JP5666894B2 - Appearance inspection apparatus and appearance inspection method - Google Patents

Description

  The present invention relates to an appearance inspection apparatus and an appearance inspection method for inspecting the appearance of an object.

  As a conventional appearance inspection apparatus, for example, the one described in Patent Document 1 is known. In such an appearance inspection apparatus, the surface (inspected surface) of an object is irradiated with light from a light source to form a predetermined light / dark pattern, and an image of the surface is acquired by a CCD camera. Then, based on the acquired image, surface defects (for example, uneven defects, paint defects, scratches, dirt, and the like) existing on the surface are detected, and thereby the appearance abnormality of the object is determined.

Japanese Patent Laid-Open No. 11-63959

  However, in the appearance inspection apparatus as described above, complicated image processing and computation are required, and setting conditions such as a light source may be severe. Furthermore, there is a problem that adjustment and threshold setting become difficult due to differences in objects to be inspected.

  Therefore, an object of the present invention is to provide an appearance inspection apparatus and an appearance inspection method capable of easily inspecting an appearance abnormality of an object.

  In order to achieve the above object, an appearance inspection apparatus according to the present invention is an appearance inspection apparatus for inspecting the appearance of an object, and includes an image acquisition unit that acquires an image of the object, and an image acquired by the image acquisition unit. A binarization processing unit that obtains a binarized image in which a plurality of pixels are configured by the first value or the second value, and a binarized image acquired by the binarization processing unit An abnormality determination unit that determines whether there is an abnormality in the appearance of the object, and the abnormality determination unit exists in one direction as the number of first values for each of a plurality of pixels arranged in one direction in the binarized image. Calculating the amount and the unidirectional change amount as the number of two pixels adjacent in one direction changing between the first value and the second value, respectively, and orthogonal to the one direction in the binarized image For each of a plurality of pixels arranged in the direction, the other direction existence amount as the number of first values The other direction change amount is calculated as the number of two pixels adjacent to each other in the other direction changing between the first value and the second value. The sum of the change amount, the sum of the plurality of other direction existence amounts, and the sum of the plurality of other direction change amounts are set in advance as one-way existence amount reference value, one-way change amount reference value, and other direction existence amount reference, respectively. When there is a discrepancy with respect to each of the value and the other direction change amount reference value, it is determined that there is an appearance abnormality. For example, the abundance includes “the number of pixels of either white or black”, and the variation includes, for example, “the number of times the distribution of pixels changes from white to black or from black to white”. .

  In the appearance inspection apparatus according to the present invention, a binarization process is performed on an object image, and a binarized image obtained by binarizing a plurality of pixels and obtaining a simplified image is obtained. Then, the spatial distribution for the first value and the second value of the binarized image is simply calculated as a one-way existence amount, a one-way change amount, another direction existence amount, and another direction change amount, Determine the appearance abnormality of the object. That is, complicated image processing and computation are not required, and abnormality determination relating to appearance inspection can be simplified and performed with high accuracy. In addition, it is not necessary to form a light and dark pattern on the surface of the object, and therefore, setting and adjustment of the light source and the like are not necessary. Therefore, according to the present invention, it is possible to easily inspect the appearance abnormality of the object.

  At this time, the abnormality determination unit includes a sum of a plurality of unidirectional existence amounts, a sum of a plurality of unidirectional change amounts, a sum of a plurality of other direction existence amounts, and a plurality of other directions for a binarized image of a normal object. In some cases, the total sum of the change amounts is set in advance as a one-way existence reference value, a one-way change reference value, another direction existence reference value, and another direction change reference value.

  In addition, a processing unit is formed on the object, the abnormality determination unit includes a position shift determination unit that determines the presence or absence of a positional shift abnormality of the processing unit, and the position shift determination unit is one direction in the binarized image. For each of the plurality of pixels arranged in the same direction, a one-way ratio is calculated as a value obtained by dividing the one-way existence amount or the one-way change amount by the coordinate value in the other direction for the plurality of pixels, and the other direction in the binarized image. For each of the plurality of pixels arranged in the same direction, the other direction ratio as a value obtained by dividing the other direction existence amount or the other direction change amount by the coordinate value of the one direction related to the plurality of pixels is calculated. When each of the sum and the sum of the other direction ratios does not match the preset one-way ratio reference value and each of the other direction ratio reference values, it is determined that there is a positional deviation abnormality of the processing part. preferable. In this case, it is possible to easily inspect the positional deviation abnormality of the processed part of the object.

  At this time, the abnormality determination unit calculates the one-way ratio reference value and the other-direction ratio reference value for each of the plurality of one-way ratio sums and the plurality of other-direction ratio sums for the binarized image of the normal object. May be set in advance.

  In addition, the binarization processing unit acquires a lightness distribution related to the gradation for the image acquired by the image acquisition unit, and the lightness distribution has the first gradation at the first peak value with the highest lightness and the lightness. The second gradation at the second peak value that is the second highest next to the first peak value is derived, and the value between the first and second gradations is a threshold value for performing binarization processing. As the binarization level, it is preferable to binarize the image. Here, in the image, the lightness in the original part of the surface of the object and the lightness in the processed part, surface defect, etc. (hereinafter simply referred to as “processed part etc.”) may have different peak values. Found. Therefore, as in the present invention, when the first and second gradations when the lightness is the first and second peak values are derived, and the binarization process is performed with the value between these as the binarization level, It is possible to obtain a binarized image that can appropriately identify the processing unit and the like.

