each pixel. The average value of luminance values is obtained. Thus, an average image, that is, a reference image in which each pixel has a luminance average value calculated for the pixel is obtained (FIG. 12B).

Further, the standard deviation of the luminance value for each pixel of all normal images is calculated, and values obtained by multiplying the standard deviation by a constant are set as the upper limit threshold and the lower limit threshold of the pixel. Thereby, an upper limit threshold image in which each pixel has an upper limit threshold calculated for the pixel and a lower limit threshold image in which each pixel has a lower limit threshold calculated for the pixel are obtained. In FIG. 13, Lu represents the luminance value for each pixel of the upper threshold image, and Ld represents the luminance value for each pixel of the lower threshold image.
In the conventional inspection apparatus, when the luminance value of each pixel of the inspection image is equal to or lower than the upper limit threshold value of the pixel and equal to or higher than the lower limit threshold value, the chip to be inspected is regarded as a non-defective product.

JP 2002-195958 A

  In the conventional inspection apparatus disclosed in Patent Document 1, since the average image of a plurality of normal images is used as the reference image, the pattern edge portion (E1 in FIG. 12B) of the reference image is smoothed and the inspection image is obtained. As a result, the outline is blurred, which causes excessive extraction of pattern defects and may cause erroneous determination.

  Further, in the conventional inspection apparatus disclosed in Patent Document 1, values obtained by multiplying the standard deviation of the luminance value for each pixel by a constant are used as the upper threshold value and the lower threshold value for the pixel. The threshold value of the pattern edge (E2 in FIG. 13) becomes excessive due to factors such as error of the pattern and pattern tolerance, and the upper limit threshold value becomes larger and the lower limit threshold value becomes smaller. For this reason, the interval between the upper limit threshold and the lower limit threshold is widened, and the luminance that should be determined as a defective product is within this wide interval, which may cause a defective product to be determined as a non-defective product and cause an erroneous determination. There was sex.

  Furthermore, the inspection apparatus disclosed in Patent Document 1 has a problem that it takes time to select a plurality of normal images. In order to select a normal image, the chip image taken by the CCD camera is checked on the monitor to determine whether the product is a non-defective product or a defective product.

  The present invention has been made to solve the above-described problems, and an object thereof is to provide an image inspection method and an image inspection apparatus capable of realizing high-precision inspection.

In the image inspection method of the present invention, a reference image creation procedure that employs a normal image to be inspected that has been determined as a non-defective product in advance as a reference image, and a preset upper threshold value are added to the luminance value of each pixel of the reference image. And a threshold image creating procedure for creating a lower threshold image by adding a preset lower threshold to the luminance value of each pixel of the reference image, and an image of the inspection object. And an inspection procedure for comparing the inspection image with the upper threshold image and the lower threshold image for each pixel to determine whether the inspection object is good or bad.
Further, in one configuration example of the image inspection method of the present invention, the reference image creation procedure includes a noise removal procedure for replacing image data of a portion including the noise component of the reference image with image data of a portion not including the noise component. It is characterized by including.
Further, in one configuration example of the image inspection method of the present invention, the inspection procedure includes, among the pixels of the inspection image, pixels having a higher luminance value than the pixels of the upper threshold image and pixels of the lower threshold image. A pixel having a low luminance value is a pixel having a first luminance value, and a pixel having a luminance value equal to or lower than the luminance value of the pixel of the upper threshold image and equal to or higher than the luminance value of the pixel of the lower threshold image is binary. A defect location extraction procedure for generating a binarized image, and a determination procedure for inspecting the binarized image on the basis of a predetermined determination criterion and determining pass / fail of the inspection object.

