JP2011146005A - Apparatus, system and method for constructing subject spot database - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To facilitate the construction and maintenance of a subject spot database in which spot data including a photographic image of each subject spot is grouped by the subject spot. <P>SOLUTION: Whether posting data posted by a user is on an existing subject spot or on a new subject spot is determined according to the degree of similarity between an image in the posting data and images saved in a subject spot database. If the posting data is determinedly on an existing subject spot, it is stored in the subject spot database as spot data associated with the subject spot. Meanwhile, if the posting data is determined relating to a new subject spot, it is stored in the subject spot database under a group of spot data having similar images. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an imaging target spot database construction apparatus, construction system, and construction method for storing spot data including an image obtained by photographing an imaging target spot in association with the corresponding imaging target spot.

  For example, as disclosed in Patent Document 1, an image editing apparatus that organizes image data according to contents in accordance with shooting locations is known. The image editing apparatus includes an HDD that stores a landmark database and content data. When the image data to which the position data is added is acquired from the memory card connected to the memory card reader, the CPU searches the landmark data based on the position data, and classifies the image data for each landmark from the search result. . Then, the CPU reads the content data related to the landmark, and creates a photo album or the like by combining the grouped image data and the read content data.

JP 2007-194948 A

  In the above-described image editing apparatus, the digital camera that captures an image includes a GPS device that measures the current position and generates position data, and the acquired position data is added to the acquired image data. It is assumed that it is.

  However, some digital cameras are not equipped with a GPS device, and no position data is added to image data obtained by photographing a digital camera without such a GPS device. Cannot be used. For this reason, it is desired to provide a system and service that can classify and organize image data without position data for each shooting target spot such as a landmark.

  Therefore, for example, a database (shooting target spot database) in which images obtained by shooting each shooting target spot are grouped for each shooting target spot is prepared, and input from the similarity between the input image and the image in the database is input. It is conceivable to construct a system for determining which shooting target spot is taken by the captured image. If such a system can be configured, for example, when an image of a certain landmark is transmitted to the system, a service providing detailed information on the landmark is provided, or the captured image is linked to each shooting target spot. It is also possible to provide services for organizing (album).

  However, the number of shooting spots such as landmarks is enormous, and it is very troublesome for the service provider to prepare a shooting target spot database that covers all areas. Moreover, even if a shooting target spot database can be constructed once, new shooting target spots appear every day, so that maintenance is also required.

  The present invention has been made in view of the above points, and facilitates the construction and maintenance of a shooting target spot database in which spot data including images obtained by shooting each shooting target spot is grouped for each shooting target spot. An object of the present invention is to provide a photographing object spot database construction apparatus, construction system, and construction method.

In order to achieve the above-described object, an apparatus for constructing a shooting target spot database according to claim 1 comprises:
An acquisition means for acquiring post data including an image obtained by shooting a shooting target spot from a terminal device operated by a user via a communication network;
Based on the similarity between the image in the posted data acquired by the acquisition means and the image stored in the shooting target spot database, whether the posted data relates to an existing shooting target spot or other than the existing shooting target spot A discriminating means for discriminating whether the spot is a shooting target spot;
Existing spot data storage means for storing post data determined to be related to an existing photographing target spot in the photographing means as spot data associated with the photographing target spot in the photographing target spot database;
New candidate spot data storage means for storing post data determined to be related to a shooting target spot other than the existing shooting target spot in the shooting target spot database as spot data related to a new candidate of the shooting target spot in the determination means; Prepared,
The discriminating means is stored if spot data relating to a new candidate for the shooting target spot is already stored in the shooting target spot database when the posted data is determined to be related to the shooting target spot other than the existing shooting target spot. The degree of similarity with the image of the spot data related to the new candidate spot to be photographed,
The new candidate spot data storage unit is configured to group spot data related to a new candidate for the shooting target spot based on the determination result of the determination unit for each spot data with similar images and store the spot data in the shooting target spot database. And

  In other words, according to the apparatus for constructing the shooting target spot database according to claim 1, the shooting target spot database can be enhanced by using post data including an image of the shooting target spot from the user. Therefore, it is possible to reduce the time and labor of construction and maintenance of the shooting target spot database.

