KR20210157509A - System for automatically distinguishing a cut meat part of livestock products using AI technology - Google Patents

Abstract

The present invention relates to an automatic identification system for parts of livestock products using artificial intelligence technology. To ensure that a food processing company cannot change the name of a meat part while processing carcasses brought into a meat processing plant and producing livestock products for each part of meat, after wrapping cut meat in meat wrapping paper, the meat is placed on a scale, and a picture thereof is taken to automatically acquire an image of a meat packaged portion and the weight value of the portion. By executing a machine learning artificial intelligence program that compares an automatically acquired meat packaged portion meat image and the weight value thereof with a data set including a partial meat portion image for automatic identification of the partial meat portion stored in a database and the weight value for each image, the name of the packaged portion of meat is automatically determined, and the same is automatically printed by marking on the label attached to the meat packaging. According to the present invention, by preventing an attempt to falsely indicate the name of a part of livestock product by a food processing engineer during carcass processing and production in a meat processing factory, administrative procedures such as monitoring and monitoring of cases of false labeling of part names of livestock products or illegal distribution of livestock products with false part names in administrative agencies for the management and supervision of livestock products are reduced, and it can be expected to reduce administrative costs. By complying with legal regulations and standards, livestock meat processing plants can reduce damages caused by violations.

Description

System for automatically distinguishing a cut meat part of livestock products using AI technology

The present invention relates to a system for determining a part of a livestock product, and more particularly, to a system for automatically discriminating a part of a livestock product using an artificial intelligence (AI) technology.

In general, livestock such as cattle or pigs raised for slaughter in livestock farms are identified by their unique individual numbers and, when slaughtered at the slaughterhouse, are carried out as carcasses assigned with carcass numbers corresponding to the individual number and transported to the meat processing plant.

After cutting the head and removing the intestines from the carcass brought in to the meat processing plant, the carcass is divided in half and processed into two carcasses.

For example, in the case of Korean beef, the two-part carcass is cut into partial meat parts such as sirloin, tenderloin, ethmoid, beef leg, and tail, and then packaged for meat packaging. (For example, Korean beef, beef cattle, hand pork, etc.), history number, grade, slaughterhouse, meat processing plant, etc. are attached to the meat packaging and distributed.

On the other hand, as described above, the meat processing ERP (Enterprise Resource Planning, Enterprise Resource Planning) system, which is installed to process the carcass brought into the meat processing factory to produce livestock products for each part of the meat, is an Open AP provided by the Livestock Product Quality Assessment Service. Using the data obtained through the connection with the child (Open API; Open Application Programming Interface, open application program interface), after checking the carcases that were slaughtered on the same day (e.g., Korean beef, beef cattle, Korean pigs, etc.), the production team The production instruction is issued, and basic information such as the carcass history number, breed, slaughterhouse, production date, grade, and gender are output on the screen of the production instruction display device (eg, touch monitor) installed in the meat processing plant.

If the production team confirms the production instruction for Korean beef carcases, the meat processor cuts the Korean beef carcass divided into two parts for each part, and then cuts the part meat with transparent plastic wrap to prevent contamination of the livestock products produced. After packaging the meat, put it on the scale, check the weight value, and directly input the name of the meat part produced by processing through the part name input screen displayed on the production instruction display device (eg, touch monitor).

After that, the label paper indicating the weight, part name, production date, expiration date, country of origin, livestock breed (eg, Korean beef, beef cattle, hand drawn pork, etc.), history number, grade, slaughterhouse, meat processing plant, etc. Once printed, the meat processing company attaches the label to the meat packaging and distributes it.

As described above, in the case where the meat processor processes the carcass for which the production instruction has been confirmed in the meat processing plant and produces livestock products by part, the meat processor may intentionally change the name of the part meat part of the carcass part, and in fact, the meat processing plant In Korea, there are cases of false labeling of the relevant part meat part name for parts of meat with high selling price or high preference, and livestock products with false part names are illegally distributed.

