JP2000135268A - Tablet testing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a tablet testing device of high reliability and high processing speed capable of monitoring testing data even after tablets are delivered to a patient, and reducing testing labor of a testing pharmacist. SOLUTION: Penetrating ray (X-ray) is radiated to a tablet package 12a by a penetrating ray (X-ray) generating means 14, and this penetrating ray (X-ray) is detected by a penetrating ray (X-ray) detecting means 16. Based on the result of detection by the penetrating ray (X-ray) detecting means 16, the number of tablets packaged in the tablet package 12a is determined by a tablet number determination processing means 18. Tablet number data to be packaged in the tablet package 12a is extracted from prescription data, this tablet number data is compared with the number of the tablets determined by the tablet number determination processing means 18, and it the tablets based on the prescription data are packaged or not is determined by a comparison determination processing means 18.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a tablet inspection apparatus for automatically inspecting whether a packaged tablet is packaged as prescribed.

[0002]

2. Description of the Related Art Conventionally, an inspection of whether or not a tablet packaged by a tablet packaging device is packaged according to a prescription is generally performed by a dispenser with the naked eye. A device for supporting the examination by the dispenser is disclosed in
No. 307 and Japanese Patent Publication No. 4-17666. These inspection devices are configured to inspect the type and number of tablets by imaging the packaged tablets with a camera.

[0003]

However, since these inspection apparatuses take a two-dimensional image of tablets wrapped in a wrapping paper, when a plurality of tablets are wrapped, it is possible to recognize overlapping or contact of the tablets. In some cases, it was determined that the packaging was defective. Further, when characters are printed in a band shape with white paint on the wrapping paper, tablets hidden behind the band cannot be identified.

[0004] These problems cannot be solved even if the image is taken from the front and back sides of the wrapping paper. Therefore, an apparatus for separating and conveying tablets immediately before packaging one by one and imaging and inspecting these tablets is disclosed in JP-A-5-337168 and JP-A-8-168727. However, these inspection devices are slow in processing speed because it is necessary to separate tablets one by one. In addition, since the inspection performed immediately before packaging by these inspection devices is not an inspection of the actual packaged state, even if tablets to be packaged in a certain wrapping paper may be mixed in the wrapping paper before and after the wrapping paper. , It cannot be detected and its reliability is low. For example, when the types of tablets to be taken in the morning and the afternoon are different, even if the tablets for the day are mixed in the wrapping paper in which the tablets for the morning are wrapped, they may be handed to the patient and taken as they are, and the medicinal effect may be lost. There is. In addition, even if foreign materials such as screws, springs, and other components of the device such as screws and springs, insects, and dust enter, there is a risk of being handed to the patient and taken.

The present invention has been made in view of the above-mentioned problems, and has high reliability, high processing speed, and can monitor test data even after being handed over to a patient. It is an object of the present invention to provide a tablet inspection device capable of alleviating the problem.

[0006]

According to a first aspect of the present invention, there is provided a tablet inspection apparatus for inspecting tablets packed in a tablet package based on prescription data. Transmission line generating means for generating and irradiating the tablet package, transmission line detection means for detecting the transmission line generated by the transmission line generation means, and the number of tablets packaged in the tablet package Tablet number discriminating processing means for discriminating from the detection result of the transmission line detecting means, and tablet number data to be packed in the tablet package from the prescription data, and discriminating between the tablet number data and the tablet number discriminating processing means And a comparing and judging means for judging whether or not tablets according to the prescription data are packed by comparing the number of tablets thus obtained.

The tablet package of the present invention includes a vial which is a cylindrical container with a lid, in addition to a packaging bag. Further, the term “transmission ray” is a concept including X-rays, radiation, and light rays that pass through an object, for example, infrared rays.

According to the invention as the first means, the number of tablets packaged in the tablet package is discriminated from the detection result of the transmission line detecting means, and the number of tablets to be packaged in the tablet package is determined. In comparison with the above, since it is possible to determine whether or not tablets are packed according to the prescription data, the inspection of the packed tablets becomes simple and quick, and the reliability is improved.

As a second means for solving the above-mentioned problems, the present invention relates to a tablet inspection apparatus for inspecting tablets packaged in tablet packages based on prescription data. A transmission line generating unit that irradiates the package, a transmission line detection unit that detects a transmission line generated by the transmission line generation unit, and a detection of the shape of the tablet packaged in the tablet package by the transmission line detection unit. Tablet shape discriminating processing means for discriminating from the results, storage means for storing the tablet shape data for each type of tablet, and extracting the type of tablet to be packaged in the tablet package from the prescription data; The tablet shape data corresponding to the prescription data is retrieved from the storage means, and the tablet shape data is compared with the tablet shape determined by the tablet number determination processing means. Is obtained by a comparison determination processing unit determines whether the.

According to the invention as the second means, the shape of the tablet packaged in the tablet package is determined from the detection result of the transmission line detecting means, and the shape data of the tablet packaged in the tablet package is determined. It is possible to determine whether or not tablets are packed according to the prescription data in comparison with the above, so that the inspection of the packed tablets is simple and quick, and the reliability is high, as in the first invention. .

According to a third aspect of the present invention, there is provided a tablet inspection apparatus for inspecting a tablet packaged in a tablet package based on prescription data. A transmission line generating unit that irradiates the package, a transmission line detection unit that detects a transmission line generated by the transmission line generation unit, and a detection of the shape of the tablet packaged in the tablet package by the transmission line detection unit. Tablet shape discriminating processing means for discriminating from the results, storage means for storing the tablet shape data for each type of tablet, and extracting the type of tablet to be packaged in the tablet package from the prescription data; Is called from the storage unit, and the tablet shape data is associated with the shape of the tablet determined by the tablet number determination processing unit, and the tablet packaged in the tablet package is stored. A tablet number discriminating means for discriminating the number, a type of tablet to be packed in the tablet package and tablet number data are extracted from the prescription data, and the tablet number data and the tablet number discriminating means are discriminated. A comparison / determination processing means for comparing the number of tablets with each other to determine whether or not tablets are packed according to the prescription data.

According to the invention as the third means, the number of tablets packed in the tablet package is determined from the detection result of the transmission line detecting means, and the number of tablets to be packed in the tablet package is determined. It is possible to determine whether or not tablets are packed according to the prescription data in comparison with the above, so that the inspection of the packed tablets is simple and quick, and the reliability is high, as in the first invention. .

[0013] The tablet number discriminating processing means is provided when the data detected by the transmission line detecting means includes data which cannot be converted as a number, or when the data detected by the transmission line detecting means includes data whose shape cannot be recognized. When doing so, it may be processed as defective packaging. This further increases reliability. Further, when an abnormality is determined, a marking means for marking a corresponding tablet package can be provided, whereby defective products can be easily identified.

The transmission line detecting means can detect a transmission line as a transmission level or a shadow of a tablet.

According to a fourth aspect of the present invention, there is provided a tablet inspection apparatus for inspecting a tablet packaged in a tablet package based on prescription data. A transmission line generating unit that irradiates the package, a transmission line detection unit that detects a transmission line generated by the transmission line generation unit, a storage unit that stores image data for each type of tablet, Tablet discriminating processing means for discriminating the type of the tablet by capturing the shape of the packaged tablet from the detection result of the transmission ray detecting means, and calling tablet image data corresponding to the type of the tablet from the storage means; Video data generating means for generating video data by superimposing data on the detection data of the transmission line detecting means.

According to the invention as the fourth means, the type of the tablet is determined from the detection result of the transmission line detecting means, and the tablet image data corresponding to the type of the tablet is superimposed on the transmission line detection data. Since the video data is created, the packaged tablets can be viewed as video data, and the inspection of the tablets is facilitated.

[0017] The tablet inspection apparatus may include display means for displaying the video data, and storage means for storing the video data in association with prescription data. This makes it easier and faster to inspect tablets.

In the invention as the first to fourth means, a plurality of the transmission line generation means and the transmission line detection means are provided in different directions of the irradiation direction, and the detection data of the transmission line detection means is reproduced three-dimensionally. Reproducing means can be provided. Further, the transmission line generating means may be rotatably provided around the tablet package so that the irradiation direction can be changed, and a reproducing means for three-dimensionally reproducing the detection data of the transmission line detecting means may be provided. As a result, even if the packaged tablets overlap, those tablets can be reliably determined, and the reliability of the inspection is further improved.

[0019]

Embodiments of the present invention will be described below with reference to the accompanying drawings.

<First Embodiment> FIG. 1 shows a tablet packing machine 2 provided with a tablet inspection apparatus 1 according to the present invention. The tablet packing machine 2 includes a tablet dispensing device 4, a packaging device 5, and a printing device 6 controlled by the control device 3. The tablet dispensing device 4 includes a number of cassettes 7 in which tablets are stored by type. This tablet dispensing device 4
The tablets are dispensed from the corresponding cassette 7 according to the prescription data received from the host computer 8 through the personal computer 9. The tablets dispensed by the tablet dispensing device 4 are guided to the packaging device 5 via the hopper 10. Packaging device 5
Is a wrapping paper 12 sent out from the roll 11 by a fixed amount.
Is folded in the longitudinal direction, and when one packet of tablets is supplied, the wrapping paper 12 is sealed and packaged by the heater roller 13. The feed amount of the wrapping paper 12 is detected by an encoder (not shown). On the wrapping paper 12, the taking time, the patient name, and the like are printed by the printing device 6. The tablets packaged in the wrapping paper 12 are inspected by the tablet inspection device 1 described below.

The tablet inspection apparatus 1 comprises an X-ray tube 14 as a transmission line generator and its control unit 15, a line sensor 16 as a transmission line detection unit, a marking device 17, and a control unit 18.

As the X-ray tube 14, a fixed target X-ray tube 14a shown in FIG. 2 or a target rotating X-ray tube 14b shown in FIG. 3 can be used.

The target fixed X-ray tube 14a shown in FIG.
Has a cathode 21 made of a filament covered with a cover 20 and a target 22 made of a heavy metal such as tungsten or molybdenum inclined at about 2 ° to 20 ° with respect to the cathode 21 in a glass tube 19. I have. The cathode 21 and the target 22 are each connected to a power supply terminal. Behind the target 22, a cooling nozzle 23 for ejecting cooling oil to control the surface temperature of the target 22 is provided. When a voltage is applied to the power supply terminal by the electronic circuits shown in FIGS. 4 and 5, electrons move from the filament 21 toward the target 22, and X-rays are emitted from the target 22 through the beryllium window 24.

