JP2005031069A - X-ray inspection apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To maintain high inspection accuracy while shortening the time required for inspection for counting the number of articles in an X-ray inspection apparatus. <P>SOLUTION: The X-ray inspection apparatus includes an X-ray irradiating device 13 for irradiating X-rays; an X-ray line sensor 14; and a control computer 20. The X-ray line sensor 14 receives X-rays from the X-ray irradiating device 13. The control computer 20 performs image processing on an image by the X-rays received by the X-ray line sensor 14 and inspects merchandise passing between the X-ray irradiating device 13 and the X-ray line sensor 14. In image processing or inspection, the control computer 20 computes the perimeters of lump parts on the image on the basis of the arrangement of circumference pixels forming the circumferences of the lump parts and compares them with a reference perimeter. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an X-ray inspection apparatus, and more particularly to an X-ray inspection apparatus that inspects an article by performing image processing on an X-ray image.

  In the production line of commodities such as foods, an inspection may be performed by an X-ray inspection apparatus in order to prevent shipment of such commodities when foreign matters are mixed into the commodities or the products are cracked. In this X-ray inspection apparatus, X-rays are irradiated to articles continuously conveyed, the X-ray transmission state is detected by an X-ray line sensor, and foreign matters are not mixed in the articles, or It is determined whether or not the product is cracked or missing or the quantity of unit items in the package is insufficient. In addition, the X-ray inspection apparatus may perform an inspection that counts the number of articles in the package.

When inspecting an article with such an X-ray inspection apparatus, it is common to perform pattern matching between a detected image and a reference image (see, for example, Patent Document 1).
However, pattern matching between the detected image and the reference image requires a certain amount of time because it is necessary to adapt the directions of the two images or to perform a complicated similarity determination process.
In contrast, in an X-ray inspection apparatus deployed in a production line for food or the like that requires high speed, the time required for image processing is as much as possible for inspections that count the number of articles and confirm the quantity of articles. There is a need to shorten it. Also, in an X-ray inspection apparatus that inspects an abnormality in the shape of an article, it takes time to perform a similar similarity determination process. However, it is required to reduce the inspection time as much as possible.

Of course, if expensive hardware or software is used, pattern matching processing can be performed at high speed, but in reality, there is a demand for low speed as well as high speed.
A technique for solving such a problem is proposed in Patent Document 2. In this inspection apparatus, the integrated number of pixels constituting the periphery of the inspection target article is obtained, and whether or not the inspection target article is cracked is determined based on whether or not the integrated number is within a predetermined range.
JP 2002-98652 A (right column on page 6) JP 2002-310946 A

In the inspection apparatus of Patent Document 2, it is assumed that the integrated number of pixels constituting the periphery of the inspection target article is proportional to the peripheral length of the inspection target article, and the integrated number is a determination target.
However, approximating the perimeter of the article to be inspected by such an integrated number actually increases the error and makes it difficult to perform an accurate inspection. In particular, when the shape of the inspection target article is complicated or includes many curved portions, the cumulative number of pixels that form the periphery of the inspection target article may not be proportional to the peripheral length of the inspection target article. Become more.

  An object of the present invention is to maintain high inspection accuracy while reducing the time required for inspection of an article in an X-ray inspection apparatus.

  An X-ray inspection apparatus according to a first aspect includes an X-ray source that emits X-rays, an X-ray light receiving unit, and an inspection unit. The X-ray light receiving unit receives X-rays from the X-ray source. The inspection unit performs image processing on the X-ray image received by the X-ray light receiving unit, and inspects an article passing between the X-ray source and the X-ray light receiving unit. Further, in the image processing or inspection, the inspection unit calculates the peripheral length of the block portion on the image based on the arrangement of surrounding pixels that form the periphery of the block portion, and compares it with the reference peripheral length.

Here, the X-rays from the X-ray source pass through the article when the article is present, and otherwise reach the X-ray light receiving unit without passing through the article. Image processing is performed on the X-ray image, and further inspection of the article is performed. In this way, the inspection unit performs an inspection as to whether the shape of the sample is not abnormal or whether the number of unit articles in the sample is normal.
Here, in the image processing or inspection, the inspection unit compares the peripheral length of the block portion on the image with the reference peripheral length. As described above, by comparing the perimeter of the lump portion indicating the article or estimated as the article with the reference perimeter, image processing and inspection can be performed easily, at high speed, and with relatively high accuracy. become. Thereby, in this X-ray inspection apparatus, the time required for the inspection of the article is shortened.

  Here, since the perimeter of the lump portion is calculated based on the arrangement of surrounding pixels that form the perimeter of the lump portion, the number of surrounding pixels is simply added to approximate the perimeter of the lump portion. In comparison, the perimeter of the lump portion can be obtained with high accuracy. For this reason, in this X-ray inspection apparatus, the time required for the inspection of the article can be shortened, and the inspection accuracy can be maintained high.