  An appearance inspection method according to the present invention is an appearance inspection method for inspecting the appearance of an object, and includes an image acquisition step of acquiring an image of the object, and binarization processing on the image acquired in the image acquisition step. A binarization processing step for obtaining a binarized image in which a plurality of pixels are configured with the first value or the second value, and an abnormality of the object based on the binarized image acquired in the binarization processing step An abnormality determination step for determining the presence or absence of the image, and in the abnormality determination step, for each of a plurality of pixels arranged in one direction in the binarized image, the one-way existence amount as the number of first values and adjacent in one direction For each of a plurality of pixels arranged in the other direction orthogonal to one direction in the binarized image, while calculating one direction change amount as the number of the two pixels that change between the first value and the second value In the other direction as the number of the first value and in the other direction The other direction change amount is calculated as the number of the two pixels in contact with each other changing between the first value and the second value, and the calculated sum of the one-way existence amounts and the sum of the plurality of one-way change amounts are calculated. , A sum of a plurality of other direction existence amounts, and a sum of a plurality of other direction change amounts are set in advance as a one-way existence amount reference value, a one-way change amount reference value, another direction existence amount reference value, and the like. When there is no coincidence with each of the direction change amount reference values, it is determined that there is an abnormality.

  In the appearance inspection method of the present invention, a binarization process is performed on an object image, and a binarized image obtained by binarizing a plurality of pixels and obtaining a simplified image is obtained. Then, the spatial distribution for the first value and the second value of the binarized image is simply calculated as a one-way existence amount, a one-way change amount, another direction existence amount, and another direction change amount, Determine the appearance abnormality of the object. That is, complicated image processing and computation are not required, and abnormality determination relating to appearance inspection can be simplified and performed with high accuracy. In addition, it is not necessary to form a light and dark pattern on the surface of the object, and therefore, setting and adjustment of the light source and the like are not necessary. Therefore, also in the present invention, the above-described effect that the appearance abnormality of the object can be easily inspected is exhibited.

  At this time, the unidirectional presence amount reference value, the unidirectional change amount reference value, the other direction presence amount reference value, and the first direction change amount reference value are set in advance. A total of a plurality of unidirectional existence amounts, a plurality of unidirectional change amounts, a plurality of other direction existence amounts, and a plurality of other direction change amounts for each of a normal object binarized image, There are cases where the unidirectional existence amount reference value, the unidirectional change amount reference value, the other direction existence amount reference value, and the other direction change amount reference value are set in advance.

  In addition, the processing part is formed on the object, and the abnormality determination step includes a position shift determination step for determining presence / absence of a position shift abnormality of the processing unit. In the position shift determination step, one direction in the binarized image is obtained. For each of the plurality of pixels arranged in the same direction, a one-way ratio is calculated as a value obtained by dividing the one-way existence amount or the one-way change amount by the coordinate value in the other direction for the plurality of pixels, and the other direction in the binarized image. For each of the plurality of pixels arranged in the same direction, the other direction ratio as a value obtained by dividing the other direction existence amount or the other direction change amount by the coordinate value of the one direction related to the plurality of pixels is calculated. When each of the sum and the sum of the other direction ratios does not match the preset one-way ratio reference value and each of the other direction ratio reference values, it is determined that there is a positional deviation abnormality of the processing part. Good Arbitrariness. In this case, it is possible to easily inspect the positional deviation abnormality of the processed part of the object.

  At this time, it comprises a second pre-processing step for presetting the one-way ratio reference value and the other-direction ratio reference value. In the second pre-processing step, a plurality of unidirectional ratios for a binarized image of a normal object are provided. Each of the sum and the sum of the other direction ratios may be set in advance as the one-way ratio reference value and the other direction ratio reference value.

  In the binarization processing step, a brightness distribution related to the gradation of the image acquired by the image acquisition unit is acquired, and the first gradation at the first peak value having the highest brightness in the brightness distribution and the brightness are The second gradation at the second peak value that is the second highest next to the first peak value is derived, and the value between the first and second gradations is a threshold value for performing binarization processing. As the binarization level, it is preferable to binarize the image. In this case, as described above, the lightness at the original portion of the surface of the object and the lightness at the processed portion in the image have different peak values from each other, so that the lightness is the first as in the present invention. When the first and second gradations at the time of the second peak value are derived, and the binarization process is performed with the value between them as the binarization level, the binarization that can identify the processing part etc. suitably Images can be acquired.

  According to the present invention, it is possible to easily inspect the appearance abnormality of an object.

It is a schematic block diagram which shows the structure of the external appearance inspection apparatus which concerns on one Embodiment. It is a flowchart which shows the external appearance inspection method by the external appearance inspection apparatus which concerns on one Embodiment. It is a figure which shows an example of a reference | standard image. It is a figure which shows an example of the histogram of a brightness. It is a figure which shows an example of a reference | standard binarization image. It is a figure which shows an example of the image for a test | inspection. It is a figure which shows an example of the binarized image for a test | inspection. It is a figure which shows another example of the binarized image for a test | inspection. It is a figure which shows another example of the binarized image for a test | inspection. It is a figure which shows another example of the binarized image for a test | inspection. It is a figure which shows another example of the binarized image for a test | inspection.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings. In the following description, the same or equivalent elements will be denoted by the same reference numerals, and redundant description will be omitted.

