JP7565247B2 - Area determination method, tablet inspection method, and tablet inspection device - Google Patents

Description

The present invention relates to a method for determining an area in a captured image of a tablet, a tablet inspection method for inspecting tablets, and a tablet inspection device for inspecting tablets.

Conventionally, tablet printing devices are known that use an inkjet system to print images such as letters and marks on the surfaces of tablets such as medicines and supplements. The tablet printing device prints images on the surfaces of tablets by ejecting ink droplets toward the tablets while transporting the tablets on a conveyor or the like. The tablet printing device also photographs the tablets before or after printing, and performs an appearance inspection of the tablets based on the photographed images.

A conventional tablet printing device is described, for example, in Patent Document 1.

JP 2019-124519 A

In visual inspection of tablets, the top and side surfaces of the tablet may be inspected using different parameters. Also, there may be cases where only the top surface of the tablet is inspected. For this reason, it is necessary to distinguish between the area corresponding to the top surface of the tablet and the area corresponding to the side surface of the tablet in the captured image of the tablet. In the above-mentioned Patent Document 1, the boundary between the area corresponding to the top surface of the tablet and the area corresponding to the side surface of the tablet is identified by detecting edges. However, when edge detection is used, there is a problem in that it is difficult to adjust the threshold value.

The present invention has been developed in consideration of these circumstances, and aims to provide a technology that can determine the top surface area of a tablet contained in a captured image without relying on edge detection.

In order to solve the above problem, the first invention of the present application is a method for determining an area in a photographed image of a tablet having an upper surface and a side surface, comprising: a) a step of classifying each pixel of the photographed image into three areas, a first upper surface area corresponding to the upper surface, a first side surface area corresponding to the side surface, and a first background area corresponding to the background, by clustering based on brightness values; b) a step of generating a first overall area corresponding to the entire tablet by expanding the first upper surface area toward the first side surface area after the step a); c) a step of determining a second overall area corresponding to the entire tablet by a region segmentation method that determines a boundary line with a large brightness gradient using the first overall area as a teaching label after the step b); d) a step of generating a second upper surface area corresponding to the upper surface by shrinking the second overall area toward the first upper surface area after the step c); and e) a step of determining a third upper surface area corresponding to the upper surface using a region segmentation method that determines a boundary line with a large brightness gradient using the second upper surface area as a teaching label after the step d).

The second invention of the present application is the area determination method of the first invention, further comprising the step of: f) after step e), determining a second side area corresponding to the side by taking the exclusive OR of the second overall area and the third top surface area.

The third invention of the present application is the region determination method of the first or second invention, in which in step a), each pixel of the captured image is classified into the first top surface region, the first side surface region, and the first background region by k-means clustering.

The fourth invention of the present application is a region determination method according to any one of the first to third inventions, in which in the steps c) and e), a boundary line C is found that minimizes the energy E expressed by the formula (1) described below.

The fifth invention of the present application is a tablet inspection method for inspecting tablets, comprising: A) a step of acquiring a photographed image of the tablet; B) a step of determining an inspection area in the photographed image by the area determination method according to any one of claims 1 to 4; and C) a step of inspecting the inspection area in the photographed image.

The sixth invention of the present application is a tablet inspection device for inspecting tablets having a top surface and a side surface, comprising an imaging unit for acquiring a photographed image of the tablet, an area determination unit for determining an inspection area in the photographed image, and an inspection unit for inspecting the inspection area of the photographed image, the area determination unit performing the steps of: a) classifying each pixel of the photographed image into three areas, a first top surface area corresponding to the top surface, a first side surface area corresponding to the side surface, and a first background area corresponding to the background, by clustering based on brightness values; and b) after step a), expanding the first top surface area toward the first side surface area to obtain a first background area corresponding to the entire tablet. A process for generating an entire region; c) after the process b), a process for determining a second entire region corresponding to the entire tablet using a region segmentation method that uses the first entire region as an instruction label to find a boundary line where the luminance gradient increases; d) after the process c), a process for generating a second upper surface region corresponding to the upper surface by shrinking the second entire region toward the first upper surface region; and e) after the process d), a process for determining a third upper surface region corresponding to the upper surface using a region segmentation method that uses the second upper surface region as an instruction label to find a boundary line where the luminance gradient increases. The third upper surface region is set as the inspection region.