The target rotary X-ray tube 14b shown in FIG.
Is basically a fixed target X-ray tube 14 shown in FIG.
Since they are the same as a, the corresponding parts are denoted by the same reference numerals. The target rotating X-ray tube 14b is configured so that the target 22 rotates at a high speed. Since the rise in the temperature of the target 22 is suppressed by dispersing the irradiation positions, the oil cooling device is not provided, and the angle of the target 22 is also changed from 25 ° to 50 °. The surface of the target 22 is composed of two types of metal arranged alternately. When the target 22 rotates, the two types of metal alternately receive electrons emitted from the filament 21.

Although the material of the target 22 of the X-ray tube 14 greatly affects the intensity of X-rays, molybdenum is preferable for use as the tablet detecting device 1 of the present invention.
The reason for this is that the X-ray intensity must be reduced as much as possible because the test object is a wrapping paper and a tablet, which are thin. Also,
As shown in FIGS. 6 to 8, molybdenum has a higher absorptance to an object than tungsten, and has a property of increasing photons at a low photon energy.

FIG. 9 shows a tube voltage waveform obtained by actually generating X-rays from the X-ray tube 14 using the circuit of FIG. The capacitor terminal voltage at the time of X-ray irradiation decreases with time as a result of discharge, and the voltage drop ΔV (V) is ΔV = iT
/ C, and the tube voltage waveform becomes a sawtooth waveform. This Δ
V can be reduced by increasing the capacitance of the capacitor.

Further, even if tungsten is used as the material of the target 22, the X-ray intensity can be reduced by lowering the applied voltage.
X can be detected even if transmission of the line cannot be confirmed.
A line having a high photon energy may be applied in a pulsed manner for a short time. In this case, the X-ray tube is preferably the target rotating X-ray tube 14b shown in FIG. By using two types of materials for the target 22, photon energy can be generated with two types of intensity. In particular, as shown in FIGS.
When the tube voltage is reduced by using
In the region of 上 が っ 40 KeV, the number of photons increases, the noise increases, and the peak of Ka1 also decreases. Therefore, when the photon energy outside the region reaches the line sensor 16 described below after passing through the aluminum filter 25 described below, the image is blurred. It is a factor that occurs.

X-rays generated from the X-ray tube 14 are shown in FIG.
As shown in FIG. 0, the test object is irradiated through the aluminum filter 25. The aluminum filter 25 has a thickness of about 1 mm, and has a function of removing a region where the number of photons fluctuates and irradiating X-rays in a region where the photon energy is stable. When a collimator wedge filter is used, a filter that matches the shape of the test object is used.
It is also possible to prevent the edge of the tablet from being blurred.

The X-rays irradiated on the object to be inspected are sliced by the mask 26 and enter the line sensor 16. Mask 2
6 has a slit 27 extending in a direction perpendicular to the transport direction of the packaging bag 12a, as shown in FIG. The width of the slit 27 is usually about 2 mm. As for the X-rays that have reached the mask 26, the X-rays at portions other than the slit 27 are blocked, and only the X-rays that have passed
6 is incident.

As the line sensor 16, an ionization chamber detector 16a or a semiconductor detector 16b is used. As shown in FIG. 11, the ionization chamber detection device 16a alternately arranges a plate-like bias electrode 29 and a signal electrode 30 in a container 28 at a fixed interval d, and supplies 10p of xenon gas between the electrodes 29 and 30. (Atm). When X-rays enter from the concave portion 31 on the upper surface of the container 28 and are ionized while a voltage is applied between the bias electrode 29 and the signal electrode 30, a minute current flows between the electrodes 29 and 30. It is detected numerically.

The characteristic conditions of the line sensor 16 are obtained as follows.

Assuming that ionization occurs uniformly by X-rays in the parallel plate ionization chamber, the electric field in the ionization chamber becomes zero due to space charge.
Consider the limitations of ion pair collection. n ₀ pieces // cc /
When s ion pairs are generated, the ion charge density at point x is given by:

(Equation 1)

Here, w ₊ is the flow velocity of the ions, and is expressed by Equation 2, where E is the electric field, P is the gas pressure, and μ ₊ is the moving velocity of the ions. The Poisson equation describing the electrolysis is shown in Equation 3.

(Equation 2)

(Equation 3)

Μ ₊ is a value specific to the gas, and Xe (xenon) is 0.58. Substituting Equation 1 into Equation 3 and giving a boundary condition that E = 0 at x = d due to space charge, the following equation is obtained.

(Equation 4)

The potential difference V between the electrodes can be obtained by calculating Equation 4 from x = 0 to x
= D

(Equation 5)

That is, when it is assumed that when a parallel-plate ionization chamber is irradiated with X-rays, ionization occurs uniformly in the ionization chamber, the potential difference Vs required to collect the charge of n ₀ e is obtained.
Is limited by space charge unless it is equal to or higher than Vs in Equation 5.

Equation 5 is expressed in CGS system, but μ ₊
(Cm ² / V / s) and d (cm), and the following formula is obtained by using an advantageous unit. Here, I is the ionization chamber output current (A), and D is the ionization chamber effective volume (cm ³ ).

(Equation 6)

Since the potential difference Vs increases as the output current I increases, it is necessary to set the applied voltage so as to saturate at the maximum output current. In particular, when the gas pressure is high, the charge density increases in a portion near the X-ray entrance window in the ionization chamber, and thus a potential higher than the potential Vs calculated from the average charge density is required. When irradiating X-rays in a pulse shape, the step response must be fast in addition to the saturation characteristic which is a static characteristic. Particularly, in the tablet packaging machine 2 according to the present invention, since 90 packages can be packaged per minute in recent years, it is necessary to inspect a range of 114 mm per second. To enable this condition, it is necessary to emit X-rays at 5 ms intervals. Since the rise time of the X-ray is about 0.1 to 0.2 ms, there is no problem. However, the rising characteristics of the line sensor 16 are determined by the flow velocity of ions.

In the ionization chamber with an electrode spacing d, n ₀ / c when t ≧ 0
Consider a rising characteristic of the output current I at the collecting electrode when ion pairs are generated uniformly at a rate of c / s. The rate of change of the ion density over the ionization chamber at time t is given by:

(Equation 7)

Here, the first term on the right side represents occurrence, and the second term represents the amount reaching the collecting electrode. N ₀ represents n ₀ d · S, and S
Is the area of the electrode plate. As shown by the following equation, n ₊ increases at a rate of occurring every second until ions generated at x = d at time t = 0 reach the collecting electrode at a density reaching the collecting electrode, but thereafter, Become parallel.

(Equation 8)

When Equation 8 is substituted for the boundary condition, the following equation is obtained.

(Equation 9)

By substituting equation 9 into equation 7 and obtaining N +, the following equation is obtained.

(Equation 10)

The same formula holds true for electrons, but the electron flow velocity w ₋ is much faster than ions.

[Equation 11] You can think about it.

The output current I is

(Equation 12) By substituting Equations 10 and 11 into this, the following equation is obtained.

(Equation 13)

Thus, the rise time of the output current of the line sensor 16 is equal to the ion movement time d / w ₊
It can be seen that FIG. 12 shows this measurement. Applied voltage V to equal the rise time of the output current d / w ₊ the rise time t ₀ of the X-ray
Is represented by the following equation, is proportional to the pressure p, and is 2
It is proportional to the power.

[Equation 14]

FIG. 13 shows these relationships. To increase the detection speed at a low voltage, the distance d between the electrodes is set to 1 mm, the gas pressure p is set to 10 atm, and the applied voltage V is set to 1500 V. Thereby, the rise time t can be set to 100 μs.

As the semiconductor detecting device 16b, a pn junction type can be used as shown in FIG. When an X-ray is incident between the junctions of the element, a current pulse is generated. p-
In addition to the n-junction, a pin-junction type can also be used. When such a semiconductor detector 16b is used, a multi-channel detector that is compact and has good output linearity can be obtained.

The principle of detection by the line sensor 16 will be described. The detection object includes a spatially non-uniform capsule and a spatially substantially uniform tablet. In order to transmit X-rays through the packaging bag 12a in which such an object is packaged to reproduce a plane in a certain range, generally, at all points in the plane, all points in the direction passing through the point are used. Shooting data is required. However, a reproduced image can be formed within a permissible range even from discrete finite photographing data. In this case, of course, the accuracy of the reproduced image differs between a tablet having a spatially uniform structure and a capsule having a spatially non-uniform structure. Since an object to be detected such as a tablet has a relatively simple outer shape, X-rays do not penetrate the object twice from a certain direction except for a donut-shaped troche tablet. Therefore, the image can be reproduced by taking into account the absorptivity when X-rays pass through the tablet.

As shown in FIG. 15, when the packaging bag 12a is conveyed in the z (-) direction within the horizontal xz plane in three axes (x, y, z) as shown in FIG.
They are arranged on the line aa ′ parallel to the x-axis direction. Packaging bag 12
If the tablets overlap each other in the area a, it is difficult to determine the number and shape of the tablets. Therefore, as shown in FIG.
Is preferably provided. The separating device 32 includes two sponge rollers 33 provided below the packaging bag 12a.
a and 33b, and one sponge roller 33c arranged above the packaging bag 12a. The overlapping tablets are separated by passing the packaging bag 12a between them in a corrugated manner. ing. The separating device 32 includes three sponge rollers 33a, 33b, 33c.
However, the present invention is not limited to this, and may include more sponge rollers.

As the packaging bag 12a is transported, the line sensor 16 detects p1, p2, p3, and p3 of the packaging bag 12a.
, Pn is detected at each X-ray radiation point. The pitch of the shooting points p1, p2, p3,..., Pn is 0.76 mm in this embodiment, as shown in FIG.

When the capsule agent is located at the point where the line sensor 16 strikes as shown in FIG. 17, the transmission amount of X-rays detected by the line sensor 16 is shown by a graph shown in FIG.