For example, when checking the range of weight values for each part based on 87 parts of beef determined by the Ministry of Agriculture, Food and Rural Affairs of the Republic of Korea, live weight (the weight of live cattle) and carcass weight (when divided into 2 parts after slaughtering) weight) and production weight (weight for each meat part) can be checked. , if the tenderloin was produced with a weight of 30 kg, the name of the tenderloin part can be considered as falsely indicated by the meat processing engineer during carcass processing and production at the meat processing plant.

However, in order to legally punish the illegal distribution of livestock products with false labeling of part names of livestock products as described above, or illegal distribution of livestock products with false part names, the administrative agency for the management and supervision of livestock products inspects all parts of carcass during processing and production of carcases. is practically impossible to do.

KR 10-1226460 B1

The present invention is to solve the conventional problems as described above, and an object of the present invention is to process the carcass brought into the meat processing factory to produce livestock products for each portion of the meat so that the food processing company cannot change the name of the portion of meat, After packaging the cut meat in meat wrapping paper, place it on a scale and take a picture to automatically acquire the image of the meat packaged portion and the weight value of the portion, and the automatically obtained image of the meat packaged portion and its weight value are stored in the database. The machine learning artificial intelligence program that compares the stored image of each part of meat for automatic identification of the part of meat and a dataset containing the weight value of each image is performed to automatically determine the name of the part of meat packaged in the meat package and is attached to the package of meat. It is to provide an automatic identification system for meat parts of livestock products using artificial intelligence technology that marks and automatically prints on label paper.

In order to achieve the object of the present invention as described above, the livestock product partial meat part automatic determination system using artificial intelligence technology according to the present invention is measured in a state in which the processed carcass part meat part is packed with meat packaging paper after being brought into the meat processing plant. a scale that automatically acquires the weight value of the meat-packed portion of meat when it is placed on a plate and transmits it to the kiosk; A cam that automatically acquires an image of the meat-packaged portion by automatically taking a picture of the meat-packed portion after a predetermined time elapses after the meat-packed portion is placed on the weighing plate of the scale and automatically transmits the image to the kiosk; The weight value of the meat packaged part meat part automatically acquired by the scale and the meat packaged part meat part image image automatically acquired by the cam are received by the computer installed in the housing and transmitted to the discrimination server, and automatically by the discrimination server a kiosk that receives the identified meat packaged portion meat part name to the computer and automatically prints it by marking on the label paper attached to the meat packaging paper through a label printer installed in the housing to be connected to the computer; and machine learning artificial to compare the meat packaged portion meat image and the weight value received from the computer of the kiosk with a dataset including images for each portion meat portion for automatic identification of portion meat portions stored in a database and a weight value for each image It is characterized in that it is composed of; a determination server that automatically determines the name of the meat packaged part meat portion by executing an intelligent program and transmits it to the computer of the kiosk.

In the system for automatically determining parts of livestock products using artificial intelligence technology according to the present invention, the computer of the kiosk includes a monitor and a keyboard, and a meat processing ERP (ERP; Enterprise Resource Planning, Enterprise Resource Planning, Enterprise Resource Planning) system installed in a meat processing factory Production instructions and carcass information (e.g., carcase history number, breed, slaughterhouse, production date, grade, gender, etc.) through 2 ends, etc.) is checked and displayed on the monitor.

In the system for automatically determining part of livestock products using artificial intelligence technology according to the present invention, when the computer of the kiosk receives the name of the part of meat packaged in the meat package automatically determined by the discrimination server, the corresponding part of the production item screen displayed on the monitor It is characterized in that the name of the portion of the meat packaged in the meat is displayed in a predetermined display color (eg, red).

In the system for automatically determining part of livestock products using artificial intelligence technology according to the present invention, the computer of the kiosk displays the name of the part of meat packaged in the meat package automatically determined by the discrimination server on the screen of the production item in a display color (eg, red) It is characterized in that it is automatically printed by marking on the label paper attached to the meat packaging paper of the portion of meat that is automatically identified by the label printer at the same time or after marking.