  FIG. 1 is a schematic block diagram showing the configuration of an appearance inspection apparatus according to an embodiment of the present invention. As shown in FIG. 1, the appearance inspection apparatus 1 inspects the appearance of an object 10 and determines an appearance abnormality. The appearance inspection apparatus 1 of this embodiment inspects the surface 10a of the object 10 on which the processed portion Z (see FIG. 3), which is a resistance spot welding mark, for example, is formed. The object 10 is not particularly limited, and examples thereof include a window glass and an outer wall of a vehicle (railway vehicle, track, linear motor, automobile, etc.).

  The appearance inspection apparatus 1 includes an image acquisition unit 2, an image analysis unit 3, and a notification unit 4. The image acquisition unit 2 acquires a digital image (hereinafter simply referred to as “image”) of the surface 10a of the object 10. It is an imaging part for doing. The image acquisition unit 2 is connected to the image analysis unit 3 and outputs the acquired image to the image analysis unit 3. As the image acquisition unit 2, a CCD camera, a digital camera, a scanner, or the like is used.

  The image analysis unit 3 performs binarization processing on the image acquired by the image acquisition unit 2 and acquires a binarized image (binarization processing unit) 3a, and based on the binarized image. An abnormality determination unit 3b that determines the presence / absence of an appearance abnormality of the object 10 and a position shift determination unit 3c that determines the presence / absence of a position shift abnormality of the processing unit Z are configured. The image analysis unit 3 is connected to the notification unit 4 and outputs an abnormality notification command to the notification unit 4 when it is determined that there is an abnormality. As the image analysis unit 3, an ECU of a personal computer is used, and includes, for example, a CPU, a ROM, a RAM, and the like.

  The notification unit 4 notifies the inspector of abnormality based on the abnormality notification command output from the image analysis unit 3. As the notification unit 4, for example, a speaker, a buzzer, a warning light, a personal computer display (monitor), or the like is used.

  Next, an appearance inspection method for inspecting the appearance of the object 10 by the above-described appearance inspection apparatus 1 will be described with reference to the flowchart of FIG. Here, as an example, a case where the appearance of the object 10 on which the processed portion Z is formed is inspected.

  First, as pre-processing steps, reference values for determining an appearance abnormality (X direction existence amount reference value αx, X direction change amount reference value βx, Y direction existence amount reference value αy, and Y direction change amount reference value βy). Is calculated. At the same time, reference values (X-direction ratio reference value γx and Y-direction ratio reference value γy) for determining a misalignment abnormality of the machining part Z are calculated.

  That is, the image acquisition unit 2 captures an image of the object 10 on which the processing unit Z is normally formed. Thereby, the reference image which is a digital image of the surface 10a including the normal processing part Z in the object 10 is acquired (S1). As shown in FIG. 3, in the reference image 11 here, as a normal processed portion Z, a resistance ring weld mark having a circular ring shape is formed on the surface 10 a.

  Subsequently, a histogram of brightness of the reference image 11 is derived by the image analysis unit 3 (S2). Specifically, as shown in FIG. 4, a grayscale image is created for the image 11, the brightness (brightness) of each pixel is obtained, and the brightness distribution for 256 gradations is obtained as the reference histogram 12. At this time, the lightness of each pixel is obtained by an expression such as “lightness = 0.299 × R component + 0.587 × G component + 0.144 × B component”. In the reference histogram 12 here, gradation is shown on the horizontal axis, and the gradation is expressed by 256 gradations from 0 to 255. The lightness is shown on the vertical axis, and the lightness is expressed in pixels (px).

  Subsequently, in the reference histogram 12, the first gradation 12H at the first peak value P1 having the highest lightness and the second gradation at the second peak value P2 having the lightness next to the first peak value P1. 12L is derived. Then, the median value 12M of the first and second gradations 12H and 12L is set as a binarization level serving as a threshold for performing binarization processing (S3). Subsequently, binarization processing is performed on the reference image 11 based on the set binarization level, and binarization in which a plurality of pixels are configured with only one value (first value) or zero value (second value). A reference binarized image 13 which is an image is acquired (S4).

  FIG. 5 is a diagram illustrating an example of a reference binarized image. In the figure, the reference binarized image 13 is a non-square matrix whose pixels G are 12 rows and 11 columns. In the reference binarized image 13, the upper left corner position in the figure is set as the coordinate reference of the pixel G (that is, the origin (0, 0)), the left and right direction in the figure is set to the X direction, and the vertical direction in the figure is set. It is set in the Y direction (other direction) orthogonal to the X direction. In the drawing, for convenience of explanation, it is assumed that the image of the processing portion Z is superimposed on the reference binarized image 13. These are the same in FIGS. As shown in FIG. 5, in the reference binarized image 13, it can be seen that the pixel G corresponding to the processed portion Z has a 1 value and the other pixels G have a 0 value.

  Subsequently, the image analysis unit 3 uses the X direction existence amount Ax as the number of one value for each of a plurality of pixels G arranged in the X direction in the reference binarized image 13 (hereinafter referred to as “X pixel group”), An X-direction change amount Bx is calculated as the number of changes in the two pixels G adjacent in the X direction between 1 value and 0 value (S5). In other words, among the plurality of pixels G arranged in the X column, the number of one value is the X-direction existence amount Ax, and one value is obtained when the plurality of pixels G arranged in the X column are viewed in order from one side to the other side in the X direction. A number that becomes 0 value (or 0 value → 1 value) is defined as an X-direction change amount Bx.

