JP5553483B2 - Optimal oblique photograph providing method, optimum oblique photograph providing system, and optimum oblique photograph providing apparatus - Google Patents

Description

  The present invention is a method of using a photograph in the optimum photograph providing method, in which a user requests a target feature from a plurality of photographs taken at various angles and distances with respect to the target feature. The present invention relates to an optimum oblique photograph providing method for providing an optimum photograph.

  In the geographic information retrieval system on the Web represented by the existing GIS and Google's Google Map, the satellite and aerial photographs taken almost perpendicular to the ground surface are orthorectified and displayed superimposed on the map. Yes. In addition, there are systems such as Google Earth of Google and Virtual Earth of Microsoft that display images by associating them with a map based on the coordinates of the shooting range even if they are taken from an oblique direction other than vertical.

In addition, there is a system for generating feature information by generating a bird's-eye view image in a virtual space using a map and a building shape as in Japanese Patent No. 3212113.
Patent 3212113

  However, in a dense area of high-rise buildings in urban areas, photographs taken perpendicular to the ground surface only visualize the zenith such as the rooftop and roof of the building, and the side surfaces such as the walls are not photographed .

  For this reason, when providing video information for the purpose of improving user access to features and space recognition, the sight from the viewpoint of pedestrians and vehicles as information users and the conventional method provided There is a problem that a difference is likely to occur in the scene of the vertically photographed photograph, and it is difficult to associate the feature in the real space with the feature in the photograph as compared with the oblique photograph.

  In addition, even in diagonally photographed photographs, the target feature occupies the majority of the angle of view, the searcher has a local intuition for the area, and the correspondence with a general two-dimensional map from the section shape etc. There is a problem that it is difficult to find a search target feature from an obliquely photographed photograph in which a plurality of features are photographed, except in a case where it is easy.

  Even if the photographed area is well known, if the target feature is the first time, the target feature is found from a plurality of features in the photograph using the address.

  In addition, even if there is only one target feature, if there are a plurality of oblique photographs, the photograph is taken on the largest surface of, for example, a river, sea, road, etc. Whether it is a photograph or not can not be judged immediately.

  On the other hand, due to the characteristic of oblique shooting, even if the camera is oriented in the direction of the target feature, if there is another feature in front, the target feature may be hidden or hidden behind it In all circumstances, oblique photography is not superior to vertical photography as an auxiliary means for space recognition around the target feature.

  The present invention has been made in order to solve the above-described problems. In accordance with the purpose of various search applications of geographic information search, a desired shooting direction and a desired shooting direction can be selected from photographs taken by various means. An object of the present invention is to provide an optimum oblique photograph providing method that can easily find a search target feature even when there are a plurality of features.

  The optimum oblique photograph providing method of the present invention is a method in which a target target feature is a designated feature among a plurality of oblique photographs obtained by photographing a feature at a photographing point (Pi) from a diagonal direction with different flight paths. This is an optimum oblique photograph providing method for providing an oblique photograph for each photographing point (Pi) photographed in contact with a photograph as an optimum oblique photograph.

First storage means for storing a planar map in which a planar shape of a feature including the target feature is defined, and attribute information for the feature in the planar map;
Wherein for each shot point (Pi), shooting position and the shooting point (Pi) each of the planar imaging range and their photographic point of the camera in the map camera focal length and the orientation and the camera of the CCD of the (Pi) of Second storage means for storing shooting information including an angle of view;
A third storage means for storing an oblique photograph of the camera for each shooting point (Pi) in association with the shooting information of the shooting point (Pi);
Computer
Reading the attribute information of the target feature and extracting all of the shooting information of the shooting point (Pi) having the shooting range including the target feature of the attribute information from the second storage means;
Defining a predetermined range in which the target feature is shown in common based on each of the shooting ranges included in the extracted shooting information as a common map in a fourth storage unit;
Defining the shooting position and the angle of view included in each of the extracted shooting information in the common map;
Obtaining, as a judgment boundary line (EFGH), a line where the designated feature and the target feature existing on the common map are in contact with each photographing position of the photographing point (Pi);
For each shooting position of the shooting point (Pi), a straight line is drawn from the shooting point (Pi) to the target feature, and it never passes through the inside of the target feature. Obtaining a set of points intersecting the object as collimable lines (EF) on the determination boundary line (EFGH) facing the photographing position of the photographing point (Pi);
For each shooting position of the shooting point (Pi), the width of the angle of view of the shooting point (Pi) is a projection plane on which the target feature is projected on the CCD, and the collimable line (EF) is Projecting onto the projection plane and determining the projected line as a collimable line projection plane width (wcl);
For each imaging position of the imaging point (Pi) , the value of the ratio of the collimable line projection plane width (wcl) to the projection plane width (WLl) is set as a collimable line projection ratio Wi (Wi = Wcl / ( WLl)) as a step
For each shooting position of the shooting point (Pi), the value of the ratio of the length of the collimable line (EF) of the shooting position of the shooting point (Pi) to the length of the determination boundary line (EFGH) is determined as a determination boundary. Obtaining the collimable line ratio (Li);
Comparing each of the determination boundary collimable line ratios ( Li ) for each photographing position of the photographing point (Pi ) to obtain a determination boundary collimable line ratio ( Li ′ ) having the largest value;
Comparing the collimable line projection ratios (Wi) to obtain the largest collimable line projection ratio (Wi ′);
The imaging information including the imaging position of the imaging point (Pi) where the collimable line projection ratio (Wi ′) and the determination boundary collimable line ratio (Li ′) having the largest values are obtained . Read from the storage means of the second, the oblique photograph of the third storage means associated with the shooting position of the shooting point (Pi) included in this shooting information from the shooting direction of the target feature of interest And a step of outputting as an optimum oblique photograph.

  In the optimum oblique photograph providing system of the present invention, a user terminal and a server of the optimum photograph providing service center are connected by a communication network, and the server changes a flight path and adds a feature for each photographing point (Pi). Among the plurality of oblique photographs taken from the oblique direction with the camera, the oblique photograph at each photographing point (Pi) where the target target feature is photographed in contact with the designated feature is set as the optimum photograph to the user terminal. It is an optimal oblique photograph providing system to transmit.

The server
First storage means for storing a planar map in which a planar shape of a feature including the target feature is defined, and attribute information for the feature in the planar map;
Wherein for each shot point (Pi), shooting position and the shooting point (Pi) each of the planar imaging range and their photographic point of the camera in the map camera focal length and the orientation and the camera of the CCD of the (Pi) of Second storage means for storing shooting information including an angle of view;
Third storage means for storing an oblique photograph of the camera for each photographing point (Pi) in association with the photographing information of the photographing point (Pi);
Means for storing user information from the user terminal and the search condition for the oblique photograph in a user information search condition storage means;
Means for transmitting, via the communication network, input screen information for allowing the user terminal to input search conditions for the user information and the oblique photograph;
When the attribute information of the target feature is input from the user terminal, the attribute information of the target feature is read, and the shooting point (Pi having the shooting range including the target feature of the attribute information) Means for extracting all of the shooting information from the second storage means;
Means for defining, in a fourth storage means, a predetermined range in which the target feature is shown in common based on each of the shooting ranges included in the extracted shooting information;
Means for defining the shooting position of the shooting point (Pi) included in each of the extracted shooting information and the width of the angle of view in the common map;
Means for obtaining, as a determination boundary line (EFGH), a line where the designated feature and the target feature existing on the common map are in contact with each photographing position of the photographing point (Pi);
For each shooting position of the shooting point (Pi) , a straight line is drawn with respect to the target feature from the shooting point (Pi) of the shooting point (Pi), and never passes through the inside of the target feature. Means for first obtaining a set of points intersecting the target feature as collimable lines (EF) on the determination boundary line (EFGH) facing the photographing position of the photographing point (Pi);
For each shooting position of the shooting point (Pi) , the width of the angle of view at the shooting position of the shooting point (Pi) is a projection plane on which the target feature is projected on the CCD, and the collimable line ( EF) is projected onto the projection plane, and the projected line is obtained as a collimable line projection plane width (wcl);
For each imaging position of the imaging point (Pi) , the value of the ratio of the collimable line projection plane width (wcl) to the projection plane width (WLl) is set as a collimable line projection ratio Wi (Wi = Wcl / ( WLl)) means to obtain
For each photographing position of the photographing point (Pi), a value of a ratio of the length of the collimable line (EF) of the photographing point (Pi) to the length of the judgment boundary line (EFGH) is determined as a judgment boundary collimable line. Means for obtaining the ratio (Li);
Means for comparing each of the determination boundary collimable line ratios ( Li ) for each photographing position of the photographing point (Pi ) to obtain a determination boundary collimable line ratio ( Li ′ ) having the largest value ;
Means for comparing the collimable line projection ratios (Wi) to obtain the largest collimable line projection ratio (Wi ′);
The imaging information having the imaging position of the imaging point (Pi) where the collimable line projection ratio (Wi ′) and the determination boundary collimable line ratio (L1 ′) having the largest values are obtained . Read from the storage means of the second, the oblique photograph of the third storage means associated with the shooting position of the shooting point (Pi) included in this shooting information from the shooting direction of the target feature of interest Means to make an optimal oblique photo;
Means for transmitting the optimum oblique photograph to the user terminal via the communication network,
The user terminal is
Means for receiving the input screen information from the server, and transmitting the attribute information of the target target feature and the designated feature input to the input screen information to the server as the search condition;
The gist is to have means for receiving an optimum oblique photograph from the server and displaying it on the screen.