The X-ray inspection apparatus according to claim 2 is the X-ray inspection apparatus according to claim 1, wherein the inspection unit includes a lump portion estimated as an article from an image and a background portion (a portion outside the lump portion). Then, the integration of the plurality of articles on the image due to the contact of the plurality of articles is checked by comparing the perimeter of the mass portion estimated as the article with the reference perimeter.
When inspecting an image including a plurality of articles by performing image processing, for example, when the two adjacent articles are close enough to contact each other, or the two adjacent articles are actually in contact with each other. The two articles may be recognized as one on the image. In this case, an incorrect quantity is detected in the case of inspection for counting the number of articles, and a normal article is erroneously recognized as an abnormal article when inspecting the shape of the article.

  On the other hand, in the apparatus of claim 2, the integration phenomenon of a plurality of articles on the image can be checked by distinguishing the lump portion and the background portion and comparing the peripheral length of the lump portion with the reference peripheral length. It is like that. Accordingly, it is possible to suppress the same image processing and inspection as in the case where there is no integration phenomenon despite the occurrence of the integration phenomenon, and giving erroneous inspection results.

The X-ray inspection apparatus according to claim 3 is the X-ray inspection apparatus according to claim 2, wherein the inspection unit compares the perimeter of the lump portion estimated as the article with the reference perimeter. After checking the integration phenomenon of the plurality of articles in the image processing, image processing for separating the lump portion in which the integration phenomenon is confirmed into a plurality of articles is performed.
Here, when the integration phenomenon is confirmed, a process of separating a plurality of integrated articles on the image is performed. After the separation image processing, image processing and inspection can be performed by normal logic. it can.

Note that if the separation image processing described in this section is not performed, it is necessary to change the inspection determination logic and the like. For example, when two articles are integrated on the image, it is recognized from the perimeter that two articles are in contact with each other instead of one or three, and the quantity of the articles is based on the recognition. It is necessary to inspect.
The X-ray inspection apparatus according to claim 4 is the X-ray inspection apparatus according to claim 2 or 3, wherein the inspection unit compares the perimeter of the lump portion estimated as the article with the reference perimeter. In addition, the integration phenomenon of a plurality of articles on the image is checked by comparing the area of the lump portion with the reference area.

Here, since both the perimeter and area of the lump portion on the image are compared with the reference values (reference perimeter and reference area), the accuracy of image processing and inspection becomes higher.
An X-ray inspection apparatus according to a fifth aspect is the X-ray inspection apparatus according to any one of the first to fourth aspects, wherein the inspection unit inspects the quantity of a plurality of articles.
Here, it is possible to perform an inspection in which the quantity of a plurality of articles contained in a package such as a box is measured or compared with a predetermined quantity.

An X-ray inspection apparatus according to a sixth aspect is the X-ray inspection apparatus according to the fifth aspect, wherein the plurality of articles are contents contained in a package. Then, the inspection unit obtains the position of the planar center of gravity of each of the plurality of contents on the image, compares the position of the center of gravity with the reference center of gravity position, and further inspects the arrangement of the plurality of contents in the package. To do.
Here, since the integration phenomenon of a plurality of articles (contents) on the image can be checked by comparing the peripheral length of the lump part with the reference peripheral length, the identification of the center of gravity of each content and the reference center of gravity position The arrangement of a plurality of contents in the package can be inspected.

An X-ray inspection apparatus according to a seventh aspect is the X-ray inspection apparatus according to the first aspect, wherein the inspection unit inspects an abnormality in the shape of the article.
Here, by comparing the peripheral length of the lump portion indicating the article or estimated as the article with the reference peripheral length, it is possible to inspect abnormalities in the shape of the article easily, at high speed, and with relatively high accuracy. It becomes like this.

An X-ray inspection apparatus according to claim 8 is the X-ray inspection apparatus according to claim 1, wherein the inspection unit compares the peripheral length of the lump portion with the reference peripheral length, Compare with reference area.
Here, since both the perimeter and area of the lump portion on the image are compared with the reference values (reference perimeter and reference area), the accuracy of image processing and inspection becomes higher.

The X-ray inspection apparatus according to claim 9 is the X-ray inspection apparatus according to any one of claims 1 to 8, wherein the inspection unit takes into account a difference in distance between adjacent neighboring pixels, and a lump portion Calculate the perimeter of.
Here, based on the arrangement of surrounding pixels, the perimeter of the block portion is calculated taking into account the difference in distance between adjacent surrounding pixels. For this reason, the circumference of the lump portion can be obtained with high accuracy.

