JP5405245B2 - Image inspection method and image inspection apparatus - Google Patents

Description

  The present invention relates to an image inspection method and an image inspection apparatus for determining whether or not an image to be inspected that shows the appearance of a product in order to determine whether there is a defect in the product to be inspected, for example.

  2. Description of the Related Art Conventionally, a method is known in which the presence or absence of defects in a product to be inspected is determined from an image to be inspected by photographing the appearance of the product. As one of the methods, there is a method for inspecting the quality of an inspected image using the Mahalanobis Taguchi method. The basic procedure for image inspection by the Mahalanobis Taguchi method is as follows.

1) Feature values are extracted from a plurality of product images in the most normal state according to a plurality of items representing the state. In this way, a data group of normal products and non-defective products is referred to as a reference space, and becomes a reference when performing determination, diagnosis, and prediction by the Mahalanobis Taguchi method.
2) Each feature value is normalized using the average value and standard deviation of the feature values calculated for each item.
3) Calculate all correlations between the standardized feature quantities and create a correlation matrix.
4) The Mahalanobis distance of the reference space, which is a group of products in the most normal state, is calculated using the normalized feature quantity and the inverse matrix of the correlation matrix. The average value of the Mahalanobis distance in this reference space is about 1.
5) The Mahalanobis distance is calculated using the same procedure as above for the product image (inspected image) whose state is unknown.
6) Based on a preset threshold value of hamaranobis distance (good / bad determination reference value), the state of the product whose state is unknown is determined. In the Mahalanobis Taguchi method, the Mahalanobis distance increases as the degree of abnormality increases. Therefore, if the Mahalanobis distance becomes larger than the reference value, it is determined that the product is in an abnormal state (defective). In this way, data that does not belong to the reference space is called a signal space, and its state is determined, diagnosed, and predicted based on the Mahalanobis distance.

  In this way, the Mahalanobis-Taguchi method uses a multivariate analysis method to determine the state of a product, so that various information (features) can be obtained in the same manner as a visual inspection in which an operator visually determines the state of a product. Comprehensive judgment based on this can be made. Various methods have been proposed in order to accurately inspect the quality (for example, see Patent Documents 1 and 2).

  For example, Patent Document 1 discloses that primary parameters such as luminance and density, average values of primary parameters, and standard deviations are used as image feature amounts used for calculation of Mahalanobis distance in order to accurately inspect the quality of the entire image to be inspected. A feature quantity distribution such as a luminance distribution of the entire image to be inspected is used instead of a secondary parameter such as. This is because the Mahalanobis distance is calculated based on the feature amount including the information on the entire image to be inspected, so that the quality of the entire image to be inspected can be determined more accurately than when inspecting with some feature amounts of the image to be inspected. The aim is to be able to inspect.

  Patent Document 2 discloses that an image to be inspected is divided into small regions, feature amounts are extracted from each small region, and a statistical distance measure (Mahalanobis distance, etc.) between the extracted feature amount and a predetermined feature amount reference value. Describes a method for inspecting the quality of an image to be inspected by using a plurality of sizes of small areas to be divided. This aims at accurately detecting not only local defects but also global defects.

JP-A-2005-252451 JP 2007-132757 A

  However, since the method described in Patent Document 1 calculates the Mahalanobis distance based on the luminance distribution, this method is used for the appearance inspection of products, particularly defects in products with a large change in appearance and steep feature amounts in the appearance. When used for the appearance inspection of defects in products with various changes, there is a problem that the feature quantity of the defects is hidden in the noise and cannot be inspected with high accuracy.

  In addition, although the method described in Patent Document 2 has a defect, the Mahalanobis distance may not be determined because the Mahalanobis distance is small. Therefore, a second determination is added in addition to the determination based on the Mahalanobis distance. doing. For this reason, there is a problem that the efficiency of the inspection is lowered.

  The present invention has been made in view of the above problems, and an object thereof is to provide an image inspection method and an image inspection apparatus capable of inspecting the quality of an image to be inspected accurately and efficiently.

In order to solve the above problems, the present invention is an image inspection method for determining whether an image is good or bad,
A non-defective image dividing step for dividing the non-defective image into a plurality of images;
A calculation step of calculating a Mahalanobis distance for each good product divided image that is an image divided in the good product image dividing step;
Based on the Mahalanobis distance calculated in the calculation step, a reference value setting step for setting a pass / fail determination reference value that is a Mahalanobis distance that is a reference for pass / fail determination of an image to be inspected for each of the non-defective product divided images;
A division step of dividing the image to be inspected into a plurality of images so as to be the same division as the division of the non-defective image ;
A creation step for creating a signal space consisting of Mahalanobis distances for each inspected divided image that is an image divided in the division step;
The Mahalanobis distance of the object to be inspected divided picture images, by comparing with the non-defective dividing the quality determination reference value set in the image corresponding to said inspection image segment quality performing quality determination in each of the inspection divided images and the determination step, only including,
In the non-defective image dividing step, the non-defective product divided images having a plurality of sizes are obtained by dividing the non-defective images into different sizes.
In the reference value setting step, for each of the plurality of types of sizes, the pass / fail judgment reference value for each non-defective product divided image is set,
In the pass / fail judgment step, as the pass / fail judgment, in addition to determining the inspected divided image as good / bad according to the comparison result, a defect exists in the inspected divided image without being determined as good or bad. It is a step of determining that there is a possibility of provisional failure,
A subdivision step for further dividing the inspected divided image into a plurality of images so as to be the same size as one of the plurality of sizes, when there is the inspected divided image determined to be the provisional failure;
Re-creating step for creating a signal space consisting of Mahalanobis distance again for each provisional defective divided image that is an image divided in the re-dividing step;
The Mahalanobis distance of the tentative defect divided picture images, by comparing the said quality determination reference value set in the non-defective divided images corresponding to the provisional failure divided image, the quality determination in each of the interim defect divided images again A re-defective determination step to be performed;
Including
If there is the temporary defective divided image newly determined as the provisional defect in the re-defective determination step, the re-dividing step is performed on the temporary defective divided image until it can be determined as good or defective. The re-creation step and the re-quality determination step are repeatedly executed.

