KR102654311B1 - A Smart Farm Farming Data Conversion System Using Optical Character Recognition Model Based On Artificial Intelligence - Google Patents

Abstract

The smart farm farming data conversion system using an artificial intelligence-based optical character reading model according to the present invention includes an image collection unit that collects images of farming data; an image pre-processing unit that pre-processes the farming data images collected by the image collection unit; A text conversion unit that converts the agricultural data image into text by applying the agricultural data image pre-processed in the image pre-processing unit to an artificial intelligence-based optical character recognition (AI-OCR) model; a category recommendation unit that analyzes the text converted by the text conversion unit and recommends a category by subject that matches the text; a document creation unit that stores the text in a specific subject category selected by the user among the subject categories recommended by the category recommendation unit and generates an electronic document for agricultural data including the category of the specific subject; and a document storage unit that stores the electronic farming data document generated by the document generation unit; and a document display unit that displays the electronic farming data document to the user.

Description

A Smart Farm Farming Data Conversion System Using Optical Character Recognition Model Based On Artificial Intelligence}

An embodiment of the present invention relates to a smart farm farming data conversion system using an artificial intelligence-based optical character reading model.

OCR (Optical Character Recognition) is a technology that recognizes and reads characters by shining light on printed (or handwritten) characters and detecting the strength and weakness of the reflected rays.

In the case of existing OCR technology, the recognition rate results show that the OCR project satisfies the need for automatic processing through text, but the actual recognition rate of the OCR engine is very low, so the remaining recognition rate must be supplemented by manual correction, and the results are utilized. There is a problem that use is decreasing due to decreased confidence in sex.

Meanwhile, among these OCRs, AI OCR, which incorporates artificial intelligence (AI), is a technology that has been further strengthened and developed to recognize even handwriting and unstructured information, improving accuracy and efficiency, and reducing time. Recently, AI OCR has been utilized to Although various services such as financial, medical, educational, and public are provided, there are no cases of agricultural data conversion technology being applied in the smart farm field.

In general, a farming log is a record of what happened and what work was done at the farm, and can be used to determine the next year's crop, etc. In addition, for farms receiving budget support from the government, submission of a farming log is sometimes essential. .

Currently, the number of farms using mobile phones or PCs has increased, but there are still farms that write farming logs by hand. In other words, there are many cases where farming logs, which are farming data, are prepared and stored in an analog manner, such as by public institutions providing farming notebooks (diaries) to farmers.

Republic of Korea Patent Publication No. 10-1516684 (May 11, 2015)

The present invention is an artificial intelligence-based optical character that can generate an electronic document that can secure the accuracy and reliability of the farming data by converting images of hand-written farming data (farming logbook) into text using AI OCR. The purpose is to provide a smart farm farming data conversion system using a reading model.

In addition, the purpose of the present invention is to provide a smart farm farming data conversion system using an artificial intelligence-based optical character reading model that increases the recognition rate of hand-written farming data, reduces costs due to the use of manpower, and utilizes records. do.

In addition, the present invention is an artificial intelligence-based system that saves time through quick conversion, prevents damage and loss of paper farming logs, can be organized in a consistent form, and allows for more convenient farm management using calendar and search functions. The purpose is to provide a smart farm farming data conversion system using the optical character reading model.

In addition, the present invention adds a function to certify the reliability of electronic farming data documents and provides a smart farm farming data conversion system using an artificial intelligence-based optical character reading model that can be developed for submission as evidence to national public institutions. The purpose is to provide

A smart farm farming data conversion system using an artificial intelligence-based optical character reading model according to the present invention to achieve the above-described purpose includes an image collection unit that collects images of farming data; an image pre-processing unit that pre-processes the farming data images collected by the image collection unit; A text conversion unit that converts the agricultural data image into text by applying the agricultural data image pre-processed in the image pre-processing unit to an artificial intelligence-based optical character recognition (AI-OCR) model; a category recommendation unit that analyzes the text converted by the text conversion unit and recommends a category by subject that matches the text; a document creation unit that stores the text in a specific subject category selected by the user among the subject categories recommended by the category recommendation unit and generates an electronic document for agricultural data including the category of the specific subject; and a document storage unit that stores the electronic farming data document generated by the document generation unit; and a document display unit that displays the electronic farming data document to the user.