  More specifically, when determining whether post data is related to an existing shooting target spot or a shooting target spot other than the existing shooting target spot based on the similarity of the images, the shooting target spot database is searched from various viewpoints. If an image obtained by photographing the photographing target spot is stored, the determination can be performed with higher accuracy. In the invention of claim 1, post data determined to be related to an existing shooting target spot is stored in the database as spot data associated with the shooting target spot. Therefore, it is possible to enrich the spot data of existing shooting target spots using the posted data. For this reason, in the initial shooting target spot database, for example, it is only necessary to prepare a representative image of each shooting target spot data, so that it is possible to reduce the effort for constructing the initial shooting target spot database.

  Post data determined to be related to a shooting target spot other than the existing shooting target spot is stored in the database as spot data related to a new candidate for shooting target spot. At that time, when spot data related to a new candidate for the shooting target spot is already stored in the database, the similarity to the stored image of the spot data related to the new candidate for the shooting target spot is determined, and based on the determination result The spot data relating to the new candidate spot for photographing is grouped for each spot data with similar images and stored in the database. Therefore, when adding a new shooting target spot in the shooting target spot database, the grouped spot data can be used, so that the new shooting target spot can be added relatively easily. For this reason, it is possible to facilitate the maintenance of the shooting target spot database. Further, it is not always necessary to cover all shooting target spots as the initial shooting target spot database. For example, it is possible to start from only spot data relating to main shooting target spots.

  As described in claim 2, the spot data includes document data related to the shooting target spot in addition to the image, and when the determination unit acquires the image and the comment as post data, the similarity between the images In addition, considering the similarity between the document data and the comment, it may be determined whether the post data relates to an existing shooting target spot or a shooting target spot other than the existing shooting target spot. good. In addition to the image similarity, the similarity between the document data and the comment is also taken into consideration, so that the discrimination accuracy can be improved.

  According to the third aspect of the present invention, the most frequently used text data is extracted from the document data of the spot data related to the new candidate spot to be photographed, which is stored in groups by the new candidate spot data storage means, and photographed. In addition to the name of the target spot, spot addition means for adding a new shooting target spot may be provided in the shooting target spot database. As a result, it is possible to automatically add a new shooting target spot by assigning the name of the shooting target spot from the spot data relating to the grouped new candidates.

As described in claim 4, the spot data includes, in addition to the image, position data indicating the position of the shooting target spot,
In the case where the determination means obtains the image and the position data indicating the position where the image was taken as the post data, the post data considers the similarity between the position data in addition to the similarity between the images. However, it may be determined whether the spot is related to an existing shooting target spot or a shooting target spot other than the existing shooting target spot. Considering the similarity of position data in addition to the similarity of images, the discrimination system can be further improved. For example, when there are a plurality of shooting target spots at close positions, the post data may be stored in association with an erroneous shooting target spot that should not be classified, based on the position data alone. Even in such a case, the post data can be stored in association with the correct photographing target spot by considering both the position data and the similarity between the images.

  And as described in Claim 5, the imaging | photography object spot database construction system is comprised from the imaging | photography object spot database construction apparatus in any one of Claims 1 thru | or 4, and the said terminal device operated by the user. can do.

  In addition, since the effect by the construction method of the imaging | photography object spot database described in Claims 6-9 is the same as the effect by the imaging | photography object spot database construction apparatus of Claims 1-4 mentioned above, description is abbreviate | omitted. .

It is a block diagram which shows the whole structure of the construction system of the imaging | photography object spot database by embodiment. It is a block block diagram which shows schematically the structure of the terminal device as a client. It is a figure which shows the example of the input screen for creating contribution data in a terminal device. It is a block block diagram which shows the structure of a server roughly. It is a flowchart which shows the process performed when a server receives posting data.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the present embodiment, the server mainly receives post data including an image (photograph) from a terminal device as a client, and uses the received post data to capture an image ( A system configuration and method for constructing (enriching) a shooting target spot database in which photographs are grouped for each shooting target spot will be described.

  However, when the server receives post data including an image (photo) from the client, the server determines which shooting target spot is displayed in the image, and information related to the corresponding shooting target spot. To the client, or to provide a service that organizes the posted data by shooting target spot into an album. For this reason, post data is continuously transmitted from many clients to the server.

  FIG. 1 shows the overall configuration of a system for constructing a shooting target spot database according to an embodiment. As shown in FIG. 1, the system for constructing a shooting target spot database includes a terminal device 10 as a client and a server 20 having a shooting target spot database. The terminal device 10 and the server 20 are connected to be communicable with each other via a communication network 30 such as the Internet.