Further, as shown in FIG. 15, the shooting points p1, p of the packaging bag 12a in which the capsules and tablets are packaged.
Graphs of the X-ray transmission amounts of 2, p3, p4, and p5 are shown in FIGS. For simplification of the drawing, each point is shown every two points. 22 to 23, at point p1-1, the attenuation of the amount of X-ray transmission by only the packaging bag 12a is detected, and at point p1-2,
An attenuated portion of the amount of X-ray transmission due to the end of the capsule begins to be seen, and from point p1-3 to point p3-2, a large attenuation due to the center of the capsule is observed. In addition, point p3
From -3, the part where the amount of X-ray transmission by the tablet is attenuated begins to be seen, and at point p4-1, the attenuation of the amount of X-ray transmission by the capsule decreases. Point p4-2 to point p5
Until -2, only the attenuation of the X-ray transmission amount by the tablet is detected, and at the point p5-3, the attenuation of the X-ray transmission amount by only the packaging bag 12a is detected. FIG. 24 shows the attenuation data shown in FIGS. 19 to 23 in a three-dimensional graph.

Now, as shown in FIG. 25, the packaging bag 12a
In the same way, if a missing wire or tablet, which is a defective packaging element, is packaged inside, the amount of X-ray transmission is detected in the same way, the attenuation width is obtained from the attenuation data, and the center of the attenuation width is set on the axis. When arranged, the graphs shown in FIGS. 26A to 27D are obtained. In this figure, 0000-111 on the horizontal axis
1 is a detection coordinate position of the line sensor 16, and the vertical axis is an attenuation width.

Next, the amount of attenuation by the packaging bag 12a is subtracted from the attenuation data of the amount of X-ray transmission, and noise removal, edge emphasis processing, and detection of the center value of the amount of attenuation are performed. In other words, the graphic image shown in FIG. 28 is obtained. In the present embodiment, since only the attenuation of the X-ray transmission amount in the thickness direction, that is, the y direction, can be detected, the tablet shape of the overlapping tablet is reproduced as it is.

Although the graphic image of FIG. 28 simply expresses the X-ray absorptivity of the tablet as halved, the X-ray absorptivity of the tablet is different for each type, but in the y-axis direction. Has distortion. This distortion can be corrected by multiplying the distortion rate after the tablet type is specified.

The distortion in the y-axis direction is not such a serious problem that the tablets cannot be distinguished. The reason for this is that since the identity of the packaged drug can be known in advance by the prescription data, the X-ray absorption rate data of the drug that would have been packaged is read out from the storage unit, and the X-ray absorption data is determined based on the absorption rate data. This is because shape data may be created by correcting the detection data of the line transmission amount, and it may be determined whether or not the created shape data matches the shape data of the pre-packaged tablet.
If these do not match, the packaged tablet is likely to be wrong.

At this time, actual measurement data as shown in FIG. 29 and FIG.
The accuracy is higher when comparing the criterion data with the detection data of the object to be detected than when using the data stored as three-dimensional data. The reason is that there is a high possibility that the data deviates from the true value in the process of converting the data having the distortion in the y-axis direction into three-dimensional data. FIG. 29 is a round tablet with Ebutol shown in FIG. 47, and FIG. 30 is a square tablet with an “A” groove engraved on the surface, and is measured data when irradiated from one direction of anadol shown in FIG. 47. is there.

As shown in FIG. 31, the marking device 17 supports a plurality of pens 34 of different colors corresponding to the respective defective items on a support shaft 36 via a holder 35 and is attached to a frame 37. Each pen 3 with solenoid 38
4 can be used in such a manner that the front end is pressed against the packaging bag 12a. Each pen 34 has a solenoid 3
When the package 8 does not operate, it is urged by a spring (not shown) so as to separate from the packaging bag 12a. The pen tip is stored in a groove 39 formed in the frame 37 so as not to dry. The marking device 17 is provided behind the position where the medicine in the packaging bag has passed the line sensor 16 by the time required for the control device 18 to perform the number or shape determination processing.

As shown in FIG. 1, the control device 18 comprises the storage means, the number discrimination processing means or the shape discrimination processing means, and the comparison processing means of the present invention, and can be constituted by a personal computer. Here, the control device 18 is set to 3
A table may be provided. That is, the first control device writes the data from the line sensor 16 in the storage unit in a time series and transmits the data to the second control device. The second control device determines the number of packaged medicines and performs shape determination processing and stores the data. The third control device compares the prescription data packed in the packaging bag with the corresponding data processed by the second control device,
The mark providing means 17 is operated based on the determination result. In this way, when a plurality of control devices are connected and processed for each processing step, the result of the pass / fail determination can be quickly determined, so that the processing can be performed without lowering the packaging speed.

The measurement operation and the data processing operation by the control device 18 will be described later, and another embodiment of the tablet detection device 1 will be described below. In the following embodiments, portions substantially corresponding to those of the first embodiment are denoted by the same reference numerals, and description thereof is omitted.

<Second Embodiment> FIG. 32 shows a second embodiment of the present invention.
1 shows a tablet inspection device according to an embodiment. In this tablet inspection device, an X-ray tube 14 and a line sensor 16 are arranged similarly to the tablet inspection device of the first embodiment in FIG. 10, but in addition to this, the medicine packaged in the packaging bag 12a is used. A CCD camera 40 for photographing from the y-axis direction is disposed above the packaging bag 12a.

The color data and the external dimension data of the tablet can be obtained by the CCD camera 40. CC
By obtaining the color data by the D camera 40, the tablets can be overlapped to some extent, and the discrimination of the tablet type becomes more accurate. In the X-ray data, the tablet size detection accuracy is deteriorated by the penumbra as shown in FIG.
The correction can be made using the external dimension data acquired by the CCD camera 40. In the case of performing such processing as well, if the control device is made independent, the entire processing can be performed without affecting the packaging speed.

<Third Embodiment> FIG. 34 shows a third embodiment of the present invention.
1 shows a tablet inspection device according to an embodiment. In this tablet inspection device, in addition to the X-ray tube 14 and the line sensor 16 arranged in the y-axis direction, the X-ray tube 14 and the line sensor 16 are also arranged in the x-axis direction. X-rays are simultaneously emitted from two shifted directions.
According to this device, as shown in FIG.
The data of the amount of X-ray transmission in the axial direction makes it possible to discriminate tablets packaged in an overlapping manner to some extent. Further, the cross-sectional image shown in FIG. 36 and the three-dimensional image shown in FIG. 24 can be displayed based on the data in two directions.

<Fourth Embodiment> FIG. 37 shows a fourth embodiment of the present invention.
1 shows a tablet inspection device according to an embodiment. In this inspection apparatus, an annular trajectory 41 is arranged around an object to be inspected, the X-ray tube 14 is movably mounted on the annular trajectory 41, and an annular line sensor 16 is arranged around the outside of the annular trajectory 41. Have been. Annular track 41 and annular line sensor 16
Between them, a mask 26 having a slit 27 provided between two rings is arranged. A filter 25 made of aluminum is provided between the X-ray tube 14 and the packaging bag 12a for the object to be inspected, and is turned around the object to be inspected integrally with the X-ray tube 14.

In the tablet inspection apparatus, as shown in FIGS. 37 to 40, the packaging bag 12a is sent in the z-axis direction, and the X-ray tube 14 is turned by a predetermined angle to obtain X-ray transmittance data. And from the data, the packaging bag 1
Subtracting the attenuation amount of 2a, performing noise removal, edge enhancement processing, and detection processing of the center value of the attenuation amount, aligning the centers of the attenuation amounts, corresponding to the detection angle, and joining them spirally,
The image data shown in FIG. 41 is obtained.

<Fifth Embodiment> FIG. 42 shows a fifth embodiment of the present invention.
1 shows a tablet inspection device according to an embodiment. This inspection apparatus has a configuration combining the configurations shown in FIGS. 34 and 37. That is, the two X-ray tubes 14
Are arranged, and an annular orbit 41 of the two X-ray tubes 14 and an annular line sensor 16 are arranged around a packaging bag 12a which is an object to be inspected. In this tablet inspection apparatus, since the two X-ray tubes 14 rotate around the packaging bag 12a, which is the inspection object, while maintaining the positional relationship of 90 °, the thickness data of the inspection object and the X-ray transmission amount data are simultaneously obtained. In addition to being able to detect, the outer diameter shape of the test object can be measured from multiple directions. For this reason, even if a large amount of tablets are packed in the packaging bag 12a and overlapped multiple times, the type and quantity of tablets can be accurately determined.

In the tablet inspection apparatus, the X-ray tube 14 has 10
° 360 ° while irradiating X-rays at the pitch firing point
Since the packaging bag 12a rotates and can move in the z-axis direction at a pitch of 0.76 mm each time the 10 ° assault point comes, the inspection time can be reduced.

<Sixth Embodiment> FIG. 43 shows a sixth embodiment of the present invention.
1 shows a tablet inspection device according to an embodiment. This inspection apparatus is the same as the fifth embodiment, except that the inspection object is a tablet filled in the vial 42.
The vial 42 is used in Europe and the United States, and has a plastic cap 43 for preventing an infant from opening.
And one kind of tablet for the number of prescription days is filled. As shown in FIG. 44, the vial 42
Conveyed above. The vial 42 has a state in which the tablets are multiplexed in the vial 42, and the packaging bag 1 of the above-described embodiment.
Since the amount of X-ray attenuation is larger than 2a, it is necessary to increase the number of measurements at one measurement point in the z-axis direction and determine the number and type of tablets with higher accuracy.

<Measurement Operation> Next, the measurement operation and data processing operation of the tablet inspection apparatus of the second embodiment shown in FIG. 32 will be described with reference to the flowcharts of FIGS. 45 and 46.
The same applies to other embodiments.

First, the measuring operation will be described.
As shown in FIG. 5, in step 101, the transport coordinates of the packaging bag 12a in the z-axis direction are reset, and in step 102, packaging is started. In step 103, it is determined whether or not the packaging bag 12a has reached the detection point. If it has reached, the X-ray is irradiated in step 104, and the output data of the line sensor 16 is stored in step 105. Subsequently, in step 106, it is determined whether or not data acquisition has been completed for all detection points within the range of one package, and if not completed, the flow returns to step 103 to acquire data for the next detection point. If the data acquisition at all the detection points within the range of one package has been completed, at step 107
The area of the package is photographed by the CCD camera 40, and step 10
At step 8, the color image is stored. Then, Step 10
In step 9, it is determined whether or not there is a measurement for the next packaging bag. If so, the process returns to step 101; otherwise, the process ends.