  According to this, since the pass / fail determination is performed for each divided image obtained by dividing the image to be inspected, it is possible to reduce that the feature amount of the defect is hidden in the noise and to inspect with high accuracy. In addition, in the pass / fail judgment step, there is provided a case in which it is determined that there is a possibility that a defect exists in the inspected divided image without being determined to be good or bad. It is possible to prevent a misjudgment such as a judgment or a judgment that the judgment is bad despite being good. Then, when it is determined that the provisional failure, the image is divided into small sizes and the pass / fail determination is repeated until the determined divided image is determined to be good or bad. Therefore, it is possible to perform the quality determination with higher accuracy than in the quality determination. Further, since the divided image determined to be provisional defective is re-divided and the pass / fail determination is performed again, it is not necessary to reduce the size of the divided image to be inspected from the beginning. Furthermore, since the quality determination can be made with high accuracy based on the Mahalanobis distance, it is not necessary to add a second determination in addition to the determination based on the Mahalanobis distance. Therefore, it is possible to efficiently inspect the quality of the inspected image.

In the image inspection method of the present invention, in the reference value setting step, two reference values, a first reference value and a second reference value larger than the first reference value, are used as the pass / fail judgment reference value. Set for each non-defective product split image,
The pass / fail determination step and the re-pass / fail determination step are performed when the Mahalanobis distance of the divided image to be the pass / fail determination target is smaller than the first reference value set in the corresponding non- defective product split image. The image is determined to be good, and when the Mahalanobis distance of the divided image is larger than the second reference value set for the corresponding non- defective divided image, the divided image is determined to be defective. When the Mahalanobis distance is larger than the first reference value and smaller than the second reference value, the divided image is determined as the provisional defect.

  In this way, provision of two reference values makes it possible to determine provisional defects in addition to good and bad divided images. The two reference values are provided for each of the inspected divided image and the provisional defective divided image, so that the pass / fail judgment can be made accurately for each of the divided images.

Further, in the image inspection method of the present invention, in the non- defective image dividing step, the step of dividing the non-defective images so as to be divided into the same division is executed a plurality of times while changing the non-defective image division size. , For each of the plurality of sizes, so as to obtain a plurality of non-defective product divided images having the same position in the non-defective product image,
In the reference value setting step, the pass / fail judgment reference value is set based on the maximum value among the Mahalanobis distances of the plurality of good product divided images having the same position in the good product image for each of the plurality of sizes .

Thus, good determination reference value, so determined maximum value of the Mahalanobis distance of a plurality of non-defective divided image as a reference, it is possible to further accurately quality decision.

The present invention is an image inspection apparatus for determining the quality of an image,
A non-defective image dividing means for dividing a non-defective image into a plurality of images;
Calculating means for calculating the Mahalanobis distance for each good divided image that is an image divided by the good image dividing means;
Based on the Mahalanobis distance calculated by the calculation means, a reference value setting means for setting a pass / fail judgment reference value that is a Mahalanobis distance serving as a reference for pass / fail judgment of the image to be inspected for each non-defective product divided image;
A dividing means for dividing the image to be inspected into a plurality of images so as to be the same division as the division of the non-defective image ;
Creating means for creating a signal space consisting of Mahalanobis distance for each divided image to be examined, which is an image divided by the dividing means;
The Mahalanobis distance of the object to be inspected divided picture images, by comparing with the non-defective dividing the quality determination reference value set in the image corresponding to said inspection image segment quality performing quality determination in each of the inspection divided images Determination means ,
The non-defective image dividing means divides a plurality of non-defective images into different sizes so as to obtain the non-defective product divided images of a plurality of sizes,
The reference value setting means sets the pass / fail judgment reference value for each non-defective product divided image for each of the plurality of sizes.
The pass / fail judgment means determines whether the test divided image is good / bad according to the comparison result as the pass / fail judgment, and there is a defect in the test divided image without being determined to be good or bad. It is determined as a provisional defect that there is a possibility,
When there is the inspected divided image determined to be the provisional failure, a re-dividing unit that further divides the inspected divided image into a plurality of images so as to be the same size as one of the plurality of sizes ;
Re-creating means for creating a signal space consisting of Mahalanobis distance again for each provisional defective divided image that is an image divided by the re-dividing means,
The Mahalanobis distance of the tentative defect divided picture images, by comparing the said quality determination reference value set in the non-defective divided images corresponding to the provisional failure divided image, the quality determination in each of the interim defect divided images again Re-pass / fail judgment means to perform,
In the case where there is the provisional defect divided image newly determined as the provisional defect by the re-defective / non-defective determination unit, the re-segmentation unit is applied to the provisional defect divided image until it can be determined as good or defective. The processing by the re-creation unit and the re-defectiveness determination unit is repeatedly executed.

  According to this, the same effect as the image inspection method of the present invention can be obtained.

1 is an external view of an image inspection apparatus 1. FIG. FIG. 4 is a view showing a surface of a work 4. It is a general flowchart for creating a reference space. It is a general flowchart for calculating the Mahalanobis distance of an image to be inspected and performing pass / fail judgment. It is a detailed flowchart for creating a reference space. It is a figure for demonstrating the division | segmentation method of the good quality image 41. FIG. It is the figure which showed the good quality division | segmentation image 411 in the same position between the some good quality images 41. FIG. It is a detailed flowchart for calculating the Mahalanobis distance of the image to be inspected and performing pass / fail judgment. It is the figure which showed the to-be-inspected images 42-44 as an illustration of to-be-inspected image. It is the figure which showed having further divided | segmented the to-be-inspected division | segmentation image 431 (1) and the temporary defect division | segmentation image 431 (1-1) containing the error 81. FIG. It is the figure which showed having further divided | segmented the to-be-inspected divided image 441 (5) in which the small defect 72 was included.

