JP7388871B2 - Judgment device, judgment method and judgment program - Google Patents

Description

The present invention relates to a determination device, a determination method, and a determination program.

2. Description of the Related Art Conventionally, in the field of logistics, techniques are known for determining a packaged product from image data of packaging materials. Such technology includes, for example, capturing an image of the product display on the surface of the packaging material, using a computer to identify the product name based on the obtained image data, and then transmitting product information corresponding to the product name via a network. A technique is known in which information is acquired from a server using a computer (for example, see Patent Document 1). This technology is used to create labels that display identification codes for products and attach them to packaging materials.

Japanese Patent Application Publication No. 2003-104554

However, the conventional techniques described above may not be able to determine whether the product actually contained in the packaging material is correct.

An object of one aspect of the present invention is to provide a technique that can determine whether an object contained in a packaging material is correct or not.

In order to solve the above problems, a determination device according to one aspect of the present invention includes an image data acquisition unit that acquires image data obtained by imaging a packaging material, and a determination device that determines the weight of the packaging material containing a target object. The apparatus includes a weight data acquisition section that acquires weight data, and a determination section that refers to the image data and the weight data to determine whether or not a desired object is contained in the packaging material.

In addition, in order to solve the above problems, a determination method according to one aspect of the present invention includes a step of acquiring image data obtained by imaging a packaging material, and a weight related to the weight of the packaging material containing a target object. The method includes the steps of acquiring data, and making a determination as to whether or not a desired object is contained in the packaging material by referring to the image data and the weight data.

Furthermore, in order to solve the above problems, a determination program according to one aspect of the present invention is a determination program for causing a computer to function as the above determination device, the determination program including the image data acquisition section, the weight data acquisition section. and a determination program for causing a computer to function as the determination section.

According to one aspect of the present invention, it is possible to provide a technique that can determine whether an object contained in a packaging material is correct or not.

1 is a diagram schematically showing an example of the configuration of a determination device according to Embodiment 1 of the present invention. 1 is a block diagram showing an example of a functional configuration of a determination device according to Embodiment 1 of the present invention. FIG. FIG. 3 is a diagram for explaining an example of a weight data table according to Embodiment 1 of the present invention. 3 is a flowchart illustrating an example of a process flow in a method for determining a packaged item according to Embodiment 1 of the present invention. FIG. 3 is a diagram schematically showing an example of an image displayed on a liquid crystal panel in Embodiment 1 of the present invention. FIG. 3 is a diagram schematically showing an example of the functional configuration of a determination device according to Embodiment 2 of the present invention. It is a figure which shows typically an example of the functional structure of the determination apparatus based on Embodiment 3 of this invention.

[Embodiment 1]
Hereinafter, one embodiment of the present invention will be described in detail. First, the overall configuration of a determination device according to an embodiment of the present invention will be described.

[Overall configuration of determination device]
FIG. 1 is a diagram schematically showing an example of the configuration of a determination device according to Embodiment 1 of the present invention. The determination device 10 includes a scale 1, a camera 2, a liquid crystal panel 3, a notification lamp 4, and a switch 5, as shown in FIG. Conveying devices 21 and 22 are arranged on both sides of the scale 1, respectively. The transport device 21 is a device for transporting the package 20 to the scale 1, and the transport device 22 is a device for transporting the package 20 from the scale 1.

A stopper 11 is disposed on one side edge of the scale 1 that is not adjacent to the transport devices 21 and 22. The stopper 11 is a protrusion extending along the conveying direction of the conveying devices 21 and 22, and protrudes above the top surface of the scale 1.

The scale 1 is a digital scale for measuring the weight of the package 20. The scale 1 is arranged at a position where the top surface of the scale 1 is substantially flush with the top surfaces of the transport devices 21 and 22.

Camera 2 is placed on the opposite side of scale 1 with stopper 11 in between. When the scale 1 is viewed from the front, the camera 2 is placed at a position to photograph the back of the packaged item 20.

The liquid crystal panel 3 is disposed on the other side of the stopper 11 and at a sufficiently high position above the camera 2, facing toward the scale 1 side.

The notification lamp 4 is arranged at a sufficiently high position along with the liquid crystal panel 3. The notification lamp 4 is a device for indicating information regarding the progress status of the judgment work of the packaged item 20, and is composed of three colored lamps, for example, red, yellow, and green in order from the top.

