JP4581424B2 - Appearance inspection method and image processing apparatus - Google Patents

Description

The present invention relates to an appearance inspection method and an image processing apparatus for inspecting the appearance of a liquid in a container.

Conventionally, as an inspection method for detecting a foreign matter with respect to a liquid in a container, an inspection method for detecting bubbles and foreign matters by switching illumination / imaging directions with a plurality of illuminations (2 units) and a plurality of imaging cameras (4 units) and performing differential processing. (For example, Patent Document 1) or by adjusting the amount of transmitted light and reflected light relative to each other to cause bubbles to be exposed with high brightness and to allow foreign objects to be exposed with low brightness. There is an inspection method for discriminating (for example, Patent Document 2), a comparative inspection image from which bubbles are removed using the minimum luminance of each pixel of a time-series image, and a method for detecting foreign matter from this difference image (for example, Patent Document 3). It was.
Japanese Patent Laid-Open No. 9-325122 (FIG. 1, paragraph number 0021) JP 2001-116703 A Japanese Patent No. 3295628

  Among the conventional examples, the method of Patent Document 1 requires a plurality of photographing cameras and illumination, and there is a problem in that the size and cost of the apparatus cannot be avoided. The method of Patent Document 2 requires a light source for transmitted light and a light source for reflected light, and also requires means for adjusting the amount of illumination light. Furthermore, in the method of Patent Document 3, the container containing the liquid to be inspected is placed on the rotary table, and the container is rotated and stopped repeatedly, or the container is moved to turn the foreign matter in the liquid, It is intended to capture the turning image, and means for rotating and moving the container is necessary.

The present invention has been made in view of the above points, and its object is to provide an external appearance capable of accurately detecting bubbles in a liquid in a container in a stationary state without rotating the container. An object is to provide an inspection method and an image processing apparatus .

To achieve the above object, the invention of claim 1, in the appearance inspection how the liquid inside is sealed in a transparent or translucent container capable imaging externally, by the imaging means of liquid from the vessel outside An imaging process of capturing and holding as a grayscale image, and generating a differential value image and a differential direction value image by performing spatial differential processing on each pixel in a preset inspection region in the captured grayscale image A differential processing process, a raster scan within the inspection area of the differential value image, search for a pixel having a value equal to or greater than a predetermined differential value, and a bubble candidate point extraction process using the pixel extracted in this search as a bubble candidate point; When the extracted bubble candidate point is the starting point, if there are other bubble candidate points in the surrounding pixels, these bubble candidate points are concatenated to extract the bubble outline and use it as the bubble candidate part. And outside the outline of the extracted bubble candidate part A bubble circumscribed rectangle setting process for setting a circumscribed rectangle to be set, and an inspection line is set in at least one of the vertical direction and the horizontal direction inside the set bubble circumscribed rectangle, and the bubble candidate portion where the inspection line intersects is set. The number of contours, the number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value among the pixels on the inspection line, and the total sum of pixels having a value greater than or equal to the predetermined differential value among pixels on the inspection line Or using any of the tracking lines that return to the start point in the tracking process of pixels having a value equal to or greater than a predetermined differential value among the pixels on the inspection line, to determine whether or not bubbles exist in the bubble candidate portion And a determination process.

According to the first aspect of the present invention, the inspection line set inside the outline of the bubble candidate portion is set by differential processing in the inspection region, and the bubbles are formed in the bubble candidate portion using each pixel on the inspection line. since but to determine whether there is, in comparison with such a mere binarization processing, it is accuracy better detection.

In the invention of claim 2, determined in the invention of claim 1, in the determining step, the number of the contour of the bubble candidate section inspection lines set intersect, whether bubbles are present in the bubble candidate section It is characterized by doing.

  According to the second aspect of the present invention, it is possible to detect bubbles only by counting pixels by raster scanning on the inspection line, and it is possible to detect bubbles at high speed.

According to a third aspect of the present invention, in the first aspect of the invention, in the determination process, the number of types of differential direction values of pixels having a value equal to or greater than a predetermined differential value is determined for the pixels on the set inspection line. If the value of the number is equal to or greater than a predetermined threshold, it is determined that bubbles are present in the bubble candidate portion .

According to the invention of claim 3, since it is assumed that there are a plurality of pixels having different differential direction values on the inspection line by utilizing the feature that the differential direction values are different between the inside and the outside of the outline of the bubble . It can be detected with high accuracy foam.

According to a fourth aspect of the present invention, in the first aspect of the present invention, in the determination process, the sum of pixels having a value equal to or greater than a predetermined differential value is calculated among the pixels on the set inspection line, and the total value is a predetermined value. If it is more than a threshold value, it will determine with a bubble existing in a bubble candidate part, It is characterized by the above-mentioned.

  According to the fourth aspect of the present invention, it is possible to stably detect the bubble portion by removing noise such as a portion where a locally low differential value exists within the outline of the bubble candidate portion.

According to a fifth aspect of the present invention, in the first aspect of the invention, in the determination process, the differential direction value of a pixel having a value equal to or greater than a predetermined differential value for each inspection line in the set inspection line. It is characterized in that it is determined that bubbles are present in the bubble candidate portion if the sum of the inspection lines whose value of the number of types is equal to or greater than a predetermined value is equal to or greater than a predetermined threshold.