<Data Processing Operation> Next, the operation of processing the data obtained by the measurement operation will be described with reference to the flowchart of FIG.

In step 111, first, the stored output data of the line sensor 16 is read. Step 112
Then, the attenuation width of each point of the output data is specified, and in step 113, the center of the attenuation width is aligned in correspondence with the sensor coordinates and the z-axis coordinates. In step 114, the shape of the tablet is determined from the X-ray transmission amount, which is the output data of the line sensor 16, based on the X-ray attenuation rate stored in advance.
Next, in step 115, the center position of the tablet shape is determined. This center position becomes a mark when it is made to correspond to the color data in a later step. In step 116, the rotation direction of the tablet is determined. The tablets are packaged in the packaging bag 12a in all directions. If the tablet rotation direction is determined in this way, the tablet can be identified and collated quickly.

In step 117, the color data acquired by the CCD camera 40 is read from the storage device, and in step 118, the color data of the CCD camera 40 and the X-ray transmission amount data of the line sensor 16 are adjusted to the center position of the tablet to correct the deviation. adjust. In step 119, step 11
The three-dimensional image data is created from the color data created in step 8.

Subsequently, at step 120, the judgment reference data is read. At this time, it is better to read out only the discrimination reference data corresponding to the tablets to be packed in the packaging bag 12a currently focused on than to read out the discrimination reference data corresponding to all tablets and perform the collation judgment.
This is preferable because the determination process can be performed very quickly.
In step 121, the number of tablet center positions in the three-dimensional image data is counted to determine the number of tablets of the same kind. In step 122, the type of tablet is determined in correspondence with the determination reference data read in step 120. , Step 12
In step 3, the packaged tablets are identified.

At step 124, it is determined whether or not there is detection data that does not correspond to the discrimination reference data, in order to check whether or not there is dust, foreign matter, chipped tablet or the like in the packaging bag 12a. At step 127, a defective product mark is provided, and at step 128, a defective product is displayed. In step 125, it is determined whether or not the prescription data of the medicine to be packaged matches the number and type of the determined tablets. If they do not completely match, a defective mark is added in step 127, and
Defective products are displayed at 28, and if they completely match, step 1
At 26, a pass is indicated. The above determination result is stored in association with the patient identification number.

<Defective Product Data Processing> The following items are identified as defective products. 1. 1. Include a shift before and after the loading position of the packaging bag belt. 2. Mixing of foreign matter 3. Cracking of drug 4. Excess or deficiency of drug Difference between prescription and packaged drug types

When the defective packaging is detected, it is necessary to notify the dispenser of the fact and to repeat packaging in some cases. A monitor, a display lamp, a buzzer, or the like may be used to notify a defective product. However, in order to clearly indicate what part of the packaging bag belt has failed and what kind of defect has occurred, the marking device 17 described above is used to notify the defective product. It is preferable that a mark corresponding to the defective item is directly added to the corresponding portion.

First, among the above-mentioned packaging defects, data processing in the case where there is a misalignment before and after the insertion position of the packaging bag belt will be described. For example, for a patient, tablet A 3
One tablet after each meal and one tablet B each morning and evening twice after meals
If there is a prescription for the day, the wrapping bag belt will include "morning",
"Lunch" and "Evening" are printed in order, two tablets in the packaging bag printed "Morning", one tablet in the packaging bag printed "Lunch",
Two tablets must be packaged in the packaging bag marked "evening". Here, when a shift occurs before and after the insertion position of the packaging bag belt, one tablet is added to the packaging bag printed with “morning”.
There are two tablets in the packaging bag marked "noon", which means that an appropriate number of tablets are not contained in the printed dose period.

Conventionally, when such a defect occurs,
It was necessary to cut off the two bags of "morning" and "noon" from the packaging bag band, input the prescriptions of the two bags from the terminal of the packaging device, and repackage. However, in the embodiment of the present invention,
When the defect is determined, the prescription data corresponding to the two bags of "morning" and "noon" are set as interrupt standby information of the packaging device, so that the trouble of inputting can be saved. The two defective tablets are automatically re-packaged after the completion of the patient's packaging according to the interrupt waiting information.

Next, in the case where there is a defect such as mixing of foreign matter, cracking of the medicine, and excess or deficiency of the medicine, only one packaging bag becomes a defective product. It is set as interrupt waiting information. Objects containing metals such as screws, wires, nuts, pebbles, etc., which have an X-ray transmission amount of zero, and small objects that do not conform to the discrimination reference data are all determined to be foreign substances. Further, FIGS. 62 (A) and (B)
As shown in the figure, when the detected data matches 90% of the actual measurement data, but there is a missing part within 10% and the data does not match, it is processed as a non-defective product, and it is judged that the data is smaller than that. Is determined as a crack in the tablet. Further, as shown in FIG. 62 (C), in the case where the data does not match due to a result such as a large or small size in spite of the fact that the stored data is geometrically identical to the actually measured data, the data is stored in advance. The data is stored in correspondence with the dimensional tolerance of the drug, and if the measured data is within the range of the stored dimensional tolerance, it is determined that the drug is suitable for prescription,
Those exceeding the dimensional tolerance are treated as if other chemicals were mixed.

When there is a defect in the difference between the prescription drug type and the packaged drug type, the same processing as described above is performed. Therefore, it is preferable to stop the packaging, display an error, and reset the prescription data.

<Tablet Discrimination Criteria Data> Next, the tablet discrimination reference data will be described. When performing tablet discrimination, comparative data serving as a reference is required. This tablet discrimination reference data is preferably stored in the following three forms. 1. The measured X-ray value of each tablet is stored without noise. 2. The measured X-ray value of each tablet is corrected with the transmittance and stored. 3. It is stored as a three-dimensional data size including the color information of each tablet. In addition, equations such as the volume, transmittance, shape, and size of each tablet can be stored.

FIG. 47 is a diagram in which the drug data is converted into three-dimensional data and stored. FIG. 47 shows the one corresponding to the apparatus in FIG. 10 and the one corresponding to FIG.

The data corresponding to the device shown in FIG. 10 is actually stored in the format shown in FIGS. 29 and 30, and the stored data is represented by a three-dimensional image. For this reason, since the data corresponding to the apparatus of FIG. 10 is expressed purely only by the amount of transmission of X-rays, the cross-sectional structure of the tablet cannot be determined. Therefore, it can be determined from the transmission amount data (stored in advance as actual measurement data).

The data corresponding to the apparatus shown in FIG.
Since it is stored as a three-dimensional image and appearance color data, it can be used as a key for determining a cross-sectional configuration and the like and for determining a layer thickness of an inner layer. This is shown in FIG.
This is because the X-ray tube rotates and data can be acquired from all directions, and the configuration of FIG. 42 and FIG. 43 can also be adopted.

<Coloring of Display> The data after the image processing is sometimes displayed on a monitor, and the state of the packed tablet is monitored in response to the tablet packing operation, so that there is no problem in the packing process. Or you can check.

At this time, data scanned by X-rays or the like is slice image data, and this data cannot be directly inspected. Therefore, when performing an audit, it is desirable that the appearance shape that a dispenser normally determines is three-dimensionally reproduced, and that the colors be faithfully reproduced.

Therefore, the imaging data of the tablet previously stored in the discrimination reference tablet storage data is read from the storage data matching the discrimination result from the discrimination reference tablet storage data, and the color information is reproduced three-dimensionally in the appearance. Appended to the data so that the identification of the appearance tablet can be determined.

Further, since a tablet having the same shape in white can be discriminated by comparing the inner layer structure, it is preferable to display the tablet cross-section laminated structure together with the display. In this case, in order to clearly display the stacking difference, the stacking difference is displayed by coloring with a pseudo color, a shading conversion table or the like. In the color coloring, since the direction of the light beam is determined, the color may be three-dimensionally expressed by forming a gradation from the direction of the light beam. Further, when performing the pass / fail determination, it is not necessary to perform the shape data and the color data together, and they may be determined separately.

<Identification Code Recognition> The data reproduced in the three-dimensional image can also recognize a medicine identification code predetermined according to the type of tablet. For example, in the tablets shown in FIGS. 30 and 41, since the identification code is engraved, the character portion can be reproduced as a depression.

Further, in a device equipped with a CCD camera, one side of a tablet can be caught by image data. In this case, recognition is performed only when identification information by paint other than engraving is directed to the surface. be able to. What is important here is that many of the medicines are white medicines, and the colors obtained by the CCD camera are slightly different. For this reason, when capturing the appearance image of a medicine as data with a CCD camera, unless the white adjustment is performed or the image data of the acquired white medicine is corrected, the data acquired by the CCD camera is immediately used as discrimination data. Can not. Further, the data acquired by the CCD camera is also affected by the hardware error of the camera itself, the brightness and direction of the light. Means to solve these problems include: Fix conditions such as light brightness and direction. 2. One or more reference whites are prepared, and the one or more reference whites are photographed under the fixed condition of the light, and the photographed data is stored as reference white data. 3. One or more types of reference white data stored in step 2 are compared with one or more types of previously stored reference white data, and the reference white data stored in step 2 is corrected, a light condition is changed, or a CCD is changed. Adjust camera operating conditions. 4. Steps 2 and 3 are periodically repeated. In steps 1 to 4, a reference medicine may be used instead of the reference white. By such means, the reliability of the data obtained from the CCD camera is improved.

If the color data and the identification code of the drug can be recognized by these means, the type of the drug can be determined based on this information. When determining the tablet type, the recognition of the drug identification code is first performed. If the tablet is discriminated, and the tablet that could not be discriminated is determined based on the shape, dimensions, color, inner layer configuration, and the like, it is possible to improve the determination accuracy and speed up the determination speed.

It is preferable that the identification code of the medicine is recognized by OCR.

Since a three-dimensional image or an image picked up by a CCD camera corresponds to image data, it is desirable to convert it to font information once.

In this case, a problem is that in the case of a three-dimensional image, the font information cannot be accurately read unless the front and back of the character are distinguished and the rotation of the character (vertical direction of the character) is not recognized.

Also, even in the case of the image data of the CCD camera, it is necessary to recognize the rotation of the character (vertical direction of the character).