  The best mode for carrying out the present invention will be described below with reference to the drawings. In FIG. 1, the external view of the image inspection apparatus 1 of this invention is shown. The image inspection apparatus 1 is an apparatus that determines the quality of the workpiece 4 by determining whether the workpiece 4 shown in FIG. As shown in FIG. 1, the image inspection apparatus 1 includes a camera 2, an illumination 3, an electronic calculator 5, and a display device 6.

  The camera 2 images the external appearance of the work 4, and a CCD camera, a CMOS camera, an analog camera, or the like can be used. In this example, a CCD camera is used as the camera 2.

  The illumination 3 is used to adjust the brightness of the imaged portion of the workpiece 4 when the camera 2 captures the appearance of the workpiece 4, and fiber illumination, LED illumination, ring illumination, coaxial epi-illumination, or the like can be used. it can. In this example, LED ring illumination is used as the illumination 3. The image data picked up by the camera 2 is sent to the electronic computer 5, and after the data processing is performed by the electronic computer 5, pass / fail judgment is performed, and the result is displayed on the display device 6.

  On the other hand, the workpiece 4 is a product to be inspected. In this example, the workpiece 4 is a cut metal part, and is used to inspect the surface of the workpiece 4 for defects such as scratches, burrs, and foreign matters. FIG. 2 is a view showing the surface of the workpiece 4. As shown in FIG. 2, the inspection target workpiece 4 exemplified here includes portions having various characteristics 450 (for example, patterns and shapes) in the appearance. Therefore, the work has a sharp change in the feature amount indicated by the feature 450 in the appearance.

  Then, the image inspection apparatus 1 determines the quality of the workpiece 4 using the Mahalanobis Taguchi method. Here, FIG. 3 and FIG. 4 show general flowcharts showing a method of determination using the Mahalanobis-Taguchi method. FIG. 3 is a flowchart for calculating the Mahalanobis distance of a plurality of product images in the most normal state, that is, a non-defective image group having the highest appearance frequency, that is, for creating a reference space. First, a good original image is input, and image processing such as filtering and edge enhancement is performed on the input image as necessary (S11). Then, after the input image is divided (S12), the feature amount is extracted from each divided image (S13), and the Mahalanobis distance is calculated for each divided image at the same position among the plurality of images (S14).

  On the other hand, FIG. 4 is a flowchart for calculating the Mahalanobis distance of the image to be inspected and making a pass / fail judgment. First, an inspected image of an inspection product is input, and if necessary, image processing such as filtering and edge enhancement is performed on the input image to facilitate detection of defects (S21). Then, the image to be inspected is divided (S22), the feature amount is extracted from each divided image (S23), and the Mahalanobis distance is calculated (S24). The calculated Mahalanobis distance is compared with a reference value for pass / fail determination (S25). If the calculated Mahalanobis distance is smaller than the reference value, it is determined to be a non-defective product, and the result is displayed on the display device 6 (S26). Then, the process of the flowchart of FIG. 4 is complete | finished. On the other hand, if the calculated Mahalanobis distance is larger than the reference value, it is determined as a defective product, the position where the defect exists is specified, the information is stored in a storage device (not shown) of the electronic computer 5, and the result is displayed. The data is displayed on the display device 6 (S27). Then, the process of the flowchart of FIG. 4 is complete | finished.

  As described above, the present invention determines pass / fail for each divided image. Furthermore, the present invention provides a case in which two reference values for pass / fail determination are provided and it is determined that there is a possibility that a defect exists in the inspected divided image without being determined to be good or bad. When it is determined that the provisional failure is detected, the image is divided into small sizes and the pass / fail determination is repeated until the determined divided image is determined to be good or bad. Details will be described below.

  First, a reference space creation method will be described with reference to the flowchart shown in FIG. 5 and FIGS. 6 and 7. First, product images of a plurality of workpieces 4 in the most normal state are captured by the camera 2, and the image data is stored in a storage device (not shown) of the electronic computer 5 (S31). Here, the product image in the most normal state is an image of a good product group having the highest appearance frequency in the case of image inspection. This is because if the appearance frequency is high, it can be said that the probability that the work 4 is normal and non-defective is high. In addition, the number of non-defective images to be inputted is inputted and n pieces of non-defective images are inputted.

  Then, image processing such as filtering and edge enhancement is performed on each non-defective image input in S31 (S32). Here, edge enhancement processing is performed. Originally, this image processing is intended to make it easier to detect defects when there is a defect in the image to be inspected, but it is necessary to perform inspection under the same conditions regardless of whether the product is non-defective or defective. Image processing is also applied to non-defective images that are known not to exist.

  Next, as shown in FIG. 6, the individual non-defective images 41 subjected to the image processing are divided into arbitrary sizes and arbitrary sizes (S33). The division direction may be determined in consideration of the defect occurrence direction and the division size in consideration of the size of the defect to be detected. Here, the original image 41 of 420 × 1320 pixels is converted into the non-defective product divided image 411 (42 × 1320 pixels) ( 1) to 411 (N). That is, the non-defective image 41 is divided into N parts in the vertical width.

  Next, each of the non-defective product divided images 411 (1) to 411 (N) is subjected to arithmetic processing, and a feature amount is extracted based on a predetermined item (S34). This item needs to represent the state of the work 4, but here, when an arbitrary horizontal line is drawn with respect to the luminance waveforms of the respective non-defective product divided images 411 (1) to 411 (N), the horizontal line and the luminance waveform are displayed. The feature amount based on the luminance information is used, such as a differential value that is the number of intersections of, an integral value that is the sum of the sections of the luminance waveform that are above the horizontal line, and a difference between the maximum luminance value and the minimum luminance value.

