KR20210056560A - Device and Method For Providing Map Information - Google Patents

Abstract

The present invention provides a system and method for providing map information, more specifically, map information of a shopping mall in a building. According to an embodiment of the present invention, an apparatus for providing map information using text search comprises: a data collection unit for receiving an image obtained by photographing a shopping mall in a specific area and location information; a text extraction unit for extracting a text from the collected images using an optical character reader (OCR) program; a database for classifying texts other than business names and phone numbers from the extracted text and storing the classified texts and related words; a learning unit for learning to match the photographed image by performing machine learning of the classified text and the related words; and a map extraction unit for extracting map information using the collected location information. Based on a word searched by a user terminal, the database searches for a related word, and shopping malls adjacent to the current location of the user terminal and including the related word are displayed.

Description

Device and Method For Providing Map Information}

The present invention relates to an apparatus and method for providing map information, and more particularly, map information of a building or a shopping mall.

In general, to register a map, access a platform such as Naver and enter the company name, phone number, address, and business type. Then, search and register the address of the business you want to enter. After accessing and searching for the company's address, enter the company name, phone number, address, and business type in the corresponding location.

However, since the conventional map registration method is directly registered by the company, it is difficult for older people to register the map because it is necessary to register directly, and in the case of a traditional market or in a market that sells the same type of products, individual promotions If it is determined that is not necessary, the company often does not register. Therefore, since the map service only serves a part of the companies that actually exist, there are many problems with companies that are not displayed than those displayed in the map service.

In addition, in the case of a traditional market or a special commercial district that handles all the same items in a specific area, it is generally displayed as one block rather than providing a shopping mall building or an internal map within the market. There is a problem that the information and the information on the map are different.

The present invention provides an apparatus and method for producing and providing map information, and more particularly, map information of a shopping mall in a building.

A map information providing apparatus according to the present invention is a map information providing apparatus using text search, comprising: a data collection unit for receiving an image of a shopping mall in a specific area and location information; A text extraction unit for extracting text from the collected image using an optical character reader (OCR) program; A database for classifying texts other than business names and telephone numbers from the extracted texts and storing the classified texts and related words; A learning unit that learns to match the captured image by machine learning the classified text and related words; And a map extracting unit for extracting map information using the collected location information, and based on the words searched by the user terminal, the database searches for related words, and adjacent and related words based on the current location of the user terminal It is possible to display a mall that includes.

Here, the text extracting unit may classify a business name, a phone number, and the remaining text from the photographed image of a shopping mall by machine learning, and the database may store the business name and phone number together with the classified text and related words.

In addition, the user terminal may input or search for information on the shopping street, access the database, and correct the image, extracted text, or related words of the shopping street.

In addition, the map information extraction unit creates a point cloud using LiDAR, slices the height axis into several layers, averages each of the slices to generate 2D, generates a drawing, and compares the map information with a digital map. Can be obtained.

In addition, the related word includes at least one of a technical term, an industry idiom, and a non-standard word, and the learning unit may learn and expand the related words through machine learning.

In addition, the method for providing map information according to the present invention is a method for providing map information using text search, the method comprising: receiving, by a data collection unit, an image and location information of a shopping mall in a specific area from a user terminal; Extracting text from the collected image using an optical character reader (OCR) program by a text extraction unit; Classifying, by a database, texts other than business names and telephone numbers from the extracted texts, and storing the classified texts and related words; Learning, by a learning unit, machine learning the classified text and related words to match the captured image; A step of extracting map information using the collected location information by a map extraction unit; and, when the user terminal searches for a word, the database searches for a related word, and is adjacent based on the current location of the user terminal. It is possible to display a mall containing related words.

Here, the text extracting unit may classify the business name, phone number, and the rest of the text from the photographed image of the shopping street, and the database may store the business name and phone number together with the classified text and related words.

In addition, the user terminal may input or search for information on the shopping street, access the database, and correct the image, extracted text, or related words of the shopping street.

In addition, the map information extraction unit creates a point cloud using LiDAR, slices the height axis into several layers, averages each of the slices to generate 2D, generates a drawing, and compares the map information with a digital map. Can be obtained.

In addition, the related word includes at least one of a technical term, an industry idiom, and a non-standard word, and the learning unit may learn and expand the related words through machine learning.