The switch 5 is a switch operated by a worker regarding the judgment of the package 20. The switch 5 is disposed at a position that can be operated by the worker during the determination work, for example, at a position below the notification lamp 4 and to the side of the camera 2. The switch 5 has two buttons, for example, an A button and a B button. The A button is a button for the worker to confirm the normal determination process. The B button is a button for the operator to check for abnormalities in the determination process.

The conveyance devices 21 and 22 are devices for conveying the package 20 in one direction in a plane, and are, for example, belt conveyors or roller conveyors. The direction in which the package 20 is transported from the transport device 21 to the scale 1 and the direction in which the package 20 is carried out from the scale 1 by the transport device 22 are both in the same direction on the same straight line.

Note that the package 20 is composed of a packaging material and an object accommodated therein. In this embodiment, the packaging material is a cardboard box, the target object is a valve, and the packaging object 20 is a sealed cardboard box containing a predetermined valve. The cardboard box is sealed and the valves contained within are not visible from the outside.

A label is formed on the back side of the package 20. The label is a part on which at least the product name of the object (bulb) is printed, and includes information such as the brand name of the valve, the product name of the valve, and the number of valves that should be contained in the package 20. (See FIG. 5).

[Functional configuration of determination device]
FIG. 2 is a block diagram showing an example of the functional configuration of the determination device in this embodiment. As shown in FIG. 2, the determination device 10 includes a control section 100 and a storage section 104 as functional components. The control section 100 is electrically connected to the camera 2, the scale 1, the input section 105, the liquid crystal panel 3, and the notification section 106.

The control section 100 includes an image data acquisition section 101, a weight data acquisition section 102, and a determination section 103. The image data acquisition unit 101 acquires from the camera 2 image data obtained by the camera 2 capturing an image of the packaging material of the packaged item 20 . The weight data acquisition unit 102 acquires from the scale 1 weight data obtained by the scale 1 weighing the packaged item 20 .

The determination unit 103 refers to the image data acquired by the image data acquisition unit 101 and the weight data acquired by the weight data acquisition unit 102. The determining unit 103 determines whether or not a desired valve is housed in the cardboard box, based on information or a program stored in the storage unit 104, for example.

The storage unit 104 has information regarding the determination of the package 20. The storage unit 104 is configured with a known recording medium, such as a ROM (Read Only Memory) or a RAM (Random Access Memory).

The information stored in the storage unit 104 is, for example, a model that outputs output data according to the image data when the image data is input. Further, the output data corresponding to the image data is, for example, information regarding the type of valve. The information regarding the type of valve is, for example, the information printed on the label described above, and in this embodiment, is the brand name of the valve and the product name of the valve.

The model that outputs the output data need only be able to output output data corresponding to the image data described above from the image data of the package 20. Examples of such models include neural networks and support vector machines. Examples of neural networks include convolutional neural networks (CNNs), recurrent neural networks (RNNs), and fully connected neural networks.

When using a neural network, input data may be processed in advance for input to the neural network. For such processing, in addition to one-dimensional arraying or multidimensional arraying of data, a technique such as data augmentation can be used.

In addition to neural networks and support vector machines, models include clustering, inductive logic programming (ILP), genetic programming (GP), or Bayesian networks (BN). Network) may also be used.

The model may be a trained model trained using teacher data. The teacher data includes image data captured by the camera 2 and output data corresponding to the image data. The output data included in the training data may be data that can be used for determining the package 20, and includes the number of valves that should be contained in the package 20, in addition to the above-mentioned valve brand name and valve product name. , may be included.

The trained model is formed by causing a neural network to learn a sufficient number of teacher data and determining path weights according to the image data described above. Examples of algorithms for training the model include backpropagation and ID3.

Techniques for recognizing images in image data are not limited to the above examples. The technology may be a technology such as optical character recognition (OCR) that converts image data into a character code that can be used by a computer and identifies it as a character. Alternatively, the technique may be a technique for identifying the image information itself from the characteristics of the image data.

Further, the storage unit 104 stores a collation table. The comparison table stores information on the brand name, product name, and number of valves contained in the packaging material. The data of the collation table may be stored in the storage unit 104 prior to the determination process.

Further, the storage unit 104 stores work schedule data. The work schedule data includes information on the product name and number of valves in the package 20 to be subjected to the determination process by the determination device 10. The work schedule data may be read by the control unit 100 by, for example, having the worker photograph information corresponding to the work schedule data using the camera 2 . Alternatively, the work schedule data may be directly transmitted by the administrator of the determination device 10 to the control unit 100 via wireless communication or the like.