  For example, in the X pixel group Gx1 shown in the example in FIG. 5, there are five values of one value, and thus the X-direction existence amount Ax is “5”. Further, in the X pixel group Gx1, there are one place that changes from 0 to 1 and one place that changes from 1 to 0. Therefore, the X-direction change amount Bx is “2”.

  In S5, for each of a plurality of pixels G arranged in the Y direction in the reference binarized image 13 (hereinafter referred to as “Y pixel group”), the Y direction existence amount Ay as the number of values is adjacent to the Y direction. The Y direction change amount By as the number of the two pixels G to be changed between the 1 value and the 0 value is calculated. In other words, the number of one value among the plurality of pixels G arranged in the Y column is defined as the Y-direction existence amount Ay, and one value is obtained when the plurality of pixels G arranged in the Y column are viewed in order from one side to the other side in the Y direction. A number that becomes 0 value (or 0 value → 1 value) is defined as a Y-direction change amount By.

  For example, in the Y pixel group Gy1 shown in the example in FIG. 5, there are six 1-values, and therefore the X-direction existence amount Ay is “6”. Further, in the X pixel group Gy1, there are one place that changes from 0 to 1 and one place that changes from 1 to 0, and thus the Y-direction change amount By becomes “2”.

  In S5, the X direction ratio Cx is calculated as a value obtained by dividing the X direction existence amount Ax by the coordinate value in the Y direction in the plurality of X pixel groups. Further, in the plurality of Y pixel groups, the Y direction ratio Cy is calculated as a value obtained by dividing the Y direction existence amount Ay by the X direction coordinate value. In other words, the ratio between the X direction existence amount Ax and the Y direction position information (Y coordinate) for the X direction existence amount Ax is defined as an X direction ratio Cx. Further, the ratio between the Y direction existence amount Ay and the X direction position information (X coordinate) for the Y direction existence amount Ay is defined as a Y direction ratio Cy.

  For example, in the X pixel group Gx1 shown in the example of FIG. 5, “2.5” obtained by dividing the X direction existence amount Ax “5” by the Y coordinate value “2” is the X direction ratio Cx. On the other hand, in the Y pixel group Gy1, “3” obtained by dividing the Y-direction existence amount Ay “6” by the X-coordinate value “2” is the Y-direction ratio Cy.

  Subsequently, the sum ΣAx of the plurality of X-direction existence amounts Ax is set as the X-direction existence amount reference value αx, and the sum ΣAy of the plurality of Y-direction existence amounts Ay is set as the Y-direction existence amount reference value αy. Further, the sum ΣBx of the plurality of X-direction change amounts Bx is set as the X-direction change amount reference value βx, and the sum ΣBy of the plurality of Y-direction change amounts By is set as the Y-direction change amount reference value βy. Then, the sum ΣCx of the plurality of X direction ratios Cx is set as the X direction ratio reference value γx, and the sum ΣCy of the plurality of Y direction ratios Cy is set as the Y direction ratio reference value γy (S6).

  Next, in this step, the appearance of the inspection object 10 is inspected using each reference value set in S6. That is, first, the inspection object 10 is imaged by the image acquisition unit 2, and an inspection image which is a digital image of the surface 10a including the processing unit Z is acquired (S11). In the inspection image 21 illustrated in FIG. 6, for example, due to deterioration of welding conditions, the processed portion Z has an elliptical shape that is crushed in the horizontal direction in the drawing.

  Subsequently, in the same manner as in S2 to S4, the image analysis unit 3 derives a brightness histogram of the inspection image 21, sets a binarization level, and based on the binarization level, the inspection image 21. A binarization process is performed to obtain a binarized image for inspection 23 (S12 to S14).

  FIG. 7 is a diagram illustrating an example of a binarized image for inspection. As shown in FIG. 7, the binarized image for inspection 23 is an example corresponding to the inspection image 21 of FIG. 6, and here, the pixel G corresponding to the elliptical processed portion Z becomes one value. The other pixels G have 0 values.

  Subsequently, the image analysis unit 3 calculates the X-direction existence amount Ax and the X-direction change amount Bx for each X pixel group in the inspection binary image 23 (S15). In the X pixel group Gx2 shown in the example in FIG. 7, there are five values of one value, and therefore the X-direction existence amount Ax is “5”. Further, in the X pixel group Gx1, there are one place that changes from 0 to 1 and one place that changes from 1 to 0. Therefore, the X-direction change amount Bx is “2”.

  In S15, the Y-direction existence amount Ay and the Y-direction change amount By are calculated for each Y pixel group in the inspection binarized image 23. In the Y pixel group Gy2 shown in the example in FIG. 7, there is no number of one value, and therefore the X-direction existence amount Ay is “0”. Further, in the X pixel group Gy2, there is no portion that changes from 0 to 1 and there is no portion that changes from 1 to 0. Therefore, the Y-direction change amount By is “0”.