  According to the invention of claim 5, it can be predicted that bubbles exist when there are a plurality of differential direction values on the inspection line, but by using the information of the plurality of inspection lines, detection of bubbles with higher accuracy can be performed. it can.

According to a sixth aspect of the invention, in the first aspect of the invention, in the determination process, the number of types of differential direction values of pixels having a value equal to or greater than a predetermined differential value is obtained in the pixels on the set inspection line. A gray value of pixels existing between pixels having a differential direction value is measured, and when the measured value has a value equal to or smaller than a predetermined threshold, it is determined that bubbles exist in the bubble candidate portion.

According to the sixth aspect of the present invention , stable bubble detection can be performed by using a plurality of differential direction values and density data on the contour on the inspection line.

In the invention of claim 7, in the invention of claim 1, in the determination process, the number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value is obtained in the pixels on the set inspection line. When the gray value of pixels existing between pixels having a differential direction value is measured, and the area of the region surrounded by the pixels whose measured value is equal to or greater than a predetermined value is equal to or greater than a predetermined threshold, bubbles are generated in the bubble candidate portion. It is determined that it exists.

According to the seventh aspect of the present invention , bubbles can be stably detected when the shade value of the outline of the bubble candidate part and the width of the outline of the bubble candidate part can be specified to some extent in advance.

According to an eighth aspect of the present invention, in the first aspect of the invention, in the determination process, the number of types of differential direction values of pixels having a value equal to or greater than a predetermined differential value is determined for the pixels on the set inspection line. measuring the gray value of the pixels existing between the pixels having the direction value, at which measurements are Jo Tokoro threshold above the area of the region surrounded by the pixel is greater than or equal to a predetermined value, and the outer perimeter of the area When the value is equal to or greater than a predetermined threshold, it is determined that bubbles exist in the bubble candidate portion.

According to the invention of claim 8, when the width of the gray value and the contour of the contour portion can be pre-specified to some extent, by the judgment condition along the area and perimeter can detect stable foam.

In the invention of claim 9, in the invention of claim 1, in the determination process, the number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value is obtained in the pixels on the set inspection line. Measures the gray value of pixels existing between pixels having a differential direction value, and the area of the region surrounded by the pixels whose measured value is less than or equal to a predetermined value has a value equal to or greater than a predetermined threshold, and the average of the regions When the gray value has a predetermined threshold value or less , it is determined that bubbles exist in the bubble candidate portion.

According to the invention of claim 9, when the width of the contour of the gray values and foam candidate portions of the contour of the bubble candidate parts can be pre-specified to some extent, the use of the gray value data of a plurality of pixels existing on the area and the contour Can stably detect bubbles.

According to a tenth aspect of the present invention, in the first aspect of the invention, in the determination process, the type of differential direction value of a pixel having a value greater than or equal to a predetermined differential value for each inspection line in the pixels on the set inspection line. take the number, for the value of the type number is a predetermined value or more test lines, and wherein the determining the number of the inspection lines are consecutive when having a predetermined threshold value or more, the foam is present in the bubble candidate section To do.

  According to the invention of claim 10, it is assumed that there are a plurality of pixels having different differential direction values on the inspection line by utilizing a feature having different differential direction values inside and outside the outline of the bubble, Since continuity across a plurality of inspection lines is taken into account, the detection of bubbles can be performed stably.

According to an eleventh aspect of the present invention, in the first aspect of the invention, in the determination step, the type of differential direction value of a pixel having a value greater than or equal to a predetermined differential value for each inspection line in the pixels on the set inspection line. A pixel having a differential direction value in a predetermined direction greater than a predetermined number greater than 2 is present on the previous inspection line, and the differential direction value in one direction or two directions is present on the next inspection line. Or there is a pixel having a differential direction value in one direction or two directions on the previous inspection line, and a differential direction value in the direction equal to or greater than the predetermined number on the next inspection line. When there is a pixel having, it is determined that a bubble is present in the bubble candidate portion .

  According to the invention of claim 11, it is assumed that there are a plurality of pixels having different differential direction values on the inspection line by using a feature having different differential direction values inside and outside the outline of the bubble. Since information on the continuity of a plurality of inspection lines is used, it is possible to detect bubbles with high accuracy.

In the invention of claim 12, in the invention of claim 1, in the determination process, the number of types of differential direction values of the pixels having a value equal to or greater than a predetermined differential value is obtained in the pixels on the set inspection line. The gray value of pixels existing between pixels having a differential direction value is measured, and the area of the region surrounded by pixels whose measured value is equal to or greater than a predetermined value has a value equal to or greater than a predetermined threshold, and the average of the regions When the gray value has a value equal to or greater than a predetermined threshold value for which an offset is set, it is determined that bubbles are present in the bubble candidate portion.

According to the invention of claim 12, when the width of the contour of the gray values and foam candidate portions of the contour of the bubble candidate parts can be pre-specified to some extent, using the gray value data of a plurality of pixels existing on the area and contour, In addition, flexible bubbles can be detected in consideration of the margin of the average gray value.

According to a thirteenth aspect of the present invention, in the first aspect of the invention, in the determination process, the number of types of differential direction values of pixels having a value equal to or greater than a predetermined differential value is determined for the pixels on the set inspection line. the value of the number to set the gray value measurement pixels on either side of a pixel having a differential direction value of not less than a predetermined value obtains a shading difference between two gray values the measuring pixel, the absolute value of the shading difference is Tokoro When it has a value equal to or greater than a certain threshold value, it is determined that bubbles exist in the bubble candidate portion.