  As described above, according to the present invention, an oblique photograph in the photographing direction in which the surface of the target feature in contact with the specified feature (for example, road, river, park) is photographed most greatly is searched from a plurality of oblique photographs. This is provided as an optimal photograph (screen or user terminal).

  For this reason, the image which image | photographed the surface side of the target feature in contact with the designated feature (road) designated by the user can be provided.

  For example, in an information search for the purpose of displaying and accessing the appearance of a specific feature on a map such as a restaurant or a real estate, when there are a plurality of images taken from various directions of the target feature, generally a video The user can obtain a photograph close to the sight that the user sees in the field.

  In addition, since the photo search is performed by the shooting depression angle, it is possible to provide an oblique photo in which a surface that can be observed by the user is widely taken from the three-dimensional shape of the target feature. For this reason, it is possible to provide a photograph of the side wall of the building excluding a photograph showing only the rooftop or a photograph of only the feet.

  Further, even if there are a large number of features around the target feature, the occlusion rate is used, so that it is possible to provide a photograph in which the target feature is shown most greatly without the target feature being hidden by other features.

  In addition, by highlighting the target feature on the displayed photograph, the location of the target feature can be determined in the case of an image taken from a high altitude at a high-rise building in a city. You can avoid problems that cannot be found quickly.

  Furthermore, a method of use such as a two-time image search or a past feature photo search is possible by a combination of the determination result and photographing information which is a general photo narrowing-down means.

  This embodiment has at least an internal orientation element such as the focal length of the lens, the size and number of pixels of an image sensor such as a CCD, and distortion information of the lens, and an external angle such as a three-dimensional coordinate of the photographed camera and an angle with respect to the coordinate axis. A photograph in which shooting information such as a location element and shooting date / time, a photographer, and a camera used is acquired is used.

  Further, in addition to orientation information and shooting information, a shooting information database storing geographical coordinates of a range on the ground surface where the photo is shot, a photo database storing the photo in a digitized manner, and a house map on the search screen 2D map data (planar map; also referred to as 2D information) expressed by 2D vector data that allows the user to select the feature outline to be searched for, etc., ground surface elevation data, and 3D shape data of buildings 3D data (also called 3D information) and other feature information database composed of other required feature attribute data, which is most useful for visual recognition and spatial understanding of search target features. It is an optimal photo providing system that searches for suitable photos.

  The optimum photograph providing system is a two-dimensional map on the premise that there is attribute information of a feature that the information provider wants to provide information and vector data on the two-dimensional map data associated therewith in the feature information database. When the user selects feature vector data that can provide information on the display screen, the following processing is performed.

From the coordinate information obtained by outputting the geographic coordinates of the vector data of the relevant feature, the photograph is retrieved based on the photographing range coordinates of the photographing information database for the photograph in which the relevant feature is photographed, and the image retrieval result When there are multiple photos that correspond, the spatial direction of the target feature and its neighboring features, such as the mutual positional relationship such as touching a road or touching a public facility, is the most suitable direction and scale. Means to narrow down the photos being taken,
From the 3D data of the feature information database of the target feature, the surface that enters the field of view from the assumed user's observation position is determined for the three-dimensional shape of the target feature, and the observable surface is photographed for easy viewing. Means to narrow down the photos,
From the shooting range and orientation information of the shooting information database, 3D data of the feature information database existing within the angle of view of the photograph is extracted, and a 3D model is generated in the 3D virtual space using this extracted data, and the search target location Based on the results of simulating how an object is photographed in the photo, there is a means to narrow down the optimal photo that is not hidden behind the surrounding features,
Obtained by combining the results of the above three narrowing-down means and the conventional photo search means used conventionally such as whether it is a recent feature or a past feature from shooting information, for example, date and time information. A method for adjusting the extraction conditions according to the intended use and determining the optimum photo from the photos registered in the photo database;
It is preferable that the optimum photograph includes means for highlighting the target feature in the extracted photograph and improving the recognition to the user.

  In other words, the present embodiment automatically searches for a photograph (an oblique photograph) that is most suitable for displaying a search target feature (hereinafter referred to as a target feature Hi) designated by the user from among a plurality of existing photos. , A system for providing them to a search service used in a plurality of media such as a personal computer, the Internet, a mobile phone, and CATV.

  For example, in an information search for the purpose of displaying and accessing the appearance of a specific feature on a map such as a restaurant or a real estate, when there are a plurality of images taken from various directions of the target feature, generally a video This is a search service for providing an optimal photo that preferentially searches for an image taken of the side of the target feature facing the road so that the photograph is close to the sight seen by the user.

  This search service will be described below.

The first search service
When there are multiple images of a single target feature taken from various directions, or when there are multiple photos of the target feature in multiple buildings, or the target feature has the most of the angle of view. Retrieval service (by shooting direction) that provides a preferential (larger) photo of the surface of the symmetrical feature Hi that touches the location (road, park, river, etc.) desired by the user Also called a search service).

  If the target feature occupies most of the angle of view, or if there are multiple photos of the target feature in multiple buildings and the searcher has a local intuition, the new target feature This is effective when searching. When there are multiple photos that have only one symmetric feature, a diagonal photo showing a large surface that touches the specified object such as a river, road, or park is provided. This search method is effective when a desired photograph is selected from a plurality of photographs taken of a target feature from a plurality of directions.

The second search service
This is a search service (also referred to as a shooting depression angle determination service) that preferentially searches for oblique photographs (images) obtained by broadly photographing a surface that the user can observe from the three-dimensional shape (polygon) of the target feature Hi.

  When there are multiple photos with the target feature in multiple buildings and the searcher has a local insight of the area, the new target feature will search for a photo showing the largest wall surface. It is valid.

  The third search service is a search service (also referred to as an occlusion determination rate) that preferentially searches for photos in which the target feature is not hidden by other features in consideration of the three-dimensional structure of the city. The search service and the second search service are combined.

  That is, the third search service, for example, a shooting orientation determination that preferentially searches for an image obtained by shooting the side of the target feature facing the road, and a plane that the user can observe from the three-dimensional shape of the target feature. Shooting depression angle search that preferentially searches images taken widely, occlusion determination that preferentially searches for photos in which the target feature is not hidden in consideration of the three-dimensional structure of the city, and the determination result The optimum photo determination means for determining a photograph by a combination of the narrowing-down means based on the shooting information such as the shooting date and time determines a photograph suitable for the search purpose and displays an image highlighting the building on the photograph.

  Note that the search services (1) and (2) can also perform a search service (also referred to as an optimal photo determination service for shooting date / time) that determines a photo by a combination of narrowing-down means based on shooting information such as shooting date / time.

  In the following, a search service by shooting direction will be described as a first embodiment, a determination service by shooting angle will be described as a second embodiment, and a search service by an occlusion rate will be described as a third embodiment.