Specifically, a personal computer or a mobile phone is used as the terminal device 10 that is a client. A configuration required as the terminal device 10 in the photographing target spot database construction system will be described with reference to FIG. Hereinafter, an example in which a personal computer is used as the terminal device 10 will be described. As illustrated in FIG. 2, the terminal device 10 includes a display device 11, an input device 12, a screen display unit 13, a photo acquisition unit 14, a comment. An acquisition unit 15, a shooting position acquisition unit 16, a post data creation unit 17, and a communication unit 18 are provided.

  The display device 11 is composed of a liquid crystal display, for example, and the display content is controlled by the screen display unit 13. For example, the screen display unit 13 displays information acquired from the server 20 on the display device 11 or displays a post data input screen. FIG. 3 shows an example of a post data input screen.

  In the present embodiment, the posting data includes a photograph as an image, a comment regarding the photograph, and position data indicating a position where the photograph is taken. However, although the photo is indispensable information, the comment and the position data are arbitrary information. If at least the photo is included, the server 20 accepts the post data.

  The input device 12 includes not only input operation devices such as a keyboard and a mouse, but also a memory reader for taking a photograph taken by a digital camera and stored in a memory. When the post data input screen is displayed on the display device 11, a post is taken by taking a photo by a memory reader or by selecting a desired photo if the photo is already stored in the personal computer. The photo to be used as data is determined. The determined photograph is displayed in the photograph display area on the post data input screen shown in FIG.

  Comments about photos can be entered in the comment entry field. For this input, a keyboard is mainly used, but other input devices such as a voice input device may be used. The shooting position is input using the shooting position input field. Specifically, a map is displayed in the shooting position input field, and on the map, the shooting position is specified by a pinpoint or an area having a certain size using a mouse or the like. When the shooting position is specified by pinpoint, the position data is expressed by latitude and longitude, and when specified by an area having a certain size, the position data depends on, for example, the center position of the area and the shape of the area. Is represented by the inner diameter.

  However, the input of the shooting position is not limited to this example, and an address of the shooting position and a place name including the shooting position may be input. Further, when the digital camera includes a GPS device and position data obtained by the GPS device is added to a photograph taken, the position data may be extracted as a photographing position.

  The photo acquisition unit 14 acquires a photo determined to be post data on the post data input screen. The comment acquisition part 15 acquires the input comment, when a comment is input into the comment input column of a contribution data input screen. Further, when information indicating the shooting position is input to the shooting position input field of the posting data input screen, the shooting position acquisition unit acquires the input shooting position or adds position data to the image. In some cases, the position data is extracted.

  The posting data creation unit 17 collects the photos, comments, and shooting positions acquired by the photo acquisition unit 14, the comment acquisition unit 15, and the shooting position acquisition unit 16, and generates posting data including those data. The created post data is transmitted to the server 20 via the communication network 30 by the communication unit 18.

  Next, the configuration of the server 20 will be described with reference to FIG. As illustrated in FIG. 4, the server 20 includes a storage device 21 that stores a shooting target spot database, a photograph feature value calculation unit 22, a comment feature value calculation unit 23, an existing spot acquisition unit 24, a similarity calculation unit 25, a spot classification. Unit 26 and communication unit 27.

  The communication unit 27 receives post data transmitted from the terminal device 10 of each client. When the posted data includes a shooting position, the existing spot acquisition unit 24 acquires shooting target spot data (photo, document, shooting position) from the shooting target spot database regarding the shooting target spot existing in the vicinity of the shooting position. . For example, when the shooting position is represented by latitude and longitude, shooting target spot data is extracted for a shooting target spot that exists within a radius N (m) centered on the latitude and longitude. When the shooting position is specified in an area having a certain size, a range including at least the area is set, and shooting target spot data for the shooting target spots included in the range is extracted. Note that if the posting data does not include the shooting position, all shooting target spot data (photographs, documents) are acquired from the shooting target spot database.

  Here, in the shooting target spot database, for each shooting target spot such as a landmark, a photograph of the shooting target spot, a document including descriptions and information on the shooting target spot, and the position of the shooting target spot are associated with each other. Stored (grouped).

  For example, the shooting target spot database includes a spot table and a spot detailed attribute table. The spot table stores a spot ID, which is a unique number for each spot, a spot name, and position data (latitude / longitude) indicating a position where the spot exists. The spot detailed attribute table stores a spot ID, an attribute ID, a photograph, and a document. As described above, the spot table and the spot detailed attribute table are associated with each other using the spot ID as a key.