Additionally, the image preprocessing unit may preprocess the farming data image through tilt correction, blank space removal, noise removal, and rotation correction of the farming data image.

In addition, the text conversion unit includes an object classification module that detects objects included in the farming data image and classifies the objects by type; a text conversion module that converts objects by type classified by the object classification module into text; and a text correction module that checks the display and grammar of the text converted by the text conversion module and corrects incorrectly displayed text.

In addition, the category recommendation unit includes a text classification module that analyzes the location, type, and format of the text converted by the text conversion unit and classifies the text by topic; and a category recommendation module that automatically recommends a thematic category that matches the thematic classification of the text classified by the text classification module.

Additionally, the document creation unit may include a text storage module that stores the text in a category of a specific topic selected by the user; an important word confirmation module that determines the specific word as an important word when a specific word is selected by the user in the text; and a document creation module that generates the agricultural data electronic document including a category of a specific subject in which the text is stored.

In addition, when the text storage module stores the text in the category of the specific topic, the important words determined in the important word confirmation module are displayed in a different display method from other words constituting the text in the category of the specific topic. It can be saved in .

In addition, the present invention may further include a document form storage unit in which a document form for generating the agricultural data electronic document is stored.

In addition, the document form storage unit includes a basic form storage module that stores a basic document form for generating the agricultural data electronic document; a document form modification module that modifies the basic document form according to user settings to form a modified document form; and a modified form storage module that stores a modified document form for generating the agricultural data electronic document.

In addition, the present invention may further include a document search unit that searches for a specific electronic farming data document stored in the document storage unit.

In addition, the document search unit includes a date search module that searches for a specific agricultural data electronic document stored in the document storage unit by inputting a specific date; and a word-specific search module that searches for a specific agricultural data electronic document stored in the document storage unit by inputting a specific word.

As described above, according to the smart farm farming data conversion system using the artificial intelligence-based optical character reading model according to the present invention, images of hand-written farming data are converted into text using AI OCR to provide information about the farming data. It has the effect of creating electronic documents that can ensure accuracy and reliability.

In addition, the present invention has the effect of increasing the recognition rate of hand-written farming data, reducing costs due to the use of manpower, and enabling the use of records.

In addition, the present invention saves time through quick conversion, prevents damage and loss of farming logs written on paper, organizes them in a consistent form, and enables more convenient farm management using calendar and search functions. .

In addition, the present invention has the effect of adding a function to certify the reliability of electronic agricultural data documents, which can be developed for use as evidence for submission to national public institutions.

Figure 1 is a first configuration diagram of a smart farm farming data conversion system using an artificial intelligence-based optical character reading model according to the present invention.
Figure 2 is an example of a farming data image collected by the image collection unit according to the present invention.
Figure 3 is a configuration diagram of a text conversion unit according to the present invention.
Figure 4 is a configuration diagram of a category recommendation unit according to the present invention.
Figure 5 is an example diagram of categories by subject recommended by the category recommendation unit according to the present invention.
Figure 6 is a configuration diagram of a document creation unit according to the present invention.
Figure 7 is an example of an electronic farming data document generated by the document generation unit according to the present invention.
Figure 8 is a second configuration diagram of a smart farm farming data conversion system using an artificial intelligence-based optical character reading model according to the present invention.
Figure 9 is a configuration diagram of a document form storage unit according to the present invention.
Figure 10 is a configuration diagram of a document search unit according to the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings. First of all, it should be noted that the same components or parts in the drawings are given the same reference numerals as much as possible. In describing the present invention, detailed descriptions of related known functions or configurations are omitted so as not to obscure the gist of the present invention.

Figure 1 is a first configuration diagram of a smart farm farming data conversion system using an artificial intelligence-based optical character reading model according to the present invention.