<Embodiment 1>
FIG. 1 is a schematic configuration diagram of the optimum photograph providing system according to the first embodiment. As shown in FIG. 1, this system connects a user terminal 1 such as a personal computer at home or a mobile phone and a service providing center 2 to a communication network 3, and the service providing center 2 is shown in FIG. Thus, the user terminal 1 is provided with a map, and the target feature selected on the map is shown in contact with the place requested by the user (road, park, river, railway ...). A photograph in a photographing direction is provided in a format that a user can easily recognize.

  FIG. 2A is a search service screen provided and displayed on the user terminal 1 of the user, and a photo screen in the shooting direction is displayed next to the map, showing the surface in contact with the location requested by the user. ing.

  In FIG. 2B, a photograph screen in the photographing direction in which a surface in contact with a place requested by the user is displayed on the map is displayed superimposed on the map.

  FIG. 2 (c) shows the target feature designated on the orthophoto image in contact with the place requested by the user in the omnidirectional image obtained by mounting the omnidirectional camera (0 degree to 360 degree) on the vehicle. The photo screen of the shooting direction showing the screen is displayed.

  A user terminal 1 shown in FIG. 1 includes a WWW browser, an Internet connection function, and the like, and displays a map, a photograph, and the like from the service providing center 2.

  As shown in FIG. 1, the service providing center 2 includes a Web server 4, a GIS (geographic information system) server 5, a database server 6, an optimum photo generation device (application server) 7, and the like.

  The above-mentioned geographic information search system has the same function as a system that can simultaneously display and search for map images and text attribute information provided by GIS and Web implemented on existing PCs and mobile phones. There is no particular limitation on the form of the system as long as the position information of the feature can be acquired.

(Configuration of each database)
The database server 6 includes a feature information database 10 storing the feature information Ai, a photograph information database 11 storing the photograph information Di, a shooting information database 12 storing the shooting information Bi, etc. It is also referred to as a photo-providing database.

  The feature information Ai, the photographic information Di, and the shooting information Bi store information on the shooting range for each shooting range of the shooting information Bi.

  The feature information database 10 stores feature information Ai (A1, A2, A3,...). This geographic information Ai is, for example, subject attribute data Aai (name: XX restaurant address: XX business hours: 11:00 to 15:00 recommended: △△, ...) , Spatial attribute data Aci (geographic coordinate information such as latitude / longitude coordinates, plane rectangular coordinates, etc.) for the feature to be provided, and time attribute data Abi (opened on April 1, 2008) for the feature to be provided, etc. .

  The spatial attribute data Aci includes the two-dimensional map data Cai (plan map) and the three-dimensional shape data Cbi (three-dimensional city model based on a combination of the Geographical Survey Institute digital elevation data and building polygons) to which the above-mentioned geographical coordinates are assigned. Etc.

  Among these, the two-dimensional map information Cai includes the subject attribute information such as the area of the search target feature and the area of the adjacent feature that is the determination criterion of the visibility, for example, the section shape of the building and the road. It is desirable to store it as vector data in a form distinguished by Aai.

  Similarly, it is desirable that the three-dimensional information Cbi has at least altitude data on the ground surface excluding buildings and vegetation and polygon data of three-dimensional shapes of buildings and buildings.

  The above-described feature information Ai (A1, A2,...) Is specifically in a format described below, for example.

  FIG. 3 is an explanatory diagram illustrating an example of 2D map information and 3D information. FIG. 3A shows an example of an actual two-dimensional map, FIG. 3B shows three-dimensional information, and FIG. 3C shows these actual data structures. The range of the two-dimensional map information in FIG. 3B is the same as that in FIG.

  That is, the three-dimensional information and the two-dimensional map information are the feature information Ai (A1, A2,...) Is the subject attribute information Abi (Ab1, Ab2,. .), Space attribute information Aci (Ac1, Ac2,...), Site shape indicating the position and shape of the feature (two-dimensional map information Cai (Ca1, Ca2,...), Three-dimensional Information Cbi (Cb1, cb2,...: Feature polygon defined by three-dimensional coordinates) and the like.

  The three-dimensional information is composed of the elevation data of the ground surface and the three-dimensional shape data of the features such as buildings. The elevation data is, for example, a plane map such as a 50m mesh (elevation) of the Geographical Survey Institute's numerical map. It is desirable to have height information on the ground surface, and it is desirable that the 3D shape data be created from polygon data that can be used on the Web, CAD, and GIS.

  Further, the relationship between the two-dimensional map information and the three-dimensional information has a one-to-one relationship such as a site shape B and a polygon B corresponding to the subject attribute information Ab2 of the feature information A2.

  However, since the altitude data is premised on holding the geographical coordinates due to the nature of the information, the association is easy if the range of the geographical coordinates such as the site shape of the feature is known.

  That is, the two-dimensional map information Cai (corresponding to the shooting range) is expressed by two-dimensional vector data such as land section shape (X, Y), building outer periphery shape, land use boundary, etc., and three-dimensional information (shooting). (Corresponding to the range) is information indicating the feature (building) of the two-dimensional map information with a three-dimensional polygon.

  The photograph information database 11 stores a plurality of oblique photographs (aerial photographs, satellite photographs, mobile phone photographs, digital cameras, and in-vehicle cameras). These oblique photographs are two-dimensional image data (hereinafter sometimes simply referred to as photographs) obtained by digitizing image data taken by a camera.

  As long as necessary information is acquired in the shooting information database 12, the two-dimensional data dai of the photograph Di (D1, D2,...), The photographing equipment, the photographing method, monochrome, color, Encoding formats and the like are associated and stored as photo information Dai. Furthermore, a shooting range (coordinates) is stored in association with each other.

  On the other hand, the plurality of oblique photographs are taken on the ground with, for example, four flight paths. FIG. 4 shows an example of a photograph taken in a different direction (however, there is one symmetric feature).

  The shooting information database 12 stores shooting information data Bi (location information about shooting positions and orientations, camera specifications, and data related to shooting ranges for photographs used for providing information).

  For each photo (D1, D2,...) In the photo information database 11, the shooting information Bi is obtained by associating the relevant information Bai such as the date and time of the photo, the photographer, and the camera used with each photo. Shooting range data Bci that calculates the geographical coordinates of the range on the ground surface, the three-dimensional coordinates of the shooting position of the camera, the rotation angle of the shooting axis with respect to the coordinate axis, the size and the number of pixels of the imaging element such as focal length and CCD, It is composed of an internal orientation element such as lens distortion information and an external orientation element such as a three-dimensional coordinate of the photographed camera in real space and an angle with respect to the coordinate axis. Further, the external orientation element and the internal orientation element are collectively referred to as orientation information Bbi in the present embodiment.

  In addition, as shown in FIG. 5, the shooting range is obtained by calculating the viewing angle of the camera at the time when the ground is sequentially captured from the oblique direction with the camera mounted on the aircraft in the space attribute information (2D data, 3D data). ) Are the respective ranges (defined by polygons) on the spatial attribute when projected onto. FIG. 5 shows only one shooting range.

  As shown in FIG. 5, the shooting range is defined by the three-dimensional coordinates (X, Y, Z) of the shooting position Pi of the photograph in the shooting information database 12 and the rotation angle (ω, φ, κ), focal length (f), CCD or actual pixel size on the film surface (dx, dy), number of vertical and horizontal pixels (px, py), and lens distortion and feature information Using the elevation data in the three-dimensional information of the database, for example, a numerical map 50 m mesh (elevation) issued by the Geographical Survey Institute, the geographical coordinates of the photographing range Bci are calculated. FIG. 5 shows an outline of the photographing range Bci when the lens distortion is removed.

  Since these photographs, photographing information, three-dimensional information, and two-dimensional map information are information in the photographing range, they must be linked.

  An outline of this linking will be described below using the flowchart of FIG.

  In FIG. 6, these pieces of information other than the photo information Bi, the shooting information Bi, and the feature information Ai corresponding to the shooting range will be described without adding i for distinction.

  First, information providing feature determination processing is performed (S1).