  In S15, the X direction ratio Cx is calculated for each of the plurality of X pixel groups, and the Y direction ratio Cy is calculated for each of the plurality of Y pixel groups. In the X pixel group Gx2 shown in the example in FIG. 7, “2.5” obtained by dividing the X direction existence amount Ax “5” by the Y coordinate value “2” is the X direction ratio Cx. On the other hand, in the Y pixel group Gy2, “0” obtained by dividing the X-direction existence amount Ax “0” by the X-coordinate value “2” is the X-direction ratio Cx.

  Subsequently, the appearance abnormality is determined based on the existence amounts Ax, Ay and the change amounts Bx, By. Specifically, the plurality of X-direction existence amounts Ax are summed to calculate the sum ΣAx of the X-direction existence amounts Ax, and the plurality of X-direction change amounts Bx are summed to calculate the sum ΣBx of the X-direction change amounts Bx. To do. Further, the plurality of Y-direction existence amounts Ay are summed to calculate the sum ΣAy of the Y-direction existence amounts Ay, and the plurality of Y-direction change amounts By are summed to calculate the sum ΣBy of the Y-direction change amounts By. Then, it is determined whether or not these sums ΣAx, ΣBx, ΣAy, ΣBy match the reference values αx, βx, αy, βy set in advance in 6 above (S16).

  In S16, when each of the sums ΣAx, ΣBx, ΣAy, ΣBy does not match the reference values αx, βx, αy, βy, “notification of appearance” is notified by the notification unit 4 (S17). .

  In the example of FIG. 7, the total sum ΣAy of the Y direction existence amount is “32”, which does not coincide with “36” (see FIG. 5) of the Y direction existence amount reference value αy. Further, the sum ΣBy of the Y direction change amounts is “28”, which does not coincide with “32” (see FIG. 5) of the Y direction existence amount reference value αy. Therefore, in this case, the appearance abnormality of the processed part Z is determined and notified. Thus, in the present embodiment, when there is an abnormality in the appearance of the processed part Z, the abnormality in the appearance is detected and notified with high accuracy.

  On the other hand, if the sums ΣAx, ΣBx, ΣAy, ΣBy coincide with the reference values αx, βx, αy, βy in S16, the positional deviation abnormality of the machining part Z is determined. Specifically, the plurality of X direction ratios Cx are summed to calculate the sum ΣCx of the X direction ratios Cx, and the plurality of Y direction ratios Cy are summed to calculate the sum ΣCy of the Y direction ratios Cy. Then, it is determined whether or not these sums ΣCx, ΣCy respectively match the reference values γx, γy set in advance in 6 (S18).

  If the sums ΣCx and ΣCy do not match the reference values γx and γy in S19, the notification unit 4 notifies that there is “position shift” (S19). On the other hand, when each of the sums ΣCx and ΣCy matches the reference values γx and γy, it is determined that there is no abnormality in the appearance of the object 10, and a notification that the notification unit 4 is normal is performed (S20).

  FIG. 8 is a diagram illustrating another example of the binarized image for inspection. In the object 10 here, a positional deviation occurs in the processing part Z, and as shown in FIG. 8, a positional deviation of the processing part Z toward the left side of the processing part Z occurs in the acquired binarized image 23 for inspection. In the example shown in FIG. 8, the sum ΣCy of the Y direction ratio is “12.6”, which does not coincide with “8.1” (see FIG. 5) of the Y direction ratio reference value γy. Therefore, in this case, the positional deviation abnormality of the processed part Z is determined and notified in S18 and S19. As described above, in the present embodiment, when there is a positional deviation abnormality of the processing part Z, the positional deviation is accurately detected and notified.

  As described above, in this embodiment, the inspection image 21 of the object 10 is acquired, and the binarization process is performed on the inspection image 21, thereby binarizing the plurality of pixels and simplifying the inspection. An image 23 is acquired. Then, the spatial distribution of the 1-value and 0-value of the inspection binarized image 23 is simplified and calculated as the X-direction existence amount Ax, the X-direction change amount Bx, the Y-direction existence amount Ay, and the Y-direction change amount By. By doing so, the appearance abnormality of the object 10 is determined. That is, complicated image processing and computation are not required, and abnormality determination relating to appearance inspection can be simplified and performed with high accuracy. Furthermore, it is not necessary to illuminate the object 10 with a light source or the like, it is not necessary to form a light / dark pattern on the surface 10a of the object 10, and it becomes unnecessary to set or adjust the light source or the like. Therefore, according to the present embodiment, it is possible to easily inspect the appearance abnormality of the object 10.

  In the present embodiment, as described above, the spatial distribution shift for the 1-value and 0-value of the inspection binarized image 23 is simply calculated as the X-direction ratio Cx and the Y-direction ratio Cy. Thus, the positional deviation abnormality of the processed part Z is determined. Therefore, in this case, it is possible to easily inspect the positional deviation abnormality of the processing part Z of the object 10.

  By the way, in the images 11 and 21, it is found that the lightness in the original part of the surface of the object 10 and the lightness in the other processed parts have different peak values. In this regard, in the present embodiment, as described above, when acquiring the binarized images 13 and 23, the brightness distribution relating to the gradation of the image 21 is acquired as the histograms 12 and 22, and the first and second peak values thereof are acquired. The median value 12M of the first and second gradations 12L and 12H of P1 and P2 is set to a binarization level, and binarization processing is performed on the images 11 and 21. Therefore, it is possible to obtain the binarized images 13 and 23 that can appropriately identify the processed part or the like. In other words, it is possible to acquire the binarized images 13 and 23 in which the pixels corresponding to the processing part or the like have a single value with high accuracy among a plurality of pixels. In particular, by setting the median value 12M to the binarization level in this way, in the binarized images 13 and 23, the surface 10a portion and the portion such as the processed portion can be easily distinguished.