In the case of a three-dimensional image, the data that can be identified as characters are limited to those engraved, and one line serving as a reference is used for 80% or more of the tablets to which the identification information is given by this engraving. .

The line includes a straight line having an external shape, a dividing line, a trademark mark of a pharmaceutical manufacturer, and the like, and characters are engraved along these reference lines.

For this reason, it is desirable to first identify the reference line, and to multiply and recognize the reference line by the character direction pattern.

In the case of character information having no reference plane,
For example, in the case of a carnigen tablet, “HLM” is engraved on substantially one center in the diameter direction on one side of the columnar tablet, and there is no reference line other than the circumference of the circle.

Even in the case of such a tablet, the character rotation direction can be determined based on one side of the character. According to this means, it can be adopted even in the case of image data of a CCD camera. In addition, since the identification information of these medicines is composed of alphanumeric characters and is partially provided with a trademark mark, it can be recognized by simple OCR software, and furthermore, it does not take much time to determine.

However, since the character size and the font to be used are slightly different, font information may be stored in association with the stored data in order to cope with this.

The details of each of the three-dimensional image data creation, filter processing, and image restoration processing will be described below.

<Creation of Three-Dimensional Image Data> FIGS. 34 and 4
As shown in FIG. 2, when the object is irradiated with X-rays from two directions and captured by two line sensors, the output data of the line sensor is related to the feed speed of the object in the z direction, and It can be replaced with three-dimensional image data.

When the X-ray beam of I ₀ passes through the inside of the object to be detected (tablet), the absorption rate of the X-ray varies depending on the composition, components, density and the like of each layer of the tablet.

The composition of the tablet composed of each layer is decomposed for each unit and considered, and u1, u2, u
Assuming that there is an attenuation coefficient of 3,..., Un, the X-ray dose passing through is expressed by the following equation.

[0107]

(Equation 15) Here, Δ is the length of the unit. By transforming this equation, the following equation is obtained.

[0108]

(Equation 16) Here, the left side is created by logarithmic conversion of the output of the X-ray detector (line sensor), and this is treated as the X-ray detection density.

In the above equation (16), u _i Δ may be considered as the density of each unit, and the sum thereof may be considered as the projection density. From this fact, if data in several projection directions is used, the internal density can be calculated. Can be obtained by

As shown in FIG. 48, there are four parts.
The X-ray absorption coefficient of the tissue ₁, X_Two, X_Three, X_FourWhen
And the true value of each is X₁= 1, X_Two= 3, X_Three= 2, X_Four
= 4, projection data in 6 directions from A to F and unknown
The relationship between the numbers is as follows.

[0111] Here, since the number of unknowns is four, it can be solved by using an appropriate four of the equations from A to F.

However, when this method is actually used, the number of unknown elements is large, and the number of simultaneous equations is enormous.

In consideration of the above problems, the solution of the equation can be obtained by an iterative method instead of obtaining it at once.

First, the average value of the whole is adopted as the estimated initial value of each element, and the values in each projection direction based on those estimated values are sequentially changed so as to match the data. As shown in FIG. 49 (a), the correction term, -1, 1 when comparing data in both AB directions is equally divided into two rows and added to the respective estimated values. Thereafter, if the same correction is made in comparison with the CD direction, the result is as shown in FIGS. 49 (b) and 49 (c), and the reconstruction is completed.

As another solution, there is a Fourier transform method.
The Fourier transform of f (x, y) is given by the following equation.

[0116]

[Equation 17]

When μ = 0, it is given by the following equation.

[0118]

(Equation 18)

Here, assuming that g (y) = ｘf (x, y) dx, g (y) is obtained by integrating (projecting) f (x, y) in the y-axis direction. Is the one-dimensional Fourier transform.

That is, the one-dimensional Fourier transform of the projection data in the y-axis direction is F (0, ν). This holds true for projection data in the y-axis direction.

In the case of the apparatus corresponding to FIG. 42, as shown in FIG. 50, when the projection data in the θ direction is considered, it can be expressed by the following equation.

[0122]

[Equation 19]

Here, δ (·) is a delta function, and a straight line: xcos θ + ysin θ = R indicates a straight line having the same inclination as the projection direction in which the distance from the center (origin) is R, and g ₀
(R) indicates the integration of the data on the line.

Expression 17 of the Fourier transform, when converted into polar coordinates (ρ, β), is expressed by the following expression.

[0125]

(Equation 20)

That is, the Fourier transform on the straight line in the β direction (passing through the origin) is equivalent to one-dimensional Fourier transform of the projection data of the straight line in that direction in the x, y space. Thus, by performing one-dimensional Fourier transform on the projection data and inversely transforming the data in the Fourier space, real space data can be obtained as in the following equation.

[0127]

(Equation 21)

Here, in the apparatus shown in FIG. 42, the back projection b _θ (x, y) of the projection data in the θ direction is represented by the following equation.

[0129]

(Equation 22)

The back-projected image is represented by the following equation by adding those in each of the θ directions.

[0131]

(Equation 23)

Further, g _θ (R) is a function having R as a variable when θ is fixed, and is obtained by inversely transforming the Fourier function F (ρ, θ). That is,

(Equation 24) Given by

(Equation 25) Becomes

By comparing Equations 25 and 20, it can be seen that ρ in the integrand of Equation 20 does not exist in the previous equation. From this, in order to obtain the data of Expression 20, it is sufficient to use | ρ | F (ρ, θ) instead of F (ρ, θ).

The multiplication in the Fourier space is a convolution in the real space. Also, the real space is sampled, and if the interval (pitch of the line sensor) is w,
According to the sampling theorem, R = 1/2 in Fourier space
The band is limited by w.

Therefore, when the inverse transform of the Fourier function is obtained (the real part is taken), the following equation is obtained.

[0136]

(Equation 26)

Here, r = nw, n = 0, ± 1, ± 2,
Then, the following equation is obtained.

[Equation 27]

This is a function as shown in FIG.
Further, as shown in FIG. 52, a function obtained by improving FIG. 51 can be used.

<Filter Processing> The tablet inspection data is subjected to filter processing such as noise and emphasis in order to be compared with the discrimination reference data.

Normally, the data detected by the X-ray detecting means includes noise, and when compared with the discrimination reference data, it is unlikely that they completely match. For this reason, the obtained data is compared with the outer diameter contour obtained by replacing the acquired data with three dimensions, the internal configuration pattern, and the like, and the judgment is made.

In the case of a sugar-coated tablet or the like, the outer diameter profile has a laminated structure, and 80% of the central part is a drug substance, which is protected by covering the outside with a sugar shell, or a drug substance in the intestine. In some cases, a protective agent is provided under the sugar coating to reach the skin.

In many cases, capsules and the like are formed in the form of powder such as granules.
Although the difference in line transmittance clearly appears, when compared with the discrimination reference data, the difference level increases.

[0143] Drugs having a characteristic outer diameter contour shape are so easy that they can be distinguished only by the outer diameter contour shape.

The detection data detected by the above-described device causes the following problem when the outer diameter contour is specified.

FIG. 33 shows a state where the tablet is irradiated with X-rays. A ray line radiated from the target of the X-ray tube intersects rays from upper and lower points depending on the angle of the target. These two light lines are the outline of the object to be detected, and only one of the two light lines is interrupted by the outline of the object to be detected. Is detected by For this reason, there is a real shadow where two light lines are cut off and a penumbra portion where one of the two light lines is cut off with respect to the detected object. And
This penumbra is detected as blurring of the outer diameter contour of the detected object.

In the blurred portion, the penumbra area fluctuates at the position of the object between the X-ray tube and the detector, and the penumbra area increases as the object approaches the X-ray tube. As a result, the resolution is increased, and the detection accuracy of the outer diameter of the detection object is improved.

However, when the object to be detected is detected by approaching the line sensor, the area of the penumbra is reduced and sharp image data can be reproduced. It is difficult to specify a subtle size. To compensate for this disadvantage, it is necessary to increase the resolution of the line sensor (to narrow the pitch).

In particular, the difference between a medical X-ray CT scanner and the present invention is that the size of an object to be detected is much smaller in the present invention, and that the discrimination accuracy for comparison should be set higher. . Against this background, tablets, especially sugar-coated tablets and capsules, are difficult to distinguish from each other in appearance (they have the same size and color appearance). It must be able to be determined from the thickness of the measurement, internal structure, laminated structure, etc.

Therefore, the contents required for the present invention require a high resolution and a level at which sharp external appearance data can be compared with stored data (stored discrimination data of an object to be detected).

To satisfy the above condition, the object to be detected is brought close to the X-ray tube, and the penumbra is processed to obtain sharp image data.

The penumbra varies with the thickness of the object to be detected and the transmittance, and the attenuation decreases in the direction away from the main shadow. Note that the amount of attenuation varies depending on the thickness, transmittance, and shape of the object to be detected, does not simply have a 特性 characteristic, and cannot be defined in inverse proportion.

The correction and restoration of the image data is intended to restore the image with high precision by restoring the distortion and blur generated in the process of generating and inputting the image and removing the noise. On the other hand, image enhancement is a process performed to further extract the information characteristics (the amount of information that the observer desires to acquire) of the image.

For example, the penumbra can be corrected for density conversion. In this penumbra portion, the amount of transmission of X-rays is reduced to about half, and if an image is displayed as it is, the edge portion of the detected object is blurred. Therefore, a process of correcting or darkening the density of this portion is performed. Make the outline clear.

In the case of deep correction, it is necessary to determine the boundary between the main shadow and the penumbra, and the boundary is determined by the density difference of the detection level.

In the method of thinly processing the penumbra, it is difficult to determine the boundary between the penumbra and the main shadow. However, a certain distance from the boundary between the transmissive part and the penumbra is defined as the penumbra. By doing so, the boundary between the penumbra portion and the main shadow portion can be determined. In the case where the shape of the tablet is cylindrical, the penumbra area is enlarged, and in the case of a tablet having a rugby ball-shaped cross section, the penumbra area is reduced.

In general, these density conversion processes use a conversion table.

FIG. 53 shows a simplified data table. Using such a data table, correction processing is performed by adjusting the density in the range of the penumbra.

The numerical values in the conversion table are converted into image data values into image data in which the image data values are dispersed as shown in the figure, and the image data values are rearranged to obtain the display image data values. In the case of X, the value of the conversion table is 480, and the value 1 of the image data corresponds to the numerical value of the penumbra. By processing the numerical values of the actually measured image in this manner, the captured image can be displayed clearly and more clearly than the actual image data.