In the system for automatically determining parts of livestock products using artificial intelligence technology according to the present invention, the computer of the kiosk writes the names of parts of meat packaged by the identification server automatically determined by the discrimination server in the production item screen through a label printer on the label paper. After marking and automatic printing, overproduced or missing production items are identified, or live weight (living After analyzing the yields of cattle weight when present), carcass weight (weight at two divisions after slaughter), and production weight (weight for each portion of meat), if it is confirmed that the weight value for each portion of meat produced is outside the error range, It is characterized in that the production error is alerted through the production item screen or a separate alarm device (eg, an alarm light, a speaker, a buzzer, etc.).

According to the present invention, by preventing an attempt to falsely indicate the name of a part of livestock by a food processing engineer during carcass processing and production in a meat processing factory, an administrative agency for the management and supervision of livestock products illegally displays a case of false labeling of a part of livestock or a false name of livestock Administrative procedures such as monitoring and monitoring of distribution are reduced, and the effect of reducing administrative costs can be expected.

According to the present invention, the livestock meat processing plant can reduce damage to violations by complying with legal regulations and standards.

According to the present invention, by monitoring the yield of the production weight based on the information for each part of the part of the meat whose part name is automatically determined, it is possible to collect and analyze the production item (specification) and production error for each carcase and actively utilize it for improvement of the business environment and statistical data. .

1 is an embodiment showing an automatic identification system for parts of livestock products using artificial intelligence technology according to the present invention.
Fig. 2 is a front view showing the scale in which the kiosk and cam shown in Fig. 1 are installed.
Fig. 3 is a side view of the scale with the cam shown in Fig. 2;
Fig. 4 is a side view of the kiosk shown in Fig. 2;
Figure 5 is a flow chart explaining the operation of the livestock part meat part automatic determination system using artificial intelligence technology according to the present invention.

Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the accompanying drawings.

The system for automatically determining livestock parts meat parts using artificial intelligence technology according to the present invention described below is not limited to the following examples, and has ordinary knowledge in the art without departing from the gist of the technology claimed in the claims. It has the technical spirit to the extent that anyone who grows up can change it and implement it.

1 to 4, the livestock product partial meat part automatic determination system using artificial intelligence technology according to the present invention is configured to include a scale 100, a cam 200, a kiosk 300, and a determination server 400 .

The scale 100 automatically acquires the weight value of the packaged meat portion and transmits it to the kiosk 300 when the meat portion of the processed carcass is placed on the weighing plate after being brought into the meat processing plant. .

The cam 200 automatically takes a picture of the meat-packed portion-meat portion after a predetermined time has elapsed after the meat-packaged portion is placed on the weighing plate of the scale 100, and automatically displays the image of the meat-packed portion-meat portion. It is obtained and transmitted to the kiosk 300 .

The kiosk 300 is a computer 310 installed in the housing with the weight value of the meat packaged part meat part automatically acquired by the scale 100 and the meat packaged part meat part image image automatically acquired by the cam 200 ) and transmitted to the determination server 400 .

The kiosk 300 receives the name of the meat packaged part meat part automatically determined by the determination server 400 to the computer 310, and a label printer 320 installed in the housing to be connected to the computer 310. It is automatically printed by marking on the label attached to the meat wrapper through the

The computer 310 of the kiosk 300 includes a monitor 311 and a keyboard 312 .

The computer 310 of the kiosk 300 provides production instructions and carcass information (eg, carcass history number, livestock breed, slaughterhouse) through the meat processing ERP (Enterprise Resource Planning, Enterprise Resource Planning) system installed in the meat processing plant. , production date, grade, gender, etc.) are checked and displayed on the monitor 311 by checking a predetermined specification (eg, 2 tenderloin, 2 sirloin, 2 sieve, etc.) for the confirmed carcass.

When the computer 310 of the kiosk 300 receives the name of the meat packaged part meat part automatically determined by the determination server 400, the corresponding meat packaged part meat part in the product screen displayed on the monitor 311 The name is displayed in a predetermined display color (eg, red).