  In the above description, the example in which the deformation of the processed portion Z is inspected as an appearance abnormality has been described. However, the present embodiment is not limited to this. For example, as shown below, it is also possible to inspect the appearance abnormality such as “the chipped portion Z”, “dirt or scratch”, “unprocessed”, and the like.

  FIG. 9 is a diagram illustrating still another example of the binarized image for inspection. In the object 10 here, the processing part Z is chipped, and as shown in FIG. 9, the acquired inspection binary image 23 has a processing part Z whose top right part is not shown.

  In the example shown in FIG. 9, the total sum ΣAx of the X direction existence amount is “26” and the sum total ΣAy of the Y direction existence amount is “26”, and “34” of the X direction existence amount reference value αx and the Y direction existence amount. The reference values αy do not match “36” (see FIG. 5). Further, the sum ΣBx of the X-direction change amounts is “28”, the sum ΣBy of the Y-direction change amounts is “26”, “36” of the X-direction existence amount reference value αx, and “32” of the Y-direction existence amount reference value αy. ”(See FIG. 5). Therefore, even in this case, the appearance abnormality of the processed part Z is determined and notified in S16 and S17.

  FIG. 10 is a diagram illustrating another example of the binarized image for inspection. In the object 10 here, dirt (or scratches) occurs in the processed portion Z, and as shown in FIG. 10, the acquired binarized image 23 for inspection has a dirt Za at a substantially ring-shaped central portion. A processed part Z having (or scratches Za) appears.

  In the example shown in FIG. 10, the sum ΣAx of the X direction existence amount is “35” and the sum ΣAy of the Y direction existence amount is “37”, and “34” of the X direction existence amount reference value αx and the Y direction existence amount. The reference values αy do not match “36” (see FIG. 5). Further, the total sum ΣBx of the X direction change amounts is “38” and the total sum ΣBy of the Y direction change amounts is “34”, “36” of the X direction existence amount reference value αx and “32” of the Y direction existence amount reference value αy. ”(See FIG. 5). Therefore, even in this case, the appearance abnormality of the processed part Z is determined and notified in S16 and S17.

  FIG. 11 is a diagram illustrating still another example of the binarized image for inspection. In the object 10 here, the processed part Z is not formed (unprocessed), and as shown in FIG. 10, the processed part Z does not appear in the acquired binarized image 23 for inspection.

  In the example shown in FIG. 11, the total sum ΣAx of the X-direction existence amount is “0” and the total sum ΣAy of the Y-direction existence amount is “0”, and “34” of the X-direction existence amount reference value αx and the Y-direction existence amount. The reference values αy do not match “36” (see FIG. 5). Further, the sum ΣBx of the X-direction change amounts is “0” and the sum ΣBy of the Y-direction change amounts is “0”, “36” of the X-direction existence amount reference value αx and “32” of the Y-direction existence amount reference value αy. ”(See FIG. 5). Therefore, even in this case, the appearance abnormality of the processed part Z is determined and notified in S16 and S17.

  The preferred embodiment of the present invention has been described above, but the present invention is not limited to the above embodiment. For example, in the above embodiment, the median value 12M of the first and second gradations 12L and 12H is set as the binarization level, but instead of this, the average value of the first and second peak values P1 and P2 The gradation at lightness may be set as a binarization level. In short, the binarization level may be a value between the first and second gradations 12L and 12H.

  In addition, two peak values (that is, the first and second peak values P1 and P2) appear in the lightness histograms 12 and 22. When three or more peak values appear, the peak values of these peak values are displayed. The top two are the first and second peak values. The object 10 may be a transparent body such as glass that is transparent to visible light, or may be an opaque body such as a metal that is opaque (non-transparent) to visible light. In either case, the present invention can inspect the appearance of the object 10.

  In the above embodiment, the X direction ratio Cx obtained by dividing the X direction existence amount Ax by the Y direction coordinate value is calculated. However, the X direction ratio Cx is the X direction change value Bx by the Y direction coordinate value. It is good also as the value which remove | divided. Similarly, in the above embodiment, the Y direction ratio Cy obtained by dividing the Y direction existence amount Ay by the X direction coordinate value is calculated, but the Y direction ratio Cy is the X direction coordinate value and the Y direction change amount. It may be a value obtained by dividing By. That is, the X direction ratio Cx and the Y direction ratio Cy may be calculated based on the existence amount or may be calculated based on the change amount.

  In the present invention, it is also possible to detect only surface defects such as scratches and dirt on the surface 10a using the object 10 on which the processed portion Z is not formed on the surface 10a. In this case, the inspection binarized image 23 shown in FIG. 11 is set as a reference binarized image (normal data), and a reference value is set from the reference binarized image. Moreover, the process part Z is not limited to the thing of the said embodiment, The thing of various shapes and aspects may be sufficient.

  Here, in the above embodiment, in addition to determining whether or not the processing portion Z is misaligned (S18), the processing portion is based on the calculated sizes (bias) of the X direction ratio Cx and the Y direction ratio Cy. It is possible to determine the direction of the Z misalignment.