  Here, the differential value, which is the number of intersections between the horizontal line and the luminance waveform, can be used as an index of the number of changes in the image feature amount. In other words, it can be said that the larger the differential value, the more portions in the image where the feature amount varies. Therefore, by using this differential value as a feature value, even in an image that includes many parts where the feature value varies, it is possible to create a reference space or a signal space that accurately reflects the feature value. It is considered that the quality of the inspected image can be determined with high accuracy.

  Further, the integral value, which is the section of the luminance waveform above the horizontal line, can be used as an index of the ratio of the portion having a certain feature amount or more in the image. In other words, it can be said that the larger the integral value, the more parts in the image that have a certain or larger feature amount. Therefore, by using this integral value as a feature amount, even in an image including many parts having a feature amount above a certain level, a reference space or signal space in which the feature amount is accurately reflected is created. Therefore, it is considered that the quality of the inspected image can be determined with high accuracy.

  Further, the difference between the maximum luminance value and the minimum luminance value can be used as an indicator of the magnitude of the variation of the feature amount in the image. In other words, it can be said that as the difference between the maximum luminance value and the minimum luminance value is larger, a portion where the feature amount fluctuates greatly is included in the image. Therefore, by using this difference as a feature value, a reference space or signal space in which the feature value is accurately reflected can be created even for an image including a part where the feature value greatly varies. It is considered that the quality of the inspection image can be determined with high accuracy.

  In S34, the feature amount is extracted with the position of the horizontal line being variable. As a result, even in an image in which the feature amount fluctuates in a complicated manner, the feature amount can be accurately extracted.

Then, as a result of extracting the feature quantity X _gij from the respective non-defective product divided images 411 (1) to 411 (N) in S34, the result is as shown in Table 1.

Table 1 shows the feature values X _gij of the respective non-defective product divided images 411 (1) to 411 (N) for each of the n non-defective products 41. The extracted feature amount X _gij has k types (the luminance waveform differential value, the luminance waveform integral value, the difference between the maximum luminance value and the minimum luminance value, etc.), and Table 1 shows the characteristic amount for each item. X _gij is shown. For example, as shown in Table 1, the feature quantity of the non-defective image 41 having the image No. “b”, the divided image No. “1”, and the item No. “2” is “X _b12 ”. That is, in Table 1, among the subscripts attached to the feature amount X _gij , the leftmost character g indicates the image No. of the non-defective image 41, and the middle character i indicates the No of the non-defective product divided image 411, The rightmost character j indicates item No. Hereinafter, the character g is used as the image No. of the non-defective image 41, the character i is used as the No of the non-defective product divided image 411, and the character j is used as the item No.

Next, the Mahalanobis distance of each of the non-defective product divided images 411 (1) to 411 (N) is calculated to create a basic space (S35). Specifically, the average value x _gj and the standard deviation σ _gj of the feature amount X _gij are calculated for each item j for each good product image 41. For example, the average value x _gj and the standard deviation σ _gj of the feature amount X _bi1 of the non-defective image 41 with the image No. g = “b” and the item No. j = “1” are the image No. “b”, Table 1 _Calculation is performed using the feature amounts X _{b11 to} X _bn1 of the item No “1”.

Then, the calculated average value x _gj and standard deviation σ _gj are substituted into the following equation (1) to normalize the feature amount X _gij (S35).

The normalized feature amount x _gij is as shown in Table 2. In Table 2, the feature quantities x _gij normalized for each of the divided images 411 (1) to 411 (N) at the same position among the plurality of non-defective images 41 are arranged.

For example, the divided image No of each good product image 41 is “1”, and the feature quantity x _gij of the item “1” is x _a11 , x _b11 ,..., X _n11 as shown in Table 2.

Then, as shown in FIG. 7, for each of the non-defective product divided images 411 (1) to 411 (N) in the same position among the non-defective product images 41, the relationship between the item p and the item q is expressed by the following equation (2). The number r _pq is calculated (S35). Then, the correlation coefficient r _pq is matrixed as in the following equation (3) to calculate a correlation coefficient matrix R (S35).

For example, the correlation coefficient _{r 12} scores 1 in good divided image 411 (1) (p = 1) and Item 2 (q = 2), the above equation (2), to be calculated by the following equation (4) Become. That is, in this case, the correlation coefficient r ₁₂ is calculated using the feature amounts x _gij of the items No. “1” and “2” in the divided image No. “1” in Table 2.

Thus, non-defective divided image 411 (1) ~411 (N) each at by calculating the correlation coefficient _{r pq,} good divided image 411 (1) ~411 (N) per the formula (3) shows a phase relationship A number matrix R is calculated. Then, an inverse matrix A of the correlation coefficient matrix R for each non-defective product divided image 411 (i) as shown in the following equation (5) is calculated.

Then, the Mahalanobis distance MD _ig of the g-th non-defective product image 41 of the i-th non-defective product divided image 411 (i) is calculated by the following equation (6).

Then, the Mahalanobis distance MD _ig is calculated with respect to all the non-defective product divided images 411 of all the non-defective products images 41 using Equation 6 (S35). As a result of calculating the Mahalanobis distance MD _ig for all the non-defective product divided images 411, the results are as shown in Table 3 here. In Table 3, the maximum value of the Mahalanobis distance _{MD ig} of each non-defective divided image 411 is also shown, for example, the maximum value of the Mahalanobis distance _{MD ig} non-defective divided image 411 (1) is "1.1" ing.