The apparatus and method for providing map information according to the present invention captures an image of a shopping mall through a user terminal or a traveling robot, extracts a text related to a product from the image, and stores the related word together in a database, and the location and search word of the user terminal Through this, it is possible to easily provide map information for shopping malls that handle related products.

1 is a block diagram of a system for providing map information according to an embodiment of the present invention.
2 is a flow chart illustrating a shopping mall map input step in a method for providing map information according to an embodiment of the present invention.
3A is a diagram illustrating an example screen of a shopping mall input in a map information providing system according to an embodiment of the present invention.
3B is a diagram illustrating a map display method through SLAM technology in a map information providing system according to an embodiment of the present invention.
4 is a flowchart illustrating a shopping mall map search step in a method for providing map information according to an embodiment of the present invention.
5 illustrates an example screen of a shopping mall search result in a map information providing system according to an embodiment of the present invention.

The preferred embodiments of the present invention will be described in detail with reference to the drawings to the extent that those of ordinary skill in the art can easily implement the present invention.

1 is a block diagram of a system for providing map information according to an embodiment of the present invention.

Referring to FIG. 1, a map information providing system 1 according to an embodiment of the present invention may include a first user terminal 10, a second user terminal 20, and a map information providing device 100.

Here, the first user terminal 10 and the second user terminal 20 may include devices such as a conventional PC, smart phone, tablet, or portable lidar. In addition, for example, the first user terminal 10 may be used to upload shopping street map information, and the second user terminal 20 may be used to search for shopping street map information. However, although the terms of the first user terminal 10 and the second user terminal 20 are separated, it is for convenience of explanation, and in reality, the user can use both as a single terminal to upload or search the map information. have.

The map information providing apparatus 100 may include a data collecting unit 110, a text extracting unit 120, a database 130, a learning unit 140, a map extracting unit 150, and a control unit 160.

First, the data collection unit 110 may collect an image photographing a specific region and location information of the photographed region.

Users can shoot traditional markets or special commercial areas dealing with similar items using a wide-angle camera. In addition, the user may upload the captured image to the map information providing apparatus 100 through the first device 10. Accordingly, the data collection unit 110 may collect an image uploaded from the first user terminal 10 and acquire location information of an area in which the collected image is photographed.

The text extraction unit 120 may extract text from the received image using an optical character reader (OCR) program. The image uploaded by the first user terminal 10 includes the name of the signboard of the store, the phone number, the brand of the displayed product, and a promotional text. The text extraction unit 120 extracts text included in the image. In addition, the text extracting unit 120 may separately extract the name of the signboard and the phone number from the text, and separate and extract the promotional phrase, the brand of the displayed product, the model name, and the like. In addition, the text extraction unit 120 may transmit the extracted text to the database 130.

The database 130 stores related words that have been previously inputted from various first user terminals 10 and learned with respect to the text transmitted from the text extracting unit 120. That is, words input by existing users through the first user terminal 10 are learned to form a related word, and the database 130 may store the extracted text and a related word corresponding thereto. Here, the related words may include technical terms, industry idioms, non-standard words, mixed terms, idioms, foreign languages, and the like.

The learning unit 140 performs machine learning based on text and images input from the first user device 10. In addition, the learning unit 140 may learn to output an image matching the input text when text is input to the previously constructed machine learning.

The map extraction unit 150 may extract and generate map information using location information of an image extracted through machine learning.

The controller 160 may extract a matched image by inputting the text input from the user into machine learning on which learning has been completed. In addition, the controller 160 may provide map information corresponding to the extracted image.

Hereinafter, a method of providing map information according to an embodiment of the present invention will be described.

2 is a flow chart illustrating a shopping mall map input step in a method for providing map information according to an embodiment of the present invention. 3A is a diagram illustrating an example screen of a shopping mall input in a map information providing system according to an embodiment of the present invention. 3B is a diagram illustrating a map display method through SLAM technology in a map information providing system according to an embodiment of the present invention. Hereinafter, FIGS. 2, 3A and 3B will be described with reference to FIG. 1.

1 to 3, the shopping street map input step includes a shopping street image capture and location information reception step (S11), text extraction and selection step (S12), a related word search and storage step (S13), and a map matching step (S14). ), may include a user correction step (S15).

In the step S11 of photographing a shopping mall image and receiving location information (S11), the first user terminal 10 photographs an image of the shopping mall and transmits the current location information of the first user terminal 10 to the data collection unit 110. . Here, the first user terminal 10 may be a conventional PC, smartphone, tablet, or the like, as mentioned above.