Specifically, as shown below, when the lower limit threshold Lx and the upper limit threshold Hx are set in the X direction ratio Cx, and the X direction ratio Cx is smaller than the lower limit threshold Lx, the processing part Z is positioned in the right direction. If it is determined that the X direction ratio Cx is larger than the upper limit threshold value Hx, it is determined that the processed portion Z is displaced in the left direction. In addition, when the lower limit threshold Ly and the upper limit threshold Hy are set in the Y direction ratio Cy, and the Y direction ratio Cy is smaller than the lower limit threshold Ly, it is determined that the processed portion Z has been displaced downward, and the Y direction ratio When Cy is larger than the upper limit threshold value Hy, it is determined that the processing unit Z has been displaced in the upward direction. As a result, when these are combined, it is possible to determine the position shift of the processing portion Z by eight directions of “up”, “down”, “left”, “right”, “upper left”, “upper right”, “lower left”, and “lower right”.
X direction ratio Cx <lower threshold Lx → right position shift
Upper limit threshold Hx <X direction ratio Cx → left position shift
Y direction ratio Cy <lower limit threshold Ly → downward position shift
Upper limit threshold value Hy <Y direction ratio Cy → upward misalignment

These lower limit threshold values Lx and Ly and upper limit threshold values Hx and Hy can be appropriately set based on the reference binarized image 13 and the like. For example, in the case of the reference binarized image 13 shown in FIG. 5, when the X-direction ratio Cx and the Y-direction ratio Cy are calculated based on the abundance, the X-direction ratio Cx, the Y-direction ratio Cy, the positional deviation direction, Is the relationship shown in Table 1 below. Thereby, the positional deviation direction of the processing part Z can be determined by setting the upper limit threshold Hx to 10.0, the lower limit threshold Lx to 7.0, the upper limit threshold Hy to 10.0, and the lower limit threshold Ly to 6.0. I understand.

On the other hand, in the case of the reference binarized image 13 shown in FIG. 5, when the X direction ratio Cx and the Y direction ratio Cy are calculated based on the amount of change, the X direction ratio Cx, the Y direction ratio Cy, the positional deviation direction, Is the relationship shown in Table 2 below. Accordingly, the misalignment direction of the processing portion Z can be determined by setting the upper limit threshold Hx to 8.0, the lower limit threshold Lx to 5.5, the upper limit threshold Hy to 8.0, and the lower limit threshold Ly to 5.5. I understand.

  DESCRIPTION OF SYMBOLS 1 ... Appearance inspection apparatus, 2 ... Image acquisition part, 3a ... Binarization part (binarization process part), 3b ... Abnormality determination part, 3c ... Position shift determination part, 10 ... Object, 11 ... (Image), 12 ... reference histogram (histogram), 12H ... first gradation, 12L ... second gradation, 13 ... reference binary image (binarized image), 21 ... (image), 22 ... inspection histogram (histogram), 23... Binary image for inspection (binarized image), Ax... X direction existence amount (one direction existence amount), Ay... Y direction existence amount (other direction existence amount), Bx. Change amount), By ... Y direction change amount (other direction change amount), Cx ... X direction ratio (one direction ratio), Cy ... Y direction ratio (other direction ratio), P1 ... first peak value, P2 ... second. Peak value, Z ... machined part, αx ... X direction existence amount reference value (one direction existence amount reference value), αy ... Y direction existence amount reference value ( Other direction existence reference value), βx... X direction change reference value (one direction change reference value), βy... Y direction change reference value (other direction change reference value), γx. Unidirectional ratio reference value), γy... Y direction ratio reference value (other direction ratio reference value), .SIGMA.Ax... X direction existence amount sum (one direction existence amount sum), .SIGMA.Ay. Sum of existing amounts), ΣBx... Sum of X direction change amounts (sum of one direction change amounts), ΣBy... Sum of Y direction change amounts (sum of other direction change amounts), ΣCx... (Sum of one direction ratios), ΣCy (sum of other direction ratios).

Claims (10)