Next, based on the Mahalanobis distance MD _ig calculated in S35, a Mahalanobis distance reference value (non-defective product determination reference value) that serves as a reference for determining pass / fail of the inspected image is set for each non-defective product divided image 411 (S36). At this time, two reference values (first reference value and second reference value) are set for each non-defective product divided image 411 as Mahalanobis distance reference values. Specifically, the first reference value is set based on the maximum value of the Mahalanobis distance MD _ig in each good product divided image 411, and a second reference value larger than the first reference value is set. Here, as shown in Table 4 below, the first reference value = the maximum value of the Mahalanobis distance MD _ig (see Table 3) +0.3, and the second reference value = first reference value + 0.2. Set to.

  Subsequently, the above step S33 is executed again to divide the non-defective image 41 into other sizes (21 × 1320 pixels, 7 × 1320 pixels,...) Different from 42 × 1320 pixels. And the process of said step S34-S36 is performed with respect to each non-defective product divided image of the other size, and the first and second reference values are set for each good product divided image. That is, the processes of steps S33 to S36 are repeated to set the first and second reference values for each non-defective product divided image for each size. Thereafter, the set first and second reference values are stored in a storage device (not shown) of the electronic computer 5, and the processing of the flowchart of FIG.

  Next, a method for determining pass / fail by calculating the Mahalanobis distance of the image to be inspected will be described. Here, FIG. 8 is a flowchart showing the processing executed by the electronic computer 5 when determining the quality of the inspected image. The method for calculating the Mahalanobis distance of the image to be inspected is that there is one input image, the average value for each item used for normalization, the standard deviation, and the correlation coefficient matrix R used for calculating the Mahalanobis distance. The inverse matrix A is the same as the method for creating the reference space except that the value at the time of creating the reference space is used. Based on this point, a method for determining pass / fail of an inspected image will be described below with reference to the flowchart of FIG.

  First, a product image of the workpiece 4 to be inspected is picked up by the camera 2 as an image to be inspected, and the image data is stored in a storage device (not shown) of the electronic computer 5 (S41). Then, image processing such as filtering and edge enhancement is performed on the inspection image input in S41 (S42).

  Next, the inspected image subjected to the image processing is divided in the same direction as the non-defective image 41 with the same size (S43). Here, in S33 of FIG. 5 described above, the non-defective image 41 is divided into a plurality of sizes. However, in S43, the largest size among them, that is, a size of 42 × 1320 pixels, is used. Divide the inspection image. Step S43 corresponds to the “dividing step” of the present invention, and the electronic computer 5 executing step S43 corresponds to the “dividing means” of the present invention.

Next, an arithmetic process is performed on each divided image to be inspected, and a feature amount Y _ij is extracted based on the same items as when the reference space is created (S44). That is, when an arbitrary horizontal line is drawn on the luminance waveform of the divided image to be inspected, the differential value that is the number of intersections of the horizontal line and the luminance waveform, the integral value that is the sum of the intervals of the luminance waveform that is above the horizontal line, and the maximum luminance A feature amount Y _ij based on luminance information such as a difference between the value and the minimum luminance value is extracted.

Next, the Mahalanobis distance of each divided image to be inspected is calculated by the same method (S35) as that for creating the reference space, and a signal space is created (S45). Specifically, the feature value Y _ij for each item j for each divided image to be inspected is normalized (inspected reference feature value) by the following equation (7) (S45). At this time, the average value x _j and the standard deviation σ _j when the reference space is created are used.

Then, the Mahalanobis distance MD ′ _i of the i-th inspection divided image is calculated by the following equation (8). At this time, the inverse matrix A of the correlation coefficient matrix R is the one used when the reference space is created.

  Steps S44 and S45 correspond to the “creation step” of the present invention. The computer 5 that executes steps S44 and S45 corresponds to the “creating unit” of the present invention.

Then, for each calculated Mahalanobis distance MD ′ _i , it is determined whether or not the reference value for pass / fail judgment corresponding to the same size as the inspected divided image is larger than the second reference value corresponding to each inspected divided image. Judgment is made (S46). That is, since the size of the inspected divided image is 42 × 1320 pixels, the second reference value in Table 4 is used. For example, for the inspection divided image whose divided image No. is “1”, the Mahalanobis distance MD ′ ₁ and the divided image No. 1 is compared with the second reference value “1.6”.

Here, if even one calculated Mahalanobis distance MD ′ _i is larger than the second reference value (S46: YES), it is determined as a defective product, the position where the defect exists is specified, and the information is electronic The data is stored in a storage device (not shown) of the computer 5 and the result is displayed on the display device 6 (S54). Then, the process of the flowchart of FIG. 8 is complete | finished.

On the other hand, if the Mahalanobis distance MD ′ _i is smaller than the second reference value (S46: NO), it is then determined whether the Mahalanobis distance MD ′ _i is larger than the first reference value in Table 4. (S47). For example, for the inspected divided image having the divided image No. “1”, the Mahalanobis distance MD ′ ₁ and the divided image No. The first reference value “1.4” of 1 is compared. If the Mahalanobis distance MD ′ _i is smaller than the first reference value in all the inspected divided images (S47: NO), the inspected image is determined to be non-defective and the result is displayed on the display device 6 ( S53). Then, the process of the flowchart of FIG. 8 is complete | finished. Steps S46 and S47 correspond to the “good / bad determination step” of the present invention, and the computer 5 that executes steps S46 and S47 corresponds to the “good / bad determination unit” of the present invention.

On the other hand, if at least one Mahalanobis distance MD ′ _i is larger than the first reference value (S47: YES), the inspection divided image having the Mahalanobis distance MD ′ _i is provisionally assumed to have a defect. It determines with it being inferior and performs the process after step S48. That is, first, the inspected divided image is divided into smaller-sized divided images (S48). Specifically, among the non-defective divided image of the reference space created in advance and a plurality of sizes, the image is divided into divided images of 21 × 1320 pixels, which is the next smaller size than the original 42 × 1320 pixels (S48). ). Therefore, the inspected divided image is further divided into two divided images. Step S48 corresponds to the “re-dividing step” of the present invention, and the electronic computer 5 that executes step S48 corresponds to the “re-dividing means” of the present invention.