In addition, the image inspection apparatus of the present invention includes a reference image creating unit that employs a normal image to be inspected as a non-defective product as a reference image, and an upper limit threshold value that is set in advance for the luminance value of each pixel of the reference image. A threshold image creating means for creating an upper limit threshold image by adding a lower limit threshold value set in advance to the luminance value of each pixel of the reference image and an imaging means for photographing an inspection object And an inspection unit that compares the inspection image obtained by photographing the inspection object with the upper threshold image and the lower threshold image for each pixel to determine pass / fail of the inspection object. .
In addition, one configuration example of the image inspection apparatus of the present invention further includes a noise removing unit that replaces image data of a portion including the noise component of the reference image with image data of a portion not including the noise component. It is what.
Further, in one configuration example of the image inspection apparatus according to the present invention, the inspection unit includes a pixel having a higher luminance value than a pixel of the upper threshold image and a pixel of the lower threshold image among the pixels of the inspection image. A pixel having a low luminance value is a pixel having a first luminance value, and a pixel having a luminance value equal to or lower than the luminance value of the pixel of the upper threshold image and equal to or higher than the luminance value of the pixel of the lower threshold image is binary. A defect location extracting unit that generates a binarized image; and a determination unit that performs an inspection on the binarized image based on a predetermined determination criterion and determines whether the inspection target is good or bad.

  According to the present invention, by adopting one normal image that has been determined to be non-defective in advance as a reference image, compared to a case where an average image of a plurality of normal images is used as a reference image as in a conventional inspection apparatus. It is possible to improve abnormalities such as blurring of the outline of the reference image. In the present invention, an upper limit threshold image is created by adding a preset upper threshold to the luminance value of each pixel of the reference image, and a preset lower limit threshold is added to the luminance value of each pixel of the reference image. By creating the lower threshold image, excessive distortion of the upper and lower thresholds at the pattern edge portion can be suppressed. As a result, according to the present invention, it is possible to extract defect portions to be inspected without excess and deficiency, and to realize high-precision inspection. Further, in the present invention, since one normal image has only to be selected in order to create a reference image, the time for creating a reference image is longer than in the case of selecting a plurality of normal images as in the conventional inspection apparatus. Can be shortened.

  In the present invention, the noise component of the reference image can be removed by replacing the image data of the reference image including the noise component with the image data of the portion not including the noise component. A small abnormal portion that exists is not reflected in the reference image, and a more appropriate reference image can be created.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of an image inspection apparatus according to an embodiment of the present invention.
The image inspection apparatus processes an XYZ table 2 on which a wafer 1 to be inspected is placed, a CCD camera 3 as an imaging means for imaging a chip of the wafer 1, and an inspection image photographed by the CCD camera 3 to determine pass / fail of the chip. The image processing apparatus 4 is configured to include a display apparatus 5 for displaying inspection results and the like, and an input apparatus 6 such as a keyboard and a mouse.

  The image processing apparatus 4 includes an A / D converter 400, a storage unit 401, a reference image creation unit 402, a noise removal unit 403, a threshold image creation unit 404, an inspection image creation unit 405, and an alignment unit 406. An expansion / contraction processing unit 407, a defect location extraction unit 408, a determination unit 409, an imaging control unit 410, a display control unit 411, and an operation input unit 412. The alignment unit 406, the expansion / contraction processing unit 407, the defect location extraction unit 408, and the determination unit 409 constitute inspection means.

  FIG. 2 is a flowchart showing the operation of the image inspection apparatus of FIG. First, the reference image creation unit 402 employs a normal image of a chip that has been determined to be non-defective in advance as a reference image (step S1 in FIG. 2), and the threshold image creation unit 404 presets the luminance value of each pixel of the reference image. The upper limit threshold value is added to create an upper limit threshold image, and a preset lower limit threshold value is added to the luminance value of each pixel of the reference image to create a lower limit threshold image (step S2). The inspection image obtained by photographing the chip, the upper threshold image and the lower threshold image are compared for each pixel to determine whether the chip is good or bad (step S3).

Hereinafter, each process of FIG. 2 will be described in detail. First, a method for creating a reference image will be described. FIG. 3 is a flowchart showing a method of creating a reference image by the image inspection apparatus according to the present embodiment.
The reference image creation unit 402 controls the XYZ table 2 and the CCD camera 3 via the imaging control unit 410, and causes the CCD camera 3 to take an image of a chip near the center of the wafer 1 (step S100 in FIG. 3). The image signal obtained by the CCD camera 3 is converted into image data by the A / D converter 400 and stored in the storage unit 401.

  Subsequently, the display control unit 411 displays the chip image stored in the storage unit 401 on the display device 5 in accordance with an instruction from the reference image creation unit 402 (step S101).