Furthermore, a weight data table is stored in the storage unit 104. FIG. 3 is a diagram for explaining an example of the weight data table. As shown in FIG. 3, the weight data table includes, for example, the product name of the valve, the number of valves to be accommodated in the package 20, and the weight at the time of packaging (the product of the weight of each individual valve and the number of valves). information is stored.

The input unit 105 is configured to input a signal to the determination device 10, and is, for example, the switch 5 described above. The input unit 105 may further include a device capable of inputting operator operations, such as a keyboard or a touch panel. The reporting unit 106 is configured to report information regarding the determination process, and is, for example, the aforementioned notification lamp 4. The reporting unit 106 may further include a device capable of reporting information regarding the determination process to the outside, such as a wireless device for reporting to an administrator or a speaker for generating a sound such as an alarm.

[Example of judgment of packaged items]
The transport device 21 transports the package 20 to the scale 1. The packaged item 20 is placed on the scale 1 from the transport device 21. As described above, the package 20 is a cardboard box containing a valve. The package 20 is placed on the scale 1 with the label facing the camera 2. Furthermore, the packaged item 20 is placed on the scale 1 at a position where one bottom edge of the side wall portion on which the label is printed comes into contact with the stopper 11. The transportation of the package 20 and the installation on the scale 1 may be carried out by a worker or by a robot.

The camera 2 is adjusted in advance so that it can photograph the label of the packaged item 20 as an image with sufficient resolution. The camera 2 images the above-mentioned label on the package 20 whose distance is defined by the stopper 11 . In addition, the scale 1 measures the weight of the package 20 placed at a position where it contacts the stopper 11 .

FIG. 4 is a flowchart illustrating an example of the process flow in the packaged item determination method according to the present embodiment. In step S11, the image data acquisition unit 101 acquires image data of the package 20 captured by the camera 2, that is, an image of the label.
Further, the image data acquisition unit 101 transmits the acquired image data to the liquid crystal panel 3, and the liquid crystal panel 3 displays the received image data (see FIG. 5).

In step S12, the weight data acquisition unit 102 acquires data on the weight of the packaged item 20 from the scale 1. Further, the weight data acquisition unit 102 transmits the acquired weight data to the liquid crystal panel 3, and the liquid crystal panel 3 displays the received weight data (see FIG. 5).

In addition, in this embodiment, the order of step S11 and step S12 is not limited, and the order of step S11 and step S12 may be sufficient, or the reverse order may be sufficient as it.

The control unit 100 transmits a signal corresponding to the start of the determination process to the notification unit 106 in response to the acquisition of image data by the image data acquisition unit 101 or the acquisition of weight data by the weight data acquisition unit 102. In response to the signal from the control unit 100, the notification unit 106 lights up a yellow lamp in the notification lamp 4 to notify the surroundings that the determination process is in progress.

In step S13, the determination unit 103 refers to the image data acquired by the image data acquisition unit 101 and the weight data acquired by the weight data acquisition unit 102. Then, it is determined whether the desired bulb is housed in the imaged cardboard box.

The determination unit 103 performs determination by referring to output data output from a model to which image data is input. For example, the determination unit 103 inputs the image data of the label to a predetermined model such as the trained model described above. The model outputs information on the brand name, product name, and number of bulbs according to the input image data. In this way, the output data includes the product name of the valve, which is information regarding the type of valve.

The determining unit 103 determines whether the output data based on the image data matches the information corresponding to the work schedule data among the valve-related information in the collation table stored in the storage unit 104. The determination unit 103 outputs a positive (OK) determination as a determination result based on the image data when the above output data and the information in the comparison table match. If the above output data and the information in the comparison table do not match, the determination unit 103 outputs a negative (NG) determination as a determination result based on the image data.

Further, the determination unit 103 determines whether the weight data acquired by the weight data acquisition unit 102 matches the information regarding weight in the weight data table stored in the storage unit 104. The determination unit 103 outputs a positive (OK) determination as a determination result based on the weight data when the weight data and the information regarding the weight match. When the above weight data and the above information regarding weight do not match, the determination unit 103 outputs a negative (NG) determination as a determination result based on the weight data.