Next, the newly divided divided image is used as a temporary defective divided image, and the Mahalanobis distance is calculated again for each temporary defective divided image to create a signal space. That is, for each provisional defective divided image, the feature amount Y _ij is extracted again in the same manner as in the previous step S44 (S49). Next, as in the previous step S45, the Mahalanobis distance MD ′ _i of each provisional defective divided image is calculated, and a signal space is created again (S50). Steps S49 and S50 correspond to the “re-creation step” of the present invention, and the electronic computer 5 that executes steps S49 and S50 corresponds to the “re-creation unit” of the present invention.

Next, for each calculated Mahalanobis distance MD ′ _i , it is determined whether or not it is greater than a second reference value corresponding to each provisional defective divided image among the reference values for quality determination corresponding to the same size as the temporary defective divided image. Judgment is made (S51). That is, since the size of the temporary defective divided image this time is 21 × 1320 pixels, the second reference value corresponding to the same position as the temporary defective divided image among the second reference values corresponding to the 21 × 1320 pixels. Is used.

If at least one newly calculated Mahalanobis distance MD ′ _i is larger than the second reference value (S51: YES), it is determined as a defective product, the position where the defect exists is specified, and the information Is stored in a storage device (not shown) of the electronic computer 5 and the result is displayed on the display device 6 (S54). Then, the process of the flowchart of FIG. 8 is complete | finished.

On the other hand, if the new Mahalanobis distance MD ′ _i is smaller than the second reference value (S51: NO), then whether or not the Mahalanobis distance MD ′ _i is larger than the corresponding first reference value is determined. Judgment is made (S52). If the Mahalanobis distance MD ′ _i is smaller than the first reference value in all the provisional defective divided images (S52: NO), the image to be inspected is determined to be non-defective, and the result is displayed on the display device 6 ( S53). Then, the process of the flowchart of FIG. 8 is complete | finished. Steps S51 and S52 correspond to the “re-defective / non-defective determination step” of the present invention, and the computer 5 that executes steps S51 and S52 corresponds to the “re-defective / non-defective determination unit” of the present invention.

On the other hand, if at least one Mahalanobis distance MD ′ _i is larger than the first reference value (S52: YES), the provisional defect divided image of the Mahalanobis distance MD ′ _i is provisionally assumed to have a defect. It is determined again as defective, and the process returns to step S48, and the processes after step S48 are executed again. That is, the provisional defective divided image is divided into smaller divided images (S48). Specifically, among the non-defective divided image of the reference space created in advance and the plurality of sizes, the image is divided into divided images of 7 × 1320 pixels, which is the next smaller size than the original 21 × 1320 pixels (S48). ). Therefore, the provisional defective divided image is further divided into three divided images. Thereafter, the Mahalanobis distance is calculated again for the three sub-provisional defective divided images, and the pass / fail determination is performed again (S49 to S52).

If at least one newly calculated Mahalanobis distance MD ′ _i is larger than the second reference value (S51: YES), it is determined as a defective product (S54), and all Mahalanobis distances MD ′ _i are first. If it is smaller than the reference value (S52: NO), the image to be inspected is determined to be non-defective. On the other hand, if at least one Mahalanobis distance MD ′ _i is larger than the first reference value (S52: YES), the processes of steps S48 to S52 are executed again. That is, when it is determined as a temporary defect, the image is divided into small sizes and the pass / fail determination is repeated until it is determined that the determined temporary defect divided image is good or defective. In this way, it is finally determined whether the inspected image is good or bad.

  Here, the quality determination result when the images to be inspected are inspected images 42 to 44 shown in FIG. 9 will be described. The inspection image 42 shown in FIG. 9A has a large defect 71 and is a defective product. Then, as shown in FIG. 9A, the defect 71 is included in the inspected divided images 421 (1) to (N) of the inspected image 42, and in particular, the inspected divided image 421 (5). It is assumed that many are included.

  Further, the inspected image 43 shown in FIG. 9B includes an error 81 (for example, dust attached to the inspected product) that is not a defect but is not included in the non-defective product image 41. Therefore, since no defect is included, it should be determined as a non-defective product. Then, the error 81 is included in the inspected divided images 431 (1) to (N) of the inspected image 43 as shown in FIG. 9B, and in particular, the inspected divided image 431 (1). It is assumed that many are included.

  In addition, the inspected image 44 shown in FIG. 9C has a small defect 72 and is a defective product. Then, as shown in FIG. 9C, the defect 72 is included in the inspected divided images 441 (1) to (N) of the inspected image 44, and in particular, the inspected divided image 441 (5). It is assumed that many are included.

  When the Mahalanobis distance was calculated for each of the inspected divided images 421, 431, and 441 for the inspected images 42 to 44, the results shown in Table 5 were obtained.

  As shown in Table 5, in the inspected image 42, the divided image No. Since there is a defect 71 at a position in FIG. The Mahalanobis distance of the divided image No. 5 is larger than the second reference value 2.1. Therefore, it was determined that the inspected image 42 is a defective product (S46: YES → S54).

  On the other hand, the inspected image 43 includes the divided image No. 1 because the inspected divided image 431 (1) includes the error 81. The Mahalanobis distance of 1 is divided image No. It was larger than the first reference value 1.4 of 1 and smaller than the second reference value 1.6. The other inspected divided images 431 (2) to 431 (N) were determined to be good because each Mahalanobis distance was smaller than the first reference value. Therefore, it is determined that there is a possibility that a defect exists in the inspected divided image 431 (1) (S47: YES), and the inspection process is continued. That is, as shown in FIG. 10A, the inspected divided image 431 (1) having 42 × 1320 pixels is converted into the provisional defective divided images 431 (1-1) and 431 (1-2) having a size of 21 × 1320 pixels. (S48). At this time, the error 81 is included in the provisional defective divided image 431 (1-1). Then, the Mahalanobis distance was calculated for each of the provisional defective divided images 431 (1-1) and 431 (1-2) (S49, S50).