  The operator of the image inspection device looks at the chip image displayed on the display device 5 to confirm whether there is a large abnormal portion (step S102). In particular, since it is difficult to remove noise, which will be described later, for a fine pattern, it is confirmed whether there is an abnormal portion in the fine pattern. If there is an obvious abnormal portion (determination NO in step S102), the operator is inappropriate as the reference image, and the operator operates the input device 6 to input a determination result indicating “defective product”. The reference image creation unit 402 that has received the determination result via the operation input unit 412 returns to step S100 to photograph another chip. If there is no large abnormal portion, the operator operates the input device 6 and inputs a determination result indicating “non-defective”. This completes the selection of the reference image.

  Next, the noise removing unit 403 removes a small abnormal portion (noise component) present in the selected reference image (step S103). FIG. 4 is a diagram for explaining the noise removal processing of the reference image. In the noise removal process, the image data of the part including the noise component of the reference image is replaced (masking process) with the image data of the part not including the noise component.

  In the example of FIG. 4A, there is a noise component whose luminance value is locally reduced at a location A of the reference image 50 due to, for example, insufficient pattern thickness. Therefore, by replacing the image data of the location A with the image data of the location B that does not include a noise component, the noise component of the location A can be removed as shown in FIG. As the location B not including the noise component, a location having the same brightness as the original luminance value of the location A (a location adjacent to the location A in the example of FIG. 4A) may be selected. The operator specifies the location including the noise component and the location not including the noise component by operating the input device 6. This completes the creation of the reference image. The created reference image is stored in the storage unit 401.

  In FIG. 3, the images of the chips are taken one by one. However, the present invention is not limited to this. Images of a plurality of chips are taken at the same time, and a reference image is selected from these taken images. You may do it.

Next, a method for creating an upper threshold image and a lower threshold image will be described. FIG. 5 is a flowchart showing a method of creating a threshold image by the image inspection apparatus according to the present embodiment.
The threshold image creation unit 404 detects the luminance values of all the pixels of the reference image (step S200). Subsequently, the threshold image creating unit 404 acquires a preset upper limit threshold (for example, +10) from the storage unit 401 (step S201), and adds the upper limit threshold to the luminance value of each pixel of the reference image, thereby obtaining the upper limit. A threshold image is created (step S202). If the luminance value of the pixel i of the reference image is Lref (i) and the upper threshold is THu, the luminance value Lu (i) of the pixel i of the upper threshold image is as follows.
Lu (i) = Lref (i) + THu (1)

Further, the threshold image creating unit 404 acquires a preset lower limit threshold (for example, −10) from the storage unit 401 (step S203), and adds the lower limit threshold to the luminance value of each pixel of the reference image, thereby obtaining the lower limit. A threshold image is created (step S204). When the lower threshold is THd, the luminance value Ld (i) of the pixel i in the lower threshold image is expressed by the following equation.
Ld (i) = Lref (i) + THd (2)

  This completes the creation of the threshold image. The created upper threshold image and lower threshold image are stored in the storage unit 401. FIG. 6 shows an example of luminance values of the reference image, the upper threshold image, and the lower threshold image. The upper limit threshold THu and the lower limit threshold THd may be set in advance so that pattern defects can be extracted without excess or deficiency. Specifically, the luminance variation in the lot is measured in advance, and the upper limit threshold value THu and the lower limit threshold value THd may be set based on the measured luminance variation.

Next, a method for inspecting chips on the wafer 1 will be described. FIG. 7 is a flowchart showing a chip inspection method by the image inspection apparatus of the present embodiment.
The inspection image creation unit 405 controls the XYZ table 2 and the CCD camera 3 via the imaging control unit 410, and causes the CCD camera 3 to take an image of the chip of the wafer 1 (step S300 in FIG. 7). An image signal obtained by the CCD camera 3 is converted into image data of an inspection image by the A / D converter 400 and stored in the storage unit 401. Thus, an inspection image is created.

  Next, the alignment unit 406 aligns the upper threshold image and the lower threshold image stored in the storage unit 401 with the inspection image (step S301). In this alignment process, the alignment unit 406 determines whether the position of a predetermined positioning mark (or a pattern corresponding to the positioning mark) existing in the upper threshold image and the lower threshold image exists in the corresponding inspection image (or The inspection image may be moved and rotated so as to coincide with the position of the pattern corresponding to the positioning mark. It goes without saying that the upper threshold image and the lower threshold image may be moved and rotated.