An X-ray inspection apparatus according to a tenth aspect is the X-ray inspection apparatus according to any one of the first to ninth aspects, wherein the image is composed of a large number of pixels arranged in two dimensions vertically and horizontally. Then, the inspection unit calculates the perimeter of the lump portion by taking into account whether adjacent surrounding pixels are aligned vertically or horizontally or diagonally.
Here, since the pixels are arranged two-dimensionally in the vertical and horizontal directions, the distance between adjacent peripheral pixels differs depending on whether the adjacent peripheral pixels are arranged vertically or horizontally and obliquely. coming out. Compared to the case where adjacent surrounding pixels are arranged vertically or horizontally, when the adjacent surrounding pixels are arranged obliquely, the distance between the surrounding pixels is increased. Since the perimeter of the lump portion is calculated in consideration of the above, the calculated perimeter of the lump portion is highly accurate here.

An X-ray inspection apparatus according to an eleventh aspect is the X-ray inspection apparatus according to any one of the first to tenth aspects, further comprising an input unit for inputting a dimension of an actual peripheral length of the article.
Here, the user of the X-ray inspection apparatus has only to manually input the actual perimeter of the article as it is in the input unit, and it is necessary to convert the actual perimeter of the article into the number of surrounding pixels in the X-ray image. There is no. For example, if a scale is provided near the X-ray inspection apparatus, the actual perimeter of the article can be easily measured by the user of the X-ray inspection apparatus.

  In the present invention, in image processing or inspection, the inspection unit compares the peripheral length of the block portion on the image with the reference peripheral length. As described above, by comparing the perimeter of the lump portion indicating the article or estimated as the article with the reference perimeter, image processing and inspection can be performed easily, at high speed, and with relatively high accuracy. become. Thereby, in this X-ray inspection apparatus, the time required for the inspection of the article is shortened. In the present invention, the perimeter of the lump portion is calculated based on the arrangement of surrounding pixels that form the perimeter of the lump portion. Therefore, when the peripheral length of the lump portion is approximated simply by adding the number of surrounding pixels Compared to the above, the perimeter of the lump portion can be obtained with high accuracy. For this reason, in this X-ray inspection apparatus, the time required for the inspection of the article can be shortened, and the inspection accuracy can be maintained high.

An appearance of an X-ray inspection apparatus according to an embodiment of the present invention is shown in FIG. This X-ray inspection apparatus 10 is one of the apparatuses that perform quality inspection in the production line of products such as foods. The X-ray inspection apparatus 10 transmits X-rays to the products that are continuously conveyed and transmits the products. This is a device for judging the defect of a product based on the X-ray dose.
As shown in FIG. 4, the product G, which is a specimen of the X-ray inspection apparatus 10, is conveyed to the X-ray inspection apparatus 10 by the front conveyor 60. The product G is subjected to at least one inspection of foreign matter mixing inspection, shape inspection, and content quantity inspection in the X-ray inspection apparatus 10. The determination result in the X-ray inspection apparatus 10 is sent to a distribution mechanism 70 disposed on the downstream side of the X-ray inspection apparatus 10. The distribution mechanism 70 sends the product G to the regular line conveyor 80 when the product G is determined to be a non-defective product in the X-ray inspection device 10, and the product G is determined to be a defective product in the X-ray inspection device 10. In this case, the product G is distributed to the defective product storage conveyor 90.

<Configuration of X-ray inspection apparatus>
As shown in FIGS. 1 and 2, the X-ray inspection apparatus 10 mainly includes a shield box 11, a conveyor 12, an X-ray irradiator (X-ray source) 13, and an X-ray line sensor (X-ray light receiving unit). 14, a monitor 30 with a touch panel function, and a control computer 20 (see FIG. 5).

[Shield box]
The shield box 11 has an opening 11a for carrying in / out a product on both side surfaces. In this shield box 11, a conveyor 12, an X-ray irradiator 13, an X-ray line sensor 14, a control computer 20 and the like are accommodated.
Although not shown in FIG. 1, the opening 11 a is closed with a shielding nore for suppressing leakage of X-rays to the outside of the shielding box 11. This shielding nolen is formed from rubber containing lead, and is pushed away by the product when the product is carried in and out.

In addition to the monitor 30, a key insertion slot and a power switch are arranged on the upper front portion of the shield box 11.
〔Conveyor〕
The conveyor 12 conveys goods in the shield box 11, and is driven by a conveyor motor 12a shown in FIG. The conveyance speed by the conveyor 12 is finely controlled by the inverter control of the conveyor motor 12a by the control computer 20 so as to be the set speed input by the user.