The computer 310 of the kiosk 300 displays the name of the meat packaged part meat part automatically determined by the discrimination server 400 in the display color (eg, red) in the production item screen, or at the same time Afterwards, the label printer 320 marks the part name on the label attached to the meat packaging paper of the part meat part automatically determined and automatically prints it.

After the computer 310 of the kiosk 300 marks the name of the meat packaged part meat part automatically determined by the determination server 400 on the label paper through the label printer 320 in the production item screen and automatically prints it , an overproduced or missing overproduced item is identified against a pre-determined specification for the carcass (eg 2 tenderloin, 2 sirloin, 2 loin, etc.), or As a result of analyzing the yield of carcass weight (weight at 2 divisions after slaughtering) and production weight (weight for each part meat part), if it is confirmed that the weight value for each part meat part produced is outside the error range, the above production item screen or separately A production error is alerted through an alarm (eg, alarm light, speaker, buzzer, etc.) of the

The kiosk 300 is installed when the computer 310 of the kiosk 300 collects and analyzes the production items and production errors for each carcase by monitoring the yield of the production weight based on the information for each part of the meat part where the part name is automatically determined. It can be actively used for improving the business environment and statistical data at the meat processing plant.

The determination server 400 stores the meat packaged portion meat image and the weight value of the meat packaged portion received from the computer 310 of the kiosk 300 in the database for automatic identification of the portion meat portion image and weight for each image. By executing a machine learning artificial intelligence program that compares the data set including the value, the name of the meat packaged portion is automatically determined and transmitted to the computer 310 of the kiosk 300 .

Preferably, the determination server 400 communicates with the computer 310 of the kiosk 300 through a wired communication network or a wireless communication network.

The determination server 400 provides the information received from the computer 310 of the kiosk 300, for example, the meat packaged portion meat image and its weight value, and the automatically determined name of the meat packaged portion meat portion from the Livestock Products Quality Assessment Service. It is desirable to automatically report to the livestock product history system of the Livestock Product Quality Assessment Service through linkage with the Open API (Open Application Programming Interface, open application program interface).

The determination server 400 automatically reports the information received from the computer 310 of the kiosk 300 to the livestock product history system of the Livestock Product Quality Assessment Service, an administrative agency for livestock product management and supervision (eg, Ministry of Agriculture, Food and Rural Affairs, food Ministry of Drug Safety, each local government, etc.), by sharing the information, livestock products produced by processing carcasses at the meat processing factory where the kiosk 300 is installed. Administrative procedures such as monitoring and monitoring of illegal distribution of drugs are reduced, and administrative costs can be reduced.

In addition, in the meat processing plant where the kiosk 300 is installed, damage to violations can be reduced by complying with legal regulations and standards.

The system for automatic identification of parts of livestock products using artificial intelligence technology according to the present invention configured as described above operates as follows.

5 is a flowchart for explaining the operation of the system for automatically determining livestock parts using artificial intelligence technology according to the present invention.

Hereinafter, with reference to Figure 5 will be described in detail the operation of the system for automatic identification of parts of livestock products using artificial intelligence technology according to the present invention.

First, the computer 310 of the kiosk 300 checks production instructions and carcass information through an Enterprise Resource Planning (ERP) system installed in a meat processing factory.

In this case, as the carcass information, the carcass history number, breed, slaughterhouse, production date, grade, gender, etc. may be provided.

Subsequently, the computer 310 of the kiosk 300 checks a predetermined specification for the conductor and displays it on the monitor 311 .

For reference, on the production item screen displayed on the monitor 311, production items produced by processing the corresponding conductor, such as two tenderloin, two sirloin, and two ends, are displayed.

If the screen of production items for Korean beef carcass is displayed on the monitor 311, thereafter, the meat processing company cuts the carcass divided into two parts of the Korean beef for each part, and then, in order to prevent contamination of livestock products, transparent plastic The portion of the meat cut with the wrap is packed with meat and then placed on the weighing plate of the scale (100).