  The photographic feature amount calculation unit 22 calculates the image feature amount of the photograph of the posted data, and calculates the image feature amount of the photograph of the shooting target spot data acquired by the existing spot acquisition unit 24.

  Various methods are known as a method for calculating the image feature amount of a photograph. In the present embodiment, a method of representing an image feature amount by a feature amount of an image contour line using a Gabor filter is adopted. Hereinafter, a specific image feature amount calculation method according to the present embodiment will be described.

First, in the present embodiment, the feature quantity I ^x of a certain image x is represented by the number of contour lines having a predetermined angle in the image x as shown in Equation 1.

That is, regarding the contour line in the image x, the number of contour lines having the angle is totaled for each angle to obtain the image feature amount I ^x .

The contour line angle I ^x _θi (x, y) of each pixel (x, y) in the image x can be obtained by the following Equation 2.

  The Gabor filter is defined by Equation 3 below.

Using the above-described mathematical formulas 2 and 3, the contour angle I ^x _θi (x, y) of each pixel (x, y) is obtained, and predetermined contour angles I ^x _θ1 , I ^x _θ2,. by matching the ^x _.theta.k, it is possible to calculate the image feature quantity I ^x of the image x. However, as described above, various methods are known for calculating image feature amounts, and image feature amounts (for example, SIFT feature amounts and SURF feature amounts) may be calculated according to other calculation methods.

  The comment feature quantity calculation unit 23 is calculated based on the frequency with which a predetermined keyword exists in the comment (document). That is, the comment feature quantity calculation unit 23 first decomposes a comment (document) into words by morphological analysis. Then, a comment (document) feature amount is calculated based on how many predetermined keywords are included in the decomposed word.

For example, it is assumed that the predetermined keywords are w ₁ , w ₂ ,..., W _L (L: number of keywords). If a certain comment (document) y has the number of each keyword as w ^y ₁ , w ^y ₂ ,..., W ^y _L , the comment (document) feature quantity D ^y of the comment (document) ^y is It is expressed by Equation 4.

That is, the feature amount D ^y of the comment (document) y is represented by a set of ratios of the number of appearances of each keyword to the number of appearances of all keywords set in advance.

The similarity calculation unit 25 first calculates the image feature amount I ^x of the photograph calculated by the photograph feature amount calculation unit 22 and the comment (document) feature amount D ^y calculated by the comment feature amount calculation unit 23. The data feature amount F is calculated by linear combination according to the following formula 5. Therefore, the data feature amount F is defined as a vector composed of a plurality of terms. The data feature amount F is calculated for post data received by the communication unit and shooting target spot data acquired by the existing spot acquisition unit 24.

In the above formula 5, α and β are weighting factors that are arbitrarily set in a range larger than 0 and smaller than 1. Both of these weighting factors α and β are normally set to 0.5, but it is desired to judge the degree of similarity by placing more emphasis on either the image feature quantity I ^x or the comment (document) feature quantity D ^y In addition, the weighting factor of the item to be emphasized is set to a larger value than the other weighting factor.

If the post data does not include a comment, the comment (document) feature amount is ignored. Specifically, the data feature amount F is calculated from only the image feature amount I ^{x for} all data. The data feature amount F is calculated by setting all the comment (document) feature amounts ^Dy to NULL, or the comment (document) feature amount ^Dy is replaced with the average value of all existing document data, and the data feature amount F is Or calculate.

  In addition, each extracted spot data to be photographed may include a plurality of photographs. In this case, the data feature amount F is calculated using a plurality of photos and a document (common or different) associated with each photo.

  In particular, in this embodiment, as described later, when it is determined that the posted data relates to a shooting target spot that already exists in the shooting target spot database, the posted data is incorporated into the spot data of the corresponding shooting target spot. It is. As described above, the post data once incorporated in the shooting target spot data is then handled as a part of the shooting target spot data. Therefore, each time the post data is incorporated into the shooting target spot data, a photo and a comment about the photo (if added) are added, and the photo (if added) is added. The number of comments will increase.

  When the above-described data feature amount F is calculated for the posted data and the shooting target spot data, the similarity calculation unit 25 calculates the similarity based on the inner product of the data feature amount F of the posted data and the data feature amount F of the shooting target spot data. Ask for. That is, the cosine of the two data feature amounts F defined as vectors is between 0 and 1, and the similarity between the two data feature amounts F increases as the similarity calculated by the inner product is closer to 1. Become.