  Subsequently, the expansion / contraction processing unit 407 performs expansion processing of 8 neighborhoods on the upper threshold image (step S302), and performs contraction processing of 8 neighborhoods on the lower threshold image (step S303). FIG. 8 is a diagram for explaining pixels in the vicinity of 8, and FIG. 9 is a diagram for explaining an expansion process for the upper limit threshold image. The squares in FIGS. 8 and 9 represent pixels, and the numerical values in the squares in FIG. 9 represent the luminance of the pixels.

Eight neighbors represent eight neighboring pixels adjacent to the target pixel 90 as shown in FIG. The upper threshold image expansion process is a process in which the luminance value of the target pixel is compared with the luminance value of the neighboring pixel, and the maximum value is set as the luminance value of the neighboring pixel. As a result of comparing the luminance value of the pixel of interest in the center of FIG. 9A with the luminance value of the neighboring pixels, it appears that the bright region is expanded (FIG. 9B).
On the other hand, the contraction process of the lower threshold image is a process in which the luminance value of the target pixel is compared with the luminance value of the neighboring pixel and the minimum value is set as the luminance value of the neighboring pixel. When the shrinking process is performed, the bright area appears to shrink as a result.

The reason for performing the upper threshold image expansion process and the lower threshold image contraction process is to absorb an error for one pixel of alignment so that the alignment error does not affect the result when performing a comparison process described later. It is to make it.
In the processing in steps S302 and S303, processing for one pixel is performed over the entire area of the threshold image. However, depending on the region of the image, if it is considered that the positional deviation is larger, only for a specific part, Processing may be performed with a large expansion amount and contraction amount.

  Next, the defect location extraction unit 408 compares the upper limit threshold image after the expansion process and the lower limit threshold image after the contraction process with the inspection image, and extracts a defect location (step S304). In this comparison process, the defect location extraction unit 408 compares the upper limit threshold image after the expansion process and the inspection image for each corresponding pixel, and compares the lower limit threshold image after the contraction process and the inspection image for each corresponding pixel. In comparison, when the luminance value of the pixel of the inspection image exceeds the luminance value of the pixel of the upper threshold image or falls below the luminance value of the pixel of the lower threshold image, the pixel of the inspection image is set to “1” (white) When the luminance value of the pixel of the inspection image is equal to or lower than the luminance value of the pixel of the upper threshold image and equal to or higher than the luminance value of the pixel of the lower threshold image, the pixel of the inspection image is set to “0” (black). Thereby, a binarized image is obtained in which the defective portion is white and the portion having no defect is black. This binarized image is stored in the storage unit 401.

The determination unit 409 performs a process of collectively collecting a region where white pixels are continuous by a labeling process for tracking the edge of the white paper pixel of the binarized image, and measures the area of this region, that is, the number of pixels (step). S305).
The determination unit 409 compares the measurement result of step S305 with a predetermined determination criterion, and determines whether there is an area where the number of pixels exceeds the determination criterion (step S306). The determination unit 409 determines the inspection image chip as a defective product if there is an area where the number of pixels exceeds the determination criterion, and determines that the inspection image chip is a non-defective product if the number of pixels in all the regions is within the determination criterion. Judge as.

Subsequently, the determination unit 409 generates a result image. The display control unit 411 displays the result image on the display device 5 according to an instruction from the determination unit 409 (step S307). The result image is to display the determination result (non-defective product / defective product) at the same time that the defect location is surrounded by, for example, a circle on the inspection image so that the defect location can be easily grasped visually.
Then, the determination unit 409 determines whether or not all chips on the wafer 1 have been inspected (step S308). If there is an uninspected chip on the wafer 1, the process returns to step S300. The inspection is completed when the processing of steps S300 to S307 is completed for all the chips.