  According to the invention of claim 13, by measuring the density difference between pixels located on both sides of the outline of the bubble candidate portion, the information on the boundary of the bubble candidate can be examined more accurately, and stable bubble detection is possible. Is possible.

In the invention of claim 14, in the invention of claim 1, in the determination process, the number of types of differential direction values of pixels having a value equal to or greater than a predetermined differential value is determined for the pixels on the set inspection line. A gray value measurement pixel is set on both sides of a pixel having a differential direction value when the value of the number is equal to or greater than a predetermined value, and a gray level difference between the two gray value measurement pixels is obtained. checks whether the test line is a value or more continuously present, when the inspection line values of shading difference is the predetermined value or more is present to continue the predetermined threshold value than the upper communication, foam bubble candidate section It is characterized by determining that exists.

  According to the invention of claim 14, by measuring the density difference between the pixels located on both sides of the outline of the bubble candidate portion, it is possible to check the boundary information of the bubble candidate more accurately, and a plurality of inspection lines By determining the continuity using the information, stable foam can be detected.

According to a fifteenth aspect of the present invention, in the first aspect of the invention, in the determination process, a pixel having a value greater than or equal to a predetermined differential value among the pixels on the set inspection line is subjected to tracking processing based on the differential direction value. If there are two or more tracking lines that return to the starting point, it is determined that bubbles are present in the bubble candidate portion .

According to the fifteenth aspect of the present invention, it is possible to determine a bubble by a simple process by tracking the pixels on the contour on the assumption that there are two contours of the foam candidate portion on the outer side and the inner side.
According to the sixteenth aspect of the present invention, in an image processing apparatus used for an appearance inspection apparatus for a liquid sealed in a transparent or translucent container capable of imaging the inside from the outside, together with an imaging means for imaging the liquid from the outside of the container, Unit, an image memory area, a differential image memory area, and a differential direction image memory area, and the arithmetic unit uses the grayscale image captured by the imaging means as the original image and stores the data of the original image in the image memory area And a differential image and differential direction value image are generated by performing spatial differential processing on each pixel in a preset inspection region in the captured grayscale image, and image data of the differential image Is stored in the differential image memory area, the differential processing step of storing the image data of the differential direction value image in the differential direction image memory area, and raster scanning in the inspection area of the differential value image, Search for pixels with a value greater than or equal to the fractional value, and use the extracted bubbles as candidate points to extract the bubble candidate points, and other bubble candidate points exist in the surrounding pixels starting from the extracted bubble candidate points When the bubble candidate points are connected, the bubble extraction is performed to extract the bubble outline, and the bubble candidate part outline extraction process is set, and the circumscribed rectangle that circumscribes the extracted bubble candidate part outline is set. The circumscribed rectangle setting process, the inspection line is set in at least one of the vertical direction and the horizontal direction inside the set bubble circumscribed rectangle, and the number of the contours of the bubble candidate portion where the inspection line intersects, the inspection line The number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value among the upper pixels, the sum of pixels having a value greater than or equal to the predetermined differential value among the pixels on the inspection line, or on the inspection line A pixel with a value greater than or equal to a predetermined differential value Using any of the tracking line back to the starting point in the sense, and performs processing of each process with the determination process whether bubbles are present in the foam candidate section.
According to the sixteenth aspect of the present invention, an inspection line set inside the outline of the bubble candidate portion is set by differential processing in the inspection region, and bubbles are added to the bubble candidate portion using each pixel on the inspection line. Therefore, it is possible to detect with high accuracy compared with a simple binarization process.

The present invention sets the inspection line set inside the outline of the bubble candidate part by differential processing in the inspection region, and for each pixel on the inspection line, whether or not bubbles exist in the bubble candidate part. since the determination, as compared with such a mere binarization processing, there is an effect that accuracy better detection becomes possible.

  FIG. 6 shows a schematic configuration of an appearance inspection apparatus for implementing the appearance inspection method of the present invention. In this appearance inspection apparatus, the liquid enclosed in the transparent or translucent container 1 is the object to be inspected, and the CCD is used as an imaging device for imaging the liquid inside the illumination apparatus 2 and the container 1 with the container 1 in between. A camera 3 is arranged. The image data picked up by the CCD camera is taken into the image processing device 4, and the image processing device 4 performs various arithmetic processing using the acquired image data, and the result of the determination processing using the arithmetic processing result is converted into an image. The data is output to an external device 6 such as a monitor device 5 or PLC to control the device.

  FIG. 7 shows the internal configuration of the image processing device 4 in detail. An A / D converter 7 for A / D converting analog image data from the CCD camera 3 and an image to the image monitor device 5 are shown. A VRAM 8 that holds digitized image data for projection; a D / A converter 9 that D / A converts the image data held in the VRAM 8 and outputs the converted image data to the image monitor device 5; Flash ROM 11 storing data and programs, an arithmetic unit 12 including a CPU for performing arithmetic processing, a signal for giving a clock and the like to the main memory unit 10 and the arithmetic unit 12 including VRAM 8 and RAM The control unit 13 and an external control unit 14 that sends commands for controlling the external device 6 under the control of the calculation unit 12 to the external device 6. The main memory unit 10 includes image data. Is read from the flash ROM 11, the differential image memory area 10 b for storing differential image data for storing differential images to be described later, the differential direction image memory area 10 c for storing differential direction image data, and the flash ROM 11. Each area of the image processing program memory area 10d for storing the processed image processing program and data is set.