According to the first to sixth aspects of the present application, the top surface area of the tablet contained in the captured image can be determined without relying on edge detection. In addition, the top surface area of the tablet can be determined more accurately than when it is determined by clustering alone.

In particular, according to the second invention of the present application, the side area of the tablet contained in the captured image can be determined without relying on edge detection. Furthermore, the side area of the tablet can be determined more accurately than when it is determined by clustering alone.

FIG. 1 is a diagram showing a configuration of a tablet printing device. FIG. FIG. FIG. 2 is a block diagram showing electrical connections between a computer and each part of the tablet printing device. FIG. 1 is a block diagram conceptually showing the functions of a computer related to visual inspection of tablets. 13 is a flowchart showing a procedure of an area determination process. FIG. 13 is a diagram showing an example of a captured image. 1 is a histogram showing the distribution of luminance values of a plurality of pixels constituting a captured image. FIG. 11 is a diagram showing the processing state of step S1. FIG. 11 is a diagram showing the processing state of step S2. FIG. 11 is a diagram showing the processing state of step S3. FIG. 11 is a diagram showing the processing state of step S4. FIG. 11 is a diagram showing the processing state of step S5. FIG. 11 is a diagram showing the processing state of step S6.

The following describes an embodiment of the present invention with reference to the drawings. In the following description, the direction in which the tablets are transported is referred to as the "transport direction," and the direction perpendicular and horizontal to the transport direction is referred to as the "width direction."

1. Overall configuration of the printing device
1 is a diagram showing the configuration of a tablet printing device 1 including a tablet inspection device according to an embodiment of the present invention. This tablet printing device 1 is a device that prints images such as product name, product code, weight, ingredient content, company name, logo mark, etc. on the surface of each tablet 9 while conveying a plurality of tablets 9. As shown in FIG. 1, the tablet printing device 1 of this embodiment includes a conveying mechanism 10, a first imaging unit 20, a printing unit 30, a second imaging unit 40, a defective product ejection unit 50, a non-defective product ejection unit 60, and a computer 70.

The conveying mechanism 10 is a mechanism that conveys multiple tablets 9 while holding them in an aligned state. The conveying mechanism 10 has a pair of pulleys 11 and a circular conveying belt 12 that is stretched between the pair of pulleys 11. Multiple tablets 9 fed into the tablet printing device 1 are aligned at equal intervals and supplied to the surface of the conveying belt 12 by a feed mechanism 80 that is composed of a vibrating feeder, a conveying drum, etc. One of the pair of pulleys 11 is rotated by power obtained from a conveying motor 13. As a result, the conveying belt 12 rotates in the direction of the arrow in FIG. 1. At this time, the other of the pair of pulleys 11 is rotated in conjunction with the rotation of the conveying belt 12.

Figure 2 is a partial perspective view of the conveying mechanism 10. As shown in Figure 2, the conveying belt 12 is provided with a plurality of suction holes 14. The plurality of suction holes 14 are arranged at equal intervals in the conveying direction and the width direction. Also, as shown in Figure 1, the conveying mechanism 10 has a suction mechanism 15 that sucks out gas from the space inside the conveying belt 12. When the suction mechanism 15 is operated, the space inside the conveying belt 12 becomes negative pressure, which is lower than atmospheric pressure. The plurality of tablets 9 are adsorbed and held by the suction holes 14 due to this negative pressure.

In this way, the tablets 9 are held on the surface of the conveyor belt 12 in a state aligned in the conveying direction and width direction. The conveyor mechanism 10 then rotates the conveyor belt 12 to convey the tablets 9 in the conveying direction. Below the four heads 31 described below, the tablets 9 are conveyed horizontally.

The first imaging unit 20 is a part that captures images of the tablets 9 before printing. The first imaging unit 20 captures images of the tablets 9 transported by the transport belt 12 at a first imaging position P1 that is downstream of the transport path from the loading mechanism 80 and upstream of the transport path from a printing position P2 described below. The first imaging unit 20 is, for example, a line sensor in which imaging elements such as CCD or CMOS are arranged in the width direction. However, the first imaging unit 20 may also be another imaging device such as an area sensor.

