JP7320314B1 - Seaweed grading processing system, seaweed grading method, and program - Google Patents

Abstract

A laver grading processing system, a laver grading method, and a program are provided for automating the grading of laver and the work associated with the grading to improve the efficiency of the laver grading work. Kind Code: A1 A seaweed grade determination processing system (1) includes an imaging device (200) for capturing a surface image of seaweed; and a printing device 300 for printing a first code corresponding to the determined grade on the seaweed strip. [Selection drawing] Fig. 1

Description

The present invention relates to a seaweed grade determination processing system, a seaweed grade determination method, and a program.

Since the market price of seaweed is determined by the grade, grading of seaweed is an important concern for producers. Grading is carried out by visual inspection by specific inspectors in the fisheries association in the state of dried seaweed. However, the human grading has a problem that, for example, the judgment result varies depending on the inspector, or even by the same inspector, due to fatigue or the like. In addition, there was also the problem that the shortage of successors due to the aging of inspectors has become serious. Therefore, there is a demand for automating the grade determination of dried seaweed.

For example, in Patent Document 1, the color of dried seaweed is measured with a color difference meter to obtain the b ^* value, which is the color value in the L ^* a ^* b ^* color system, and then based on the obtained b ^* value. A judgment method for judging the quality of dried seaweed is disclosed. In addition, Patent Document 2 discloses a method of inspecting the quality of dried seaweed based on the RGB values of each pixel of the image obtained by imaging dried seaweed with a color imaging device and A/D converting it.

JP-A-2002-207012 Patent No. 3230217

By the way, the work of determining the grade of seaweed includes unpacking the cardboard box containing the dried seaweed, determining the grade, stamping the grade, repackaging, recording, etc. The work is done from early morning to late at night, and the labor is enormous. There was also a problem. Therefore, in addition to automating the grade determination, it is also desired to automate the work associated with the grade determination.

Therefore, in view of the above problems, the present invention provides a laver grading processing system, a laver grading method, and a laver grading system that automates the grading of dry laver and the work associated with the grading to make the laver grading work more efficient. and provide programs.

The present invention comprises: an imaging unit that captures a surface image of seaweed; a grade determination unit that determines the grade of the surface image of seaweed based on a learning model generated based on the surface image of seaweed that has been graded; and a printing unit for printing a first code corresponding to the grade on the seaweed strip.

Further, the present invention provides a seaweed grade determination processing system comprising a conveying device for conveying the seaweed, wherein the imaging unit sequentially captures surface images of the conveyed seaweed.

In addition, the present invention is provided with a reading device for reading production stage information relating to the production of the laver attached to the box in which the laver was stored, and the conveying device conveys the laver taken out from the box. The printing unit provides a seaweed grade determination processing system in which a second code based on the production stage information is printed on the seaweed strip together with the first code.

Further, the present invention includes a detection unit for detecting the seaweed being transported, and the imaging unit performs seaweed grade determination processing for capturing a surface image of the seaweed in response to the detection of the seaweed by the detection unit. provide the system.

The present invention also provides a seaweed grade determination processing system that adds the surface image of the seaweed graded by the grade determination unit to the teacher data of the learning model.

The present invention also provides a seaweed grading processing system in which the learning model is a deep learning model.

Further, the present invention is a seaweed grade determination processing system, comprising: a step of capturing a surface image of seaweed; A method is provided comprising the steps of determining a grade, and printing a first code corresponding to the determined grade on the strip of seaweed.

Further, the present invention provides a seaweed grade determination processing system, an imaging unit that captures a surface image of seaweed, and grades the surface image of seaweed based on a learning model generated based on the surface image of seaweed that has been graded. A program is provided that functions as a grade determination unit for determination and a printing unit for printing a first code corresponding to the determined grade on the seaweed strip.

According to the present invention, it is possible to automate the grading of dried seaweed and the work associated with the grading, thereby making the work of grading laver more efficient.