In step S14, the control unit 100 refers to the image determination information and weight determination information by the determination unit 103, generates a screen that displays the determination result based on the image data, and displays it on the liquid crystal panel 3 ( (See Figure 5). FIG. 5 is a diagram schematically showing an example of an image displayed on the liquid crystal panel in this embodiment. The screen includes a product name 31, first determination information 32, and second determination information 33, as shown in FIG.

The product name 31 is information to be determined, and is information that can specify, for example, the package 20 or the valve contained therein. The information may be information regarding the type of valve. The product name 31 represents the product name of the valve to be accommodated in the packaging material, and is, for example, "AAAA".

The first determination information 32 is information regarding determination based on the image data acquired by the image data acquisition unit 101, and includes, for example, a label image 321 and image determination result information 322. The label image 321 is image data acquired by the image data acquisition unit 101, and includes, for example, information regarding the type of bulb. The information regarding the type of valve may be information on the output data output by the above-mentioned model. For example, the label image 321 includes the valve brand name "XXX", the valve product name "AAAA", and the number information "1". ” is included. The image determination result information 322 is information on the determination result based on the image data, and is, for example, either "OK" or "NG".

The second determination information 33 is information related to determination based on the weight data acquired by the weight data acquisition unit 102, and includes, for example, weight data information that displays the weight data, and information on determination results based on the weight data (weight determination result information). The weight data information may be information displayed numerically, may be information displayed relative to a measurable range, or may be both.

For example, the second determination information 33 includes a gauge 331, a weight value 332, and weight determination result information 333. The gauge 331 has a semicircular shape, and the measurable range of the scale 1 is represented by the entire arc of the semicircle. Weight data is displayed by rotating the needle based on the center of the semicircle and coloring the range in which the needle rotates, depending on the proportion of the weight value of the weight data to the measurable range of scale 1. be done. The weight value 332 is the weight value indicated by the weight data.

The weight determination result information 333 is information on the determination result based on the weight data, and is, for example, either "OK" or "NG". If the determination result is "correct", it is "OK", and if the determination result is "fail", it is "NG".

The image shown in FIG. 5 shows that both the determination result based on the image data and the determination result based on the weight data are "correct". That is, it can be seen that the package 20 contains desired valves in a desired quantity.

If both the determination result based on the image data and the determination result based on the weight data are "correct", the control unit 100 transmits a signal according to the determination result to the reporting unit 106. The notification unit 106 switches the lamp lit in the notification lamp 4 from a yellow lamp to a green lamp in response to the signal from the control unit 100. This indicates to those around you that the determination process is positive.

If at least either the determination result based on the image data or the determination result based on the weight data is "no", the control unit 100 transmits a signal according to the determination result to the reporting unit 106. The notification unit 106 switches the lamp lit in the notification lamp 4 from a yellow lamp to a red lamp in response to the signal from the control unit 100. This indicates to those around you that the determination process is negative.

Control unit 100 resets the determination result in response to an input signal from input unit 105. The input signal is, for example, a signal generated by a worker pressing either button A or B on the switch 5.

If the determination result is positive, the package 20 on the scale 1 is placed on the transport device 22. The transport device 22 transports the packaged object 20 to a workplace for the next process. When the worker presses the A button on the switch 5, the input section 105 outputs a signal indicating that the A button has been pressed. In response to the signal from the input unit 105, the image data acquisition unit 101 and the weight data acquisition unit 102 each discard the acquired data. The determination unit 103 may discard the information on the previous determination result in response to the signal, but may also store the information on the previous determination result in the storage unit 104, or may store the information on the previous determination result in the reporting unit 106. The result information may be reported to the administrator.

If the determination result is negative, the worker who has confirmed the abnormality presses the B button on the switch 5. When the control unit 100 receives a signal caused by the input unit 105 pressing the B button and the notification unit 106 acquires the signal, the notification unit 106 stops lighting the red lamp. The worker removes the package 20 for which the judgment result is negative from the scale 1. Then, when the A button is pressed, the acquired data is appropriately discarded, stored, or reported as described above.

In this way, when preparations for the determination processing of the next packaged item 20 are completed, the transport device 21 transports the next packaged item 20 to be determined to the scale 1. The determination device 10 then determines the next packaged item 20 as described above.

[Effect]
In this embodiment, the distance between the package 20 (label) to be imaged and the camera 2 is easily kept constant by the stopper 11 . Therefore, it is possible to accurately and easily obtain image data of a label printed on the surface of a cardboard box.