The tablets 9 have circular upper and lower surfaces and cylindrical side surfaces. The lower surfaces of the tablets 9 are attracted to the suction holes 14 of the conveyor belt 12. The first imaging unit 20 captures images of the tablets 9 from diagonally above. Therefore, the images of the tablets 9 captured by the first imaging unit 20 include the upper and side surfaces of the tablets 9.

The captured image is transmitted from the first imaging unit 20 to the computer 70. Based on the captured image transmitted from the first imaging unit 20, the computer 70 checks whether or not there is a tablet 9 in each suction hole 14, the position of the tablet 9, the rotation angle of the tablet 9, and whether or not there are any defects such as chips in the tablet 9.

The printing unit 30 is a processing unit that prints an image on the surface of the tablet 9 by an inkjet method at a printing position P2 that is downstream of the first imaging position P1 on the conveying path and upstream of the second imaging position P3 described later on the conveying path. As shown in FIG. 1, the printing unit 30 of this embodiment has four heads 31. The four heads 31 are located above the conveying belt 12 and arranged in a row along the conveying direction of the tablet 9. The four heads 31 eject ink droplets of different colors (e.g., cyan, magenta, yellow, and black) toward the surface of the tablet 9. Then, a multi-color image is printed on the surface of the tablet 9 by superimposing the monochromatic images formed by each of these colors. Note that the ink ejected from each head 31 is edible ink manufactured from raw materials approved by the Japanese Pharmacopoeia, the Food Sanitation Act, etc.

Figure 3 is a bottom view of one head 31. In Figure 3, the conveyor belt 12 and multiple tablets 9 held by the conveyor belt 12 are indicated by two-dot chain lines. As shown enlarged in Figure 3, the discharge surface 310, which is the bottom surface of the head 31, is provided with multiple nozzles 311 capable of discharging ink droplets. In this embodiment, the multiple nozzles 311 are arranged two-dimensionally in the conveying direction and the width direction on the bottom surface of the head 31. Each nozzle 311 is arranged with its position shifted in the width direction. In this way, by arranging the multiple nozzles 311 two-dimensionally, the positions of each nozzle 311 in the width direction can be made close to each other. However, the multiple nozzles 311 may also be arranged in a line along the width direction.

The method of ejecting ink droplets from the nozzle 311 is, for example, the so-called piezo method, in which a voltage is applied to a piezo element to deform it, thereby pressurizing the ink in the nozzle 311 and ejecting it. However, the ink droplets may also be ejected using the so-called thermal method, in which electricity is passed through a heater to heat and expand the ink in the nozzle 311, thereby ejecting it.

The second imaging unit 40 is a part that captures images of the tablets 9 after printing. The second imaging unit 40 captures images of the tablets 9 transported by the transport belt 12 at a second imaging position P3 that is downstream of the printing position P2 on the transport path and upstream of the defective product discharge position P4 described below on the transport path. The second imaging unit 40 is, for example, a line sensor in which imaging elements such as CCD or CMOS are arranged in the width direction. However, the second imaging unit 40 may also be another imaging device such as an area sensor.

The second imaging unit 40 photographs the tablets 9 from diagonally above. Therefore, the photographed image of the tablets 9 obtained by the second imaging unit 40 shows the top and side surfaces of the tablets 9. The photographed image is transmitted from the second imaging unit 40 to the computer 70. The computer 70 inspects the quality of the printed image formed on the surface of the tablets 9 based on the photographed image transmitted from the second imaging unit 40.

The defective product discharge section 50 is a section that discharges tablets 9 determined to be defective at a defective product discharge position P4 downstream of the second imaging position P3 and upstream of the non-defective product discharge position P5 described later. The defective product discharge section 50 has a defective product blow mechanism 51 arranged inside the conveyor belt 12 and a defective product collection section 52 arranged below the conveyor belt 12 from the defective product blow mechanism 51. The defective product blow mechanism 51 blows gas only to the suction hole 14 that holds the tablet 9 determined to be defective among the multiple suction holes 14 of the conveyor belt 12. Then, the suction hole 14 becomes a positive pressure higher than atmospheric pressure. As a result, the suction of the tablet 9 at the suction hole 14 is released, and the tablet 9 determined to be defective falls from the conveyor belt 12 to the defective product collection section 52.