The conversion table is applied according to the set intensity function, and the types shown in FIGS. 54 (a) to 55 (g) can be used as needed. When processing is performed using the linear function table shown in FIG. 54A, processing is performed in which lightness and darkness are further emphasized.
Data having no X-ray transmission difference is subjected to correction processing of the light-dark difference once using this function. Using the inversion function of FIG. 54 (b),
The light and dark areas are reversed. That is, white portions are replaced with black, and black is processed as white. When the piecewise linear function table of FIG. 54C is used, the penumbra can be excluded as a real image and the outline of the tablet can be emphasized as described above.
This is effective in removing blur and noise of a predictable image in advance. The pseudo contour function shown in FIG.
When the contour display of (g) is used, the image density difference can be continuously laminated stepwise. In addition, in the threshold processing of FIG.
The band processing is used when the depth of a three-dimensional object is expressed or the outline is emphasized by emphasizing only a certain brightness portion of the image color.

FIG. 56 shows the input / output relationship corresponding to these function tables, and shows that the r-level X-ray measurement data has been converted into the output Tr corresponding to the curve in the function table. When this function table is used, output is performed so as not to process the intermediate contrast while suppressing the output of the bright and dark portions.

In the output detected by the X-ray detection device and the line sensor, the level detected at each measurement terminal is not constant, and the detection level decreases toward both ends of the line sensor. Therefore, it is possible to correct the detection level difference using the function table.

Further, the output detected by the line sensor contains noise. This noise is partially emphasized due to irregular reflection or refraction of X-rays, etc., and is detected or not. As a result, the detection result greatly differs from the detection level of the neighboring sensor.
The result is different from the actual shape.

Therefore, it is necessary to correlate the actual shape with the detected data by processing the noise. The following methods are available as means for noise processing.

The first method is simple smoothing. If the noise gets on the condition that the output span is fine, the curve shape becomes rough,
If the output span is rough, the curve shape looks smooth even with some noise. When comparing noise and signal components,
Generally, the latter has the effect of attenuating components. Since the linear filter is given in the form of a convolution operation, it is expressed by the following equation.

[0165]

[Equation 28]

In general, the smoothing filter sets the value of h (x) to be large at the center (x = 0) and smaller as the distance from the center increases, so that the value of h (x) generally becomes 0 when the value exceeds a certain range (± W). Often do. In this case, the following equation is obtained.

[0167]

(Equation 29)

When averaging pixels simply, h (i, j)
Is represented by the following equation when represented by a 5 × 5 matrix.

[0169]

[Equation 30]

The case where the weighted average is used is as follows.

[0171]

(Equation 31)

In addition to these filters, there is a mode filter. The mode filter takes a density histogram of the filter section and outputs its mode value (mode value within the size of the filter).

There is also a median filter that arranges the density values of pixels in a filter section in order and outputs an intermediate value. The median filter is extremely effective in removing dot type noise. The median filter may use a value other than the intermediate value, and may take the maximum value or the minimum value.

In the case of an image having many noise components, it is difficult to process the image only with the filter. Therefore, noise can be removed without blurring the edge portion by using a dispersion filter.

The principle of such filtering can be realized by smoothing the edge portion without averaging over both sides, and it is necessary to detect the edge portion. As a means for detecting the edge portion, the sample variance value of the image density in the local region is effective as a means for detecting the presence of the edge.

Now, in FIG. 57, consider two adjacent areas, area 1 and area 2, which follow different density distributions in an image. The density distribution of the area 1 follows the probability density function f ₁ (x) of average μ ₁ and variance σ ₁ ² , and the density distribution of area 2 follows the probability density function f ₂ (x) of average μ ₂ and variance σ ₂ ^2. Shall be.

By taking a local small area in this image, FIG.
As shown in the figure, the proportions of the small areas including the areas 1 and 2 are represented by P ₁ and P ₂ (where 0 ≦ P ₁ , P ₂ ≦ 1, P ₁ + P ₂ =
Assuming 1), the density distribution of the small area follows the following equation.

[0178]

(Equation 32)

The average μ and the variance σ ² of f (x) are expressed by the following equations, respectively.

[0180]

[Equation 33]

[0181]

(Equation 34)

When μ ₁ ≠ μ ₂ , using P ₁ + P ₂ = 1,
By transforming equation 34, the following equation is obtained.

[0183]

(Equation 35)

Therefore, as P ₁ changes from 0 to 1, ie, as the relative positional relationship between the edge and the small area changes, σ ² changes as shown in FIG.

[0185] If

[Equation 36]

That is,

(37)

If the condition (1) is satisfied, the maximum point of σ ² exists in the section of 0 <P ₁ <1. From this,
When the small area is on the edge, the variance value increases, and the presence of the edge can be detected.

Without blurring the edge portion, as shown in FIG. 57, as a smoothing filter, a small area ABCD having a size of k × k around each point of interest (i, j) in the image.
(In the following formulas, to the R ₁ and R _4.) Take, consider the following types of filter that determines the output value according to the average value mu _i and sample variance value sigma _i ² concentration in each area .

The input image is [σ **] and the output image is [σ *].
*],

(38)

Here, the average value and variance in each neighboring area are
It is expressed by the following equation.

[Equation 39]

(Equation 40)

W _j is a weight determined by the value of the variance σ _j ² in R _j , and the sum is 1.

Several filters can be considered by the selection method of W _{j. As} a simple method, the average value of the minimum variance region is used as the filter output, and the average value of the minimum variance region is used as the output from the filter. Is what you do. That is, Wk = 1; sk2 ≦ sm2, and all m∈ (1, 2, 2,
3, 4), 0: k other than the above is selected.

Since it is sufficient that at least one of the four small areas does not include the edge of the image area, this area is selected based on the criterion of minimum variance, and the average value of the area is used as the filter output at the point of interest. It is assumed that. Since the switching of the selected small area is clearly performed on both sides of the edge, an extremely sharp edge output can be obtained,
On the other hand, even if the dispersion is slightly changed, the area is switched, and spike-like noise is likely to be generated near the edge. Therefore, a combination with a median filter is desirable.

The outline and edge portions of the image pickup body can be sharpened by emphasizing high frequency components. The differential operator used for this purpose indicates a local change in the density value of the image, and is obtained by the following equation.

[0195]

[Equation 41]

This operator is a local differential operator which is linear and does not exist at a position.

Usually, n is only 1 or 2, and when n = 1, 2 of (∂ / ∂x) and (∂ / ∂y) is used.
And the differential value Δ obtained by combining the two and θ in the direction
Is represented by the following equation.

[0198]

(Equation 42)

[Equation 43]

In the case of a digital image, these differential calculations are performed using differences, and the first derivatives of x and y can be expressed by the following equations.

[0200]

[Equation 44]

[Equation 45]

Δ is expressed by the following equation.

[Equation 46]

The second derivative is defined by the following equation in each of the x and y directions.

[0203]

[Equation 47]

[Equation 48]

In addition, the following definitions can be used.

[Equation 49]

[Equation 50]

(Equation 51)

Further, as a first order differential value Δ,

(Equation 52)

(Equation 53)

(Equation 54) Can be used instead of equation 46.

Laplacian is

[Equation 55] Where Laplacian in a digital image is

[Equation 56] Is defined as

By transforming this equation,

[Equation 57] It can be seen that the difference between the average value of the neighboring area and the pixel value is represented.

Since these differential operations are for detecting a change in density, high-frequency noise is also emphasized. Therefore, a reduction in high-frequency noise is measured in combination with a smoothing filter to detect a change in density value of a signal component. For this purpose, the following calculation using 3 × 3 neighboring pixels is used.

[0209]

[Equation 58]

[0210]

[Equation 59]

[Equation 60]

[Equation 61]

(Equation 62)

The following 5 and 6 are for obtaining differential images in two directions of x and y.

[Equation 63]

[Equation 64]

[0212] When the filtering processes 5 and 6 are performed,
Negative values often occur as pixel values,
In displaying, it is preferable to take an absolute value or add an appropriate bias value.

Further, an image obtained by adding the absolute values of the differential values in both the x and y directions can be used as the differential image.

An example of Laplacian corresponding to the second derivative is

[Equation 65] Are used. On the other hand, when displaying a processed image, appropriate bias addition and absolute value processing can be performed.

Since the Laplacian emphasizes a wide-range component, it is possible to change the image to a sharp one by subtracting the Laplacian image from the original image, and there is also a blur recovery effect.

A filter for realizing this is:

[Equation 66] Are used.

In addition to these filters, there are low-pass or high-pass filters. In addition, the meaning of the low band and the high band here means that, for example, a white portion of the black-and-white original image is a low band and a black portion is a high band.

The low-pass filter is set so as to reduce the high-frequency component H (μ, ν) black component, and includes the following short filters and Butterworth filters.

The short filter has a pass band and a non-pass band in one.
Divided by frequency, R ₀Is the cutoff frequency
When

[Equation 67] Thus, the filter function has a shape as shown in FIG.

A Butterworth filter does not have a sharp cutoff like a short filter, but gradually reduces a pass band. The filter characteristic is expressed by the following equation.

[Equation 68]

FIG. 58 shows the form of the filter function when n = 1.
(B).

Next, the high-pass filter will be described. H (μ, υ) is set so as to reduce the low frequency band. As described above, there are a short filter and a Butterworth filter.

The rectangular filter is represented by the following equation.

[Equation 69]

The Butterworth filter is represented by the following equation.

[Equation 70]

The filter functions are shown in FIG.
(B). (B) showing a Butterworth filter is a case where n = 1.

These high-pass and low-pass filters make the imaging screen dark or too bright depending on the thickness, shape, and absorptance of the object when the sensor detects a constant dose emitted from the X-ray tube. Such factors make it difficult to determine the type of drug. However, even if such a filter is taken out of a certain amount of contrast range, it can be corrected to a state in which the medicine can be identified.

<Image Restoration> Since the X-ray imaging data corrected by the filter is finally output as monitor or printer data, the blurred image is not preferable for judging the type of tablet.

The path for monitoring an image follows the path as shown in FIG. 60. However, since it is normal for the monitor to be blurred or superimposed with noise, the above-described filter or the like is used. It must be at a level that can be monitored.