[X-ray irradiator]
As shown in FIG. 2, the X-ray irradiator 13 is disposed above the conveyor 12 and radiates fan-shaped X-rays (see the hatched area X in FIG. 2) toward the lower X-ray line sensor 14. To do.
[X-ray line sensor]
The X-ray line sensor 14 is disposed below the conveyor 12 and detects X-rays that pass through the product G and the conveyor 12. As shown in FIG. 3, the X-ray line sensor 14 includes a large number of pixels 14 a that are horizontally arranged in a straight line in a direction orthogonal to the conveying direction by the conveyor 12.

[LCD monitor]
The monitor 30 is a full dot display liquid crystal display. The monitor 30 also has a touch panel function, and displays a screen that prompts input of parameters relating to initial settings and defect determination.
[Control computer]
In the CPU 21, the control computer 20 executes a foreign substance inspection routine 27, a shape inspection routine 28, a quantity inspection routine 29 and the like included in the control program. The control computer 20 also stores and accumulates images and inspection results used for inspection of defective products in a storage unit such as the HDD 25.

  As a specific configuration, as shown in FIG. 5, the control computer 20 includes a CPU 21, and a ROM 22, a RAM 23, and an HDD (hard disk) 25 as a main storage unit controlled by the CPU 21. The HDD 25 stores a reference value file 25a for storing reference values such as a reference perimeter and a reference area for each product, which will be described later, and an inspection result log file 25b for storing inspection images and inspection results. The control computer 20 also has an FDD (flexible disk drive) 24 that performs input / output with the flexible disk.

Further, the control computer 20 includes a display control circuit that controls data display on the monitor 30, a key input circuit that captures key input data from the touch panel of the monitor 30, and an I / O for controlling data printing in a printer (not shown). It has a port.
The CPU 21, ROM 22, RAM 23, FDD 24, HDD 25, and the like are connected to each other via a bus line such as an address bus or a data bus.

The control computer 20 is connected to a conveyor motor 12a, a rotary encoder 12b, a photoelectric sensor 15, an X-ray irradiator 13, an X-ray line sensor 14, and the like.
The rotary encoder 12b is mounted on the conveyor motor 12a, detects the conveying speed of the conveyor 12, and sends it to the control computer 20.
The photoelectric sensor 15 is a synchronous sensor for detecting the timing at which the product G as a specimen comes to the position of the X-ray line sensor 14, and is composed of a pair of light projectors and light receivers arranged with a conveyor interposed therebetween.

<Determination of product defects by control computer>
[Create X-ray image]
The control computer 20 receives a signal from the photoelectric sensor 15, and when the product G passes through the fan-shaped X-ray irradiation unit (see FIG. 2), an X-ray fluoroscopic image signal (see FIG. 3) by the X-ray line sensor 14. ) At fine time intervals, and an X-ray image of the product G is created based on these X-ray fluoroscopic image signals. That is, data at each time is obtained from each pixel 101 of the X-ray line sensor 14 with a fine time interval, and a two-dimensional image is created from these data. This two-dimensional image is, for example, an image 61 shown in FIG.

[Foreign matter contamination inspection]
The foreign matter inspection routine 27 executed by the CPU 21 of the control computer 20 performs image processing on the X-ray image obtained as described above, and determines whether the product is good or bad (no foreign matter is mixed) by a plurality of judgment methods. Please). Examples of the determination method include a trace detection method, a binarization detection method, and a mask binarization detection method. As a result of the determination by these determination methods, if even one item is determined to be defective, the product G is determined to be defective.

In the trace detection method and the binarization detection method, a determination is made on an unmasked area of an image. On the other hand, in the mask binarization method, a determination is made on a masked area of an image. The mask is set for the container portion of the product G and the like.
In the trace detection method, a reference level (threshold value) is set along the rough thickness of the object to be detected, and it is determined that foreign matter is mixed in the product G when the image becomes darker than that. It is a method. In this method, a relatively small foreign object can be detected.

The binarization detection method and the mask binarization method are methods in which a reference level is set at a constant brightness, and it is determined that foreign matter is mixed in the product G when the image becomes darker than that. . This binarization detection method is set to detect a relatively large foreign object.
The reference level and mask area in each determination method are set and changed by an input from the user using the touch panel function of the monitor 30.

[Quantity inspection]
The quantity inspection routine 29 executed by the CPU 21 of the control computer 20 performs the following image processing on the obtained X-ray image, and then the quantity of the contents for a product in which a plurality of contents are contained in a box. To check whether the product is normal or not.
First, the quantity inspection routine 29 divides the X-ray image 61 (FIG. 6A) obtained from the X-ray fluoroscopic image signal from the X-ray line sensor 14 into a background and a non-background by binarization processing. An image 62 shown in (b) is created. The black portion in the image 62 is the background. Here, the contents appear white, and the package box also appears white.