  As described above, in the present embodiment, by adopting a single normal image determined in advance as a non-defective product as a reference image, an average image of a plurality of normal images is used as a reference image as in a conventional inspection apparatus. Compared to the case, it is possible to improve abnormalities such as blurring of the outline of the reference image. In the present embodiment, an upper limit threshold image is created by adding a preset upper limit threshold to the luminance value of each pixel of the reference image, and a lower limit threshold set in advance to the luminance value of each pixel of the reference image. Is added to create a lower threshold image, so that excessive distortion of the upper and lower thresholds at the pattern edge portion can be suppressed. As a result, in the present embodiment, it is possible to extract a defect portion to be inspected without excess or deficiency, and to realize a highly accurate inspection.

  Further, in the present embodiment, since one normal image may be selected in order to create a reference image, the reference image is compared with a case where a plurality of normal images are selected as in a conventional inspection apparatus. Creation time can be shortened. In the present embodiment, since the noise component of the reference image can be removed, a small abnormal portion existing in the normal image is not reflected in the reference image, and a more appropriate reference image can be created. it can.

  FIG. 10 is a diagram comparing a reference image and a threshold image created by the conventional inspection device disclosed in Patent Document 1 and the image inspection device of the present embodiment, and FIG. 10A shows an example of the inspection image. FIG. 10B is a diagram showing the luminance values of the reference image and threshold image created by the conventional inspection device, and FIG. 10C is the reference image and threshold image created by the image inspection device of the present embodiment. It is a figure which shows the luminance value. FIGS. 10B and 10C show the luminance values of the reference image and the threshold image created for the position of the I-I ′ line of the inspection image of FIG. Lref is the luminance value of the reference image, Lu is the luminance value of the upper threshold image, and Ld is the luminance value of the lower threshold image.

  As is clear from the comparison between FIG. 10B and FIG. 10C, according to the present embodiment, blurring of the outline of the reference image (dullness of the pattern edge portion) is improved, and the upper limit at the pattern edge portion is improved. It can be seen that excessive distortion of the threshold and the lower threshold is suppressed.

  FIG. 11 is a diagram comparing the defect location extraction results of the conventional inspection apparatus disclosed in Patent Document 1 and the image inspection apparatus of the present embodiment. FIG. 11A shows an inspection image of a chip in which a linear defect 121 of the pattern is generated. FIG. 11B shows a defect location extraction result obtained by a conventional inspection apparatus from the inspection image of FIG. FIG. 11C shows a defect location extraction result obtained from the inspection image of FIG. 11A by the image inspection apparatus of the present embodiment.

  FIG. 11D shows an inspection image of a chip in which a substrate defect 122 has occurred, FIG. 11E shows a defect location extraction result obtained by a conventional inspection apparatus from the inspection image of FIG. 11D, and FIG. F) is a defect location extraction result obtained by the image inspection apparatus of the present embodiment from the inspection image of FIG.

  As is apparent from the comparison of the defect location extraction results, according to the present embodiment, it is possible to extract a defect location with a smaller contrast difference from the background, and it is clear that the outline of the defect location is also clearer. .

In FIG. 7, the images of the chips are taken one by one. However, the present invention is not limited to this, and inspection images of a plurality of chips are taken at the same time, and these inspection images are inspected in order or in parallel. Needless to say.
In this embodiment, the sensor chip formed on the wafer is an inspection target. However, the present invention is not limited to this, and the present invention provides stable luminance variation within the lot (luminance variation <defect and background to be extracted). Can be applied to various inspection objects.

  The image processing apparatus 4 according to the present embodiment can be realized by a computer having a CPU, a storage device, and an external interface. A program for realizing the image inspection method of the present invention in such a computer is provided in a state of being recorded on a recording medium such as a flexible disk, a CD-ROM, a DVD-ROM, or a memory card. The CPU writes the program read from the recording medium into the storage device, and executes processing as described in this embodiment according to the program.

  The present invention can be applied to a technique for inspecting a fine defect to be inspected with an image.