  In the present embodiment, as described above, when the posting position includes the shooting position, the shooting target spot data is acquired only for the shooting target spot existing in the vicinity of the shooting position. In other words, the shooting target spot data that is determined to be similar according to the shooting position is narrowed down, and the similarity between the posted data and the shooting target spot data is determined in consideration of the shooting position included in the posted data. Is equivalent to

However, when the shooting position is represented by latitude and longitude, the data feature amount F may be calculated using the shooting position. This is because the shooting target spot closer to the shooting position represented by the latitude and longitude is more likely to correspond to the posted data. In this case, when the data feature amount F is calculated, the image feature amount I ^x and the comment (document) feature amount D ^y may be linearly combined while multiplying the shooting position by the weighting coefficient γ.

  Based on the similarity calculated by the similarity calculation unit 25, the spot classification unit 26 classifies whether the posted data is related to an existing shooting target spot or a new candidate for shooting target spot.

  Specifically, among the similarities calculated by the similarity calculating unit 25, when the highest similarity is a predetermined threshold (for example, 0.8) or more, the posted data has the highest similarity. It is determined that the spot is related to an existing shooting target spot. In this case, the post data is incorporated into the spot data associated with the shooting target spot and becomes a part thereof.

  Here, when determining whether the posted data relates to an existing shooting target spot or a new candidate for a shooting target spot other than the existing one based on the similarity of images or the like, various shooting is performed in the shooting target spot database. If an image obtained by photographing the spot to be photographed from the viewpoint is stored, the discrimination can be performed with higher accuracy. In this embodiment, post data determined to be related to an existing shooting target spot is stored in the shooting target spot database as part of spot data associated with the shooting target spot. Therefore, it is possible to enrich the shooting target spot data with respect to the existing shooting target spot using the posted data. Therefore, in the initial shooting target spot database, for example, even if a representative image of each shooting target spot data is prepared, it can be used as the shooting target spot database. Can be reduced.

On the other hand, when the highest similarity calculated by the similarity calculation unit 25 is less than the threshold,
The posted data is classified as related to a new candidate spot for photographing. In this case, the post data is stored in the shooting target spot database as spot data relating to a new candidate for shooting target spot.

  At this time, if spot data related to the new candidate for the shooting target spot is already stored in the shooting target spot database, the similarity calculation unit 25 uses the similarity to the stored spot data related to the new candidate for the shooting target spot. The degree is calculated. Based on the calculated similarity, the spot classification unit 26 groups the spot data related to the new candidates for the shooting target spot for each similar spot data, and stores them in the shooting target spot database. Therefore, when adding a new shooting target spot in the shooting target spot database, the grouped spot data can be used, so that the new shooting target spot can be added relatively easily. For this reason, it is possible to facilitate the maintenance of the shooting target spot database. Further, it is not always necessary to cover all shooting target spots as the initial shooting target spot database. For example, it is possible to start from only spot data relating to main shooting target spots.

  As described above, according to the system for constructing the shooting target spot database according to the present embodiment, it is possible to enhance the shooting target spot database by using post data including an image obtained by shooting the shooting target spot from the user. Therefore, it is possible to reduce the time and labor of construction and maintenance of the shooting target spot database.

  Next, processing executed when the server 20 receives post data will be described with reference to the flowchart of FIG.

  First, in step S100, it is determined whether post data has been received. At this time, it is determined that the post data is received if the post data includes at least an image. In subsequent step S110, shooting target spot data associated with the existing shooting target spot is extracted from the shooting target spot data, and the similarity to the posted data is calculated. At this time, if the posting data includes the shooting position, the shooting target spot data to be subjected to similarity determination is narrowed down in advance based on the shooting position.

  In step S120, it is determined whether or not it is similar to the existing shooting target spot data. At this time, the similarity is calculated as described above, and it is determined whether the similarity is similar or dissimilar depending on whether the highest similarity among the calculated similarities is equal to or greater than a predetermined threshold. If it is determined in step S120 that they are similar, the process proceeds to step S130, and the posted data is stored in the shooting target spot database as part of the spot data of the existing shooting target spots determined to be similar.