In this embodiment, the character information in the label is identified by analyzing the image data captured by the camera 2 and recognizing the characteristics of the image. Therefore, if the label can be imaged with a resolution that allows it to be identified as image data, it is possible to determine the package 20 with sufficiently high accuracy. Therefore, there are fewer restrictions on the camera and fewer restrictions on the imaging conditions. Therefore, it is possible to configure the determination device inexpensively and simply. Furthermore, although it is based on the above-mentioned model, it is also possible to identify image information from the characteristics of the image. Therefore, even images that include images that cannot be converted into character codes, such as logo marks, can be identified.

Note that in this embodiment, when referring to image data, it is also possible to convert the image data into a character code and then identify it as a character. In this case, in order to capture image data that can be converted into a character code, the camera 2 for that purpose, adjustment of the focal position to capture an image with desired precision, and desired imaging conditions such as illuminance around the label are required. Adjustments may be made as appropriate.

In this embodiment, the determination device 10 further includes a notification lamp 4 for notifying the surroundings. For this reason, the normal progress and occurrence of an abnormality in the determination process for the package 20 are reported to the surrounding area. Therefore, it is possible to smoothly proceed with the determination process for the packaged item 20.

Further, in this embodiment, the determination device 10 further includes a switch 5 as an input section 105. Therefore, the worker can proceed with the determination process for the packaged item 20 while checking it, and recovery from abnormality determination can also be reliably performed through the worker.

As is clear from the above description, the determination device 10 in this embodiment includes an image data acquisition unit 101 that acquires image data obtained by imaging a packaging material, and a packaging material containing a target object (package item 20). a weight data acquisition unit 102 that acquires weight data regarding the weight of the item; and a determination unit 103 that refers to the acquired image data and weight data to determine whether or not the desired object is contained in the packaging material. We are prepared.

Further, the determination method in this embodiment includes a step (S11) of acquiring image data obtained by imaging the packaging material, a step (S12) of acquiring weight data regarding the weight of the packaging material containing the object, The method includes a step (S13) of determining whether or not the desired object is contained in the packaging material by referring to the acquired image data and weight data.

Therefore, according to this embodiment, it is possible to determine whether the object contained in the packaging material is correct or not.

The determination unit may perform the above determination by referring to output data output from a model to which image data is input. This configuration is even more effective from the viewpoint of accurately acquiring information indicated by image data and making accurate determinations.

The output data may include information regarding the type of object (such as the product name of the valve). This configuration is even more effective from the viewpoint of accurately acquiring information about the packaged object from image data and making accurate judgments.

Further, the output data may include information regarding the number of objects. This configuration is
Even when a plurality of objects are housed in the packaging material, it is even more effective from the viewpoint of making accurate judgments.

[Embodiment 2]
Other embodiments of the invention will be described below. For convenience of explanation, members having the same functions as the members described in the above embodiment are given the same reference numerals, and the description thereof will not be repeated.

FIG. 6 is a diagram schematically showing an example of the functional configuration of the determination device of this embodiment. The determination device 20 includes a control section 200 and a storage section 204, as shown in FIG. The control unit 200 has the same functional configuration as the control unit 100 in the first embodiment except that it includes the determination unit 203.

The determination unit 203 further includes a reference weight identification unit 207. The reference weight identifying unit 207 identifies the reference weight of the object by referring to the reference data and the output data output by the above-described model.

The storage unit 204 further stores reference data. The reference data is data in which the type of object and the reference weight of the object are associated with each other. As described above, the type of object is, for example, the brand name of the valve or the product name of the valve. The reference weight of the object is, for example, the weight of a valve depending on the type (product name).

When casting valves, even good valves tend to have some variation in weight. For example, depending on the type of bulb, there may be an error of about 10%, but the way the weight varies may be unique depending on the type of bulb.

In this embodiment, the determination unit 203 performs determination based on image data, and then performs determination based on weight data.

The determination unit 203 acquires the image data of the label from the image data acquisition unit 101. Then, the image data is input to the above-mentioned model, and data on the brand name, type, and number of bulbs corresponding to the image data is obtained as output data. Then, the determining unit 203 determines whether or not the acquired output data matches the information corresponding to the work schedule data among the data in the collation table stored in the storage unit 204. In this way, the determination unit 203 first outputs a determination result based on the image data.