  Using the appearance inspection apparatus configured as described above, it has the same brightness as the liquid inside, except for bubbles that appear grayer than the liquid and have a diameter smaller than a predetermined value, and double in the differential direction image. An embodiment of the appearance inspection method of the present invention for identifying bubbles and foreign matters using the feature of bubbles and the feature of foreign matters that are gray to black rather than liquid and not bright inside is explained by an embodiment. .

  In the present embodiment, the appearance inspection is advanced by the procedure according to the flowchart shown in FIG.

  The first step S1 is a step in which the CCD camera 3 images the container 1 in which the liquid to be inspected is sealed, and this imaging is performed by an operator's operation.

  An original image composed of grayscale images obtained by this imaging is as shown in FIG. In the original image shown in FIG. 2A, a rectangular portion 20 is a boundary portion between the container 1 and the background, and if there are bubbles or foreign matter in the liquid, a portion 21 having a dark outline with respect to the liquid is generated. Will be present in the image.

  In the image processing apparatus 4, under the control of the arithmetic unit 12 operating along the arithmetic processing program, the original image data is held in the VRAM 8, the original image is displayed on the image monitor apparatus 5, and the image memory area 10a is displayed. To store.

  The operator sets the inspection area 22 in the rectangular portion 20 while viewing the original image as shown in FIG. For example, if the setting is set on the screen using a pointing device such as a mouse using a GUI environment created under the control of the calculation unit 12, the setting information is taken into the calculation unit 12.

  When the operator first instructs differentiation processing for each pixel in the inspection region 22 in accordance with the procedure presented through the GUI environment, the calculation unit 12 performs differentiation processing (step S2) for each pixel in the inspection region 22. Thus, a differential value image and a differential direction value image are generated, and these image data are stored in the corresponding memory areas 10b and 10c.

When an operation of enlarging the area 23 of 16 × 16 pixels including the contour portion 21 is performed, the gray value of each pixel in the area 23 of the original image shown in FIG. 2B is shown in FIG. In such a case, the differential value of each pixel of the differential image is as shown in FIG. 3B, and the differential value of each pixel of the differential image is as shown in FIG. The gradation value is 256 gradations, 2 5 5 indicates black, and 0 indicates white.

  When the differentiation process is completed, the arithmetic unit 12 performs raster scanning in the inspection area 22 as shown in FIG. 2C (step S3). As shown in the figure, the raster scanning process scans from the upper left position to the lower right end of the inspection area 22 to search for pixels having a predetermined differential value or more (step S4), and to extract the pixels extracted by this search. It is set as a bubble candidate point (step S5).

  In this raster scan, thin line processing is performed to extract pixels located outside and outside the contour of pixels having a predetermined differential value or more as bubble candidate points.

  In this way, when the raster scan in the inspection region 22 is performed and it is determined in step S6 that the raster scan has been completed, the bubble candidate point connecting process is performed in the next step S7 and subsequent steps.

  In other words, raster scanning in the inspection region 22 is started again, and the previously extracted bubble candidate points are searched (step S8). If bubble candidate points exist, other bubble pixels around the bubble candidate points are located within the surrounding eight pixels. If a bubble candidate point exists, a process of connecting as a bubble outline is performed in step S9. If it is determined in step S10 that the raster scanning in the inspection area 22 has been completed, the process proceeds to step S11.

  In step S11, the computing unit 12 sets a circumscribed rectangle 30 as shown in the figure for each contour of the bubble candidate portion shown in FIG. 4, and in the next step S12, as shown in FIG. Inspection lines L1... Ln are set, and the characteristics of the pixels on the inspection line set in step S13 are examined based on the differential value or differential direction value of each pixel (indicated by ◯ in FIG. 5). Then, it is checked whether or not the feature satisfies the above-described bubble feature (step S14). If the feature is satisfied, it is determined that the bubble is present in the inspection region 22 (step S15). It determines with a bubble not existing (step S16). That is, when it is determined that there is no bubble, it can be determined that the cause of the contour 21 is not due to the bubble, that is, due to foreign matter.

  The calculation unit 12 displays the determination result on the image monitor device 5 and outputs a control output based on the determination result to the external device 6. In this way, it is possible to reliably detect foreign matter by determining whether or not the above-described contour 21 is caused by bubbles.

Next, an example of the bubble determination process will be described.
(Example 1)
As shown in FIG. 4, the outline of the foam candidate portion obtained by connecting the foam candidate points includes an inner outline and an outer outline as shown in FIG. In this example, bubbles are detected by utilizing the fact that the outline of the bubble candidate portion appears on the outside and the inside. That is, when the inspection lines L1... Ln are set in the horizontal direction, if the outlines of the bubble candidate portions indicated by the connecting lines of the hatched pixels are present on both the inner side and the outer side as shown in FIG. There are four pixels on the outline of the bubble candidate portion that intersects the line L5. Therefore, a predetermined number of pixels that intersect with a predetermined inspection line, for example, L5 is set as a reference value when the calculation unit 12 determines bubbles, and the number of pixels that intersect when the inspection line L5 is scanned in raster scanning is set. Counting is performed, and when the count value matches the reference value, it is determined as a bubble.