As described above, when a portion of meat that is processed and produced from a carcass by a meat processor and packaged with transparent plastic wrap is placed on the weighing plate of the scale 100, the scale 100 is the weight value of the portion of the meat packaged with the meat. is automatically acquired and transmitted to the computer 310 of the kiosk 300 .

In addition, when the meat packaged portion is placed on the weighing plate of the scale 100, after a predetermined time (eg, 1 to 2 seconds) has elapsed, the cam 200 automatically takes a picture of the meat packaged portion An image of a portion of meat packaged in a photograph is automatically acquired and transmitted to the computer 310 of the kiosk 300 .

Then, the computer 310 of the kiosk 300 automatically acquires the weight value of the meat packaged part meat part obtained by the scale 100 and the meat packaged part meat part image automatically acquired by the cam 200. It is transmitted to the determination server (400).

Next, the determination server 400 stores the meat packaged portion meat image and the weight value of the meat packaged portion received from the computer 310 of the kiosk 300 in the database for automatic identification of the portion meat portion image and each image By executing a machine learning artificial intelligence program that compares with a dataset including a star weight value, the name of the meat packaged portion is automatically determined and transmitted to the computer 310 of the kiosk 300 .

For reference, the dataset stored in the database of the server 400 for each fan is configured to include a range of weight values as a comparison standard for each product and 1000 or more cut meat parts images for automatic discrimination. For example, it may include images of more than 1000 relief parts for automatic identification of relief parts corresponding to the variable range of relief weight values of 8 kg to 12 kg for Korean beef relief parts.

If the determination server 400 receives the image (eg, image presumed to be tenderloin) and the weight value of 8 kg from the computer 310 of the kiosk 300 by processing Korean beef carcass Ramen, the determination server 400 is a data set for identifying safe portions including images of relief portions for automatic determination of more than 1000 pieces of relief corresponding to the variable range of safe weight values of 8kg to 12kg and partial meat stored for each other production item Executes a machine learning artificial intelligence program that utilizes all datasets for determining the part, and automatically determines the name of the part of meat packaged in the meat packaged part image (eg, image presumed to be tender) as safe by performing the 8 kg weight value. do.

As described above, when the name of the portion of meat packaged by the determination server 400 is automatically determined and delivered to the computer 310 of the kiosk 300 , the computer 310 displays the product screen displayed on the monitor 311 . At the same time or after displaying the name of the meat packaged part meat part corresponding to the one in a predetermined display color (eg, red), the name of the part meat packaged part meat part name automatically determined through the label printer 320 is displayed on the meat packaging paper It is automatically printed by marking on the attached label paper.

At this time, on the label paper, the weight of the part meat part, automatically determined part name, production date, expiration date, country of origin, livestock breed (eg, Korean beef, beef cattle, hand drawn pork, etc.), history number, grade, slaughterhouse, meat processing plant, etc. are marked together is printed, and the meat processing company distributes the label paper by attaching it to the meat packaging paper.

On the other hand, as described above, the computer 310 of the kiosk 300 marks the name of the meat packaged part meat part automatically determined by the discrimination server 400 on the label paper through the label printer 320 in the production item screen. After automatic printing, if an over-produced or missing-produced item is identified in comparison with a predetermined production item for the conductor (eg, 2 tenderloin, 2 sirloin, 2 sieve, etc.), the computer of the kiosk 300 ( 310) alerts a production error through the production item screen or a separate alarm device (eg, an alarm lamp, a speaker, a buzzer, etc.).

For example, in the case of producing livestock products by processing one Korean beef carcass into partial meat parts at a meat processing plant, 2 tenderloin, 2 sirloin, and 2 stalks must be produced from one Korean beef carcass according to a predetermined production item. If 3 pieces are produced, an excess production error may be derived by mixing one more sirloin produced from Korean beef carcass with a different history number to the Korean beef carcass product of the relevant history number at the meat processing plant.