BRIEF DESCRIPTION OF THE DRAWINGS It is a figure explaining the outline|summary of the seaweed grade determination processing system which concerns on the 1st Embodiment of this invention. BRIEF DESCRIPTION OF THE DRAWINGS It is a figure which shows the functional structure of the seaweed grade determination processing system which concerns on the 1st Embodiment of this invention. It is a figure which shows typically grade DB which concerns on the 1st Embodiment of this invention. 4 is a flow chart showing seaweed grade determination processing of the seaweed grade determination processing system according to the first embodiment of the present invention. It is a figure which shows the functional structure of the seaweed grade determination processing system which concerns on the 2nd Embodiment of this invention. FIG. 10 is a flow chart showing a seaweed grade determination process of the seaweed grade determination processing system according to the second embodiment of the present invention; FIG.

EMBODIMENT OF THE INVENTION Hereinafter, with reference to an accompanying drawing, the form (henceforth, embodiment) for implementing this invention is demonstrated in detail. In subsequent figures, the same numbers or symbols are attached to the same elements throughout the description of the embodiments.
<First Embodiment>

[Basic Concept/Basic Configuration]
FIG. 1 is a diagram for explaining an overview of a seaweed grade determination processing system 1 according to a first embodiment of the present invention. As shown in FIG. 1 , the seaweed grade determination processing system 1 includes a seaweed grade determination device 100 , an imaging device 200 , a printer 300 and a transport device 400 .

The nori laminate packed in the packing box P by the producer is taken out from the packing box P and put into the conveying device 400 (S1). The seaweed laminate is made by bundling 10 sheets of seaweed and closing them with a paper strip called a seaweed strip (hereinafter simply referred to as a strip) B, and generally 18 pieces are packed in one packing box P. . Hereinafter, the seaweed layered body will be simply referred to as seaweed N, and the seaweed conveyed by the conveying device 400 will be referred to as N _i (where i=1, 2, . . . n). The feeding of laver N into the conveying device 400 may be performed by an operator, or may be performed automatically using a robot arm or the like.

When the seaweed _Ni put into the conveying device 400 advances in the conveying direction to a predetermined position, the imaging device 200 captures a surface image of the seaweed _Ni , and the captured surface image of the seaweed _Ni is transferred to the seaweed grade determination device 100. (S2). The seaweed grade determination device 100 determines the grade of the seaweed _Ni based on the received surface image of the seaweed _Ni and the learning model created using the graded surface image (S3).

The seaweed grade determination device 100 transmits the first code corresponding to the determined grade to the printer 300 (S4). The printing device 300 prints the first code on the strip of seaweed _Ni advanced to a predetermined position in the conveying direction by the conveying device 400 (S5).

According to such a seaweed grade determination processing system, the grade of seaweed is automatically determined based on the learning model created using the surface image of the seaweed that has been graded and the surface image of the seaweed taken. , the determined grade can be automatically printed on the seaweed strip. This makes it possible to improve the efficiency of laver grade determination work. In addition, while the laver is being conveyed by the conveying device, it is possible to sequentially judge the grade of the laver and print on the strip of the laver of the judged grade. As a result, it is possible to continuously perform the grading of seaweed and the work associated with the grading, so that the work of grading laver can be made more efficient.

[Function configuration]
FIG. 2 is a diagram for explaining the functional configuration of the seaweed grade determination processing system 1 according to the first embodiment of the present invention. As shown in the figure, the seaweed grade determination processing system 1 includes a seaweed grade determination device 100 , an imaging device 200 , a printer 300 and a transport device 400 . The imaging device 200 and the printing device 300 are connected to the laver grade determination device 100 by wire or wirelessly. In this embodiment, the imaging device 200 and the printing device 300 are devices independent of the laver grade determination device 100, but may be provided in the laver grade determination device 100 and function as respective functional units. Further, in this embodiment, the conveying device 400 is independent from the laver grade determination device 100 and is not connected to the laver grade determination device 100 .

The conveying device 400 is, for example, a belt conveyor that conveys the seaweed N. Specifically, the conveying device 400 receives seaweed N packed in a packing box P by a worker or the like, and conveys the fed seaweed N from upstream to downstream.