As described above, in this example, bubbles can be detected only by counting pixels by raster scanning on the inspection line using the characteristics of the bubbles, and bubbles can be detected at high speed. It is what.
(Example 2)
In this example, as described above, the outline of the bubble candidate portion appears on the outside and the inside, and bubbles are detected using the feature of the bubbles that have different differential direction values on the inside and outside.

  That is, for example, when the inspection line L5 is set in the horizontal direction as shown in FIG. 9, if the contour of the bubble candidate portion exists on both the inner side and the outer side, it exists on the contour of the bubble candidate portion that intersects the inspection line L5. In the illustrated example, there are four types of differential direction values of pixels: 6, 2, 8, and 4. Therefore, for example, the calculation unit 12 counts the number of types of differential direction values of the pixels that intersect the inspection line L5, and determines that bubbles are present if the number is greater than a predetermined threshold value (for example, 4) set in advance. .

As described above, in this example, it is assumed that there are a plurality of pixels having different differential direction values on the inspection line by using the feature having different differential direction values inside and outside the outline of the bubble. This makes it possible to detect bubbles with high accuracy based on the above characteristics.
(Example 3)
In this example, as shown in FIG. 10, the density difference is smaller than the outline X1 and X2 of the bubble candidate part inside the outline X1 and X2 of the bubble candidate part, but the density difference occurs due to unevenness α of the light amount, It uses the feature of bubbles that there is a possibility that a portion having a certain size of the differential value may appear.

  In other words, in the case of this example, when raster scanning is performed on the set inspection lines L1... Ln, the arithmetic unit 12 sets a predetermined value larger than the differential value of the unevenness α on the inspection line L shown in FIG. The number of pixels exceeding the value is counted, and after the raster scan is completed, the sum of the count values is compared with a predetermined threshold value, and if the sum is equal to or greater than the threshold value, a bubble is determined. .

As described above, in this example, it is possible to stably detect bubbles by removing noise (light quantity unevenness) such as a portion having a locally low differential value inside the contour of the bubble candidate portion. is there.
(Example 4)
In this example, the bubble is detected by utilizing the fact that the outline appears on the outside and the inside, and the differential direction values are different between the inside and the outside, as in the case of Example 2.

  That is, in FIG. 9, when the inspection lines L1... Ln are set in the horizontal direction, as shown in the figure, the outline of the bubble candidate portion exists on both the inner side and the outer side, and there are multiple types of differential direction values. It can be predicted that bubbles are present on the line.

And although only utilized for Example 2, the inspection line L5, for example, with L5 similarly to the four differential direction value even in the inspection line L6 Then, plurality of satisfying test line having four differential direction value Will do.

  Therefore, in this example, the calculation unit 12 counts the number of inspection lines having, for example, four types of differential direction values, and determines that bubbles are present when the count value is equal to or greater than a preset threshold value.

As described above, in this example, it is possible to detect bubbles more accurately by using the information that can be predicted that bubbles exist when there are multiple differential direction values on the inspection line and information on the plurality of inspection lines. It is what.
(Example 5)
In this example, as in the case of Example 2 and Example 4, the detection of the bubble is performed by utilizing the feature of the bubble that the outline appears on the outside and the inside and the differential direction value is different between the inside and the outside. Do.

  That is, when the inspection lines L1... Ln are set in the horizontal direction, as shown in FIG. 9, if the outlines of the bubble candidate portions exist on both the inner and outer sides, as described in Example 2, for example, intersect with the inspection line L5. There are four types of differential direction values of 6, 2, 8, and 4 for the pixels existing on the outline of the bubble candidate portion to be performed.

  When the grayscale values of the pixels whose differential direction values are 6 and 2 and 8 and 4 are measured, the grayscale values of the pixels A to D (see FIG. 9) on the inspection line L5 are the pixel A as shown in FIG. , B and between C and D become less than a predetermined value (the contour portion becomes dark).

  Thus, in this example, in order to make use of the above characteristics, the calculation unit 12 measures the gray value of the differential direction value, and compares the measured value with a predetermined threshold value which is set in advance. When the value is less than or equal to the threshold value, it is determined as a bubble.

As described above, in this example, by using the differential direction value on the inspection line and the gray value data of the pixels existing on the outline of the bubble candidate portion, it is possible to detect bubbles stably using the characteristics of the bubbles. It is a thing.
(Example 6)
In this example, the calculation unit 5 performs a process of counting the number of pixels in which the gray value in Example 5 is equal to or less than a predetermined value. In this process, in addition to the inspection line L5, the gray value is measured in other inspection lines (L6 in FIG. 9) having four or more kinds of differential direction values of the pixels on the inspection line, and the number of pixels equal to or smaller than the predetermined value is counted. Perform the process. And by taking the sum of the number of pixels in a plurality of inspection lines have been processed in this manner, that is, the area value of the enclosed at a predetermined value or less of the pixel area.
Then, this area value is compared with a preset threshold value, and when the area value is equal to or greater than the threshold value, it is determined as a bubble.

As described above, in this example, it is possible to stably detect bubbles when the gradation value of the outline of the bubble candidate portion and the width of the outline of the bubble candidate portion can be specified to some extent in advance.
(Example 7)
In this example, the processing unit 12 extracts the region described in Example 6 as the region Y as shown in FIG. 12, and performs processing for measuring the area in the extracted region Y and the outer perimeter of the region Y. When these measured values are equal to or greater than the preset threshold values, it is determined that bubbles are present.