Next, labeling is performed on the white block portion surrounded by the background in the image 62. In this labeling, as shown in FIG. 7, group numbers are sequentially assigned to each white region (each block portion) as group G1, group G2, group G3,..., Group Gn.
Next, the quantity inspection routine 29 calculates the perimeter and area of the region for each group. Here, a method of calculating the peripheral length and area of the area of the group will be described by taking one white lump portion (one group) in the binarized X-ray image shown in FIG. 16 as an example. Note that the group shown in FIG. 16 is a sample for explaining the calculation method of the perimeter and area, and does not coincide with any of the groups shown in FIG. 7 (as compared to each group shown in FIG. The group shown in Figure 2 has a small perimeter and area.)

The area of the group region is calculated by multiplying the total number of white pixels constituting the group by the area (unit area) of one pixel. In the case of the group shown in FIG. 16, the area of the group of 4.64 square millimeters is obtained by multiplying 0.16 square millimeters, which is the area of one pixel corresponding to 0.4 millimeter square, by 29 of the number of pixels. Calculated.
The perimeter of the group area is calculated based on the arrangement of surrounding pixels that form the perimeter of the area. Specifically, considering the difference in distance between adjacent surrounding pixels, whether the adjacent surrounding pixels are aligned vertically or horizontally, or whether they are aligned diagonally, The perimeter is calculated. A method of calculating the perimeter of this group area will be described in detail with reference to FIGS. First, the upper left pixel among the white pixels constituting the group is set as a target pixel, and its coordinates are stored. When coordinates are set as shown in FIG. 17 for two-dimensionally arranged pixels in the vertical and horizontal directions, the pixel of interest at the upper left is the pixel at the coordinates of “Ki 24”. Next, a pixel adjacent to this pixel of interest is examined counterclockwise (see arrow A101 in FIG. 17), and a pixel that changes from black to white is specified. Here, the pixel changes from black to white when moving from the pixel at the coordinates of “Ki 23” to the pixel at the coordinates of “K 23”, of which the white pixel (the latter is at the coordinates of “K 23”). Pixel) is specified. This identified pixel becomes the second pixel of interest. Next, the pixel adjacent to the pixel at the coordinates of “K23”, which is the second target pixel, is examined counterclockwise (see arrow A102 in FIG. 18), and the pixel that similarly changes from black to white is specified. To do. If this is repeated, the target pixel returns to the first target pixel. At this time, as shown in FIG. 19, the pixel group (“K 24”, “K 23”, “K 22”, “K 22”, “S 23”, “S 24”, “ 25 ”,“ S26 ”,“ K27 ”,“ K28 ”,“ K28 ”,“ Ki27 ”,“ Ki26 ”,“ Ki25 ”pixels) around the group area It coincides with all the surrounding pixels that constitute. When the surrounding pixels of the group area are specified in this way, the quantity inspection routine 29 is inclined with the number of times M1 moved up and down (vertical) or left and right (horizontal) while making a round around the group area. The perimeter length SL of the group area is calculated from the number M2 of movements and the arrangement pitch PL of the vertical and horizontal pixels by the following formula.

SL = PL × (M1 + M2 × √2)
In this way, by multiplying the amount of movement by √2, it can be expected that the peripheral length SL becomes a value close to the actual peripheral length as compared with the case where the number of movements is simply integrated. Here, the number of times M1 moved up and down or left and right is 6 as shown by arrows A3, A6, A10, A12, A13, and A14 in FIG. 20, and the number of times M2 moved obliquely is the arrows A1, A2, and A2 in FIG. Since it is 8 as shown by A4, A5, A7, A8, A9, A11,
SL = 0.4 mm × (6 + 8 × √2) = 0.4 mm × 17.3 = 6.9 mm
It becomes.

When the perimeter and area are calculated for each group by the above calculation method, for example, a result as shown in FIG. 8 is obtained.
Next, the perimeter and area of each group are compared to the reference perimeter and reference area. The reference perimeter and the reference area are set in advance for each product requiring quantity inspection, and values manually input from the monitor 30 with a touch panel function at the time of initial setting are stored in the reference value file 25a of the HDD 25. The manually input reference perimeter and reference area are actual dimensions of one content among commodities composed of a plurality of contents contained in a box. The quantity inspection routine 29 calls the reference perimeter length and reference area of the contents of the target product from the reference value file 25a to the RAM 23, and compares them with the perimeter length and area of each white region to which the group number is assigned. Here, the reference peripheral length is set in a range of 150 millimeters to 250 millimeters and the reference area is set in a range of 2000 square millimeters to 2500 square millimeters. Therefore, as can be seen with reference to FIG. 8, the perimeter and area of the group G1 and the group Gn which are considered to be part of the package box are out of the range of the reference perimeter and reference area, and the two contents The perimeter and area of the group G4 where objects are considered to be integrated on the image by contact are out of the range of the reference perimeter and reference area.