  On the other hand, if it is determined in step S120 that they are dissimilar, in step S140, it is determined whether spot data relating to a new candidate for the shooting target spot is already stored in the shooting target spot database. At this time, if it is determined that spot data related to the new candidate for the shooting target spot is not stored, the process proceeds to step S150, and the posted data is stored in the shooting target spot database as spot data of the new candidate for the shooting target spot. To end the process. On the other hand, if it is determined in step S140 that spot data related to the new candidate for the spot to be photographed has already been stored, the process proceeds to step S160, and the degree of similarity with the spot data of the new candidate is calculated. The method for calculating the similarity is the same as the method for calculating the similarity with the existing shooting target spot data.

  In step S170, it is determined whether or not the posted data is similar to spot data of a new candidate spot for photographing based on whether or not the calculated similarity is equal to or greater than a predetermined threshold. If it is determined in the determination process in step S170 that the images are similar, the process proceeds to step S180, grouped with similar spot data, and stored in the shooting target spot database. On the other hand, when it is determined that they are dissimilar, the process proceeds to step S190, and the posted data is stored in the shooting target spot database as spot data independent of the spot data relating to the new candidate shooting target spot already stored.

  When multiple spot data related to new candidates for the shooting target spot are grouped and stored in the shooting target spot database, the most frequently used text data is extracted from the document of the spot data (comments in the posted data). Using this as the name of the shooting target spot, a new shooting target spot can be added. Thereby, it is possible to automatically give a new shooting target spot name from the spot data related to the grouped new candidates and to automate the addition of shooting target spots.

  However, the server administrator (service provider) may add a new shooting target spot. Even in this case, spot data of new candidates for the shooting target spot are grouped, and these data can be used as spot data. Therefore, it is possible to easily add new shooting target spots.

  When the server administrator adds a new shooting target spot, check whether the spot data classified as the new candidate spot data relates to an existing shooting target spot or other shooting target spot Can do. In addition, when adding a new shooting target spot based on new candidate spot data, the server administrator may supplement the position data if the post position has not been posted even once. it can.

  In addition, it is possible to have the user decide whether or not to add a new shooting target spot and how to use the spot name and position when adding a new spot. For example, a community service is provided to the user as a server function. In the community service, the spot data of a new candidate spot to be photographed is presented to the user, and a place for discussing the spot data is provided. This allows the user to change the name and position of the new candidate spot data when it relates to an existing shooting target spot, other shooting target spot, or a new shooting target spot. You can discuss what to do. Through this discussion, it is finally classified as an existing shooting target spot, or its name and position are determined as relating to a new shooting target spot.

  When the name and position of a new shooting target spot are determined, the server 20 adds a new shooting target spot to the shooting target spot database using the name, position, and spot data.

  The preferred embodiments of the present invention have been described above. However, the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the spirit of the present invention.

  For example, in the above-described embodiment, an example in which a personal computer is used as the terminal device 10 as a client has been mainly described. However, when a mobile phone with a digital camera is used as the terminal device 10, the input device 12 for reading a photograph is not necessary as the terminal device 10. If the mobile phone is equipped with a GPS device, the shooting position can be automatically added to the posted data without the user inputting the shooting position.

  In the above-described example, the screen display unit 13 of the terminal device 10 displays the posted data input screen, and the terminal device 10 acquires the photograph determined as the posted data and the input comment. , 15 has been described. However, for example, the server 20 provides a site for accepting posted data on the web, and the user accesses the site and creates posted data including a photo, a comment, and a shooting position on the site. It is also possible.

DESCRIPTION OF SYMBOLS 10 ... Terminal device, 11 ... Display device, 12 ... Input device, 13 ... Screen display part, 14 ... Photo acquisition part, 15 ... Comment acquisition part, 16 ... Shooting position acquisition part, 17 ... Posting data creation part, 18 ... Communication , 20 ... server, 21 ... shooting target spot database, 22 ... photographic feature value calculation unit, 23 ... comment feature value calculation unit, 24 ... existing spot acquisition unit, 25 ... similarity calculation unit, 26 ... spot classification unit, 27 ... communication department

Claims (9)