As described above, in this example, as the feature of the bubble, when the gray value and width of the outline of the bubble candidate portion can be specified to some extent in advance, the region surrounded by the pixels whose measured value of the gray value of the differential direction value is a predetermined value or less By combining the area of Y and the outer perimeter of the region Y as a determination condition, stable bubble detection is possible.
(Example 8)
In Example 7, the area of the region Y described above and the outer peripheral length thereof are used as the bubble determination condition. However, in this example, instead of the outer peripheral length, the region Y, that is, the A pixel-B pixel in FIG. An average gray value between the C pixel and the pixel D is obtained, and when the average gray value is equal to or less than a preset threshold value, it is determined that the contour is a bubble.

As described above, in this example, when the gray value of the bubble candidate contour portion and the width of the contour of the bubble candidate portion can be specified to some extent in advance, by using the gray value data of the plurality of pixels existing on the area and the contour, Stable detection is possible.
(Example 9)
As described above, as shown in FIG. 9, when the outline of the bubble candidate part exists on both the inner and outer sides, for example, the differential direction value of the pixel existing on the outline of the bubble candidate part intersecting the inspection line L5 is 6, There are four types, 2, 8, and 4. In Example 2, it is determined that bubbles are present, assuming that the number of types is equal to or greater than a preset threshold value (for example, 4).

  On the other hand, in this example, the calculation unit 12 performs the same processing for the inspection lines L4, L6, and L7 before and after the inspection line 5, and counts the number of consecutive inspection lines determined as the presence of bubbles, If the count value is equal to or greater than a preset threshold value (for example, 4), it is determined that bubbles are present.

As described above, in this example, it is assumed that there are a plurality of pixels having different differential direction values on the inspection line, using the feature that the differential direction values are different between the inside and the outside of the outline of the bubble, By considering the continuity across a plurality of inspection lines, it is possible to stably detect bubbles.
(Example 10)
As described above, as shown in FIG. 9, when the contour of the bubble candidate portion exists on both the inner and outer sides, for example, in the inspection line L5, the differential direction value of the pixel existing on the contour of the intersecting bubble candidate portion is 6, There are four types, 2, 8, and 4. In Example 2, it is determined that bubbles are present, assuming that the number of types is equal to or greater than a preset threshold value (for example, 4).

  On the other hand, in FIG. 9, for example, there are three types of pixel differential direction values of 6, 5, and 4 in the inspection line L1, but there are only two types of 6 and 4 in the next inspection line L2. The inspection line L10 has only two types of 8, 2 but the next inspection line 11 has three types of 8, 1, 2.

  In this example, the determination of bubbles is performed using data on the continuity of such an inspection line. For example, the calculation unit 12 has only one or two types of differential direction values in the previous inspection line. In the next inspection line, for example, there are three or more types of differential direction values, or there are three or more types of differential direction values in the previous inspection line, and then only one type or two types in the next inspection line. If any condition in the absence is satisfied, it is determined that bubbles are present.

As described above, in this example, it is assumed that there are a plurality of pixels having different differential direction values on the inspection line by using a feature having different differential direction values inside and outside the outline of the bubble. By using the information on the continuity of the inspection line, it is possible to detect bubbles with high accuracy.
(Example 11)
In Example 8, if the average gray value is equal to or greater than the predetermined threshold value, it is determined as a bubble. However, in this example, an offset can be set to the threshold value under the control of the calculation unit 12, and the calculation unit 12 sets the offset. The set threshold value is compared with the average gray value obtained in the same manner as in Example 8, and when the average gray value is equal to or larger than the threshold value, it is determined as a bubble.

As described above, in this example, when the gradation value of the contour portion and the width of the contour portion can be specified to some extent in advance, the average gradation value of a plurality of pixels existing on the area and the contour is used, and the margin of the average gradation value is considered. By using the threshold value, flexible bubbles can be detected.
(Example 12)
As described above, when the inspection lines L1... Ln are set in the horizontal direction as shown in FIG. 9, if the contours of the bubble candidate portions exist on both the inner and outer sides, for example, the contour of the bubble candidate portion that intersects the inspection line L5. There are four types of differential direction values of the pixels existing above: 6, 5, and 4.

  In this example, as shown in FIG. 13, for example, the calculation unit 12 sets the grayscale measurement pixels (pixels inside the O portion are hatched) a1 and a2 on both sides of the pixel having the differential direction value 6 on the inspection line L5. Then, the density difference between the pixels a1 and a2 is measured, and when the absolute value is greater than or equal to a preset threshold value, it is determined that bubbles are present.

As described above, in this example, by measuring the difference in density between pixels located on both sides of the outline of the bubble candidate portion, it is possible to examine the information on the boundary of the bubble candidate more accurately, and to detect stable bubbles. It is possible.
(Example 13)
In Example 12, a single inspection line, for example L5 sets the density measurement pixels a1, a2 to measure the shading difference between the pixel a1,2, when the absolute value thereof is equal to or greater than a preset threshold value Although it has been determined that bubbles are present, in this example, the calculation unit 12 determines the presence of bubbles in, for example, the inspection line L5, and then the inspection line, for example, L6, as shown in FIG. Set the density measurement pixels b1 and b2, measure the density difference between the pixels b1 and b2, and if the density difference is greater than or equal to a predetermined threshold, perform a process of determining the presence of bubbles, When the inspection lines determined as bubbles are continuous, the presence of bubbles is finally determined.