As shown in FIG. 1, the smart farm farming data conversion system using an artificial intelligence-based optical character reading model according to the present invention includes an image collection unit 100, an image preprocessing unit 200, and a text conversion unit 300. , a category recommendation unit 400, a document creation unit 500, a document storage unit 600, and a document display unit 700.

The image collection unit 100 may collect images of farming data.

Figure 2 is an example of a farming data image collected by the image collection unit according to the present invention.

Specifically, the image collection unit 100 is an image of farming data, and as shown in FIG. 2, can collect images of a farming diary, which is unstructured data written by hand, where the farming diary image is a photo taken. and can be generated through various other methods.

The image preprocessing unit 200 may preprocess the agricultural data images collected by the image collection unit.

Specifically, the image preprocessing unit 200 corrects the tilt existing in the farming log image, removes the blank space and noise of the image, and Through the existing rotation correction, text conversion from the corresponding image can be facilitated by the text conversion unit described later.

The text conversion unit 300 can convert the farming data image into text by applying the agricultural data image pre-processed in the image pre-processing unit to an artificial intelligence-based optical character recognition (AI-OCR) model.

Figure 3 is a configuration diagram of a text conversion unit according to the present invention.

Specifically, the text conversion unit 300 may include an object classification module 310, a text conversion module 320, and a text correction module 330, as shown in FIG. 3.

The object classification module 310 can detect objects included in the farming data image and classify the objects by type.

Specifically, the object classification module 310 can classify the object by detecting the location and type of the object included in the farming data image. The object classification module 310 can classify the object by detecting the location and type of the object included in the farming data image. , shapes, etc.

The text conversion module 320 can convert objects by type classified by the object classification module into text.

Specifically, as described above, the text conversion module 320 can convert objects by type classified in the object classification module into text through an artificial intelligence-based optical character recognition (AI-OCR) model.

The text correction module 330 can correct incorrectly displayed text by checking the display and grammar of the text converted by the text conversion module.

Specifically, the text correction module 330 can perform AI-OCR conversion to check the display format and grammar of the text, then detect incorrectly written typos in the converted text and correct the typos.

The category recommendation unit 400 may analyze the text converted by the text conversion unit and recommend a topic-specific category that matches the text.

Figure 4 is a configuration diagram of a category recommendation unit according to the present invention.

Specifically, the category recommendation unit 400 may include a text classification module 410 and a category recommendation module 420, as shown in FIG. 4 .

The text classification module 410 can classify the text by subject by analyzing the location, type, and format of the text converted by the text conversion unit.

Specifically, the text classification module 410 analyzes the location of the written text through AI-OCR conversion, the type of text (Korean, numeric, English, etc.), and the display format (date, amount, sentence, etc.) Among the multiple thematic categories included in the data (farming log form), it can be classified into a specific topic related to the text.

The category recommendation module 420 can automatically recommend thematic categories that match the thematic classification of the text classified in the text classification module.

Figure 5 is an example diagram of categories by subject recommended by the category recommendation unit according to the present invention.

Specifically, the category recommendation module 420, for example, when the thematic classification of the text classified in the text classification module is related to farming work, as shown in FIG. 5, the category recommendation module 420 recommends multiple items included in the farming data (farming log form). Among the thematic categories of , “Work steps (task name)” and “Detailed work details” can be recommended as the thematic categories that match the relevant farming work topic.

The document creation unit 500 may store related text in a category of a specific subject selected by the user among the subject categories recommended by the category recommendation unit and generate an electronic document about farming data including the category of the specific subject.

Figure 6 is a configuration diagram of a document creation unit according to the present invention.

Specifically, the document creation unit 500 may include a text storage module 510, a key word confirmation module 520, and a document creation module 530, as shown in FIG. 6 .

The text storage module 510 may store related text in a category of a specific topic selected by the user.

Specifically, the text storage module 510, for example, allows the user to select “detailed work details” among the thematic categories “work steps (task names)” and “detailed work details” recommended by the category recommendation module 420. In this case, text related to the relevant farming work can be stored in the detailed work details.