The reference weight identifying unit 207 refers to the reference data in the storage unit 204 and identifies reference weight information associated with the product name of the valve identified by the determination result based on the image data. For example, assume that the reference weight is 4.5 to 5.1 kg in the case of the above-mentioned product name "AAAA". In other words, the "AAAA" valve has an error that tends to be large when the value is smaller, and is sufficiently small when the value is larger.

The determination unit 203 refers to the reference weight specified by the reference weight identification unit 207, the information on the number of valves in the output data described above, and the weight data acquired by the weight data acquisition unit 102. Then, the determination unit 203 determines whether the information regarding each valve in the weight data satisfies the reference weight condition. In this way, the determination unit 203 outputs a determination result based on the weight data.

Note that the reference data may be further associated with information on the number of valves in addition to the product name of the valve and the reference weight of the valve. In this case, the determination unit 203 refers to the weight data in which the information on the number of valves is reflected and the reference weight of the valve that is associated with the information on the number of valves. Therefore, in a determination based on weight data, by directly referencing and comparing the weight data and the reference weight, it is possible to output a determination result based on the weight data. Therefore, compared to the above-described determination based on weight data, it is possible to simplify the determination process.

Furthermore, in the case where the reference data further includes information on the number of valves, as in the first embodiment, the order of determination based on image data and determination based on weight data may be in this order, or may be reversed. You can.

As is clear from the above description, in the present embodiment, the determination unit refers to the reference data in which the type of the object and the reference weight of the object are associated with each other, and the above-mentioned output data, to A reference weight is specified, and a determination is made with reference to the specified reference weight and weight data. Therefore, even if the objects contained in the packaging material have specific variations in weight, such as valves manufactured by casting, it is possible to determine whether the packaging material is correct or not with sufficiently high accuracy.

In this embodiment, the determination unit determines the reference weight of the object by referring to the above-mentioned output data and reference data in which the type of object, the number of objects, and the reference weight of the object are associated with each other. The determination may be made with reference to the specified reference weight and the weight data. With this configuration, the weight data and reference weight are referenced while reflecting the information on the number of valves, so even when multiple objects with unique weight variations are housed in the packaging material, This is even more effective from the standpoint of simply determining whether the packaged item is correct or not with sufficiently high accuracy.

[Embodiment 3]
Other embodiments of the invention will be described below. For convenience of explanation, members having the same functions as the members described in the above embodiment are given the same reference numerals, and the description thereof will not be repeated.

FIG. 7 is a diagram schematically showing an example of the functional configuration of the determination device of this embodiment. As shown in FIG. 7, the determination device 30 has the same functional configuration as the determination device of Embodiment 1, except that it includes a control unit 300 and a storage unit 304.

The control unit 300 includes an image data acquisition unit 101, a weight data acquisition unit 302, a determination unit 203, a timer 310, a teacher data acquisition unit 308, and a learning unit 309. The image data acquisition unit 101 has the same function as the image data acquisition unit of the first embodiment. The determination unit 203 has the same functional configuration as the determination unit of the second embodiment. The storage unit 304 is configured to further receive information output from the learning unit 309.

This embodiment differs from the second embodiment in the method of acquiring image data and weight data, and in that it is configured to be able to generate a trained model.

In this embodiment, image data is acquired after acquiring weight data. When the package 20 is placed on the scale 1 at the above-mentioned position, the scale 1 measures the weight of the package 20. Then, the weight data acquisition unit 102 acquires the weight data of the package 20 measured by the scale 1.

The weight data acquisition unit 102 transmits a signal to the timer 310 indicating that the weight data has been acquired. The timer 310 measures a predetermined period of time, such as 5 seconds, in response to the signal from the weight data acquisition unit 102. When the timer 310 finishes measuring time, the timer 310 outputs a signal to the camera 2 indicating that the predetermined time has been measured.

The camera 2 starts photographing the label of the packaged item 20 in response to the signal from the timer 310, and the image data acquisition unit 101 acquires the image data of the label photographed by the camera 2.

In this manner, in this embodiment, the camera 2 images the packaging material after a predetermined period of time has passed after the weight data acquisition unit 102 acquires the weight data.

Further, in this embodiment, the control unit 300 includes a teacher data acquisition unit 308 and a learning unit 309. The teacher data acquisition unit 308 acquires teacher data from the input unit 105, for example. The teacher data includes, for example, image data of a sufficient number of labels, and information on the brand name of the bulb, product name of the bulb, number of bulbs, and reference weight of the bulb corresponding to the label.