  Next, among groups (here, groups G1, G4, and Gn) in which the perimeter and area are out of the range of the reference perimeter and reference area, the perimeter and area are larger than the reference perimeter and reference area. The separation process by filtering is performed. That is, here, the separation processing is performed on the group G4. When the separation processing is performed on the group G4 in which the two contents are integrated on the image as shown in FIG. 9A, the two contents are separated as shown in FIG. 9B. . In this separation process, as shown in FIG. 10, at each point in the white area of the group G4 surrounded by the background (in FIG. 10, it is shown by hatching instead of black for ease of viewing), the point is When a circle with a predetermined radius at the center is drawn, the number of times the boundary between the background region and the group G4 is passed is counted, and a point where the number is 3 or more is changed to the background (changed to black). For example, at the point P1 in FIG. 10, the circle C1 drawn passes the above boundary four times, at the point P2, the circle C2 drawn passes the above boundary twice, and at the point P3 the circle C3 drawn passes the above boundary. Since the number of times is 0, the point P1 is changed from white to black, and the points P2 and P3 are not changed. By performing such processing on the entire white area of the group G4, the group G4 is separated into two groups G4a and G4b as shown in FIG. 9B.

  As described above, when the separation process by the filtering is completed, the group configuration illustrated in FIG. 7 is changed to the group configuration illustrated in FIG. Then, when the perimeter and the area are compared with the reference perimeter and the reference area for each group shown in FIG. 11, a group out of the range of the reference perimeter and the reference area is a group G1 that is considered to be a part of the package box. Focused on group Gn. Here, if only groups whose perimeter and area are in the range of the reference perimeter and reference area are extracted, the image 63 shown in FIG. 12 is obtained, and the group whose perimeter and area are not in the range of the reference perimeter and reference area. If only this is extracted, an image 64 shown in FIG. 13 is obtained. The image 63 is an image obtained by extracting only the contents to be counted, and the image 64 is an image obtained by extracting only the package portion.

Based on the image 63 of FIG. 12 obtained by the image processing as described above, the quantity inspection routine 29 counts the quantity of contents and determines whether or not it is normal.
The control computer 20 can also inspect the package based on the image 64 of FIG. 13 showing only the package portion.
[Shape inspection]
The shape inspection routine 28 executed by the CPU 21 of the control computer 20 determines whether or not the outer shape of the product is normal, and distributes defective products having cracks to the defective product storage conveyor 90. For example, it is inspected whether the product G, which is a round rice cracker as shown in FIG. 14 (a), has a chip E as shown in FIG. 14 (b).

Here, as in the quantity inspection, first, the X-ray image obtained from the X-ray fluoroscopic image signal by the X-ray line sensor 14 is divided into a background and a non-background by binarization processing.
Next, the perimeter and area of the region indicating the product surrounded by the background are calculated, and the perimeter and area are compared with the reference perimeter and reference area. Here, in the case of the normal product G shown in FIG. 14A, the perimeter and the area coincide with the reference perimeter and the range of the reference area, but the product Ge having the chip E shown in FIG. For example, the perimeter and area are out of the range of the reference perimeter and reference area.

Then, the shape inspection routine 28 determines that a product whose perimeter and area are out of the range of the reference perimeter and reference area is a defective product having a crack or a chip, and sends a sorting instruction to the sorting mechanism 70.
〔Display control〕
The control computer 20 causes the monitor 30 to display the X-ray image of the obtained product G and information related to the determination by each determination method when a normal inspection is performed. Further, at the time of initial setting or testing, the monitor 30 is displayed on the monitor 30 for specifying the product and inputting the inspection parameters including the reference perimeter and reference area.

<Main features of X-ray inspection equipment>
(1)
The inspection in the X-ray inspection apparatus 10 has the following advantages compared to the prior art.
[Conventional X-ray inspection equipment]
A conventional X-ray inspection apparatus employs a count method based on pattern matching, and compares the density patterns of each part of the reference pattern image and the captured pixels. Specifically, first, a similarity is determined by comparing a reference image narrower than an image in the entire region or a predetermined region with individual images. Then, the number of objects that are similar to or match the reference image in the region is counted.

However, when the individual contents have directionality, it is necessary to match the directionality with the reference image. Furthermore, when matching the shapes, complicated processing (such as rotation of the reference image) for determining similarity is required.
Then, since all the quantities cannot be known unless all the pattern matchings are completed, it takes a long time in the quantity inspection for counting the quantities of a plurality of contents.