When a specific word is selected by a user in the text, the important word confirmation module 520 can confirm the specific word as an important word.

Specifically, the important word confirmation module 520 is linked with the text storage module 510, and when a specific word selected by the user is confirmed as an important word, the important word can be provided to the text storage module.

Therefore, the text storage module 510 can store the relevant important words so that they can be selected separately in the process of storing the relevant text in the category of a specific topic. The text storage module 510 can store the important words determined in the important word confirmation module. Words can be stored in a category of a specific topic with a different display method (color, font, thickness, etc.) from other words that make up the text.

That is, the text storage module 510 can save text in the category location of a designated form when the user selects a desired category, and important text can be stored separately.

The document creation module 530 can generate an electronic farming data document containing a category of a specific subject in which text is stored.

Figure 7 is an example of an electronic farming data document generated by the document generation unit according to the present invention.

Specifically, as shown in FIG. 7, the document creation module 530 can generate an electronic farming data document in which related content is stored in a plurality of thematic categories included in the farming data (farming log form).

The document storage unit 600 may store the electronic farming data document generated in the document creation unit.

The document display unit 700 can display electronic farming data documents to the user.

Specifically, the document display unit 700 can display an electronic agricultural data document to the user to monitor whether the specific text selected by the user is properly entered in a specific position in the agricultural data document form.

Figure 8 is a second configuration diagram of a smart farm farming data conversion system using an artificial intelligence-based optical character reading model according to the present invention.

Meanwhile, as shown in FIG. 8, the smart farm farming data conversion system using the artificial intelligence-based optical character reading model according to the present invention includes the image collection unit 100, image preprocessing unit 200, and text conversion. In addition to the unit 300, the category recommendation unit 400, the document creation unit 500, the document storage unit 600, and the document display unit 700, its configuration includes a document form storage unit 800, and a document search unit. It may further include (900).

The document form storage unit 800 may store a document form for creating an electronic farming data document.

Figure 9 is a configuration diagram of a document form storage unit according to the present invention.

Specifically, the document form storage unit 800 may include a basic form storage module 810, a document form modification module 820, and a modified form storage module 830, as shown in FIG. 9.

The basic form storage module 810 can store a basic document form for creating an electronic farming data document.

Specifically, the basic form storage module 810 can store a basic electronic document form with a basic form that matches various farming data written by hand, and this basic form storage module 810 includes a category recommendation unit 400. ), the category recommendation section can recommend specific thematic categories that exist in the basic document form.

The document form modification module 820 can form a modified document form by modifying the basic document form according to user settings.

Specifically, the document form modification module 820 can modify the basic document form into a user-customized document form by providing various tools that can modify the basic document form according to the user's needs.

The modified form storage module 830 can store a modified document form for creating an electronic farming data document.

Specifically, the modified form storage module 830 can store a modified electronic document form modified by a user that matches each of the various farming data. This modified form storage module 830 is also linked to the category recommendation unit 400. The category recommendation section can recommend categories for each specific topic that exist in the modified document form.

The document search unit 900 can search for a specific electronic farming data document stored in the document storage unit.

Figure 10 is a configuration diagram of a document search unit according to the present invention.

Specifically, the document search unit 900 is linked with the document storage unit 600 and can be displayed on the document display unit 700. As shown in FIG. 10, the document search unit 900 is a search module by date. (910), and may include a word-specific search module (920).

The search module 910 by date can search for a specific agricultural data electronic document stored in the document storage unit by inputting a specific date.

Specifically, the search module 910 by date can allow the user to view the farming log by date using the calendar function.

The word-specific search module 920 can search for a specific agricultural data electronic document stored in the document storage unit by inputting a specific word.

Specifically, the word-by-word search module 920 allows the user to view the farming log with the content desired by the user using a search function for a specific word, and the user searches using a specific word selected as an important word. By performing , you can easily search for farming logs containing relevant important words.