The learning unit 309 trains a model using teacher data for each label image data. The model is, for example, a neural network. The learning unit 309 stores the trained model in the storage unit 304. The learned model stored in the storage unit 304 is used for determining the package 20 as a model for outputting the above-mentioned output data.

In this way, the present embodiment further includes a learning unit that learns the model using training data, and the training data includes image data obtained by imaging the packaging material and information about the type associated with the image data. Contains information.

In this embodiment, the determination device 30 acquires image data after acquiring weight data. In general, determination processing based on weight data is simpler than determination processing based on image data that involves certification of image data. If the determination result based on the weight data is "fail", the determination result for the packaged item 20 will be "fail" regardless of whether the determination result based on the image data is correct or not. Therefore, determination processing based on image data can be omitted. Therefore, in the present embodiment, when the determination process based on the weight data yields a "fail" determination result, it is possible to further reduce the processing load related to the determination.

In addition, a worker may carry out the work of loading the package 20 from the transport device 21 onto the scale 1 and transferring the package 20 from the scale 1 to the transport device 22. In this case, in this embodiment, when the package 20 is placed on the scale 1 at the above-mentioned position, the weight of the package 20 is automatically measured and the label is photographed by the camera 2. Therefore, it is possible to omit the work of measuring the weight by the worker and taking an image with the camera 2.

Further, in this embodiment, a teacher data acquisition section 308 and a learning section 309 are further provided. Therefore, it is possible to use data actually acquired by the determination device 30 (for example, an image of a label captured by the camera 2) as the teacher data. Therefore, training data under substantially the same conditions as those in actual use can be used for model learning, which is advantageous from the viewpoint of increasing the accuracy of the determination of the package 20.

As is clear from the above description, the determination device further includes a learning section that learns the model using training data, and the training data includes image data obtained by capturing an image of the packaging material and associations with the image data. The information regarding the type of information received may also be included. This configuration is even more effective from the viewpoint of increasing the accuracy of the determination of the package 20.

The determination device further includes an imaging unit (camera 2) that images the packaging material and supplies the captured image data to the image data acquisition unit, and the imaging unit, after the weight data acquisition unit acquires the weight data, The packaging material may be imaged after a predetermined period of time has elapsed. This configuration is even more effective from the viewpoint of reducing the burden of judgment processing when making a negative judgment, and from the perspective of saving the labor of the operator in the work related to photographing and weight measurement.

Note that in this embodiment, the teacher data may be data that is further associated with a determination result. For example, the determination device further includes the aforementioned learning section, and the teacher data includes image data obtained by imaging the packaging material, information regarding the weight associated with the image data, and information regarding the image data and the weight. and a determination result associated with the. This configuration is even more effective in terms of reducing the load of the determination process, since it is possible to determine the package 20 without character recognition in the image data of the label.

[Example of implementation using software]
The control blocks of the determination devices 10, 20 and 30 (particularly the determination units 103 and 203) may be realized by a logic circuit (hardware) formed on an integrated circuit (IC chip) or the like, or may be realized by software. Good too.

In the latter case, the determination devices 10, 20, and 30 are equipped with a computer that executes instructions of a program that is software that implements each function. This computer includes, for example, one or more processors and a computer-readable recording medium that stores the above program. The program is, for example, a determination program for causing a computer to function as the determination devices 10, 20, and 30, and is a determination program for causing the computer to function as an image data acquisition section, a weight data acquisition section, and the determination section. It's fine.

In the computer, the processor reads the program from the recording medium and executes the program, thereby achieving the object of the present invention. As the processor, for example, a CPU (Central Processing Unit) can be used. As the recording medium, in addition to "non-temporary tangible media" such as ROM (Read Only Memory), tapes, disks, cards, semiconductor memories, programmable logic circuits, etc. can be used. Further, the computer may further include a RAM (Random Access Memory) for expanding the above program.

Furthermore, the program may be supplied to the computer via any transmission medium (communication network, broadcast waves, etc.) that can transmit the program. Note that one aspect of the present invention can also be realized in the form of a data signal embedded in a carrier wave, in which the program is embodied by electronic transmission.

[Modified example]
The present invention is not limited to the embodiments described above, and various modifications can be made within the scope of the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. are also included within the technical scope of the present invention.

For example, in the embodiment described above, the reporting unit 106 may notify the progress status of the determination process by sound in addition to lighting the reporting lamp 4.