In addition, it is desirable that the seaweed N conveyed by the conveying device 400 does not overlap and the plane horizontal to the stacking direction is the upper part. This is for increasing the accuracy of determination by the grade determining unit 130, which will be described later. Moreover, it is desirable that the seaweed N conveyed by the conveying device 400 is put in at predetermined intervals. This is to prevent surface images of adjacent seaweed N from entering when photographed by the imaging device 200 described later, and to prevent clogging when printing by the printer 300 described later.

The imaging device 200 includes a detection unit 210 and an imaging unit 220 and captures a surface image of the seaweed _Ni conveyed by the conveying device 400 . Specifically, first, the detection unit 210 detects that the seaweed _Ni conveyed by the conveying device 400 has entered the photographing range. Then, in response to the detection, the imaging unit 220 captures a surface image of the seaweed _Ni within the imaging range.

The detection unit 210 may be any device that can detect whether seaweed has entered the imaging range, such as a photoelectric sensor. When the conveying device 400 is a belt conveyor, the surface color of the belt is set to a light color such as white so that the detection unit 210 can easily detect that the seaweed _Ni conveyed by the conveying device 400 has entered the imaging range. It is desirable to Since the seaweed N is black or dark green, the detector 210 can easily detect the seaweed N by increasing the contrast between the seaweed N and the surface of the belt.

The seaweed grade determination device 100 includes a storage unit 110 , an image acquisition unit 120 , a grade determination unit 130 and a learning unit 140 . In this embodiment, the laver grade determination device 100 is assumed to be on-premise, but may be a cloud server.

The storage unit 110 includes an image database (hereinafter also referred to as DB) 111 that stores surface images of seaweed _Ni in association with determined grades of seaweed _Ni , a grade DB 112 that stores grade information of seaweed, and a grade DB 112, which will be described later. and a model DB 113 that stores the learning model generated by the learning unit 140 .

The image acquisition unit 120 acquires the surface image of the seaweed _Ni acquired by the imaging device 200 .

The grade determination unit 130 determines the grade of the seaweed _Ni based on the acquired surface image of the seaweed _Ni and the learning model created based on the surface image of the graded seaweed. Specifically, the grade determination unit 130 performs image analysis of the surface image of the seaweed _Ni by AI based on the learning model stored in the model DB 113, and determines the grade of the surface image of the seaweed _Ni . The grade determination unit 130 refers to the grade DB 112 to acquire the grade code of the determined grade, and transmits the acquired grade code to the printer 300 .

FIG. 3 is a diagram schematically showing the grade DB according to this embodiment. As shown in FIG. 3, the seaweed grade information stored in the grade DB includes a grade code for identifying the grade of seaweed N and grade name information that is the name of the grade.

The grade determination unit 130 associates the surface image of the seaweed _Ni whose grade has been determined with the determined grade, and stores them in the image DB 111 .

The learning unit 140 generates a learning model used for grade determination of laver N _i based on the surface image of the graded laver. Specifically, the learning unit 140 uses information such as the color and color intensity used in the eye color inspection by deep learning using surface images of seaweed with grades associated from the image DB 111 as teacher data. generate

For example, the learning unit 140 learns teacher data using a convolutional neural network (CNN) to generate a learning model. When the teacher data includes training data for training a learning model and verification data for evaluating the trained learning model, the learning unit 140 learns the training data by a convolutional neural network to generate a learning model. do. Then, the learning unit 140 inputs verification data to the generated learning model and determines the generated learning model.

Among the seaweed surface images associated with the grades stored in the image DB 111, the seaweed surface images associated with the grades determined by the grade determination unit 130 are used as training data, and inspectors and the like It is desirable to use the surface image of the seaweed associated with the determined grade as verification data.

Specifically, the learning unit 140 uses the generated learning model to determine the grade of the seaweed surface image that has been graded and set in advance as verification data, and determines the grade that has already been associated with the seaweed surface image. is equal to or higher than a predetermined value, for example, 70% or higher, the generated learning model is determined to be reliable and stored in the model DB 113 .