In the case of additive noise, the original image is represented by f (x,
y), noise is n (x, y), and the observation / transmission arithmetic unit is B
Then, the monitor image is formally obtained by the following equation.

[Equation 71]

When B is linear,

[Equation 72] It can be expressed as.

Here, b (x, α, y, β) is a point spread function, and is a response of an impulse at a point (α, β).
Furthermore, when the position (α, β) of this function is irrelevant,
It can be expressed by the following equation.

[Equation 73]

The restoration problem means that the monitor image g is converted to the original image f
Therefore, if the above equation 63 is Fourier-transformed,

[Equation 74] In order to estimate F (μ, ν) from G (μ, ν), it is necessary to determine the following equation M (μ, ν).

[Equation 75]

If the object is moving, the monitor image data indicates that the x and y components of the displacement are α
(T), β (t), and T as the exposure time, when there is no noise,

[Equation 76] Becomes

If this is Fourier-transformed,

[Equation 77] And the image conversion function B is obtained by the following equation.

[Equation 78]

When moving at a constant speed a in the x direction, that is, when α (t) = at and β (t) = 0,

[Expression 79] Becomes

If B (μ, ν) is known and no noise is present, the estimated value of F (μ, ν) is

[Equation 80] And it is sufficient.

However, when the value of B (μ, ν) is 0 or close to 0, it is often inappropriate to perform a division operation with that value, and this is a problem especially when noise is present.
That is,

[Equation 81] In general, B (μ, ν) has a larger attenuation when (μ, ν) leaves the origin and is closer to 0 (in a high frequency component region), compared to N (μ, ν). Since the noise does not depend much on the frequency, when B (μ, ν) approaches 0, the value of N (μ, ν) / B (μ, ν) becomes relatively large, and the result is not desirable. For this reason, in the high frequency range, the value is used as it is (ie, B (μ, ν) = 1) without performing division by B (μ, ν) so that the result does not become inconvenient.

[0238] That is, 1 for the inverse filter function, as the reciprocal until the frequency range of a radius R _0, uses the inverse to the frequency range of a radius R _0, more frequencies
Is preferable.

(Equation 82)

Further, when there is noise, a linear restoration function is used, and as a measure of optimal estimation restoration, a measure that minimizes the mean square error is used. Is calculated.

[Equation 83] Here, S _nn (μ, ν) is the Fourier spectrum of the noise, and S _ff (μ, ν) is the Fourier spectrum of the signal.

When there is no noise, that is, when S _nn = 0,

[Equation 84] And matches the inverse filter.

This means that when noise is present, 1 / B (μ, μ,
ν) has been modified.

If the information can be measured for correcting the non-uniformity of the recording / observing system and correcting geometric distortion (the center and both sides of the line sensor), restoration and restoration are not difficult.

There is non-uniformity due to the position of the measurement system,
Assuming that it is i (x, y), originally f (x, y)
Should be observed, because of the non-uniformity,

[Equation 85] Is obtained.

In order to correct this non-uniformity, for example, if an image f (x, y) = c (constant) having a white constant density value is fetched and g _c (x, y) is obtained, an input For image g (x, y),

[Equation 86] It should be corrected as

The geometric linear distortion only needs to be able to provide the correspondence between the measured pixel point and the correct position without distortion, so that the point (x, y) = (r _j , s _j ) is the original coordinate. Assuming that points (x ′, y ′) = (μ _j , v _j ), these points are

[Equation 87] , And each coefficient is obtained, the distortion can be corrected.

However, since a point exists only at a digital position on a grid in a digital image, a corresponding point often does not exist on the grid. In this case, correction or approximation is performed based on the density value of the grid point. Adjustments such as using values are required.
These typical distortions are as shown in FIG.

<Embodiment using infrared rays as transmission rays>
In the above embodiment, X-rays are used as transmission rays.
Infrared light can be used. The infrared transmission characteristics
It is much weaker than X-rays and affects the material of the substance to be transmitted. For example, in the case of paper, the infrared transmission limit is 20 m in thickness.
m. The packaging band in which the tablets are packaged by the tablet packaging device can transmit infrared rays. By measuring the amount of infrared light attenuated through the packaged tablet, the number, shape, and the like of the packaged tablet can be determined.

Examples of the infrared ray generating device include a near infrared ray irradiating strobe, an infrared ray lamp for irradiating light from a halogen lamp through a filter transmitting only an infrared ray component, a near infrared ray emitting LED array composed of a large number of LEDs, and the like. Can be used.

FIG. 63 shows an infrared ray generator 101 composed of an infrared lamp. In FIG. 63, reference numeral 102 denotes a halogen lamp. The halogen lamp 102 includes a reflector 103 for reflecting light in one direction. The near-infrared transmission filter 1
04 and a slit plate 105. As the near-infrared filter 104, Si, GaAs, InP, Ga
It can be configured using a substance that is transparent in the infrared region, such as P, ZnSe, and ZnS. The slit plate 105 is formed by forming a slit in an aluminum plate. Infrared ray generator 1
The fan 01 is provided with a fan 106 for radiating heat generated by the halogen lamp 102 and cooling it so that the internal temperature does not rise abnormally. Halogen lamp 102
A photodiode 108 and a control device 109 are provided via a heat-insulating reflection plate 107 to keep the amount of light of the halogen lamp 102 constant.

FIGS. 64 and 65 show an infrared detector 111 for detecting infrared rays emitted from the infrared generator 101 and transmitted through the tablet. This infrared detector 1
Reference numeral 11 includes a large number of sensor elements 112 arranged in a predetermined arrangement shape. The sensor element 112 has a P-type silicon semiconductor substrate 113 and a Schottky junction photoelectric conversion layer 114 formed on the substrate 113. As the photoelectric conversion layer 114, a metal such as platinum, palladium, and iridium, or a metal formed with a metal silicide compound with the metal is used. A guard ring 115 is provided around the photoelectric conversion layer 114 by an n − -type region for alleviating electric field concentration around the photoelectric conversion layer 114 and preventing dark current. 116 is the photoelectric conversion layer 1
14 is an n + type region of a transfer gate which transfers signal charges from the vertical shift register 117 to the vertical shift register 117. Gate electrode 118 and n-type buried channel 119 are CSD
Of the vertical shift register 117 of FIG. Reference numeral 120 denotes a field insulating film made of a silicon oxide film for element isolation and insulation. Reference numerals 121 and 122 denote interlayer insulating films, and these interlayer insulating films are formed of an insulator such as an oxide film. Behind the photoelectric conversion layer 114, the aluminum reflective film 1
The light receiving sensitivity is improved by reflecting infrared light transmitted through the photoelectric conversion layer 114 without being absorbed.

The gate electrode 118 of each sensor element 112 on the horizontal line is connected to the transfer gate scanner 12.
4 and to the CSD scanner 125. As a result, the gate electrode 118
The electrode of the transfer gate 116 is also used as the transfer electrode of the CSD. Each sensor element 112 on the vertical line
Of the vertical shift register 1
17. Each vertical shift register 117
Are connected to a horizontal shift register 126, which is connected to an output unit 127.

The operation of the infrared detector having the above configuration will be described.
Light incident from the surface side reaches the Schottky junction photoelectric conversion layer 114 and is photoelectrically converted. The generated optical signal charges are accumulated in the Schottky junction. One of the scanning lines 128 is selected by the transfer gate scanner 124, and the transfer gate scanner 12 is connected to the gate electrode 118 of one horizontal line connected to the scanning line 128.
4, a read pulse is applied. As a result, the optical signal charges accumulated in the Schottky junction are transferred to the n-type buried channel 119. At the same time, the photoelectric conversion layer 11
4 is reset and accumulates newly generated optical signal charges until the next read pulse is applied. CSD
The gate electrode 1 from the scanning line 128 by the scanner 125
When a vertical transfer pulse is applied to 18, the optical signal charges are transferred in the vertical direction and input to the horizontal shift register 126. In the horizontal shift register 126, the optical signal charges are transferred in the horizontal direction, and are read out from the output unit 127 as video signals of one horizontal line. Subsequently, a horizontal pulse selected by the transfer gate scanner 124 is shifted one step at a time, a read pulse is applied, and the same operation is repeated to obtain a desired video output.

As described above, when the gate electrode 118 also serves as the electrode of the transfer gate 116 for reading out signal charges and the transfer gate of CSD for transferring signal charges, a vertical transfer pulse is applied to the gate electrode 118. In order to prevent the transfer gate 116 from opening during the transfer, the threshold voltage of the transfer gate 116 is set so as to be at least the high-level voltage of the vertical transfer pulse.

The photoelectric conversion layer 114 formed of a Schottky junction can detect a light component having energy equal to or higher than the height of the Schottky barrier. For example, in the case of a Schottky junction between platinum silicide (ptSi) and p-type silicon, a light component having a wavelength of about 5.6 μm or less can be detected.

In order to prevent the conversion efficiency of the photoelectric conversion layer 114 and the output from being saturated with incident light, the photoelectric conversion layer 11
It is preferable to provide an n-type impurity introduction region 129 at the Schottky junction between P.4 and p-type silicon semiconductor substrate 113. As a result, even when detecting an object having a large amount of infrared radiation, the light sensitivity is reduced by simply reducing the voltage of the signal readout pulse, and the output saturation can be prevented to enable photographing. As a result, the light sensitivity of the infrared detector can be adjusted very easily so that the output is not saturated in accordance with the amount of incident light. Phosphorus (P), arsenic (As), or the like is used as the n-type impurity.

FIG. 66 compares the relationship between the reset voltage and the output signal in the case where the Schottky junction between the photoelectric conversion layer 114 and the p-type silicon semiconductor substrate 113 is provided with and without the n-type impurity introduction region 129. Things.
In the same figure, a curve a shows a case where the n-type impurity introduction region 129 is not provided, and a curve b shows a case where the n-type impurity introduction region 129 is provided.

When capturing the amount of infrared light transmitted through the packaged tablet, the higher the sensitivity of conversion by the photoelectric conversion layer 114, the more advantageous the depth and sharpness of the captured image.

Further, since infrared light has a low passing power, if a plurality of packed tablets overlap, a captured image is blurred or the identification of tablets is hindered. Is preferred.