An apparatus for constructing a shooting target spot database that stores spot data including an image of a shooting target spot in association with the corresponding shooting target spot,
An acquisition means for acquiring post data including an image obtained by shooting a shooting target spot from a terminal device operated by a user via a communication network;
Based on the similarity between the image in the posted data acquired by the acquisition means and the image stored in the shooting target spot database, the posted data relates to an existing shooting target spot or an existing shooting target A discriminating means for discriminating whether the spot is a spot to be photographed other than a spot;
Existing spot data storage means for storing post data determined to be related to an existing shooting target spot in the determination means as spot data associated with the shooting target spot in the shooting target spot database;
New candidate spot data storage means for storing post data determined to be related to a shooting target spot other than an existing shooting target spot in the determination means as spot data related to a new candidate for the shooting target spot in the shooting target spot database; With
When the determination unit determines that the posted data is related to a shooting target spot other than the existing shooting target spot, if spot data related to a new candidate for the shooting target spot is already stored in the shooting target spot database, The similarity with the image of the spot data related to the new candidate spot for shooting is also determined,
The new candidate spot data storage unit groups spot data related to a new candidate for a shooting target spot based on the determination result of the determination unit for each spot data having similar images, and stores the spot data in the shooting target spot database. An apparatus for constructing a shooting target spot database. The spot data includes document data related to the shooting target spot in addition to the image,
In the case where an image and a comment are acquired as the post data, the determination means considers the similarity between the document data and the comment in addition to the similarity between the images, and the post data is an existing one. 2. The apparatus for constructing a shooting target spot database according to claim 1, wherein it is determined whether the spot is related to a shooting target spot or a shooting target spot other than an existing shooting target spot.   By extracting the most frequently occurring text data from the document data of spot data relating to new candidates for the shooting target spot stored in groups by the new candidate spot data storage means, the name of the shooting target spot is obtained. 3. The apparatus for constructing a shooting target spot database according to claim 2, further comprising spot adding means for adding a new shooting target spot in the shooting target spot database. The spot data includes, in addition to the image, position data indicating the position of the shooting target spot,
In the case where the determination means acquires the image and the position data indicating the position where the image was taken as the post data, in addition to the similarity between the images, the similarity between the position data is also considered, The apparatus for constructing a shooting target spot database according to claim 1 or 2, wherein the posted data is related to an existing shooting target spot or a shooting target spot other than the existing shooting target spot. .   An imaging target spot database construction system comprising the imaging target spot database construction device according to claim 1 and the terminal device operated by a user. A method for constructing a shooting target spot database that stores spot data including an image of a shooting target spot in association with the corresponding shooting target spot,
An acquisition step of acquiring post data including an image of a shooting target spot from a terminal device operated by a user via a communication network;
Based on the similarity between the acquired image in the posted data and the image stored in the shooting target spot database, whether the posted data relates to an existing shooting target spot or shooting other than the existing shooting target spot A determination step for determining whether the spot is related to the target spot;
An existing spot data storing step of storing post data determined to be related to an existing shooting target spot in the determination step as spot data associated with the shooting target spot in the shooting target spot database;
A new candidate spot data storing step of storing post data determined to be related to a shooting target spot other than the existing shooting target spot in the determination step as spot data related to a new candidate of the shooting target spot in the shooting target spot database; With
In the determination step, when it is determined that the posted data is related to a shooting target spot other than the existing shooting target spot, if spot data regarding a new candidate for the shooting target spot is already stored in the shooting target spot database, The similarity with the image of the spot data related to the new candidate spot for shooting is also determined,
In the new candidate spot data storage step, the spot data relating to the new candidate of the shooting target spot is grouped for each similar spot data based on the determination result in the determining means, and stored in the shooting target spot database. A method for constructing a featured target spot database. The spot data includes document data related to the shooting target spot in addition to the image,
In the acquisition step, when an image and a comment are acquired as the posted data, in the determination step, the similarity between the document data and the comment is considered in addition to the similarity between the images. 7. The method for constructing a shooting target spot database according to claim 6, wherein it is determined whether the posted data relates to an existing shooting target spot or a shooting target spot other than the existing shooting target spot.   In the new candidate spot data storage step, the most frequently used text data is extracted from the document data of spot data related to the new candidate for the shooting target spot stored in groups, and the name of the shooting target spot is obtained. 8. The method of constructing a shooting target spot database according to claim 7, further comprising a spot expansion step of adding a new shooting target spot in the shooting target spot database. The spot data includes, in addition to the image, position data indicating the position of the shooting target spot,
In the acquisition step, when the post data includes an image and position data indicating the position where the image was captured, in the determination step, in addition to the similarity between the images, the similarity between the position data The photographing according to claim 6 or 7, wherein it is determined whether the posted data relates to an existing photographing target spot or a photographing target spot other than the existing photographing target spot in consideration of the degree. How to build a target spot database.

Images