The printing device 300 includes a detection unit 310 and a printing unit 320, converts the grade code corresponding to the grade determined by the grade determination unit 130 into a bar code, and prints a strip of seaweed _Ni transported by the transport device 400. to print. Barcodes include one-dimensional barcodes and two-dimensional barcodes. Note that the grade code may be a code other than the bar code.

Specifically, first, the detection unit 310 detects that the seaweed _Ni conveyed by the conveying device 400 has entered the printing range. Then, in response to the detection, the printing unit 320 converts the grade code received from the grade determining unit 130 into a barcode, and prints the barcode on the seaweed _Ni band within the printing range. Note that the printing unit 320 may convert information such as the date and time when the grade was determined and the place where the grade was determined into a barcode together with the grade code.

The functional configuration of the system described above is merely an example, and one functional block (database and functional processing unit) may be divided, or a plurality of functional blocks may be collectively configured as one functional block. Each function processing unit is stored in a storage device (storage unit) such as a ROM (Read Only Memory), a flash memory, an SSD (Solid State Drive), or a hard disk. It is realized by a computer program executed by the CPU by reading out a computer program (for example, basic software or an application that causes the CPU to execute the various types of processing described above). That is, each functional processing unit reads and writes necessary data such as a table from a database (DB; Data Base) stored in a storage device or a storage area on a memory. It is realized by controlling hardware (for example, an input/output device, a display device, a communication interface device). In addition, the database (DB) in the embodiment of the present invention may be a commercial database, but it also means a mere collection of tables and files, and the internal structure of the database itself does not matter.

The seaweed grade determination processing of the seaweed grade determination processing system 1 according to the embodiment of the present invention will be described with reference to FIG. The seaweed grade determination process starts when the seaweed N starts to be loaded from the packing box P into the conveying device 400 .

When the imaging device 200 enters the photographing range of the seaweed _Ni conveyed by the conveying device 400, it photographs the surface image of the seaweed _Ni (S11). Next, the image acquisition unit 120 acquires a surface image of the seaweed _Ni from the imaging device 200 (S12).

Next, the grade determination unit 130 determines the grade of the seaweed _Ni based on the surface image of the seaweed _Ni acquired in S12 and the learning model in the model DB 113 (S13). The learning model is trained in advance so as to determine the grade of laver _Ni from the surface image of laver _Ni photographed in S11 using teacher data in which grades and surface images of laver are associated.

The grade determination unit 130 associates the surface image of the seaweed _Ni whose grade was determined in S13 with the determined grade, and stores them in the image DB 111 (S14). The printing device 300 acquires the grade code corresponding to the grade determined in S13 from the grade determination unit 130, converts it into a barcode, and prints it on the seaweed _Ni strip being transported by the transport device 400 (S15 ).

The image acquiring unit 120 determines whether or not the surface image of the seaweed Ni ₊₁ has been acquired from the imaging device 200 (S16). If the surface image of laver Ni ₊₁ has been obtained (YES), the process returns to step S13. On the other hand, if the surface image of seaweed Ni ₊₁ has not been acquired (NO), the process ends.

As described above, according to the seaweed grade determination processing system of the first embodiment, based on the learning model created using the surface image of the seaweed that has been graded and the surface image of the seaweed taken, , the grade of seaweed can be automatically determined, and the determined grade can be automatically printed on the strip of seaweed. This makes it possible to improve the efficiency of laver grade determination work. In addition, while the laver is being conveyed by the conveying device, it is possible to sequentially judge the grade of the laver and print on the strip of the laver of the judged grade. As a result, it is possible to continuously perform the grading of seaweed and the work associated with the grading, so that the work of grading laver can be made more efficient.

Deep learning is performed using a plurality of graded seaweed surface images that have been prepared in advance to generate a learning model that is used for grade determination of ungraded seaweed surface images. As a result, the grade of seaweed is judged using a learning model obtained by deep learning, so the judgment accuracy is improved.

<Second embodiment>
The seaweed grade determination processing system 2 of the second embodiment will be described below with reference to FIGS. The seaweed grade determination processing system 2 is configured in the same manner as the seaweed grade determination processing system 1 of the first embodiment except for the following points. In addition, in FIG. 5, the same reference numerals are assigned to the functional units configured in the same manner as in the first embodiment. In addition, in FIG. 6, steps with the same processing contents as those in the first embodiment are given the same reference numerals.