The good product discharge section 60 is a section that discharges tablets 9 that have been determined to be good products at a good product discharge position P5 downstream of the defective product discharge position P4 on the conveying path. The good product discharge section 60 has a good product blow mechanism 61 that is arranged inside the conveying belt 12, and a good product recovery section 62 that is arranged below the good product blow mechanism 61 across the conveying belt 12. The good product blow mechanism 61 blows gas toward the multiple suction holes 14 of the conveying belt 12. This causes the suction holes 14 to be at a positive pressure higher than atmospheric pressure. This releases the suction of the tablets 9 at the suction holes 14, and the tablets 9 that have been determined to be good products fall from the conveying belt 12 to the good product recovery section 62.

<2. About Computers>
The computer 70 is an information processing device that controls the operation of each part in the tablet printing device 1 and inspects the appearance of the tablet 9 based on the captured image. FIG. 4 is a block diagram showing the electrical connection between the computer 70 and each part of the tablet printing device 1. As conceptually shown in FIG. 4, the computer 70 has a processor 701 such as a CPU, a memory 702 such as a RAM, and a storage unit 703 such as a hard disk drive. A computer program CP for executing the printing process and the appearance inspection is stored in the storage unit 703. The captured images of the tablet 9 transmitted from the first imaging unit 20 and the second imaging unit 40 are divided for each tablet 9 and stored in the storage unit 703.

The computer 70 is connected to the above-mentioned transport mechanism 10, the first imaging unit 20, the four heads 31, the second imaging unit 40, the defective product blowing mechanism 51, the non-defective product blowing mechanism 61, and the loading mechanism 80 via wired or wireless communication. The computer 70 controls the operation of each of these units according to the computer program CP.

Figure 5 is a block diagram conceptually illustrating the functions of computer 70 related to visual inspection of tablet 9. As shown in Figure 5, computer 70 has a first area determination unit 71, a first inspection unit 72, a print instruction unit 73, a second area determination unit 74, a second inspection unit 75, and a discharge instruction unit 76. Each function of first area determination unit 71, first inspection unit 72, print instruction unit 73, second area determination unit 74, second inspection unit 75, and discharge instruction unit 76 is realized by the processor 701 of computer 70 operating in accordance with computer program CP.

The first area determination unit 71 determines the area to be inspected from the captured image (hereinafter referred to as the "first captured image I1") transmitted from the first imaging unit 20. In this embodiment, the top and side surfaces of the tablet 9 are inspected before printing. Therefore, the first area determination unit 71 determines the area of the first captured image I1 that corresponds to the top surface of the tablet 9 and the area that corresponds to the side surface of the tablet 9 as the inspection area. The first area determination unit 71 then outputs first inspection area information D1 indicating the determined inspection area to the first inspection unit 72. A detailed method of area determination will be described later.

The first inspection unit 72 inspects the tablets 9 before printing based on the first captured image I1. Specifically, the first inspection unit 72 inspects the presence or absence of tablets 9 in each suction hole 14, the position of the tablets 9, the rotation angle of the tablets 9, and whether or not the tablets 9 have defects such as chips, based on the first captured image I1. At this time, the first inspection unit 72 inspects the presence or absence of defects in the inspection area specified by the first inspection area information D1 in the first captured image I1. The first inspection unit 72 then outputs the inspection result D2 to the print instruction unit 73 and the discharge instruction unit 76.

The print instruction unit 73 controls the operation of the printing unit 30. Based on the inspection results D2 of the first inspection unit 72, the print instruction unit 73 determines whether printing can be performed on the tablet 9, the position to print, and the angle of printing. Based on the determination result, the print instruction unit 73 then outputs print instructions D3 to the multiple heads 31. The multiple heads 31 eject ink droplets from nozzles 311 specified by the print instruction D3 at the specified timing. This causes an image to be printed on the top surface of the tablet 9.