  In this information provision determination process, when the feature information Ai has already been prepared regardless of the provision of the optimum photograph, the geographical coordinates of the feature are inside the photographing range of the photograph photographing information Bi. Are extracted as the optimal photo-providing features.

  The above procedure is not necessary when a database is newly constructed for the photograph information Bi, the photograph information Bi, and the feature information Ai in the photographing range.

  Next, the information providing feature information is associated with the two-dimensional map information Cai (S2).

  Generally, it is a premise that the feature information database is combined with a two-dimensional map like GIS, but when you want to replace the map displayed on the terminal screen when providing information with another map Or, if the prepared feature information and the two-dimensional map do not match, it is necessary to associate the newly prepared two-dimensional map information Cai with the feature information.

  Therefore, when the geographic coordinates of the feature are compared with the coordinates of the polygon information (for example, site shape) of the prepared 2D map information Cai, the mutual data are correlated and the feature is selected on the 2D map. , So that relevant feature information can be called immediately.

  Next, the feature information Ai to be provided is associated with the three-dimensional shape data (building polygon) of the three-dimensional information of the photographing range (S3).

  Similar to the association of the two-dimensional map, the feature and the three-dimensional information Cbi for each photographing range are associated with each other.

  Next, the three-dimensional shape data in the photograph in the photographing range is extracted from the three-dimensional information Cbi (S4).

  For each prepared photo data, calculate the 3D shape data (building polygon) of the features in the photo from the shooting position and orientation of the photo shooting information and camera specifications (shooting range) The same as the calculation of

  Next, an overlay display of these photographs and the three-dimensional shape data is performed (S5).

  This is a projective transformation of the wire frame of the three-dimensional shape data extracted above based on the orientation information, and is superimposed on the photograph and displayed simultaneously.

  Next, overlay determination processing is performed (S6).

  About the said superimposition result, the outline of the feature on a photograph and the shift | offset | difference of a wire frame are test | inspected visually.

  If there is no deviation, the feature information Ai, the shape data on the two-dimensional map, the association of the three-dimensional shape data, and the association of the feature in each photo are registered in the database, and the optimum photo search processing Data used at each step can be quickly extracted.

  If there is a shift, the process proceeds to the orientation information adjustment step, and the result is repeated again so that the shift amount of the wire frame displayed in a superimposed manner is minimized.

  Then, the orientation information is adjusted (S7).

  If there is a deviation between the contour of the feature on the photograph and the three-dimensional wire frame projectively transformed by the orientation information in the overlay determination, the orientation information shooting position and orientation, camera specifications (mainly lens The focal length and the distortion coefficient are changed, the three-dimensional wire frame is projectively transformed based on the readjusted orientation information, and the process proceeds to the overlay display step again.

  If it is determined in step S6 that the overlay is OK, the photographic information Di, 3D shape data, 2D map data, feature information, and the like are associated (S8). That is, the photograph information Di, the feature information Ai within the photographing range of the photograph information Di, and the geographical information of the photographing range are associated with each other and stored in the optimum photograph information database.

(Description of each part)
On the other hand, as shown in FIG. 1, the first optimum photo generation device 7 includes a user determination / condition input screen providing unit 14, a map providing / designated feature reading unit 16, and a feature-containing shooting information acquisition unit 18. The common 2D map information creating unit 20, the first optimum photograph determining unit 22, the providing unit 24, etc. are provided to the user terminal 1 to provide 2D map information including the desired features to the user terminal 1. Then, the photograph information Dip showing the largest target feature in contact with the place desired by the user is provided.

  When there is an access from the user terminal 1, the user determination / condition input screen providing unit 14 determines whether or not the access is valid. For example, “distance to be around”, “contact with road”, “largest”, “periphery distance 50 m”), and the like.

The map providing / symmetric feature reading unit 16 reads the address (may be a district) from the user terminal 1, reads the feature information Ai of this area from the database server 6 by the GIS 5, and transmits it to the user terminal 1. . Then, the geographical coordinates Ai of the target feature Hi on the map designated by the user terminal 1 are read out and stored in the memory 27 (Aip).

When the user selects feature vector data that can be provided by the user on the two-dimensional map screen of the user terminal 1, the feature-containing shooting information acquisition unit 18 includes all the coordinate information of the target feature. The image capturing range Bci is retrieved (photographing range Bcip) and stored in the memory 27, all the two-dimensional map information Cai including the target feature Hi is extracted (Cip), and the three-dimensional information is extracted (Cbip). All the photographic information Di having the photographing range Bcip including the target feature Hi is extracted (Dip) and stored in the memory 27. Specifically, the extracted information is an identification number of the information, and is obtained via the GIS.

  The common two-dimensional map information creation unit 20 can be used in common by using all the two-dimensional map information Cai extracted in the memory 27, a two-dimensional map of a certain range centered on the correction value and the target feature Hi, and the like. 2D map information Gi is generated and stored in the memory 28.

  The first optimum photograph determination unit 22 captures the distance from the target feature Hi and the shooting point Pi for each shooting information Bip having the shooting range Bchip in which the target feature Hi extracted in the memory 27 is shot. Based on the direction, the shooting information Bip that satisfies the user's search condition (for example, the wall closest to the road, the largest image is shown) is narrowed down. Information (Wi, Li) of the result is stored in the memory 29, and the target feature Hi of the optimum photograph Dip corresponding to the narrowed-down optimum photographing information Bip is made identifiable and stored in the memory 30.

Each time the providing unit 31 transmits the optimal photo Dip determined by the first optimal photo determining unit 22 to the user terminal 1, the memory 31 associates the date, the number of times with the user code, the user terminal number, and the like. 31.

(Description of operation)
The operation of the system configured as described above will be described below. FIG. 7 is a sequence diagram until an optimum photograph of the present embodiment is obtained. In the present embodiment, a sequence diagram between the service center and the user terminal will be described.

  The present invention includes at least an internal orientation element such as a focal length of a lens, a size and the number of pixels of an image sensor such as a CCD, and distortion information of a lens, and an external orientation element such as an angle with respect to a three-dimensional coordinate and a coordinate axis of a real space of a photographed camera. In addition, a photograph in which photographing information such as photographing date / time, photographer, and camera used is acquired is used.

  As shown in FIG. 7, the user operates the user terminal 1 to access (URL) the service providing center 2 (d1).

  The web server 4 of the service providing center 2 transmits an authentication code input screen (not shown) for user authentication from the user determination / condition input screen providing unit 14 to the user terminal 1, and from the user terminal 1. The input information on the input screen is sent to the user determination / condition input screen unit 14.

  The user determination / condition input screen unit 15 authenticates by comparing the input user information (authentication code) with pre-stored user information (not shown). 25 (d2).

  Then, the user determination / condition input screen unit 14 outputs the object input screen shown in FIG. 8A and the condition input screen shown in FIG. 8B to the web server 4 and transmits them to the user terminal 1, The object information (address or district, etc.) from the user terminal 1 and the search condition Ei are stored in the memory 26 (d3).

  In the present embodiment, as shown in FIG. 9, the search condition Ei is input as “contacting the road”, “photo showing the largest symmetrical feature”, and “peripheral distance 50 m” and stored in the memory 26. Suppose that The target feature is an address. Further, the shooting date may be input. When a shooting date (XX year / month / day to XX year / month / day) is input, it becomes possible to search for an optimum photograph from shooting information on the shooting date using this as a search condition.

  Then, the map providing / target feature reading unit 16 allocates the geographic information Ai including the address of the symmetric feature Hi, reads out the two-dimensional map information Cai (for example, FIG. 10), and transmits it to the user terminal 1. (D4).

  When the map providing / designated feature reading unit 16 determines that the user can use this system, it reads the theme (for example) of the feature from the user terminal 1, passes it to the GIS, and includes this feature. The geographical information (map) of the district is transmitted to the user terminal 1 by the web server 4.

  The user terminal 1 displays a map on the screen and designates a desired feature (two-dimensional polygon) on the map. At this time, attribute information (such as a building name) of the designated feature may be displayed.