The seaweed grade determination processing system 2 includes a seaweed grade determination device 101 , a reader 500 connected to the seaweed grade determination device 101 , an imaging device 200 , a printer 301 , and a transport device 400 . The reading device 500 is connected to the laver grade determination device 101 by wire or wirelessly. In this embodiment, the reading device 500 is an independent device from the laver grade determination device 101, but may be provided in the laver grade determination device 101 and function as each functional unit.

The reading device 500 is, for example, a barcode reader that reads the production stage information from the barcode attached to the packing box P in which the seaweed N is packed. The bar code attached to the packing box P in which the seaweed N is packed includes information on the production stage such as the name of the producer, the fishing ground and the manufacturing date and time, and is created by the seaweed producer and attached to the packing box P. In this embodiment, in order to distinguish between the barcode attached to the packing box P in which the seaweed N is packed and the barcode printed on the strip of seaweed, the former barcode is referred to as a second barcode, The latter barcode is called the first barcode. It should be noted that a code other than a bar code may be applied to the packing box P.

The seaweed grade determination device 101 includes a storage unit 150 , a production stage information acquisition unit 160 , a production stage information transmission unit 170 , an image acquisition unit 120 , a grade determination unit 130 and a learning unit 140 . The storage unit 150 includes an image DB 111, a grade DB 112, a model DB 113, and a production stage information DB 154.

The production stage information acquisition unit 160 acquires the production stage information read by the reading device 500 and stores it in the production stage information DB 154 .

The production stage information transmission unit 170 prints the production stage information of the seaweed _Ni obtained by referring to the production stage information DB 154 when the grade determination unit 130 transmits the grade code of the seaweed _Ni to the printing device 301. Send to device 301 .

The printing device 301 includes a detection section 310 and a printing section 321 . The printing unit 321 converts the grade code received from the grade judging unit 130 into a first barcode, and prints the first barcode on the seaweed _Ni band within the printing range. Note that the printing unit 321 may convert information such as the date and time when the grade was determined and the place where the grade was determined into the first bar code together with the grade code. The printing unit 321 also converts the production stage information received from the production stage information transmission unit 170 into a second barcode, and prints the second barcode on the seaweed _Ni band within the printing range.

Note that the printing unit 320 may print the first barcode and the second barcode simultaneously, or may print them in any order. Also, the printing unit 321 may combine the grade code and the production stage information to convert one third bar code and print it on the seaweed _Ni band.

The seaweed grade determination processing of the seaweed grade determination processing system 2 according to the embodiment of the present invention will be described with reference to FIG. Note that this processing is a processing flow showing the seaweed grade determination processing for each packing box P.

The reader 500 reads the production stage information from the barcode attached to the packing box P in which the seaweed N is packed (S21). Next, the production stage information acquisition unit 160 acquires the production stage information read in S11 and stores it in the production stage information DB 154 (S22). Next, the processing of steps S11 to S14 is executed.

Next, the printing device 300 acquires the grade code corresponding to the grade determined in S13 from the grade determination unit 130, converts it into a first bar code, and converts it into a first bar code _. Print on the top (S23). Next, the printing device 301 acquires the production stage information of the seaweed _Ni from the production stage information transmission unit 170, converts it into a second barcode, and prints it on the strip of seaweed _Ni conveyed by the conveying device 400. (S24). Then, step S16 is executed.

As described above, according to the seaweed grade determination processing system of the second embodiment, a first code containing seaweed grade information and a second code containing seaweed production stage information are placed on a seaweed strip. can be printed. As a result, it is possible to add production stage information to each seaweed, and by attaching the production stage information of the seaweed to each individual seaweed along with the grade, it is convenient for work after the grade determination process, such as sorting and bidding. improves.

<Modification>
(1) For example, in the first and second embodiments, the image DB 111 is provided in the laver grade determination devices 100 and 101, but is provided on the cloud so that it can be shared by a plurality of laver grade determination devices. may As a result, the image DB can store the grade determination results of a plurality of seaweed grade determination devices, the amount of teaching data used for creating a learning model increases, and the grade determination accuracy can be improved.