Furthermore, in the embodiment described above, the weight data in the weight data table may be the weight data of the valve measured before the package 20 is determined. The period before the determination may be, for example, at the time of manufacturing the valve, or at the time of housing the valve in a packing material (packaging). The data of the actual value of the weight of the valve in advance is managed, for example, for each manufacturing lot of the valve. Then, by referring to the weight data of the valve corresponding to the manufacturing lot and the weight data acquired for the package 20, if there is a match, a positive determination may be made as a determination result based on the weight data.

In the embodiment described above, when the package 20 is housed in a further package, a similar determination device may be used to determine whether the desired package 20 is housed in the further package. . This is effective from the perspective of preventing further packaging mistakes.

In the embodiment described above, the determination device may transmit information on the determination result to the administrator of the determination device. A positive determination result indicates the production performance of the package 20. By displaying the production schedule of the packaged item 20 in association with the information of the positive determination result, the administrator can immediately and easily check the progress of work in the production of the packaged item 20. Further, in the case where a worker records the judgment result of the packaged item 20, it is possible to omit the worker's recording work. Therefore, it is effective from the viewpoint of reducing the work of the worker.

In the third embodiment described above, the teacher data acquisition unit and the learning unit may be included in a device separate from the determination device, and only the learned model may be input to the determination device via the input unit, for example. good.

1 Scale 2 Camera 3 Liquid crystal panel 4 Notification lamp 5 Switch 10, 20, 30 Judgment device 11 Stopper 20 Package 21 Conveyance device 22 Conveyance device 31 Product name 32 First judgment information 33 Second judgment information 100, 200, 300 Control section 101 Image data acquisition unit 102, 302 Weight data acquisition unit 103, 203, 204 Judgment unit 104, 204, 304 Storage unit 105 Input unit 106 Notification unit 207 Reference weight identification unit 308 Teacher data acquisition unit 309 Learning unit 310 Timer 321 Image 322 Image Judgment result information 331 Gauge 332 Weight value 333 Weight judgment result information

Claims (6)

an image data acquisition unit that acquires image data obtained by imaging the packaging material;
a weight data acquisition unit that acquires weight data regarding the weight of the packaging material containing the valve ;
a determination unit that refers to the image data and the weight data to determine whether a desired number of valves of a desired type is housed in the packaging material;
Equipped with
The determination unit makes the determination with reference to output data output from a model to which the image data is input,
The output data includes information regarding the type of the valve and information regarding the number of the valves,
The determination unit includes:
Valve type and
The number of valves and
A reference weight of a valve, which is a weight within a predetermined numerical range of a valve that can be considered as a good valve, depending on the type of valve;
identify a reference weight of the valve by referring to reference data associated with each other and the output data;
making the determination with reference to the specified reference weight and the weight data;
Judgment device. further comprising a learning unit that learns the model using teacher data,
The teacher data includes:
Image data obtained by imaging packaging materials,
The determination device according to claim 1 , further comprising: information regarding a type associated with the image data. further comprising a learning unit that learns the model using training data,
The teacher data includes:
Image data obtained by imaging packaging materials,
Information regarding weight associated with the image data;
a determination result associated with the image data and the information regarding the weight;
The determination device according to claim 1 or 2, characterized in that the determination device includes: further comprising an imaging unit that images the packaging material and supplies the imaged image data to the image data acquisition unit,
The packaging according to any one of claims 1 to 3 , wherein the imaging unit photographs the packaging material after a predetermined period of time has passed after the weight data acquisition unit acquires the weight data. A device for determining things. acquiring image data obtained by imaging the packaging material;
obtaining weight data regarding the weight of the packaging containing the valve ;
referring to the image data and the weight data to determine whether a desired number of valves of a desired type are housed in the packaging material;
including;
In the step of making the determination, the determination is made with reference to output data output from a model into which the image data is input;
The output data includes information regarding the type of the valve and information regarding the number of the valves,
In the step of making the determination,
Valve type and
The number of valves and
A reference weight of a valve, which is a weight within a predetermined numerical range of a valve that can be considered as a good valve, depending on the type of valve;
identify a reference weight of the valve by referring to reference data associated with each other and the output data;
making the determination with reference to the specified reference weight and the weight data;
Judgment method. A determination program for causing a computer to function as the determination device according to claim 1, the determination program for causing the computer to function as the image data acquisition section, the weight data acquisition section, and the determination section.

Images