  The map providing / designated feature reading unit 16 reads designated designated feature information (pixel coordinates), obtains geographic coordinates corresponding to the designated feature information via the GIS, and includes a feature-containing shooting information obtaining unit. 18 to send.

  Next, the user designates a desired target feature Hi from the map displayed on the user terminal 1 and transmits it to the service providing center 2 (d5).

  When the target feature Hi is received by the map providing / symmetric feature reading unit 16 and stored in the memory 27, the feature-containing shooting information acquisition unit 18, the common two-dimensional map creation unit 20, and the first optimum photograph determination unit 22 are included. The providing unit 24 performs an optimum photo providing process for obtaining an optimum photo of the symmetric feature under the conditions desired by the user (d6). This optimum photograph providing process will be described later in detail.

  Then, the obtained optimum photograph is provided to the user terminal (d7). At this time, the providing unit 22 stores user information (name, user code, address, account information, etc.) and the number of times of provision in the memory 31.

  Next, the optimum photo providing process will be described in detail.

  FIG. 11 is a flowchart for explaining the operation of the photographic information acquisition unit including a feature.

  The feature-included shooting information acquisition unit 18 reads the feature information Ai having the geographical coordinates (coordinates on the two-dimensional map of the spatial attribute information) of the target feature Hi from the map providing / target feature reading unit 16 (S21). ). This read feature information Ai is referred to as feature information Aip in the present embodiment.

  Next, the feature-included shooting information acquisition unit 18 extracts all the shooting information Bi having the shooting range Bci including the target feature Hi (see FIG. 12A) and stores it in the memory 27 (S22).

  Specifically, the extracted shooting information Bi is an identification number of shooting information. In addition, the extracted shooting information Bi is referred to as shooting information Bip in the present embodiment.

  As shown in FIG. 5, these photographing ranges Bci are obtained by viewing the viewing angle information (view angle) of the camera at the time when the ground is sequentially photographed obliquely with the camera mounted on the aircraft from the spatial attribute information ( Respective ranges (defined by polygons) on the space attribute when projected onto the virtual space coordinates of the 2D map and 3D information).

  The feature-included shooting / information acquisition unit 18 searches for all the photograph information Di linked to all the extracted shooting information Bip, extracts the searched shooting information Dip, and stores it in the memory 27. (S23). The photograph information Di stored in the memory 27 is specifically an identification number of the photograph information. The extracted photographic information Di is referred to as photographic information Dip.

  Next, the feature-included shooting information acquisition unit 18 extracts all the space attribute information Aci (two-dimensional map information, three-dimensional information) corresponding to the extracted shooting information Bip and stores it in the memory 27 (S24). The space attribute information stored in the memory 27 is specifically an identification number of the space attribute information.

  Further, the extracted space attribute information Aci is referred to as space attribute information Acip.

  Next, the feature-containing shooting information acquisition unit 18 determines whether or not the extracted spatial attribute information Acip includes a polygon (three-dimensional shape data) of three-dimensional information corresponding to the target feature Hi specified by the user. Determine (S25).

  If it is determined in step S25 that a three-dimensional polygon corresponding to the target feature Hi exists, the polygon of the three-dimensional information is acquired (hereinafter referred to as Cbip) and stored in the memory 27 (S26).

  That is, in the memory 27, as shown in FIG. 13, the extracted shooting information Bip (for example, B1p, B2p,...) Including the target feature Hi is linked to the feature information Api of the target feature Hi. Photograph information Dip (D1p, D2p,...) And spatial attribute information Acip (Ac1p, Ac2p,...) Are linked and stored in the shooting information Bip (for example, B1p, B2p,...). It will be.

  Then, when these pieces of information are stored in the memory 27, the common two-dimensional map creation unit 20 reads a distance value (for example, 50 m) as the peripheral distance of the target feature stored in the memory 26, and stores it in the memory 27. Each two-dimensional map information Caip linked to each stored photographing information Bip is expanded based on this distance value, these are combined, and the map information Gi including the common surroundings using a predetermined correction value or the like. (Simply referred to as a common map Gi) is generated and stored in the memory 28 (S27: see FIGS. 12B and 12C).

  For example, the distance around the designated feature (east, west, south, north, or OOm or radius OOm from the designated feature) and each spatial attribute information (two-dimensional map information) linked to each shooting range , Generating spatial attribute information (common map Gi) that can be used in common for optimum photo search from a predetermined range of a planar map including the correction value and the target feature Hi, and the geography of the designated feature Associate with coordinates.

  This common map Gi is, for example, in the case where the image is taken horizontally with respect to the ground surface, when the depth direction coordinates of the shooting range are infinite and the amount of information to be extracted becomes enormous, the target location is centered on the shooting position. By ignoring data only in the range farther than the object, it is possible to reduce the load according to the hardware capability.

  If it is determined in step S25 that each spatial information Aci does not have the three-dimensional information (polygon) of the target feature Hi, the shooting information, photograph, and geographic information corresponding to the spatial information are deleted. The process proceeds to step S27 (S28).

  Next, the operation of the first optimum photograph determination processing unit 22 will be described. The first optimum photograph determination processing unit 22 compares each piece of shooting information with the common map Gi, and calculates an easy-to-see index (an index of visibility according to the shooting direction) based on the shooting direction and the distance from the shooting point Pi. calculate.

  In the present embodiment, using the target feature, the common map Gi, and the subject attribute information (roads, buildings), a photograph in the optimal shooting direction is determined from among a plurality of photos of the target feature. The index calculation procedure will be described by taking as an example the case of selecting a photograph taken of the side of the target feature facing the road.

  However, in the feature information database 10 to be used, the combination of the features used for determination is distinguished from the region of the search target feature and the region of the adjacent feature serving as the determination reference by some subject attribute information. The content is not particularly limited. For example, a combination of a building facing a park and a public facility facing a river is also conceivable. It is also conceivable to use a building shape line other than the partition shape.

  14 and 15 are flowcharts for explaining the operation of the first optimum photograph determination processing unit 22.

  When the common map information Gi is generated in the memory 28, the first optimum photograph determination processing unit 22 groups the common map information Gi according to the feature (S31). The grouped data is stored in the memory 28.

  For example, as shown in FIG. 3 (c), it is divided into groups (for example, roads, building sites, buildings... With coordinates).

  For example, when the target feature is a building and the user displays a wall facing the road as a search condition, a road composed of a road center line, road boundary line, etc. using two-dimensional data (polygon) of the common map Gi The target feature is separated from the feature that determines the status of spatial continuity in the surrounding area, such as a site portion including a building shape line, a block line, and a site boundary line.

  Next, the shape of the target feature Hi (see FIG. 16 (a): EFGHI) is read from the common map Gi and the search condition of the memory 26 (the surface in contact with the road and the largest target feature is shown). Is read (S32). In the present embodiment, the shape of the target feature Hi is EFGHI as shown in FIG.

  Next, on the common map Gi, a determination boundary line Lj (EFGH) of the target feature Hi in contact with the road is obtained and stored in the memory 29 (S33).

  Then, the number bi (for example, the first number) of the shooting information Bip (B1p, B2p,...) Of the shooting point Pi stored in the memory 27 is set (S34).

  Next, the shooting point Pi (P1) corresponding to the set shooting information Bip and the projection plane qi (q1: corresponding to the angle of view) of the camera are read, and the shooting point P1 and the projection plane q1 are defined in the common map Gi. (S35: See FIG. 16 (a)). As shown in FIG. 16A, an imaging point, a width WL1 of the projection plane, and the like are defined.

In particular,
From the orientation information of this photographing information, there is a CCD of the photograph at a position that is centered on the projection plane of the camera (the coordinates (X1, Y1, Z1) of the photographing point P1) and is separated from the central point by the focal length f of the camera. Is the rotation angle (ω) with respect to the coordinate axis XYZ, the line connecting the center of the film surface pixel scale (dx, dy) and the center of the projection plane obtained from the number of vertical and horizontal pixels (px, py) and the shooting point P1 , φ, κ), the coordinates of the intersections A, B, C, D of the perpendicular line down to the ground surface with a height of 0 from the three-dimensional coordinates of the four corners of the projection plane are obtained. Similarly, a point P′1 (X1, Y1,0) where the height is 0 for the photography point P is obtained.