In addition, as another example of the first user terminal 10, a traveling robot equipped with a lidar may be used. In the case of a traveling robot, it is possible to move regardless of the presence or absence of a person, and it is also possible to acquire an image of a shopping street by moving the shopping area inside the building in the early morning hours when there is no person.

In addition, the first user terminal 10 may utilize a wide-angle camera. As is known, the wide-angle camera is highly versatile because it is possible to take close-up pictures and to take an entire image inside a shopping mall.

The first user terminal 10 may take an image in this way and transmit it to the data collection unit 100 together with location information.

In the text extraction and selection step (S12), the text extraction unit 120 extracts text from an image and extracts a related word. The photographed image may include a business name, items on display, advertisement text, and the like. The text extraction unit 120 extracts text included in an image. For example, as shown in FIG. 3, the text extraction unit 120 may extract text T such as "pump", "nozzle", "air conditioner", and "electric water heater" from the captured image. have. In this case, the extracted text may be in the form of a word or a sentence. In addition, the text extraction unit 120 extracts only nouns through morpheme analysis.

In the related word search and storage step S13, the text extraction unit 120 of the map information providing apparatus 100 may perform a related word search using the extracted noun. The map information providing apparatus 100 is used by users of various ages. Users of various ages can enter different search words, even if they search for the same object according to their age and interests. Accordingly, the text extracting unit 120 performs a word search related to the extracted text so as to provide map information matching according to a search word input from a user. Here, the related word includes at least one of technical terms, industry idioms, and non-standard words. In addition, the learning unit 140 may learn related words through machine learning, and accordingly, may gradually expand the related words.

In addition, the text extraction unit 120 transfers the extracted text and related words to the database 130, and the database 130 stores the text and related words. In addition, the database 130 stores the image of the shopping mall, the location, the extracted text, and the related word together, so that the second user terminal 20 can easily classify it to be searched later.

In the map matching step (S14), the map extractor 150 retrieves map information from the database 130, and may designate and map the location of the shopping mall according to the current location information of the first user terminal 10. In addition, in the mapping, it is possible to store the image of the shopping street, as well as the name and address of the shopping street, as well as the text extracted from the image and related words.

Meanwhile, as shown in FIG. 3B, when displaying on a map, a Simultaneous localization and mapping (SLAM) method may be used. The map extraction unit 150 (1) first creates a point cloud using a lidar, (2) slices the height axis into several layers, and (3) generates 2D by averaging each layer of the slice. (4) Finally, you can create a drawing. Here, the point cloud may be an RGB point cloud, contains depth information and color information at the same time, and has a feature similar to that of a low-resolution photo.

In addition, the map extraction unit 150 learns to match the image so that the outline of the drawing made in this way is matched by designating a number of points with a digital map that is distributed by the National Geographic Information Institute, and then overlapping the correct answer first and cutting it into pieces. I can make it. In this process, it is possible to distinguish between illegal extensions that are part of the building from temporary kiosks or automobiles, and on the map of the National Geographic Information Institute, it is possible to superimpose the approximate interior of the building even on the part marked as a simple hollow, square building. . In addition, when the learning is completed, a rough map can be obtained from a 3D picture in the form of a point cloud through the lidar, and it is possible to obtain map information by interlocking with the actual GIS data.

The user correction step S15 is a step in which the user corrects the map information through the first user terminal 10. For example, the first user terminal 10 may move the current location on a map and set it again, modify a desired term among text or related words, or may replace an image of a shopping mall with another image. In addition, when the first user terminal 10 uses the image of the walking robot mentioned above, it is possible to check the photographing position and image of the image, perform mapping on the map, and correct it.

4 is a flowchart illustrating a shopping mall map search step in a method for providing map information according to an embodiment of the present invention. 5 illustrates an example screen of a shopping mall search result in a map information providing system according to an embodiment of the present invention.

Referring to FIG. 4, the shopping street map search step of the map information providing method according to an embodiment of the present invention includes a shopping street search step (S16), the adjacent image display step (S17), a shopping mall selection step (S18), and a location on a map. It may include a step (S19) and a user correction step (S20).

Referring to FIG. 5 together, in the shopping mall search step (S16), the second user terminal 20 accesses the map information providing system 100 and searches for a desired shopping mall in the search window 22 on the display screen 21. Step. In this case, as mentioned above, the second user terminal 20 may perform a search through a word to be searched even if the name of the mall is not known. In addition, even if the second user terminal 20 is not an exact word, a shopping mall related to a desired product can be searched by simply inputting a related word already stored in the database 130.