The second area determination unit 74 determines the area to be inspected from the captured image (hereinafter referred to as the "second captured image I2") transmitted from the second imaging unit 40. In this embodiment, the top and side surfaces of the tablet 9 are inspected after printing. Therefore, the second area determination unit 74 determines the area of the second captured image I2 that corresponds to the top surface of the tablet 9 and the area that corresponds to the side surface of the tablet 9 as the inspection area. The second area determination unit 74 then outputs second inspection area information D4 indicating the determined inspection area to the second inspection unit 75. A detailed method of area determination will be described later.

The second inspection unit 75 inspects the tablet 9 after printing based on the second captured image I2. Specifically, the second inspection unit 75 inspects whether the printed image formed on the top surface of the tablet 9 is normal and whether there are any new defects on the tablet 9 based on the second captured image I2. At this time, the second inspection unit 75 inspects the quality of the printed image and the presence or absence of defects for the area of the second captured image I2 specified by the second inspection area information D4. The second inspection unit 75 then outputs the inspection result D5 to the discharge instruction unit 76.

The discharge instruction unit 76 controls the operation of the defective product discharge unit 50. The discharge instruction unit 76 judges whether the tablet 9 is a good product or a defective product based on the inspection result D2 of the first inspection unit 72 and the inspection result D5 of the second inspection unit 75. If the discharge instruction unit 76 judges that the tablet 9 is a defective product, it outputs a discharge instruction D6 to the defective product discharge unit 50. This causes the defective product blow mechanism 51 to operate, and the tablet 9 judged to be a defective product is discharged to the defective product collection unit 52. If the discharge instruction unit 76 judges that the tablet 9 is a good product, it does not output the discharge instruction D6 to the defective product discharge unit 50. In this case, the tablet 9 judged to be a good product is conveyed to the good product discharge position P5 through the defective product discharge position P4. Then, the good product blow mechanism 61 operates, and the tablet 9 judged to be a good product is discharged to the good product collection unit 62.

<3. Area determination of captured image>
As described above, the first area determination unit 71 and the second area determination unit 74 of the computer 70 determine the area to be inspected in the captured image. Details of this area determination process will be described below. In the following, the captured image acquired by the first imaging unit 20 or the second imaging unit 40 is referred to as the "captured image I." The captured image I may be the first captured image I1 or the second captured image I2.

Figure 6 is a flowchart showing the procedure for the region determination process. As shown in Figure 6, the computer 70 first classifies the multiple pixels that make up the captured image I into three regions by clustering based on the brightness values (step S1).

Figure 7 shows an example of a captured image I. Figure 8 is a histogram showing the distribution of luminance values of multiple pixels constituting the captured image I. In the histogram in Figure 8, the horizontal axis represents luminance value, and the vertical axis represents the number of pixels. As shown in Figures 7 and 8, in the captured image I of the tablet 9, the luminance value of the top surface of the tablet 9 tends to be the highest (part marked L1 in Figure 8). Also, in the captured image I of the tablet 9, the luminance value of the side surface of the tablet 9 tends to be the second highest after the top surface (part marked L2 in Figure 8). Also, in the captured image I of the tablet 9, the luminance value of the background tends to be the lowest (part marked L3 in Figure 8).

In step S1, this tendency of brightness values is used to classify the multiple pixels included in the captured image I into three regions. Specifically, the computer 70 classifies the multiple pixels included in the captured image I into three regions using k-means clustering: a first top region A11 corresponding to the top surface of the tablet 9, a first side region A21 corresponding to the side of the tablet 9, and a first background region A31 corresponding to the background.

To explain k-means clustering in more detail, the computer 70 first classifies the multiple pixels included in the captured image I into three clusters randomly or according to a predetermined rule. Next, the computer 70 calculates the average brightness value of the multiple pixels belonging to each cluster. Then, the computer 70 calculates the difference between the brightness value of each pixel and the above average value of the three clusters. Then, each pixel is reclassified into the cluster with the smallest difference. The computer 70 repeats this process until the pixel classification results no longer fluctuate or a predetermined termination condition is met. As a result, as shown in FIG. 9, the multiple pixels included in the captured image I are classified into the first top surface area A11, which is the pixel group with the highest brightness value, the first side surface area A21, which is the pixel group with the second highest brightness value, and the first background area A31, which is the pixel group with the lowest brightness value.