It is a block diagram which shows the structure of the image inspection apparatus which concerns on embodiment of this invention. It is a flowchart which shows operation | movement of the image inspection apparatus which concerns on embodiment of this invention. It is a flowchart which shows the production method of the reference | standard image by the image inspection apparatus which concerns on embodiment of this invention. It is a figure for demonstrating the noise removal process of the reference | standard image in embodiment of this invention. It is a flowchart which shows the preparation method of the threshold value image by the image inspection apparatus which concerns on embodiment of this invention. It is a figure which shows an example of the luminance value of the reference | standard image in the embodiment of this invention, an upper limit threshold image, and a lower limit threshold image. It is a flowchart which shows the test | inspection method of the chip | tip by the image inspection apparatus which concerns on embodiment of this invention. It is a figure for demonstrating the pixel of 8 vicinity in an expansion process and a contraction process. It is a figure for demonstrating the expansion process of the upper limit threshold value image in embodiment of this invention. It is the figure which compared the reference | standard image and threshold value image which were produced with the conventional test | inspection apparatus and the image test | inspection apparatus of embodiment of this invention. It is the figure which compared the defect location extraction result of the conventional inspection apparatus and the image inspection apparatus of embodiment of this invention. It is a figure for demonstrating the production method of the reference | standard image in the conventional inspection apparatus. It is a figure which shows an example of the luminance value of the upper limit threshold image in the conventional test | inspection apparatus, and a lower limit threshold image.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Wafer, 2 ... XYZ table, 3 ... CCD camera, 4 ... Image processing apparatus, 5 ... Display apparatus, 6 ... Input device, 400 ... A / D converter, 401 ... Memory | storage part, 402 ... Reference | standard image creation part, 403 ... Noise removal unit, 404 ... Threshold image creation unit, 405 ... Inspection image creation unit, 406 ... Positioning unit, 407 ... Expansion / contraction processing unit, 408 ... Defect location extraction unit, 409 ... Determination unit, 410 ... Imaging control unit 411: Display control unit, 412: Operation input unit.

Claims (6)

A reference image creation procedure for adopting a normal image to be inspected as a reference image in advance as a non-defective product;
An upper limit threshold image is created by adding a preset upper threshold value to the luminance value of each pixel of the reference image, and a lower limit threshold value is added to the luminance value of each pixel of the reference image. A threshold image creation procedure for creating an image;
An image inspection method comprising: an inspection procedure in which an inspection image obtained by photographing an inspection object is compared with the upper threshold image and the lower threshold image for each pixel to determine pass / fail of the inspection object. The image inspection method according to claim 1,
The image inspection method according to claim 1, wherein the reference image creation procedure includes a noise removal procedure in which image data of a portion including a noise component of the reference image is replaced with image data of a portion not including a noise component. The image inspection method according to claim 1,
The inspection procedure includes:
Among the pixels of the inspection image, a pixel having a luminance value higher than that of the pixel of the upper limit threshold image and a pixel having a luminance value lower than that of the pixel of the lower limit threshold image are set as pixels having a first luminance value, and the upper limit threshold value A defect location extraction procedure for generating a binarized image with pixels having a luminance value equal to or lower than the luminance value of the pixel of the image and equal to or higher than the luminance value of the pixel of the lower threshold image;
An image inspection method comprising: a determination procedure for performing an inspection on the binarized image based on a predetermined determination criterion and determining whether the inspection object is good or bad. A reference image creating means that adopts as a reference image a normal image to be inspected that has been previously determined to be non-defective,
An upper limit threshold image is created by adding a preset upper limit threshold value to the luminance value of each pixel of the reference image, and a lower limit threshold value is added to the luminance value of each pixel of the reference image. Threshold image creating means for creating an image;
Photographing means for photographing the inspection object;
An image inspection apparatus comprising: an inspection unit that compares an inspection image obtained by photographing an inspection object with the upper threshold image and the lower threshold image for each pixel and determines whether the inspection object is good or bad. The image inspection apparatus according to claim 4, wherein
The image inspection apparatus further comprises noise removing means for replacing the image data of the reference image including the noise component with the image data of the portion not including the noise component. The image inspection apparatus according to claim 4, wherein
The inspection means includes
Among the pixels of the inspection image, a pixel having a luminance value higher than that of the pixel of the upper limit threshold image and a pixel having a luminance value lower than that of the pixel of the lower limit threshold image are set as pixels having a first luminance value, and the upper limit threshold value A defect location extracting means for generating a binarized image using a pixel equal to or lower than the luminance value of the pixel of the image and equal to or higher than the luminance value of the pixel of the lower threshold image as a pixel of the second luminance value;
An image inspection apparatus comprising: a determination unit that performs an inspection on the binarized image based on a predetermined determination criterion and determines whether the inspection target is good or bad.