In addition, the shape of the package or the like is often excluded from the inspection target as a non-processing target. In the case of an inspection target, the same complicated pattern matching process as described above is required.
[X-ray inspection apparatus of this embodiment]
In the X-ray inspection apparatus 10, the control computer 20 uses the peripheral length of the lump portion (white area other than the background after binarization) on the image as the reference peripheral length in the image processing in the shape inspection routine 28 and the quantity inspection routine 29. Compare with In this way, by comparing the perimeter of the lump portion indicating the contents in the product or the product or estimated to be the content with the reference perimeter, a plurality of contents can be obtained simply, at high speed and with relatively high accuracy. It can count the quantity of goods and inspect the shape of products. Thereby, in this X-ray inspection apparatus 10, the time required for the inspection of the product is considerably shortened compared with the conventional case.

In addition, the X-ray inspection apparatus 10 specifies not only the comparison between the perimeter of the lump portion on the image and the reference perimeter, but also the comparison of the area of the lump portion on the image with the reference area to identify the contents. Therefore, sufficient accuracy of quantity inspection and shape inspection is ensured.
Since quantity inspection and shape inspection are performed by such a method, in the X-ray inspection apparatus 10, the time until the inspection result is obtained is shortened compared to the conventional one.

  Further, the X-ray inspection apparatus 10 calculates the perimeter of the lump portion based on the arrangement of surrounding pixels that form the periphery of the lump portion. More specifically, in consideration of whether adjacent neighboring pixels are arranged vertically, horizontally, or diagonally so as to take into account the difference in distance between neighboring neighboring pixels, The perimeter of the part is calculated. For this reason, compared with the case where the number of surrounding pixels is simply added to approximate the circumference of the chunk, the circumference of the chunk can be determined with high accuracy.

(2)
When performing quantity inspection by performing image processing on an X-ray image of a product including a plurality of contents, for example, when two adjacent contents are close enough to contact each other or two adjacent contents When an object is actually in contact, two contents may be recognized as one on the image as shown in FIGS. 7 and 8A. If this happens, there will be problems such as counting the wrong quantity in the quantity inspection of the contents.

  On the other hand, in the X-ray inspection apparatus 10, in the quantity inspection routine 29, after distinguishing the lump portion and the background portion by binarization processing, the perimeter length and area of the lump portion (white area portion) are set as the reference perimeter length and the reference length. The integration phenomenon of a plurality of contents on the image is checked by comparing with the area. Therefore, the problem that the same image processing and inspection as those in the case where there is no integration phenomenon despite the occurrence of the integration phenomenon is continued and an erroneous inspection result is output is avoided. Specifically, when an integration phenomenon of a plurality of contents is found, a process of separating the plurality of integrated contents on the image is performed (see FIG. 9).

<Modification>
(A)
In the above embodiment, in the quantity inspection routine 29, based on the image of FIG. 11 showing the group structure after the separation process by filtering, only the contents of the image 63 (see FIG. 12) extracted from the contents to be counted or only the package portion. An image 64 (see FIG. 13) is extracted, and the quantity of the contents is counted and the package is inspected.

  In addition to this, based on the image of FIG. 11, the position of the planar center of gravity of each of the plurality of contents can be obtained on the image, and the arrangement of the plurality of contents in the package can be further inspected. A centroid position (position indicated by x in FIG. 15) is calculated for the regions of the groups G2, G3, G4a, G4b, G5,. Can be inspected with respect to the arrangement of a plurality of contents in the package.

(B)
In the above-described embodiment, a configuration is adopted in which the reference perimeter of the contents of the product is manually input by the user at the time of initial setting from the monitor 30 with a touch panel function. In this case, it is desirable to provide a scale near the monitor 30 of the X-ray inspection apparatus 10 so that the user can easily measure the actual perimeter length of the contents. Since the monitor 30 serves as an input unit and is configured to receive the actual content dimensions as the reference perimeter of the product contents, the user converts the actual dimensions of the contents measured on the scale. Can be input without any change.

In the above embodiment, the reference perimeter has a predetermined range, but the actual contents input by the user can be obtained without having the user input the upper limit value and the lower limit value, for example. A predetermined range of + 10% to −10% can be automatically calculated by the control computer 20 based on the actual perimeter.
(C)
In the above embodiment, the configuration is such that the reference perimeter of the contents of the product is manually input by the user at the time of initial setting from the monitor 30 with a touch panel function. It is also possible.

  For example, the content of the product may be photographed with a CCD camera and the reference perimeter may be automatically calculated by image processing, or the product content that is a good product with no cracks may be flowed to the X-ray inspection apparatus 10 in advance. The reference perimeter may be automatically calculated from the obtained X-ray image by image processing. Then, if an arithmetic expression is prepared in the control computer 20 so that, for example, the range of the reference peripheral length is + 10% to −10% with respect to the measurement peripheral length, the trouble of manual input by the user is reduced. It can be omitted.

(D)
In the above embodiment, the quantity inspection routine 29 creates the image 62 shown in FIG. 6B by dividing the X-ray image 61 (FIG. 6A) into a background and a non-background by binarization processing, When the group numbers are sequentially assigned to the white lump portions and the perimeter of each group is calculated, the perimeter of the group area is configured by the method described above with reference to FIGS. The surrounding pixels are specified.