An appearance inspection apparatus for inspecting the appearance of an object,
An image acquisition unit for acquiring an image of the object;
A binarization processing unit that performs binarization processing on the image acquired by the image acquisition unit, and acquires a binarized image in which a plurality of pixels are configured with a first value or a second value;
An abnormality determination unit that determines the presence or absence of an appearance abnormality of the object based on the binarized image acquired by the binarization processing unit;
The abnormality determination unit
For each of the plurality of pixels arranged in one direction in the binarized image, the unidirectional existence amount as the number of the first values, and the two pixels adjacent in the one direction are the first value and the second And calculating the unidirectional change amount as the number that changes between the values,
For each of the plurality of pixels arranged in the other direction orthogonal to the one direction in the binarized image, the other direction existence amount as the number of the first values and the two pixels adjacent to the other direction are the first number. Calculating the amount of change in the other direction as a number that changes between 1 value and the second value,
Each of the calculated sum of the plurality of one-direction existence amounts, the sum of the plurality of one-direction change amounts, the plurality of other direction existence amounts, and the plurality of other direction change amounts is preset. Appearance characterized by determining that there is an appearance abnormality when there is a discrepancy with respect to each of the one-way existence amount reference value, the one-direction change amount reference value, the other direction existence amount reference value, and the other direction change amount reference value Inspection device.   The abnormality determination unit includes a sum of a plurality of the one-direction existence amounts, a plurality of the one-direction change amounts, a plurality of the other-direction existence amounts, and a plurality of the binarized images of the normal object. Are set in advance as the one-way existence reference value, the one-way change reference value, the other-direction existence reference value, and the other-direction change reference value, respectively. The visual inspection apparatus according to claim 1. A processed part is formed on the object,
The abnormality determination unit includes a position shift determination unit that determines presence / absence of a position shift abnormality of the processing unit,
The positional deviation determination unit
For each of the plurality of pixels arranged in the one direction in the binarized image, a one-way ratio as a value obtained by dividing the one-way existence amount or the one-way change amount by the coordinate value in the other direction related to the plurality of pixels. Respectively,
For each of the plurality of pixels arranged in the other direction in the binarized image, the other direction ratio as a value obtained by dividing the other direction existence amount or the other direction change amount by the coordinate value of the one direction related to the plurality of pixels. Respectively,
When the calculated sum of the one-way ratios and the sum of the other direction ratios do not match the preset one-way ratio reference value and the other direction ratio reference value, respectively, The appearance inspection apparatus according to claim 1, wherein it is determined that there is a positional deviation abnormality.   The positional deviation determination unit includes a plurality of one-way ratio sums and a plurality of other direction ratio sums for the binarized image of the normal object, the one-way ratio reference value, and the The appearance inspection apparatus according to claim 3, wherein each of the other direction ratio reference values is preset. The binarization processing unit
Obtaining a lightness distribution relating to the gradation of the image obtained by the image obtaining unit;
In the lightness distribution, a first gradation when the lightness is the highest first peak value and a second gradation when the lightness is the second highest peak value after the first peak value are derived. And
The binarization process is performed on the image by using a value between the first and second gradations as a binarization level that serves as a threshold for performing the binarization process. The external appearance inspection apparatus as described in any one of 1-4. An appearance inspection method for inspecting the appearance of an object,
An image acquisition step of acquiring an image of the object;
A binarization processing step of performing binarization processing on the image acquired in the image acquisition step, and acquiring a binarized image in which a plurality of pixels are configured with a first value or a second value;
An abnormality determination step of determining the presence or absence of an appearance abnormality of the object based on the binarized image acquired in the binarization processing step,
In the abnormality determination step,
For each of the plurality of pixels arranged in one direction in the binarized image, the unidirectional existence amount as the number of the first values, and the two pixels adjacent in the one direction are the first value and the second And calculating the unidirectional change amount as the number that changes between the values,
For each of the plurality of pixels arranged in the other direction orthogonal to the one direction in the binarized image, the other direction existence amount as the number of the first values and the two pixels adjacent to the other direction are the first number. Calculating the amount of change in the other direction as a number that changes between 1 value and the second value,
Each of the calculated sum of the plurality of one-direction existence amounts, the sum of the plurality of one-direction change amounts, the plurality of other direction existence amounts, and the plurality of other direction change amounts is preset. Appearance characterized by determining that there is an appearance abnormality when there is a discrepancy with respect to each of the one-way existence amount reference value, the one-direction change amount reference value, the other direction existence amount reference value, and the other direction change amount reference value Inspection method. A first preprocessing step of presetting the one-way existence reference value, the one-direction change reference value, the other-direction existence reference value, and the other-direction change reference value;
In the first pretreatment step,
The sum of a plurality of the one-direction existence amounts, the sum of the plurality of one-direction change amounts, the sum of the plurality of other direction existence amounts, and the plurality of other direction change amounts for the binarized image of the normal object. Are set in advance as the one-way existence reference value, the one-direction change reference value, the other-direction existence reference value, and the other-direction change reference value, respectively. The appearance inspection method according to claim 6. A processed part is formed on the object,
The abnormality determination step includes a position deviation determination step for determining presence / absence of a position deviation abnormality of the processing unit,
In the positional deviation determination step,
For each of the plurality of pixels arranged in the one direction in the binarized image, a one-way ratio as a value obtained by dividing the one-way existence amount or the one-way change amount by the coordinate value in the other direction related to the plurality of pixels. Respectively,
For each of the plurality of pixels arranged in the other direction in the binarized image, the other direction ratio as a value obtained by dividing the other direction existence amount or the other direction change amount by the coordinate value of the one direction related to the plurality of pixels. Respectively,
When the calculated sum of the one-way ratios and the sum of the other direction ratios do not match the preset one-way ratio reference value and the other direction ratio reference value, respectively, The appearance inspection method according to claim 6 or 7, wherein it is determined that there is a positional deviation abnormality. A second pre-processing step of presetting the one-way ratio reference value and the other-direction ratio reference value;
In the second pretreatment step,
The sum of a plurality of the one-way ratios and the sum of the plurality of other-direction ratios for the binarized image of the normal object are respectively set to the one-way ratio reference value and the other-direction ratio reference value. The visual inspection method according to claim 8, wherein the visual inspection method is set in advance. In the binarization process,
Obtaining a lightness distribution relating to the gradation of the image obtained by the image obtaining unit;
In the lightness distribution, a first gradation when the lightness is the highest first peak value and a second gradation when the lightness is the second highest peak value after the first peak value are derived. And
The binarization process is performed on the image by using a value between the first and second gradations as a binarization level that serves as a threshold for performing the binarization process. The appearance inspection method according to any one of 6 to 9.