In this step S1, the captured image I is roughly classified into three regions: the top surface of the tablet 9, the side of the tablet 9, and the background. However, some pixels may be classified into the wrong cluster due to the influence of light reflection, etc. In particular, pixels corresponding to the side of the tablet 9 are often classified into the wrong cluster. Therefore, in order to more accurately distinguish between the three regions, the computer 70 performs the processes of the following steps S2 to S6.

When step S1 is completed, the computer 70 generates a first overall region A41 corresponding to the entire tablet 9 (step S2). Here, the first overall region A41 is generated based on the first upper surface region A11, which is easier to obtain a relatively stable classification result among the three regions A11, A21, and A31 described above. FIG. 10 is a diagram showing the processing of step S2. In FIG. 10, the first upper surface region A11 before expansion is shown by a dashed line, and the first overall region A41 after expansion is shown by a solid line. As shown in FIG. 10, the computer 70 expands the first upper surface region A11, which has a substantially elliptical shape, toward the first side region A21 along the minor axis direction. At this time, the computer 70 expands the first upper surface region A11 using an expansion rate previously stored in the storage unit 703. As a result, the first overall region A41 that generally follows the overall shape of the tablet 9 is generated.

Then, the computer 70 determines a second overall region A42 that corresponds to the entire tablet 9 and is more precise than the first overall region A41 (step S3). Here, the first overall region A41 obtained in step S2 is used as a teaching label, and the second overall region A42 is determined by a region segmentation method that finds a boundary line with a large luminance gradient. Figure 11 is a diagram showing the processing of step S3. In Figure 11, the first overall region A41 that is the teaching label is shown by a dashed line, and the second overall region A42 determined by the region segmentation method is shown by a solid line.

In this step S3, the computer 70 determines a boundary line C, which is a closed curve in the vicinity of the contour of the first entire area A41 of the captured image I, for which the energy E expressed by the following formula (1) is minimum.

In formula (1), ∇l is the rate of change of the luminance value (luminance gradient), length(C) is the length of boundary line C, and α is a constant. The function g(x) is defined as g(x) = exp(-√x). Therefore, as the magnitude of the luminance gradient |∇l| increases, g(|∇l|) decreases. That is, in step S3, boundary line C with the maximum luminance gradient is found while restricting the length of boundary line C so that it does not become excessively large. Then, computer 70 determines the area enclosed by boundary line C, which is the found closed curve, as second entire area A42.

Next, the computer 70 shrinks the second overall region A42 obtained in step S3 to generate a second upper surface region A12 corresponding to the upper surface of the tablet 9 (step S4). FIG. 12 is a diagram showing the processing of step S4. In FIG. 12, the second overall region A42 before shrinkage is shown by a dashed line, and the second upper surface region A12 after shrinkage is shown by a solid line. As shown in FIG. 12, the computer 70 shrinks the second overall region A42, which is approximately elliptical, toward the first upper surface region A11 along the minor axis direction. At this time, the computer 70 shrinks the second overall region A42 using a shrinkage rate previously stored in the storage unit 703. As a result, the second upper surface region A12 that generally follows the shape of the upper surface of the tablet 9 is generated.

Next, the computer 70 determines a third top surface area A13 that corresponds to the top surface of the tablet 9 and is more precise than the second top surface area A12 (step S5). Here, the second top surface area A12 obtained in step S4 is used as a teaching label, and the third top surface area A13 is determined by a region segmentation method that finds a boundary line where the luminance gradient becomes large. Figure 13 is a diagram showing the processing of step S5. In Figure 13, the second top surface area A12, which is the teaching label, is shown by a dashed line, and the third top surface area A13 determined by the region segmentation method is shown by a solid line.