In this way, according to the present invention, by converting images of hand-written farming data into text using AI OCR, the accuracy and reliability of the farming data can be secured, and it is possible to reduce costs due to the use of manpower and utilize records. , saves time, prevents damage and loss of paper farming logs, can be organized in a consistent form, and allows for more convenient farm management using calendar and search functions.

As described above, the smart farm farming data conversion system using the artificial intelligence-based optical character reading model according to the present invention has been described with reference to the drawings, but the present invention is not limited to the embodiments and drawings disclosed in this specification. It goes without saying that various modifications can be made by those skilled in the art within the scope of the present invention.

100: Image collection unit 200: Image pre-processing unit
300: Text conversion unit 310: Object classification module
320: Text conversion module 330: Text correction module
400: Category recommendation unit 410: Text classification module
420: Category recommendation module 500: Document creation unit
510: Text storage module 520: Important word confirmation module
530: Document creation module 600: Document storage unit
700: Document display unit 800: Document form storage unit
810: Basic form storage module 820: Document form editing module
830: Modification form storage module 900: Document search unit
910: Search module by date 920: Search module by word

Claims (10)

An image collection unit that collects images of farming data;
an image pre-processing unit that pre-processes the farming data images collected by the image collection unit;
A text conversion unit that converts the agricultural data image into text by applying the agricultural data image pre-processed in the image pre-processing unit to an artificial intelligence-based optical character recognition (AI-OCR) model;
a category recommendation unit that analyzes the text converted by the text conversion unit and recommends a category by subject that matches the text;
a document creation unit that stores the text in a specific subject category selected by the user among the subject categories recommended by the category recommendation unit and generates an electronic document for agricultural data including the category of the specific subject;
a document storage unit that stores the electronic farming data document generated in the document creation unit;
a document display unit that displays the electronic farming data document to the user;
a document form storage unit storing a document form for generating the agricultural data electronic document; and,
It includes a document search unit that searches for a specific agricultural data electronic document stored in the document storage unit,
The document creation unit,
a text storage module that stores the text in a category of a specific topic selected by the user; an important word confirmation module that determines the specific word as an important word when a specific word is selected by the user in the text; And, a document creation module for generating the agricultural data electronic document containing a category of a specific subject in which the text is stored,
The text storage module is,
When storing the text in the category of the specific topic, the important words confirmed in the important word confirmation module are stored in the category of the specific topic in a display method different from other words constituting the text,
The document form storage unit,
A basic form storage module that stores a basic document form for generating the agricultural data electronic document; A document form modification module that forms a modified document form by modifying the basic document form into a user-customized document form by providing a tool (TOOL) that can modify the basic document form according to user settings; And, a modified form storage module that stores a modified document form for generating the agricultural data electronic document,
The document search unit,
A search module by date that searches for a specific agricultural data electronic document stored in the document storage unit by inputting a specific date; And, a word-specific search module that searches for a specific agricultural data electronic document stored in the document storage unit by inputting a specific word; a smart farm agricultural data conversion system using an artificial intelligence-based optical character reading model, comprising: .
According to clause 1,
The image preprocessing unit,
A smart farm farming data conversion system using an artificial intelligence-based optical character reading model, characterized in that the farming data image is pre-processed through tilt correction, blank space removal, noise removal, and rotation correction of the farming data image.
According to clause 1,
The text conversion unit,
an object classification module that detects objects included in the farming data image and classifies the objects by type;
a text conversion module that converts objects by type classified by the object classification module into text; and
A smart farm farming data conversion system using an artificial intelligence-based optical character reading model, comprising a text correction module that checks the display and grammar of the text converted from the text conversion module and corrects incorrectly displayed text.
According to clause 1,
In the category recommendation section,
a text classification module that analyzes the location, type, and format of the text converted by the text conversion unit and classifies the text by topic; and
A smart farm farming data conversion system using an artificial intelligence-based optical character reading model, comprising a category recommendation module that automatically recommends a thematic category that matches the thematic classification of the text classified in the text classification module. .
delete delete delete delete delete delete

Images