  Among the straight lines connecting point P'1 and points A, B, C, and D, a combination of straight lines with the largest angle is a temporary image on the ground surface (common map and 3D information 2D data). Set as a corner. In the figure, a straight line P'1-A and a straight line P'1-B are shown.

  Further, a straight line connecting the points A and B used for the temporary angle of view is set as a temporary projection plane W1 on the ground surface, and the width WL1 is obtained. Of course, the four points selected depend on the rotation angle (ω, φ, κ) of the photo.

  Next, the line length (collimable line projection plane width wci) of the designated feature obtained from the point P1 on the temporary projection plane W1 is obtained (S36).

In particular,
In the angle of view on the ground surface, the point connecting the point P'1 and any point on the judgment boundary line (EFGH) has never passed through the section shape (EFGHI) in the straight line section. It is assumed that the determination boundary line can be collimated from the shooting point. The set becomes the collimable line (EF) of the point P′1.

  However, although a vertex with a partition shape is determined as a collimable point, if there is no continuous collimable line between adjacent vertices, the point is not recognized as collimable.

  Next, the ratio Wi in the image of the determination boundary line Lj that can be collimated within the angle of view from the photographing point P1 is obtained. That is, the length (Lci) of the wall surface (EF) when the target feature is viewed at the angle of view (WL1) of the camera at the photographing point P1 is obtained (S37).

  The aforementioned ratio Wi is referred to as a collimated visible ray projection ratio Wi, and is obtained as Wi = Wci / Wli and stored in the memory 29.

Specifically, a search is made for a combination that maximizes the angle formed by a straight line connecting any two points on the collimable line (EF) of the point P'1 and the point P'1 (points E and F in the figure). ), A length collimable line projection plane width wc1 between the points J and K at which each straight line connecting the two points and the point P′1 intersects the temporary projection plane W1 on the ground surface is obtained.

In addition, as a method for determining the temporary projection plane, as shown in FIG. 16B, m, for example, in a photograph taken from directly above such as an aerial photograph, the coordinates of the four corners of the projection plane are lowered onto the ground surface. When the shooting point P is within the range surrounded by the intersections A, B, C, D, four areas (area AB, area) formed by the straight lines connecting the intersections A, B, C, D and the shooting point P AC, region CD, region BD), and there is a method in which the intersection points B and D constituting the region BD including the target feature are set as temporary projection planes. Next, collimation of the point P′1 As a ratio of the possible line projection plane width wc1 and the provisional projection plane width WL1 on the ground surface, the collimable line projection ratio wc1 / WL1 of the point P′1 is obtained. That is, the ratio of the actual extension distance of the judgment boundary line and the extension distance of the portion that can be collimated in the photograph is calculated. These two values are used as an easy-to-see index for determining the shooting direction, and a photograph of a part that faces a feature serving as a determination condition is determined.

  Next, it is determined whether there is another shooting information Bip in the memory (S38). If it is determined in step S38 that there is other shooting information, the next shooting information is updated and the process proceeds to step S34.

  By performing such processing, the memory 29 is in contact with the road surface when the object is imaged at the angle of view of the camera of the imaging information of each imaging point P1, P2, P3 with respect to the target feature Hi. The ratio of the determination boundary line of the wall surface (the ratio of the wall surface visible from the shooting point on the ground surface) is stored as shown in FIG.

  Next, when it is determined that there is no other shooting information Bip extracted in step S38, the determination boundary line Lj stored in the memory 29 and the length of the collimable line (EF or EFGH) from each shooting point The ratio Li to the length Lci (Lc1, Lc2) is sequentially obtained, and these are stored in the memory 29 as the determination boundary line collimable ratio Li (L1, L2,...) For each photographing point (S41).

For example, at the shooting point P1,
A ratio Lc1 / Lj between the length Lj of the determination boundary line (EFGH) and the length Lc1 of the collimable line (EF) at the point P′1 is obtained as the determination boundary line collimable ratio L1. Such processing is performed for each photographing point. Therefore, as shown in FIG. 17, the target feature Hi, the shooting point Pi, and the determination boundary line collimable ratio Li are stored in the memory 29 in association with each other.

  Then, the collimable line projection ratio Wi of each photographing point stored in the memory is compared (S42).

  Taking FIG. 16 (a) as an example, by using these collimable line projection ratios Wi, for example, the same camera as point P′1 in FIG. There are photographing points P′2 and P′3 that are equal in distance from the camera. The width of the projection plane at these three photographing points is equal to WL1 = WL2 = WL3.

In the figure, the shooting point P'2 is wc2 = 0 because all the straight lines connecting the points on the judgment boundary line (EFGH) and the shooting point pass inside the section shape (EFGHI), and the collimable line projection ratio wc2 / WL2 = 0. At the point P′3, the collimable line is wider (EFGH) than (EF) of P′1, so that the collimable line projection plane width wc3 is larger than P′1.

  Therefore, the collimable line projection ratio wc / WL at the three shooting points is P'3> P'1> P'2, and the photograph taken from the point P'3 is used as the judgment boundary line within the angle of view. It can be determined that the road side of the target feature is photographed the most.

  Of course, since the projection plane width WL changes depending on the focal length of the camera of the photograph to be compared and the distance from the target feature, the collimable line projection ratio also changes, and the image of the judgment boundary line Lj considering various shooting conditions The size within the corner can be compared.

  Further, the sizes of the determination boundary line collimable ratio Li are compared (S43).

  By using the determination boundary line collimable ratio Li (L1, L2, L3) taking FIG. 16A as an example, for example, the photographing points P′1, P′2, and P ′ in FIG. When the judgment boundary line collimable ratio Li (L1, L2, L3) of 3 is compared, Lc1 / Lj = (EF) / (EFGH) <1 at P′1, and Lc2 / Lj at point P′2. = 0 / (EFGH) = 0, and at the point P′3, Lc3 / Lj = (EFGH) / (EFGH) = 1.

For this reason, the judgment boundary line collimable ratio is P'3> P'1> P'2, so the photograph taken from the point P'3 shows the road side of the target feature as the judgment boundary line. It is possible to determine that the images are being taken evenly.

Next, when the two indexes of the collimable line projection ratio Wi and the determination boundary line collimable ratio Li are maximized, the photographing information of the photographing point Pi is extracted, and the photograph information Dip of the photographing information is obtained as the user but the optimal photo information to satisfy the user's search criteria for the specified object (S44).

And it memorize | stores in the memory 30 (S45). In other words, the ratio of the size of the determination boundary line to the camera angle of view of the shooting point and the ratio of the portion that can be collimated in the photograph are calculated. These two values are used as an easy-to-see index for determining the shooting direction, and a photograph of a portion that faces a feature serving as a determination condition is determined.

  As a result, even when it is determined that the photograph taken very close to the target feature Hi by the determination boundary line collimable ratio Li is the maximum in the collimable line projection ratio Wi, the determination boundary line vision If the quasi-possible ratio Li is small, it is possible to select other determination boundary lines as photographing information that has been photographed uniformly.

  Next, the polygon of the three-dimensional information of the target feature corresponding to the extracted shooting information is superimposed on this photographic information with outline enhancement (for example, red, pink) (S46). Then, the photograph information is transmitted to the user terminal (S47).

That is,
The above two indicators, the collimable line projection ratio wc / WL and the judgment boundary line collimable ratio Lc / Lj, photograph the side facing the feature specified by the user from the attributes of the surrounding space of the target feature. It is a numerical value indicating the condition of ease of viewing that changes depending on the shooting direction and distance in order to extract a photograph that is being taken.

  Here, the description of the first optimum photograph determination unit 22 will be supplemented. FIG. 17 is an explanatory diagram for explaining the processing operation of the first optimum photograph determination unit in association with the memory.

  The collimable line projection ratio Wi for each photographing point Pi is stored in the memory 29a, the judgment boundary line collation possibility ratio Li for each photographing point Pi is stored in the memory 29b, and the judgment boundary line Lj is stored in the memory 29c. Remembered. Further, the extracted feature information Aip, shooting information Bip, and photo information Dip are stored in the memory 27. The common map Gi is stored in the memory 28.