In the adjacent image display step (S17), the map information providing system 100 may provide the second user terminal 20 with information on a shopping mall as a result of the search. In this case, based on the current location of the second user terminal 20, adjacent shopping malls may be displayed on the map. Specifically, the information on the shopping street may be displayed together with the image 23 of the shopping street, the extracted words and related words 24, and the location 25 on the map on the search screen 21.

The shopping center selection step S18 is a step of selecting a desired shopping area from among several shopping areas on which the second user terminal 20 is displayed. In this case, the second user terminal 20 may select a shopping area in which a desired product is located through the location of the shopping area and the extracted text and related words.

In the step of displaying the location on the map (S19), the path from the current location of the second user terminal 20 to the selected shopping mall is displayed on the map. Following the path of the second user terminal 20, the user can reach the actual shopping mall with the most convenient path.

In the user correction step S20, the second user terminal 20 corrects the searched shopping mall information. For example, when an image of an actual shopping mall has changed, has been closed, or it is necessary to add a new text, the second user terminal 20 may correct the information of the shopping mall. Accordingly, the shopping street information of the map information providing system 100 may be updated by participation of users.

What has been described above is only one embodiment for implementing the system and method for providing map information according to the present invention, and the present invention is not limited to the above embodiment, and as claimed in the claims below, Without departing from the gist, anyone of ordinary skill in the field to which the present invention pertains will have the technical spirit of the present invention to the extent that various changes can be implemented.

One; Map information providing system 10; 1st user terminal
20; A second user terminal 100; Map information providing device
110; Data collection unit 120; Text extraction unit
130; Database 140; Learning Department
150; Map extraction unit 160; Control unit

Claims (10)

In the map information providing apparatus using text search,
A data collection unit for receiving an image of a shopping mall in a specific area and location information;
A text extraction unit for extracting text from the collected image using an optical character reader (OCR) program;
A database for classifying the remaining texts from the extracted texts except for a business name and a telephone number through machine learning, and storing the classified texts and related words;
A learning unit that learns to match the captured image by machine learning the classified text and related words; And
Including a map extraction unit for extracting map information using the collected location information,
A map information providing apparatus for the database to search for a related word based on the word searched by the user terminal, and to display a shopping mall adjacent to the user terminal and including the related word based on the current location of the user terminal. The method of claim 1,
The text extracting unit classifies the business name and phone number and the rest of the text from the photographed image of the shopping street,
The database is a map level providing device for storing the business name and phone number together with the classified text and related words. The method of claim 1,
The user terminal may input or search for information on a shopping street, access the database, and correct the image, extracted text, or related words of the shopping street. The method of claim 1,
The map information extraction unit creates a point cloud using LiDAR, slices the height axis into several layers, averages each layer of the slice to generate 2D, generates a drawing, and obtains map information by comparing it with a digital map. Map information providing device. The method of claim 1,
The related word includes at least one of technical terms, industry idioms, and non-standard words, and the learning unit learns and expands the related words through machine learning. In the method of providing map information using text search,
Receiving, by a data collection unit, an image of a shopping mall in a specific area and location information from a user terminal;
Extracting text from the collected image using an optical character reader (OCR) program by a text extraction unit;
Classifying, by a database, texts other than business names and telephone numbers from the extracted texts, and storing the classified texts and related words;
Learning, by a learning unit, machine learning the classified text and related words to match the captured image;
Including, a map extraction unit extracting map information using the collected location information,
When the user terminal searches for a word, the database searches for a related word, and displays a shopping street adjacent to the user terminal and including the related word based on a current location of the user terminal. The method of claim 6,
The text extracting unit classifies the business name and phone number and the rest of the text from the photographed image of the shopping street,
The method for providing a map degree in which the database stores the business name and telephone number together with the classified text and related words. The method of claim 6,
The user terminal may input or search for information on a shopping street, access the database, and correct the image, extracted text, or related words of the shopping street. The method of claim 6,
The map information extracting unit receives an image of a shopping mall through a traveling robot or a portable lidar equipped with a lidar, and provides a photographing position and an image of the traveling robot to the data collection unit together. The method of claim 6,
The related word includes at least one of technical terms, industry idioms, and non-standard words, and the learning unit learns and expands the related words through machine learning.

Images