As described above, in this example, by measuring the difference in density between pixels located on both sides of the outline of the bubble candidate portion, the information on the boundary of the bubble candidate can be examined more accurately, and more than one inspection line By determining continuity using information, stable bubble detection is possible.
(Example 14)
In this example, the outline of the bubble candidate portion appears on the outer side and the inner side, and the feature of the bubbles that the direction values on the inner side and the outer side are different is used. When the inspection line is set in the horizontal direction, FIG. As shown in the figure, when raster scanning is performed by the calculation unit 12, the pixels of the outline of the bubble candidate portion are tracked in order from the start point I as indicated by arrows, and returned to the start point I via the end point II. In the case where two tracking lines exist outside and inside (possibly there may be three or more due to the influence of noise or the like), it is determined that bubbles are present.

  As described above, according to this example, assuming that there are two outlines of the foam candidate part, the outer and inner sides, it is possible to determine the foam by a simple process by tracking pixels on the outline. Is.

It is a flowchart of the basic processing of one Embodiment of this invention. (A) is an example of an original image taken by a CCD camera of an appearance inspection apparatus using the same as above, (b) is an example of an image when an inspection area is set, and (c) is a state diagram of raster scanning for the set inspection area. It is. (A) is an enlarged view showing the density value of each pixel of the bubble part, (b) is an enlarged view showing the differential value of each pixel of the bubble part, (c) is a picture of each pixel of the bubble part. It is an enlarged view which shows a differential direction value. It is an enlarged view which shows the outline of a foam candidate part same as the above. It is setting explanatory drawing of an inspection line same as the above. It is a block diagram of the external appearance inspection apparatus used for the same as the above. It is a block diagram of the image processing apparatus of the external appearance inspection apparatus used for the same as the above. It is explanatory drawing of Example 1 of the bubble determination method used for the same as the above. It is explanatory drawing of Example 2, Example 4, Example 5, Example 6, Example 9, and Example 10 of the bubble determination method used for the same as the above. It is explanatory drawing of the example 3 of the bubble determination method used for the same as the above. It is explanatory drawing of Example 5 and Example 8 of the bubble determination method used for the same as the above. It is explanatory drawing of the example 7 of the bubble determination method used for the same as the above. It is explanatory drawing of the example 12 and the example 13 of the bubble determination method used for the same as the above. It is explanatory drawing of Example 14 of the bubble determination method used for the same as the above.

Claims (16)