  The first optimum photograph determination processing unit 22 uses the common map Gi stored in the memory 28 to extract the shooting information Bip (including the shooting position Pi) extracted from the memory 27 and the projection plane qi (q1: image) of the camera at the shooting position Pi. (Corresponding to the corner) is defined (S35: see FIG. 16A). As shown in FIG. 16A, an imaging point, a width WL1 of the projection plane, and the like are defined.

  Then, the line length (collimable line projection plane width wci) of the designated feature obtained on the temporary projection plane W1 from each photographing point Pi (P1, P2, P3) of the common map Gi is obtained.

  Next, the ratio Wi of the determination boundary line Lj (EFGH in FIG. 16A) of the surface in contact with the road that is one of the search conditions designated by the user within the angle of view from the photography point Pi to the image. Ask for.

  Specifically, as the ratio of the collimable line projection plane width wc1 of the point P′1 and the width WL1 of the provisional projection plane on the ground surface, the collimable line projection ratio Wi ( wc1 / WL1) is obtained and stored in the memory 29a in correspondence with the photographing information of the photographing point Pi.

  Further, a ratio Li between the determination boundary line Lj stored in the memory 29c and the length Lci (Lc1, Lc2) of the collimable line (EF or EFGH) from each photographing point is sequentially obtained, and these are photographed. The determination boundary line collimable ratio Li (L1, L2,...) For each point Pi is stored in the memory 29b.

  Then, the collimable line projection ratio Wi of each photographing point stored in the memory 29a is compared, and the determination boundary line collimable ratio Li (L1, L2, L3) of the memory 29b is compared. The comparison result is stored in the memory 29d.

  Next, when the two indexes of the collimable line projection ratio Wi and the judgment boundary line collimable ratio Li are maximized, the photographing information Bip of the photographing point Pi is extracted, and the photographic information Dip of the photographing information Bip is extracted. The optimum photo information Dip that satisfies the user's search condition for the object specified by the user is used.

  Then, the image data obtained by superimposing the polygon of the three-dimensional information of the target feature Hi on the optimum photographic information Dip and enhancing the outline (for example, red, pink) on the photographic information is stored in the memory 30. Then, the photograph information is transmitted to the user terminal 1.

  FIG. 18 is an explanatory diagram for explaining the overlapping process. The first optimum photograph providing unit 22 uses three-dimensional shape information in the same range as the two-dimensional map information where the target feature Hi exists as shown in FIG. 18A (the feature is defined in the coordinate system of polygons and XYZ). Except for the polygon of the target feature Hi, other polygons are deleted (FIG. 18B). Then, the outline of the polygon is colored with pink, red, yellow, etc. (S40).

  Then, the photograph information Dip linked to the shooting information Bip having the shooting range including the target feature Hi and the polygon layer of only the target feature Hi are synthesized (FIGS. 18C and 18B). Is transmitted to the user terminal 1 as the optimum photograph (FIG. 18E).

  On the other hand, in the example shown here, numerical values are calculated in a two-dimensional space, but may be implemented in a three-dimensional space.

  FIG. 19 is an explanatory diagram for explaining a method of generating a determination boundary line when the section shape of the target feature is only point coordinates.

  In the calculation of the legibility index based on the shooting direction, a method of generating a determination boundary line in a pseudo manner when there is only a point coordinate and no section shape of the target feature in the two-dimensional map information will be described.

  For example, in 2D map information composed only of operational space data such as roads, railways, rivers, and public facilities, when the site boundary of individual general buildings is not included, the point coordinates of the target feature are the center A circle with a certain radius is generated, and the line information of the service space within the circle is used as a judgment boundary line. In FIG. 19, a road line is used as the determination boundary line Lj.

  For the radius of the circle to be generated, change the length of the functional information of the subject attribute information of the target feature, for example, by classifying large-scale / small-scale facilities etc. according to the use of land such as commercial facilities and houses It is also possible.

  In addition, when there are two or more judgment boundary lines in a circle, a method of adopting the longest section among the existing judgment boundary lines, or a threshold with a distance between two judgment boundary lines on the same line In the following cases, there is a method of joining as one judgment boundary line including a section outside the circle in between.

<Embodiment 2>
The second embodiment is a second search service, which is a search service that preferentially searches for images obtained by photographing a wide range of surfaces that can be observed by the user from the three-dimensional shape (polygon) of the target feature Hi (shooting angle determination service). It is also called).

  In the second embodiment, the visibility index based on the shooting depression angle is calculated using the shooting information Bip and the extracted three-dimensional information Cbip (three-dimensional polygon of the building).

For example, when the target feature is observed from the ground surface, except for shapes that can be seen up to the zenith such as a conical shape or a triangular pyramid, the rooftop portion of a building such as a cube or a rectangular parallelepiped is put into view. I can't. In this way, features that have shapes that do not enter the field of view when observed from a certain point change the way they appear depending on the height of the shooting position and the angle of depression. The next index is calculated so that the selected photograph is selected, and the optimum photo determined based on this index is provided to the user terminal 1.

  FIG. 20 is a schematic configuration diagram of the second embodiment. In the present embodiment, description of the same parts as those in FIG. 1 is omitted. In the second embodiment, only the configuration of the application server 7 is shown.

  The second embodiment has a second optimum photo determination unit 35 instead of the first optimum photo determination unit 22 of the first embodiment.

  The second optimum photograph determination unit 35 is based on the elevation data of the ground surface and the three-dimensional shape data of the target feature of the three-dimensional information Cbip including the target feature Hi acquired by the feature-including shooting information acquisition unit 18. Then, a three-dimensional model of the city is generated in the virtual space (memory 38).

  A virtual collimation position (XYZ) preliminarily stored in the memory 39 so as to simulate the observer's viewpoint on the ground surface that is a specific distance away from the target feature in the virtual space. From this, the directly observable surface of the three-dimensional surface of the target feature is stored in the memory 36 as a collimable surface (KYZ).

  Next, the area of the collimable surface of the target feature Hi and the area of the other surface are obtained. Next, a two-dimensional virtual bird's-eye view photographed under the same conditions as the orientation information of the photograph for which the three-dimensional model generated in the virtual space is to be determined is generated, and the target feature in the virtual bird's-eye view can be collimated The area on the image of the surface and other surfaces is obtained.

The relationship between the area ratio in the virtual space of the collimable surface and the other surface and the virtual overhead view is determined as an easy-to-see index in the photographing depression angle determination, and a photograph of the collimable surface being easily captured is determined. .

  The processing of the second optimum photograph determination unit 35 will be described below.

  FIG. 21 is an explanatory diagram schematically illustrating a method for obtaining a collimable surface. FIG. 22 is a flowchart of a method for obtaining a collimable surface. FIG. 23 is an explanatory diagram for explaining an easy-to-see index based on a shooting depression angle using a collimable surface. FIG. 24 is a flowchart for obtaining an easy-to-see index based on a shooting depression angle using a collimable surface.

  First, FIG. 22 will be described.

In the calculation of the legibility index of the target feature, the photographing depression angle determination uses a surface Rpi that can be observed by the user among the three-dimensional surface of the feature.

  This collimable surface Rpi assumes that the user observes the target feature from the ground surface, such as a vehicle or a sidewalk, and uses the three-dimensional shape data of the target feature Hi and the altitude data of the ground surface. A point on the ground surface separated by a collimation distance Ri set from the contact point with the ground surface is set as a virtual collimation point Mi, and the three-dimensional shape data (polygon) of the target feature Hi that can be seen from the virtual collimation point Mi Let the surface be a collimable surface Rpi. However, the surface in contact with the ground surface is not included in either of them.

Ideally, the virtual collimation point Mi is set only on the same plane as the determination boundary line Lj used in the imaging orientation determination if possible. In addition, when using very detailed three-dimensional information, since it is affected by the fine terrain near the virtual collimation point Mi, the virtual collimation point Mi is equivalent to the average height and vehicle height from the ground surface. A method of setting with a high offset may also be considered.