  As a result, the Mahalanobis distance of the provisional defective divided image 431 (1-1) is larger than the first reference value and smaller than the second reference value. The provisional defective divided image 431 (1-2) was determined to be good because its Mahalanobis distance was smaller than the first reference value. Therefore, the provisional defect divided image 431 (1-1) is determined as a provisional defect because there is a possibility that a defect exists (S47: YES), and the inspection process is further continued. That is, as shown in FIG. 10B, the provisional defective divided image 431 (1-1) of 21 × 1320 pixels is converted into the provisional defective divided images 431 (1-1-1), 431 (7 × 1320 pixels in size. It was divided into 1-1-2) and 431 (1-1-3) (S48). At this time, the error 81 is included in the provisional defective divided image 431 (1-1-2). Then, the Mahalanobis distance was calculated for each of the provisional defective divided images 431 (1-1-1), 431 (1-1-2), and 431 (1-1-3) (S49, S50). As a result, the Mahalanobis distances of all the provisional defective divided images 431 (1-1-1), 431 (1-1-2), and 431 (1-1-3) were equal to or less than the first reference value. The inspected image 43 was determined to be a non-defective product and the inspection was completed.

  In addition, since the inspected image 44 includes the small defect 72 in the inspected divided image 441 (5), the divided image No. The Mahalanobis distance of 5 was larger than the first reference value 1.9 in Table 4 and smaller than the second reference value 2.1. The other inspected divided images 441 (1) to 441 (N) were determined to be good because each Mahalanobis distance was smaller than the first reference value. Therefore, the inspection divided image 441 (5) is determined as a provisional defect because there is a possibility that a defect exists (S47: YES), and the inspection process is continued. That is, as shown in FIG. 11, the inspected divided image 441 (5) of 42 × 1320 pixels is divided into provisional defective divided images 441 (5-1) and 441 (5-2) having a size of 21 × 1320 pixels. (S48). At this time, the defect 72 is included in the provisional defective divided image 441 (5-1). Then, the Mahalanobis distance was calculated for each of the provisional defective divided images 441 (5-1) and 441 (5-2) (S49, S50). As a result, since the Mahalanobis distance of the provisional defective divided image 441 (5-1) is larger than the second reference value (S46: YES), the provisional defective divided image 441 (5-1) has a defect. Assume that the inspected image 44 is defective (S54), and the inspection is completed.

  As described above, in the present invention, since the pass / fail determination is performed for each divided image obtained by dividing the image to be inspected, it is possible to reduce the feature amount of the defect hidden in the noise and to inspect with high accuracy. . In addition, in the pass / fail judgment, there is a case where it is determined that the defect is present in the inspected divided image without being determined to be good or bad, so it is determined to be good despite being defective. Or misjudgment that it is judged to be bad despite being good can be prevented. Then, when it is determined that the provisional failure, the image is divided into small sizes and the pass / fail determination is repeated until the determined divided image is determined to be good or bad. Therefore, it is possible to perform the quality determination with higher accuracy than in the quality determination. Further, since the divided image determined to be provisional defective is re-divided and the pass / fail determination is performed again, it is not necessary to reduce the size of the divided image to be inspected from the beginning. Therefore, it is possible to efficiently inspect the quality of the inspected image.

  In addition, since two reference values are provided as the pass / fail judgment reference values, provisional failures can be determined in addition to the pass / fail of the divided image. The two reference values are provided for each of the inspected divided image and the provisional defective divided image, so that the pass / fail judgment can be made accurately for each of the divided images.

  In addition, since the reference space is created in advance, it is possible to efficiently determine whether it is good or bad. Then, the reference space is created from divided images at the same position among a plurality of non-defective images, and the pass / fail judgment reference value is determined based on the maximum value among the Mahalanobis distance of the reference space, so that pass / fail is more accurate. Judgment can be made.

  The image inspection apparatus and the image inspection method of the present invention are not limited to the above-described embodiment, and can be variously modified without departing from the scope of the claims. For example, the image size for dividing the non-defective image or the image to be inspected is not limited to the above-described embodiment, and is appropriately set in consideration of the feature amount of the product, the assumed defect type, size, and the like. be able to. Also, the first reference value and the second reference value are not limited to the above-described embodiment, and can be set as appropriate in consideration of the required level of quality. For example, when high quality is required, the first and second reference values are reduced.

  The Mahalanobis-Taguchi method has a method called item selection, and this method may be applied to the present invention to select items that contribute to inspection accuracy. Thereby, the inspection accuracy can be further improved. This item selection is to select an item that contributes to the inspection accuracy based on the SN ratio (the degree that the function does not vary), which is the concept of quality engineering. The basic procedure for selecting items is as follows.

1) Assign a plurality of pre-defined items to the orthogonal table as Level 1: Use item, Level 2: Do not use item.
2) A feature amount is extracted from the inspected image based on each condition of the orthogonal table.
3) The Mahalanobis distance of each condition of the orthogonal table is calculated from the extracted feature quantity.
4) The SN ratio of each item is calculated from the calculated Mahalanobis distance.
5) The item whose level 1 SN ratio is larger than the level 2 SN ratio is selected as an item contributing to the inspection accuracy. Further items are selected from the selected items as required (according to the inspection specification).