(2) For example, in the first and second embodiments, the conveying device 400 was not connected to the laver grade determining devices 100 and 101, but may be connected by wire or wirelessly. Thereby, the conveying device 400 can be controlled from the seaweed grade determining devices 100 and 101 . For example, when the imaging device 200 captures an image of the seaweed _Ni , and when the printing device 300 prints on a band of the seaweed _Ni , the conveying speed of the conveying device 400 can be slowed down or stopped. As a result, it is possible to prevent the image quality of the surface image of the seaweed from deteriorating, and to prevent the print from becoming distorted or blurred. As a result, it is possible to reduce the number of redoing of photographing and printing, and improve work efficiency.

(3) For example, in the second embodiment, the seaweed grade determination process is performed for each packing box P. It is necessary for the production stage information transmission unit 170 to change the production stage information acquired from the production stage information DB 154 to the production stage information of the next packaging box P upon detecting that the packaging box P has changed to the next packaging box P. For example, when the number of seaweed to be packed in the packing box P is predetermined, the production stage information transmission unit 170 transmits the production stage information together with the grade code several times to the printing device 301 . A process for switching to obtain the next production stage information in the acquisition order in the stage information DB 154 is performed.

In addition, by leaving a predetermined time interval before loading the seaweed of the next packing box P, the production stage information transmission unit 170 will , a change of the packaging box P may be detected, and processing may be performed to switch to acquire the next production stage information in the acquisition order in the production stage information DB 154 . As a result, even if the packaging box P is changed, the production stage information transmission unit 170 can acquire the production stage information of the packaging box P in which the seaweed _Ni was packed from the production stage information DB 154 . Then, a second bar code obtained by converting the production stage information of the packing box P in which the seaweed Ni is packed can be printed on the band of the seaweed _Ni .

Although the present invention has been described using the embodiments, it goes without saying that the technical scope of the present invention is not limited to the scope described in the above embodiments. For example, in addition to seaweed, various changes or improvements can be made to the above embodiments, such as assigning grades to foods in general (vegetables, fruits, marine products, meat), grades related to the quality of machines such as automobiles and robots, etc. It is clear to those skilled in the art that In addition, it is clear from the description of the scope of the claims that forms with such modifications or improvements can also be included in the technical scope of the present invention. In the above embodiment, the seaweed grade determination processing system was described as the invention of the present invention. It can also be regarded as the invention of a program that makes

1 seaweed grade determination processing system 100, 101 seaweed grade determination device 110 storage unit 120 image acquisition unit 130 grade determination unit 140 learning unit 150 storage unit 160 production stage information acquisition unit 170 production stage information transmission unit 200 imaging device 210 detection unit 220 imaging Parts 300, 301 Printer 310 Detector 320, 321 Printer 400 Conveyor 500 Reader

Claims (2)

a conveying unit for conveying seaweed;
A detection unit that detects seaweed being transported,
an imaging unit that sequentially captures surface images of the seaweed in response to detection of the seaweed by the detection unit;
a grade determination unit that determines the grade of the seaweed surface image based on a deep learning model generated based on the graded seaweed surface image;
a printing unit that prints a first code corresponding to the determined grade on the seaweed strip;
with
Add the surface image of the seaweed graded by the grade judgment unit to the training data of the deep learning model,
Controlling the conveying unit, controlling the conveying of the seaweed,
When the imaging unit captures the surface image of the seaweed and/or when the printing unit prints on the seaweed band, the conveying speed of the conveying unit is controlled to slow down or stop,
Equipped with a reading device that reads production stage information related to the production of the seaweed attached to the box containing the seaweed,
The printing unit prints a second code based on the production stage information on the seaweed strip together with the first code. Detecting that the box has changed in response to the fact that the surface image of the seaweed has not been acquired for a predetermined time,
2. The laver grade determination processing system according to claim 1, wherein the production stage information relating to the production of the laver given to the box in which the laver was stored is switched .

Images