  The surrounding pixels may be specified by other methods instead of such a surrounding pixel specifying method. For example, a general method of specifying an edge (peripheral pixels) by differentiation may be used. In this case, only the pixels that are the edge portions of the group are specified, and for example, only the pixels of the edge portions are extracted white on the image.

  The X-ray inspection apparatus according to the present invention has features that it can shorten the time required for inspection of an article and can maintain high inspection accuracy, and is useful as an inspection apparatus for an article.

1 is an external perspective view of an X-ray inspection apparatus according to an embodiment of the present invention. The simple block diagram inside the shield box of a X-ray inspection apparatus. The schematic diagram which shows the principle of a X-ray inspection. The figure which shows the structure before and behind an X-ray inspection apparatus. The block block diagram of a control computer. (A) The figure which shows the X-ray image based on the X-ray fluoroscopic image signal by an X-ray line sensor.

(B) The figure which shows the image after a binarization process.
The figure which shows the labeling in the image after a binarization process. The figure which shows the result of having calculated the perimeter and the area with respect to each labeled group. (A) The figure before the separation process of the group in which the two contents are integrated on the image.

(B) The figure after the separation process of the group in which two contents are integrated on the image.
The figure which shows the separation | separation processing method of the group in which two contents are integrated on the image. The figure which shows the labeling in the image after a separation process. The figure which shows the image which extracted only the content which is a count object. The figure which shows the image which extracted only the part of the package. (A) The figure which shows the goods which are not missing.

(B) The figure which shows the goods with a chip.
The figure which shows the labeling and center-of-gravity position display in the image after a separation process. The figure for demonstrating the calculation method of the perimeter for each group. The figure for demonstrating the calculation method of the perimeter for each group. The figure for demonstrating the calculation method of the perimeter for each group. The figure for demonstrating the calculation method of the perimeter for each group. The figure for demonstrating the calculation method of the perimeter for each group.

Explanation of symbols

10 X-ray inspection equipment 13 X-ray irradiator (X-ray source)
14 X-ray line sensor (X-ray detector)
20 Control computer (inspection department)
25a Reference value file 28 Shape inspection routine 29 Quantity inspection routine G1, G2, G3, G4,..., Gn group

Claims (11)

An X-ray source that emits X-rays;
An X-ray light receiving unit for receiving X-rays from the X-ray source;
An inspection unit that performs image processing on an X-ray image received by the X-ray light receiving unit and inspects an article that passes between the X-ray source and the X-ray light receiving unit;
With
In the image processing or the inspection, the inspection unit calculates the perimeter of the block portion on the image based on the arrangement of surrounding pixels that form the periphery of the block portion, and compares it with a reference peripheral length.
X-ray inspection equipment. The inspection unit distinguishes the lump portion and the background portion outside the lump portion from the image, and then compares the perimeter of the lump portion with the reference perimeter, thereby contacting a plurality of the articles. Checking the integration phenomenon of the plurality of articles on the image;
The X-ray inspection apparatus according to claim 1. The inspection unit, after checking the integration phenomenon, performs image processing for separating the lump portion in which the integration phenomenon is confirmed into a plurality of the articles.
The X-ray inspection apparatus according to claim 2. In addition to comparing the peripheral length of the lump portion with the reference peripheral length, the inspection unit compares the area of the lump portion with a reference area, thereby integrating the plurality of articles on the image. Check
The X-ray inspection apparatus according to claim 2 or 3. The inspection unit inspects the quantity of the plurality of articles;
The X-ray inspection apparatus according to claim 1. The plurality of articles are contents contained in a package,
The inspection unit obtains a position of a planar center of gravity of each of the plurality of contents on the image, compares the position of the center of gravity with a reference center of gravity position, and arranges the plurality of contents in the package Inspect further,
The X-ray inspection apparatus according to claim 5. The inspection unit inspects the shape of the article for abnormality;
The X-ray inspection apparatus according to claim 1. In addition to comparing the perimeter of the lump portion with the reference perimeter, the inspection unit compares the area of the lump portion with a reference area.
The X-ray inspection apparatus according to claim 1. The inspection unit calculates the perimeter of the lump portion, taking into account the difference in distance between the adjacent surrounding pixels,
The X-ray inspection apparatus according to claim 1. The image is composed of a large number of pixels arranged in two dimensions, vertically and horizontally,
The inspection unit calculates the perimeter of the lump portion by taking into account whether the adjacent surrounding pixels are aligned vertically or horizontally or diagonally.
The X-ray inspection apparatus according to claim 1. An input unit for inputting a dimension of an actual perimeter of the article;
The X-ray inspection apparatus according to claim 1.

Images