[0259]

As is apparent from the above description, according to the invention as the first to third means, the work of collating and inspecting the contents of the tablet package to be handed to the patient with the dispensing data and the medicine information is performed. In this case, the inspection of packaged tablets is simple and quick, and the reliability is high. In addition, the inspection labor of the dispenser can be reduced. Further, according to the invention as the fourth means, since a means for creating video data of the packaged tablets is provided, the visual check is easier than the direct check of the actual drug, and the accuracy of the inspection is improved. Can be enhanced.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of a tablet packaging machine provided with a tablet inspection device according to a first embodiment of the present invention.

FIG. 2 is a cross-sectional view of a fixed target X-ray tube.

FIG. 3 is a cross-sectional view of a target rotary X-ray tube.

FIG. 4 is a circuit diagram of a continuous X-ray generation circuit.

FIG. 5 is a pulse X-ray generation circuit diagram.

FIG. 6 is a photon spectrum diagram of a tungsten target.

FIG. 7 is a photon spectrum diagram of a tungsten target.

FIG. 8 is a photon spectrum diagram of a molybdenum target.

FIG. 9 is a tube voltage waveform diagram.

FIG. 10 is a perspective view of the tablet inspection device of FIG. 1;

FIG. 11 is an exploded perspective view of a line sensor (ionization chamber detection device).

FIG. 12 is a waveform diagram showing ionization chamber pulse response.

FIG. 13 is a graph showing the relationship between gas pressure and applied voltage in an ionization chamber.

FIG. 14 is a schematic configuration diagram of a line sensor (semiconductor detection device).

FIG. 15 is a perspective view of a packaging bag illustrating measurement points.

FIG. 16 is a perspective view of the tablet detection device of FIG. 10 including a separation device.

FIG. 17 is a plan view for explaining pitches of assault points.

FIG. 18 is a graph showing transmission amount output data of a line sensor.

FIG. 19 is a graph showing transmission amount output data of a line sensor.

FIG. 20 is a graph showing transmission amount output data of a line sensor.

FIG. 21 is a graph showing transmission amount output data of a line sensor.

FIG. 22 is a graph showing transmission amount output data of a line sensor.

FIG. 23 is a graph showing transmission amount output data of a line sensor.

FIG. 24 is a three-dimensional graph showing X-ray attenuation characteristics.

FIG. 25 is a plan view showing measurement points in the case of defective packaging.

FIG. 26 is a graph showing an attenuation width of transmission amount output data of a line sensor.

FIG. 27 is a graph showing the attenuation width of transmission amount output data of a line sensor.

FIG. 28 is a diagram showing three-dimensional image data.

FIG. 29 is a three-dimensional graph showing an example of determination reference data.

FIG. 30 is a three-dimensional graph showing an example of determination reference data.

FIG. 31 is a perspective view of a mark providing device.

FIG. 32 is a perspective view of a tablet inspection device according to a second embodiment of the present invention.

FIG. 33 is a perspective view illustrating the situation of shadows caused by X-rays.

FIG. 34 is a side view of the tablet inspection device according to the third embodiment of the present invention.

FIG. 35 is a graph showing transmission amount output data in the X and Y directions of the line sensor.

FIG. 36 is a graph showing three-dimensional image data.

FIG. 37 is a side view of the tablet inspection device according to the fourth embodiment of the present invention.

FIG. 38 is a side view of the tablet inspection device at the time of 5 ° rotation.

FIG. 39 is a side view of the tablet inspection device at the time of rotation by 10 °.

FIG. 40 is a side view of the tablet inspection device at the time of rotation by 15 °.

FIG. 41 is a diagram showing three-dimensional image data.

FIG. 42 is a side view of the tablet inspection device according to the fifth embodiment of the present invention.

FIG. 43 is a side view of the tablet inspection device according to the sixth embodiment of the present invention.

FIG. 44 is a perspective view of the tablet inspection device of FIG. 43.

FIG. 45 is a flowchart showing a measurement operation.

FIG. 46 is a flowchart showing a data processing operation.

FIG. 47 is a diagram showing an example of determination reference data.

FIG. 48 is a view showing an X-ray absorption coefficient of a tissue composed of four tissues.

FIG. 49 illustrates an iterative method.

FIG. 50 is a diagram illustrating a Fourier transform method.

FIG. 51 is a diagram showing a function of g (k, w).

FIG. 52 is a view showing a function obtained by improving FIG. 51;

FIG. 53 is a view showing a method of density conversion processing.

FIG. 54 is a view showing various function tables.

FIG. 55 is a view showing various function tables following FIG. 54;

FIG. 56 illustrates an input / output relationship.

FIG. 57 is a diagram showing two adjacent small regions in an image and their variance values in density;

FIG. 58 is a view showing a filter function.

FIG. 59 is a view showing another filter function.

FIG. 60 is a diagram showing a route for monitoring an image.

FIG. 61 is a diagram showing an example of distortion.

62A is a perspective view showing (A) stored data of a tablet, (B) is measured data of a tablet having a defective portion, and (C) is a perspective view showing measured data of a tablet smaller than the stored data. FIG.

FIG. 63 is a cross-sectional view showing an infrared ray generating device in an embodiment using infrared rays as transmission rays.

FIG. 64 is a plan view of an infrared inspection apparatus according to an embodiment using infrared rays as transmission rays.

65 is a cross-sectional view of the infrared inspection apparatus of FIG. 64 taken along line AA.

FIG. 66 is a graph showing a relationship between a reset voltage and an output signal in a case where an Schottky junction is provided with an n-type impurity introduction region (a) and in a case where it is not provided (b).

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Tablet inspection apparatus 2 Tablet packing machine 12a Packaging bag (inspection object) 14 X-ray tube (transmission line generation means) 16 Line sensor (transmission line detection means) 17 Marking device (marking means) 18 Controller (number of tablets) Determination processing means, comparison determination processing means, tablet shape determination processing means, storage means) 40 CCD camera (imaging means) 101 infrared ray generating device (transmitted ray generating means) 111 infrared ray detecting device (transmitted ray detecting means)

Claims (13)

[Claims] 1. A tablet inspection apparatus for inspecting a tablet packaged in a tablet package based on prescription data, comprising: a transmission line generating means for generating a transmission line and irradiating the tablet package; Transmission line detection means for detecting the transmission line generated by the means, tablet number determination processing means for determining the number of tablets packed in the tablet package from the detection result of the transmission line detection means, and from the prescription data The tablet number data to be packaged in the tablet package is extracted, the tablet number data is compared with the number of tablets determined by the tablet number determination processing means, and the tablets are packaged according to the prescription data. A tablet inspection device comprising: a comparison determination processing unit that determines whether or not the tablet is present. 2. A tablet inspection apparatus for inspecting tablets packaged in a tablet package based on prescription data, comprising: a transmission line generating means for generating a transmission line and irradiating the tablet package with the transmission line; Transmission line detection means for detecting the transmission line generated by the means, tablet shape determination processing means for determining the shape of the tablet packaged in the tablet package from the detection result of the transmission line detection means, for each type of tablet A storage unit for storing the tablet shape data; extracting a type of tablet to be packaged in the tablet package from the prescription data; calling tablet shape data corresponding to the type of tablet from the storage unit; Comparing the shape data with the shape of the tablet determined by the tablet shape determination processing means, and comparing and determining whether or not the tablet is packaged according to the prescription data. For example was the tablet inspection apparatus. 3. A tablet inspection apparatus for inspecting a tablet packaged in a tablet package based on prescription data, comprising: a transmission line generating means for generating a transmission line and irradiating the tablet package with the transmission line; Transmission line detection means for detecting the transmission line generated by the means, tablet shape determination processing means for determining the shape of the tablet packaged in the tablet package from the detection result of the transmission line detection means, for each type of tablet A storage unit for storing the tablet shape data; extracting a type of tablet to be packaged in the tablet package from the prescription data; calling tablet shape data corresponding to the type of tablet from the storage unit; Tablet number discriminating processing means for discriminating the number of tablets packaged in the tablet package by associating shape data with the shape of the tablet discriminated by the tablet shape discriminating processing means; The tablet type and the number of tablets to be packaged in the tablet package are extracted from the tablet data, and the tablet number data is compared with the number of tablets determined by the tablet number determination processing means. A tablet inspection device comprising: a comparison determination processing unit that determines whether the same tablet is packaged. 4. The tablet according to claim 1, wherein the tablet number discriminating processing means performs processing as defective packaging when data detected by the transmission ray detecting means includes data that cannot be converted into a number. Inspection equipment. 5. The tablet according to claim 2, wherein the tablet shape discrimination processing means processes the data as a defective package when there is data whose shape cannot be recognized in the data detected by the transmission line detection means. Inspection equipment. 6. The tablet inspection apparatus according to claim 4, further comprising a marking means for marking a corresponding tablet package when a defective package is determined. 7. The tablet inspection device according to claim 1, wherein said transmission line detection means detects a transmission line transmission amount. 8. The tablet inspection device according to claim 1, wherein the transmission line detection unit detects the shadow as a tablet shadow. 9. A tablet inspection apparatus for inspecting tablets packaged in a tablet package based on prescription data, comprising: a transmission line generating means for generating a transmission line and irradiating the tablet package with the transmission line; Transmission line detection means for detecting transmission lines generated by the means, storage means for storing image data for each type of tablet, and detection results of the shape of the tablets packaged in the tablet package by the transmission line detection means Tablet discrimination processing means for discriminating the type of the tablet from the storage means, tablet image data corresponding to the type of the tablet is called from the storage means, and the image data is superimposed on the detection data of the transmission line detection means to produce video data A tablet inspection device comprising: video data creation means for creating a tablet. 10. The tablet inspection device according to claim 9, further comprising display means for displaying the video data. 11. The tablet inspection apparatus according to claim 9, further comprising storage means for storing the video data in association with prescription data. 12. A transmission means comprising a plurality of said transmission line generation means and said transmission line detection means provided in different directions of irradiation, and reproduction means for reproducing detection data of said transmission line detection means in three dimensions. The tablet inspection device according to any one of claims 3 and 9. 13. The transmission line generating means is rotatably provided around a tablet package so that an irradiation direction can be changed,
The tablet inspection device according to any one of claims 1, 2, 3, and 9, further comprising a reproducing unit that stereoscopically reproduces data detected by the transmission line detecting unit.