In the visual inspection method of liquid sealed in a transparent or translucent container that can be imaged from the outside,
An imaging process in which a liquid is imaged from outside the container by an imaging means and held as a grayscale image;
Differentiating process of generating a differential value image and a differential direction value image by performing a spatial differential process for each pixel in a preset inspection region in the captured grayscale image,
A raster scan within the inspection area of the differential image, searching for pixels having a value greater than or equal to a predetermined differential value, and a bubble candidate point extraction process using the pixel extracted in this search as a bubble candidate point;
When the extracted bubble candidate point is the starting point, if there are other bubble candidate points in the surrounding pixels, these bubble candidate points are concatenated to extract the bubble outline and use it as the bubble candidate part. When,
A bubble circumscribed rectangle setting process for setting a circumscribed rectangle circumscribing the outline of the extracted bubble candidate part;
An inspection line is set in at least one of the vertical direction and the horizontal direction inside the set bubble circumscribed rectangle, the number of the contours of the bubble candidate portion where the inspection line intersects, the predetermined number of pixels on the inspection line The number of types of differential direction values of pixels having a value greater than or equal to the differential value, the sum of pixels having a value greater than or equal to a predetermined differential value among pixels on the inspection line, or a predetermined number of pixels on the inspection line And a determination process for determining whether or not bubbles are present in the bubble candidate portion using any of the tracking lines that return to the starting point in the tracking processing of pixels having a value greater than or equal to the differential value. Method.   The appearance inspection method according to claim 1, wherein in the determination process, it is determined whether or not bubbles exist in the bubble candidate portion based on the number of contours of the bubble candidate portion where the set inspection lines intersect.   In the determination process, the number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value is taken for pixels on the set inspection line, and if the value of the number of types is equal to or greater than a predetermined threshold, bubble candidates The appearance inspection method according to claim 1, wherein it is determined that bubbles are present in the part.   In the determination process, the pixels on the set inspection line are summed up with pixels having a value greater than or equal to a predetermined differential value, and if the sum is greater than or equal to a predetermined threshold, bubbles are present in the bubble candidate portion. The appearance inspection method according to claim 1, wherein determination is made.   In the determination process, the number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value for each inspection line in the pixels on the set inspection line is taken, and the value of the number of types is a predetermined value The appearance inspection method according to claim 1, wherein if the sum of the above inspection lines is equal to or greater than a predetermined threshold value, it is determined that bubbles are present in the bubble candidate portion.   In the determination process, the pixels on the set inspection line take the number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value, and the gray value of pixels existing between pixels having the differential direction value The appearance inspection method according to claim 1, wherein when the measured value has a value equal to or less than a predetermined threshold value, it is determined that bubbles are present in the bubble candidate portion.   In the determination process, the pixels on the set inspection line take the number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value, and the gray value of pixels existing between pixels having the differential direction value 2. The method according to claim 1, wherein when the area of a region surrounded by pixels whose measured values are equal to or greater than a predetermined value is equal to or greater than a predetermined threshold, it is determined that bubbles are present in the bubble candidate portion. Appearance inspection method.   In the determination process, the number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value among the pixels on the set inspection line is taken, and the gray value of the pixel existing between the pixels having the differential direction value is obtained. A bubble is present in the bubble candidate portion when the area of a region surrounded by pixels whose measured values are equal to or greater than a predetermined value is equal to or greater than a predetermined threshold and the outer perimeter of the region is equal to or greater than a predetermined threshold. The visual inspection method according to claim 1, wherein the visual inspection method is determined.   In the determination process, the pixels on the set inspection line take the number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value, and the gray value of pixels existing between pixels having the differential direction value When the area of a region surrounded by pixels whose measured values are equal to or smaller than a predetermined value has a value equal to or larger than a predetermined threshold value, and the average gray value of the region has a predetermined threshold value or less, The appearance inspection method according to claim 1, wherein it is determined that bubbles exist in the candidate portion.   In the determination process, for each pixel on the set inspection line, the number of types of differential direction values of pixels having a value equal to or greater than a predetermined differential value is taken for each inspection line, and the value of the number of types is equal to or greater than a predetermined value. The appearance inspection method according to claim 1, wherein when the number of consecutive inspection lines has a predetermined threshold value or more, it is determined that bubbles are present in the bubble candidate portion.   In the determination process, the number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value is obtained for each pixel on the set inspection line, and a value larger than 2 is previously set on the previous inspection line. When there is a pixel having a differential direction value in a direction equal to or greater than a predetermined number set, and there is only a pixel having a differential direction value in one direction or two directions on the next inspection line, or the previous inspection line When a pixel having a differential direction value in one direction or two directions is present on the upper side and a pixel having a differential direction value in the direction equal to or larger than the predetermined number is present on the next inspection line, a bubble is generated in the bubble candidate portion. The appearance inspection method according to claim 1, wherein it is determined that an image exists.   In the determination process, the pixels on the set inspection line take the number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value, and the gray value of pixels existing between pixels having the differential direction value The area of a region surrounded by pixels whose measured values are equal to or greater than a predetermined value has a value equal to or greater than a predetermined threshold, and the average gray value of the region is equal to or greater than a predetermined threshold for which an offset is set. The appearance inspection method according to claim 1, wherein it is determined that bubbles are present in the bubble candidate portion.   In the determination process, the number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value is taken for pixels on the set inspection line, and the differential direction value when the value of the number of types is equal to or greater than the predetermined value. If the grayscale value measurement pixel is set on both sides of the pixel to obtain the grayscale difference between the two grayscale value measurement pixels, and the absolute value of the grayscale difference has a value equal to or greater than a predetermined threshold value, The appearance inspection method according to claim 1, wherein it is determined that bubbles are present.   In the determination process, the number of types of differential direction values of pixels having a value greater than or equal to a predetermined differential value is taken for pixels on the set inspection line, and the differential direction value when the value of the number of types is equal to or greater than the predetermined value. Whether or not there are continuous inspection lines in which the density difference between the two density value measurement pixels is determined by setting the density value measurement pixels on both sides of the pixel, and the density difference value is a predetermined value or more. 2. The appearance according to claim 1, wherein when the inspection line having a density difference value equal to or greater than a predetermined value exists continuously for a predetermined threshold value or more, it is determined that bubbles are present in the bubble candidate portion. Inspection method.   In the determination process, when there are two or more tracking lines that return to the starting point by performing tracking processing based on the differential direction value of pixels on the set inspection line that have a value greater than or equal to a predetermined differential value 2. The appearance inspection method according to claim 1, wherein it is determined that bubbles are present in the bubble candidate portion. In an image processing apparatus used together with an imaging means for imaging a liquid from outside the container, in a liquid appearance inspection apparatus enclosed in a transparent or translucent container capable of imaging the inside from the outside,
A calculation unit, an image memory area, a differential image memory area, and a differential direction image memory area;
The calculation unit
The process of storing the grayscale image captured by the imaging means as the original image and storing the original image data in the image memory area;
A differential image and a differential direction value image are generated by performing spatial differential processing on each pixel in a preset inspection region in the captured grayscale image, and the differential image memory stores the differential image image data. A differential process for storing the image data of the differential direction value image in the differential direction image memory area,
A raster scan within the inspection area of the differential image, searching for pixels having a value greater than or equal to a predetermined differential value, and a bubble candidate point extraction process using the pixel extracted in this search as a bubble candidate point;
When the extracted bubble candidate point is the starting point, if there are other bubble candidate points in the surrounding pixels, these bubble candidate points are concatenated to extract the bubble outline and use it as the bubble candidate part. When,
A bubble circumscribed rectangle setting process for setting a circumscribed rectangle circumscribing the outline of the extracted bubble candidate part;
An inspection line is set in at least one of the vertical direction and the horizontal direction inside the set bubble circumscribed rectangle, the number of the contours of the bubble candidate portion where the inspection line intersects, the predetermined number of pixels on the inspection line The number of types of differential direction values of pixels having a value greater than or equal to the differential value, the sum of pixels having a value greater than or equal to a predetermined differential value among pixels on the inspection line, or a predetermined number of pixels on the inspection line In the tracking process of pixels having a value equal to or greater than the differential value, the process of each process including the determination process for determining whether or not a bubble exists in the bubble candidate portion is performed using any of the tracking lines that return to the starting point.
An image processing apparatus.

Images