In this step S5, the computer 70 finds a boundary line C, which is a closed curve near the contour of the second upper surface area A12 of the captured image I, where the energy E expressed by the above formula (1) is minimum. However, the constant α in step S5 may be a value different from the constant α in the above step S3. The computer 70 finds the boundary line C where the brightness gradient is maximum, while limiting the length of the boundary line C so that it is not excessively large. The computer 70 then determines the area enclosed by the found boundary line C, which is a closed curve, as the third upper surface area A13.

Then, the computer 70 determines a second side area A22 that corresponds to the side of the tablet 9 and is more precise than the first side area A21 (step S6). FIG. 14 is a diagram showing the processing of step S6. In this step S6, the computer 70 determines the second side area A22 by taking the exclusive OR of the second entire area A42 determined in step S3 and the third upper surface area A13 determined in step S5. In other words, the computer 70 determines the remaining portion of the second entire area A42 excluding the third upper surface area A13 as the second side area A22.

Then, the computer 70 determines the third top surface area A13 determined in step S5 and the second side surface area A22 determined in step S6 as the inspection areas (step S7).

As described above, the computer 70 first classifies the pixels of the captured image I into three regions by clustering based on the brightness values (step S1). Next, the computer 70 generates the first entire region A41 by expanding the first top surface region A11, which is one of the three regions from which a relatively stable classification result can be obtained (step S2). Next, the computer 70 determines the precise second entire region A42 by a region division method using the first entire region A41 as a teaching label (step S3). Next, the computer 70 generates the second upper surface region A12 by contracting the second entire region A42 (step S4). Next, the computer 70 determines the precise third upper surface region A13 by a region division method using the second upper surface region A12 as a teaching label (step S5). After that, the computer 70 generates the precise second side region A22 by taking the exclusive OR of the second entire region A42 and the third upper surface region A13 (step S6).

In this way, the top and side regions of tablet 9 contained in captured image I can be determined without relying on edge detection. Furthermore, the top and side regions of tablet 9 can be determined more accurately than when the top and side regions of tablet 9 are determined by clustering alone. Therefore, the top and side regions of tablet 9 can be inspected with high accuracy.

In particular, in the photographed image I of the tablet 9, the boundary between the entire region of the tablet 9 and the background region has a relatively large difference in brightness value. For this reason, in the above step S3, the computer 70 determines the entire region of the tablet 9 using a region segmentation method before the top surface region of the tablet 9. This allows the computer 70 to accurately determine the second entire region A42. The computer 70 then applies a region segmentation method to the highly reliable second top surface region A12 generated by shrinking the second entire region A42, thereby determining a precise third top surface region A13. This allows the third top surface region A13 to be determined more accurately.

4. Modifications
Although the embodiment of the present invention has been described above, the present invention is not limited to the above embodiment.

In the above embodiment, k-means clustering is performed in step S1. However, k-medoids clustering may be performed instead of k-means clustering. That is, in step S1, instead of calculating the average brightness value for each cluster, the median brightness value may be calculated. Then, the difference between the brightness value of each pixel and the above median value of the three clusters may be calculated, and each pixel may be reclassified into the cluster with the smallest difference.

In the above embodiment, the top and side surfaces of the tablet 9 are inspected. However, of the top and side surfaces of the tablet 9, only the top surface may be inspected. In that case, the process of step S6 described above may be omitted.

In addition, in the above embodiment, the tablet 9 was photographed at two locations: the first imaging position P1 before printing, and the second imaging position P3 after printing. That is, in the above embodiment, inspection was performed twice, before printing and after printing. However, the tablet 9 may be inspected only either before printing or after printing.

In the above embodiment, a tablet printing device 1 that prints on the top surface of a tablet 9 has been described. However, the present invention may also be applied to a tablet inspection device that does not have a printing function.

In addition, in the above embodiment, the printing unit 30 is provided with four heads 31. However, the number of heads 31 included in the printing unit 30 may be one to three, or may be five or more.

In addition to the configuration of the above embodiment, the tablet printing device 1 may also have a drying mechanism for drying the ink adhering to the tablets 9. The drying mechanism may be provided downstream of the printing unit 30 in the conveying direction and upstream of the defective product ejection unit 50 in the conveying direction. The drying mechanism may be, for example, a mechanism for blowing hot air toward the tablets 9 conveyed by the conveyor belt 12.