DESCRIPTION OF SYMBOLS 1 Image inspection apparatus 2 Camera 3 Illumination 4 Work 5 Electronic computer 6 Display apparatus 41 Non-defective image 411 Non-defective product divided image 42, 43, 44 Inspected image 421, 431, 441 Inspected divided image 431 (1-1), 431 (1 -2), 431 (1-1-1), 431 (1-1-2), 431 (1-1-3), 441 (5-1), 441 (5-2) provisional defective divided image 71, 72 Defect 81 Error S43 Division step, division means S44, S45 Creation step, creation means S46, S47 Pass / fail judgment step, pass / fail judgment means S48 Re-segmentation step, re-segmentation means S49, S50 Re-creation step, re-creation means S51, S52 Pass / fail judgment step, re-pass / fail judgment means

Claims (4)

An image inspection method for determining the quality of an image,
A non-defective image dividing step for dividing the non-defective image into a plurality of images;
A calculation step of calculating a Mahalanobis distance for each good product divided image that is an image divided in the good product image dividing step;
Based on the Mahalanobis distance calculated in the calculation step, a reference value setting step for setting a pass / fail determination reference value that is a Mahalanobis distance that is a reference for pass / fail determination of an image to be inspected for each of the non-defective product divided images;
A division step of dividing the image to be inspected into a plurality of images so as to be the same division as the division of the non-defective image ;
A creation step for creating a signal space consisting of Mahalanobis distances for each inspected divided image that is an image divided in the division step;
The Mahalanobis distance of the object to be inspected divided picture images, by comparing with the non-defective dividing the quality determination reference value set in the image corresponding to said inspection image segment quality performing quality determination in each of the inspection divided images and the determination step, only including,
In the non-defective image dividing step, the non-defective product divided images having a plurality of sizes are obtained by dividing the non-defective images into different sizes.
In the reference value setting step, for each of the plurality of types of sizes, the pass / fail judgment reference value for each non-defective product divided image is set,
In the pass / fail judgment step, as the pass / fail judgment, in addition to determining the inspected divided image as good / bad according to the comparison result, a defect exists in the inspected divided image without being determined as good or bad. It is a step of determining that there is a possibility of provisional failure,
A subdivision step for further dividing the inspected divided image into a plurality of images so as to be the same size as one of the plurality of sizes, when there is the inspected divided image determined to be the provisional failure;
Re-creating step for creating a signal space consisting of Mahalanobis distance again for each provisional defective divided image that is an image divided in the re-dividing step;
The Mahalanobis distance of the tentative defect divided picture images, by comparing the said quality determination reference value set in the non-defective divided images corresponding to the provisional failure divided image, the quality determination in each of the interim defect divided images again A re-defective determination step to be performed;
Including
If there is the temporary defective divided image newly determined as the provisional defect in the re-defective determination step, the re-dividing step is performed on the temporary defective divided image until it can be determined as good or defective. An image inspection method, wherein the re-creation step and the re-defectiveness determination step are repeatedly executed. In the reference value setting step, two reference values of a first reference value and a second reference value larger than the first reference value are set as the pass / fail judgment reference value for each of the non-defective product divided images.
The pass / fail determination step and the re-pass / fail determination step are performed when the Mahalanobis distance of the divided image to be the pass / fail determination target is smaller than the first reference value set in the corresponding non- defective product split image. The image is determined to be good, and when the Mahalanobis distance of the divided image is larger than the second reference value set for the corresponding non- defective divided image, the divided image is determined to be defective. 2. The step according to claim 1, wherein when the Mahalanobis distance is larger than the first reference value and smaller than the second reference value, the divided image is determined as the provisional defect. Image inspection method. In the non-defective image dividing step, the step of dividing the plurality of non-defective images so as to be the same as each other is performed a plurality of times while changing the division size of the non-defective images, so that So as to obtain a plurality of the non-defective product divided images having the same position in the non-defective product image,
In the reference value setting step, for each of the plurality of sizes, the pass / fail judgment reference value is set with reference to a maximum value among Mahalanobis distances of the plurality of good product divided images having the same position in the good product image. The image inspection method according to claim 1 or 2 , characterized in that An image inspection apparatus that performs image quality determination,
A non-defective image dividing means for dividing a non-defective image into a plurality of images;
Calculating means for calculating the Mahalanobis distance for each good divided image that is an image divided by the good image dividing means;
Based on the Mahalanobis distance calculated by the calculation means, a reference value setting means for setting a pass / fail judgment reference value that is a Mahalanobis distance serving as a reference for pass / fail judgment of the image to be inspected for each non-defective product divided image;
A dividing means for dividing the image to be inspected into a plurality of images so as to be the same division as the division of the non-defective image ;
Creating means for creating a signal space consisting of Mahalanobis distance for each divided image to be examined, which is an image divided by the dividing means;
The Mahalanobis distance of the object to be inspected divided picture images, by comparing with the non-defective dividing the quality determination reference value set in the image corresponding to said inspection image segment quality performing quality determination in each of the inspection divided images Determination means ,
The non-defective image dividing means divides a plurality of non-defective images into different sizes so as to obtain the non-defective product divided images of a plurality of sizes,
The reference value setting means sets the pass / fail judgment reference value for each non-defective product divided image for each of the plurality of sizes.
The pass / fail judgment means determines whether the test divided image is good / bad according to the comparison result as the pass / fail judgment, and there is a defect in the test divided image without being determined to be good or bad. It is determined as a provisional defect that there is a possibility,
When there is the inspected divided image determined to be the provisional failure, a re-dividing unit that further divides the inspected divided image into a plurality of images so as to be the same size as one of the plurality of sizes ;
Re-creating means for creating a signal space consisting of Mahalanobis distance again for each provisional defective divided image that is an image divided by the re-dividing means,
The Mahalanobis distance of the tentative defect divided picture images, by comparing the said quality determination reference value set in the non-defective divided images corresponding to the provisional failure divided image, the quality determination in each of the interim defect divided images again Re-pass / fail judgment means to perform,
In the case where there is the provisional defect divided image newly determined as the provisional defect by the re-defective / non-defective determination unit, the re-segmentation unit is applied to the provisional defect divided image until it can be determined as good or defective. An image inspection apparatus that repeatedly executes processing by the re-creating unit and the re-defective / non-defective determining unit.

Images