On the other hand, if one sirloin is produced, a production error may be derived in which one sirloin is omitted from the product of the Korean beef carcass of the relevant history number at the meat processing plant.

On the other hand, as described above, the computer 310 of the kiosk 300 writes the name of the meat packaged part meat part automatically determined by the discrimination server 400 in the production item screen through the label printer 320 on the label paper. After automatic printing, if the yield of live weight (weight of Korean beef when alive), carcass weight (weight at the time of two divisions after slaughter) and production weight (weight of each piece of meat) were analyzed, the produced portion of meat When it is confirmed that the star weight value is out of the error range, the computer 310 of the kiosk 300 alerts the production error through the production item screen or a separate alarm (eg, alarm light, speaker, buzzer, etc.).

For example, according to the pre-determined production item, the portion of tenderloin must be produced in an amount of 8 to 12 kg from one Korean beef carcass. can be deduced as an error.

100: scale 200: cam
300: kiosk 310: computer
311: monitor 312: keyboard
320: label printer 400: determination server

Claims (5)

After being brought into the meat processing plant, when the processed meat portion of the carcass is put on the weighing plate in a state of packaging with meat packaging paper, the scale 100 automatically acquires the weight value of the meat packaged portion and transmits it to the kiosk 300;
After a predetermined time elapses after the meat-packaged portion is placed on the weighing plate of the scale 100, a picture of the meat-packaged portion is automatically acquired to automatically acquire an image of the meat-packaged portion of the meat portion, and the kiosk (300) Cam 200 to transmit to;
The weight value of the meat packaged part meat part automatically acquired by the scale 100 and the meat packaged part meat part image image automatically acquired by the cam 200 are received by the computer 310 installed in the housing to determine the server A label printer 320 installed in the housing to be transmitted to 400, and to be connected to the computer 310 by receiving the name of the meat packaged portion meat part automatically determined by the determination server 400 to the computer 310. A kiosk 300 for automatically printing by marking on the label attached to the meat packaging paper through the; and
A data set including the image of each piece of meat for automatic identification of the portion of meat and the weight value of each image, which are stored in the database for the packaged meat image and the weight value received from the computer 310 of the kiosk 300; A determination server 400 that performs a machine learning artificial intelligence program to compare and automatically determines the name of the meat-packaged portion of meat and transmits it to the computer 310 of the kiosk 300;
Automatic identification system for parts of livestock products using artificial intelligence technology, characterized in that it consists of. According to claim 1, wherein the computer (310) of the kiosk (300) includes a monitor (311) and a keyboard (312), the meat processing ERP (Enterprise Resource Planning; Enterprise Resource Planning) installed in the meat processing plant. Automated identification system for part meat parts of livestock products using artificial intelligence technology, characterized in that a predetermined production item (specification) is checked for the carcass for which the production instruction and carcass information have been confirmed through the system and displayed on the monitor (311). The production according to claim 1 or 2, wherein the computer 310 of the kiosk 300 is displayed on the monitor 311 when receiving the name of the meat packaged portion meat portion automatically determined by the determination server 400 A system for automatic identification of parts of livestock products using artificial intelligence technology, characterized in that the name of the part of meat packaged in the meat package is displayed in a predetermined display color on the item screen. The method according to claim 3, wherein the computer 310 of the kiosk 300 displays the name of the meat packaged part meat part name automatically determined by the discrimination server 400 in the production item screen as a display color, or at the same time Automated identification system for parts of livestock products using artificial intelligence technology, characterized in that the label printer 320 automatically marks and prints the part names attached to the meat packaging of the part meat parts through the label printer 320. 5. The method of claim 4, wherein the computer 310 of the kiosk 300 writes the meat packaged partial meat part name automatically determined by the discrimination server 400 on the label paper through the label printer 320 in the production item screen. After marking and automatic printing, excess or omission production is confirmed compared to the predetermined production items for carcasses Automated identification system for parts of livestock products using artificial intelligence technology, characterized in that when it is confirmed that the range is out of range, a production error is alerted through the production item screen or a separate alarm.

Images