In addition, the "tablet" in the present invention may refer not only to a pharmaceutical tablet, but also to a supplement tablet or a tablet candy such as Ramune.

In the above embodiment, the value of each pixel in the captured image is described as "brightness." However, it goes without saying that the present invention can be applied to values obtained by converting the value of each pixel in the captured image, such as "density values."

Other details of the tablet printing device can be modified as appropriate without departing from the spirit of the present invention. In addition, the elements appearing in the above embodiments and variations can be combined as appropriate without causing any contradictions.

LIST OF SYMBOLS 1 Tablet printing device 9 Tablet 10 Conveying mechanism 20 First imaging section 30 Printing section 40 Second imaging section 50 Defective product ejection section 60 Good product ejection section 70 Computer 71 First area determination section 72 First inspection section 73 Print instruction section 74 Second area determination section 75 Second inspection section 76 Ejection instruction section 80 Carrying-in mechanism A11 First top surface area A12 Second top surface area A13 Third top surface area A21 First side surface area A22 Second side surface area A31 First background area A41 First entire area A42 Second entire area D1 First inspection area information D2 Inspection result D3 Print instruction D4 Second inspection area information D5 Inspection result D6 Ejection instruction I Photographed image I1 First photographed image I2 Second photographed image

Claims (6)

A method for determining an area in a captured image of a tablet having an upper surface and a side surface, comprising:
a) classifying each pixel of the captured image into three regions, a first top region corresponding to the top surface, a first side region corresponding to the side surface, and a first background region corresponding to a background, by clustering based on brightness values;
b) after step a), expanding the first upper surface region toward the first side surface region to generate a first entire region corresponding to the entire tablet;
c) after the step b), a step of determining a second overall region corresponding to the entire tablet by a region segmentation method for determining a boundary line with a large luminance gradient using the first overall region as a teaching label;
d) after step c), shrinking the second overall region toward the first upper surface region to generate a second upper surface region corresponding to the upper surface;
e) after step d), determining a third upper surface region corresponding to the upper surface by a region division method for determining a boundary line with a large luminance gradient using the second upper surface region as an instruction label;
The region determination method includes the steps of: The area determination method according to claim 1 ,
f) after step e), determining a second side surface area corresponding to the side by taking an exclusive OR of the second overall area and the third top surface area. The area determination method according to claim 1 or 2,
In the step a), each pixel of the captured image is classified into the first top surface region, the first side surface region, and the first background region by k-means clustering. The area determination method according to any one of claims 1 to 3,
In the steps c) and e), the luminance gradient is ∇l, the function g(x) is g(x)=exp(−√x), the length of the boundary is length(C), and α is a constant,

The method for determining an area is to find a boundary line C that minimizes the energy E represented by the following equation: A tablet inspection method for inspecting tablets, comprising:
A) acquiring an image of a tablet;
B) determining an inspection area in the captured image by a region determination method according to any one of claims 1 to 4;
C) inspecting the inspection area of the captured image;
A tablet inspection method comprising the steps of: 1. A tablet inspection apparatus for inspecting tablets having a top surface and a side surface, comprising:
An imaging unit for acquiring a photographed image of a tablet;
an area determination unit that determines an inspection area in the captured image;
an inspection unit that inspects the inspection area of the captured image;
Equipped with
The area determination unit is
a) classifying each pixel of the captured image into three regions, a first top region corresponding to the top surface, a first side region corresponding to the side surface, and a first background region corresponding to a background, by clustering based on brightness values;
b) after step a), expanding the first upper surface region toward the first side surface region to generate a first entire region corresponding to the entire tablet;
c) after the step b), a step of determining a second overall region corresponding to the entire tablet by a region segmentation method for determining a boundary line with a large luminance gradient using the first overall region as a teaching label;
d) after step c), shrinking the second overall region toward the first upper surface region to generate a second upper surface region corresponding to the upper surface;
e) after step d), determining a third upper surface region corresponding to the upper surface by a region division method for determining a boundary line with a large luminance gradient using the second upper surface region as an instruction label;
and the third upper surface area is the inspection area.

Images