  The collimating distance Ri includes a method of setting a specific numerical value in advance and a method of setting the collimating distance Ri according to the feature used for determining the determination boundary line Lj used in the shooting direction determination. For example, when a road is used to determine the determination boundary line Lj, it is conceivable to use the road width as the collimation distance Ri.

  The second optimum photograph determination unit 35 stores the feature information Aip, the shooting information Bip, and the photograph information Dip including the target feature Hi in the memory 27 by the feature-included shooting information acquisition unit 18, and the common two-dimensional map. When the common map Gi is generated in the memory 28 by the information creating unit 20, the shooting information Bip in the memory 27 is read (S51).

  Next, the three-dimensional shape Ri (polygon: given by XYZ) of the target feature Hi included in the photographing range Bcip of the photographing information Bip is used based on the three-dimensional information Cbip (ground surface and altitude data) in the memory 27. Then, a three-dimensional model Gp of the city is generated in the masquerade space (memory 38) (S52).

  Next, in this three-dimensional model Gp, the determination boundary line Lj of the target feature Hi is searched from the memory 29 (S53).

  Then, a collimation position Mi set in advance in the memory 39 is set in the vicinity of the searched three-dimensional shape of the target feature Hi (S54: see FIG. 21).

  Next, a collimation plane Rpi and a collimation impossible plane Rqi are determined by drawing a collimation line from the collimation position Mi to a plane constituting the three-dimensional shape of the target feature Hi (S55). Next, the target feature Hi, the photographic information Bip, the photographic information Dip, and the like linked to the determination boundary line Lj stored in the memory 29 are read from the collimable surface Rpi or the non-collimable surface Rqi, and the memory 36 is read. Are linked to and stored (S56).

  The collimable surface Rpi and the non-collimable surface Rqi are determined by comparing the three-dimensional surface conditions of the target feature Hi. The surface in contact with the ground surface is a non-collimable surface Rqi.

  Next, it is determined whether or not another shooting information Bip is stored in the memory 27 (57). If the other shooting information Bip is stored, it is updated to the next shooting information Bip and the process proceeds to step S51 (S58). ). That is, the collimable surface Rpi and the non-collimable surface Rqi from the photographing direction are obtained for each collimation position Mi installed on the ground surface of the photographing range of the camera from all photographing points where the target feature Hi is photographed. And stored in the memory 36 (see FIG. 25).

  Next, a process for determining an easy-to-see index based on a shooting depression angle using a collimable surface will be described.

This determination process generates a virtual overhead view of the target feature photographed by the camera for each shooting point, and the area ratio between the collimable surface and the non-collimable surface of the virtual overhead view (index of visibility). ) And the area ratio (index of visibility) of the collimable surface Rpi and the collimable surface Rqi of the target feature photographed by the camera at the same photographing point is determined.

Next, FIG. 23 will be described.

  After determining the collimable surface and non-collimable surface of the 3D shape data of the target feature for each shooting point, the total area Sc of the collimable surface and the area of the non-collimable area in the three-dimensional space Find the total Sn.

  Next, using the same 3D shape data and elevation data, generate a 2D virtual overhead view as if it were shot with the same camera specifications, position, and orientation based on the orientation information in the shooting information database for each photo. Then, the total area Ac of the collimable surface and the total area An of the non-collimable surface on the image are obtained.

  For the sum of the obtained areas, the area ratio (Sc · An) / (Sn · Ac) on the photograph with respect to the actual area ratio is obtained as an easy-to-see index of the collimable area by the shooting depression angle. It is determined that the smaller the value is, the easier it is to see the collimable surface within the angle of view.

  For example, in FIG. 23, there are virtual overhead views generated for photographs 1 and 2, and the total area Sc of the collimable surfaces and the total area Sn of the non-collimable areas are the same because the object is the same feature. The sum of the collimable areas on the virtual overhead view of Photo 1 has two regions, and is Ac1 = Ac11 + Ac12, and the sum of the collimable areas An1 = An11 + An12.

  Similarly, in the virtual overhead view of Photo 2, Ac2 = Ac21 + Ac22, and the non-collimable area is An2, because there is only one area. Comparing the area ratios of these, (Sc · An1) / (Sn · Ac1) <(Sc · An2) / (Sn · Ac2), so the photo directly faces the collimable surface as shown in Photo 1. Judging that it is easier to see.

  Next, a description will be given using the flowchart of FIG. The second optimum photograph determination unit 35 reads the shooting information Bip of the target feature Hi stored in the memory 29 (S61).

  Next, a two-dimensional virtual bird's-eye view Fi as if the target feature was photographed with the camera camera specifications, position and orientation of the orientation information of the read photographing information Bip is generated and stored in the memory 37 (S62).

  For example, the three-dimensional shape of the target feature in the memory 38 having the determination boundary line Lj stored in the memory 29 is photographed with the camera camera specifications, position, and posture of the orientation information of the read photographing information Bip. A two-dimensional virtual overhead view Fi as described above is generated. That is, in the memory 37, the virtual overhead view Fi (F1, F2,...) When shooting is performed with the shooting information of the camera at each shooting point, the determination boundary line Lj, the shooting information Bip (B1p, B2p,...), Etc. Are linked and stored.

  Next, it is determined whether there is another shooting information Bip (S63). If there is another shooting information Bip, the process returns to step S61 (S64). That is, in the memory 37, the virtual overhead view Fi (F1, F2,...) When shooting is performed with the shooting information of the camera at each shooting point, the determination boundary line Lj, the shooting information Bip (B1p, B2p,...), Etc. Are linked and stored.

  When it is determined in step S63 that there is no other shooting information Bip, the total area Ac of the collimable surface and the total area An of the non-collimable surface on the image are obtained for each virtual overhead view Fi ( S65).

  Next, the area ratio Ji (Sc · An) / (Sn · Ac) on the photograph with respect to the actual area ratio is obtained as an easy-to-see index of the collimable area based on the shooting depression angle (S66). ).

  Next, it is determined whether there is another virtual overhead view Fi (S67). If it is determined in step S67 that there is another virtual bird's-eye view Fi, the next virtual bird's-eye view Fi is updated and the process proceeds to step S65 (S68).

  That is, as shown in FIG. 25, in the memory 37, the area ratio Ji (Ji = (Sc · An ) / (Sn · Ac)) is stored.

  In step S67, when it is determined that there is no other virtual overhead view Fi, on the photograph with respect to the actual area ratio as an easy-to-see index of the collimable area by the shooting depression angle for each virtual overhead view Fi in the memory 37. The photo information Dip corresponding to the shooting information Bip at the shooting point where the area ratio Ji (Ji = (Sc · An) / (Sn · Ac)) is the smallest value is determined as the optimum photo information Dip in the memory 40. Store (S69).

  Then, the contour enhancement of the target feature of the optimal photo information is performed (S70), and the photo information is transmitted to the providing unit 24 and transmitted to the user terminal 1 (S71).

<Embodiment 3>
The third search service is a search service (also referred to as occlusion determination) that preferentially searches for photos in which the target feature is not hidden behind other features in consideration of the three-dimensional structure of the city. , A service combining the second search service.

  In the occlusion determination, an easy-to-see index based on occlusion is calculated using the photographing information Bip and the extracted three-dimensional information.

  FIG. 26 is a schematic configuration diagram of the optimum photograph providing system according to the third embodiment. In the system according to the present embodiment, the optimal photo determination process for each shooting direction and distance is performed by the first optimal photo determination unit 22, the optimal photo determination process by the shooting depression angle is performed by the second optimal photo determination unit 35, and occlusion determination. The processing will be described as the third optimum photograph providing unit 50.

  The provision unit 55 of the third embodiment is stored in the memory 52, the optimum photograph for each shooting direction distance stored in the memory 30, the optimum photograph for shooting depression angle determination stored in the memory 40, and the memory 52. It has a function of allocating an optimal photo for occlusion determination and transmitting an optimal photo that has been narrowed down by shooting date (set by the user or the center).

  Further, it has a function of providing the optimum photos of the memory 30, the memory 40, and the memory 52 in the order corresponding to the preset priority level. At this time, the type and the number of sheets are changed depending on the threshold.

  The third optimum photograph providing unit 50 determines whether or not there are a plurality of building polygons in the three-dimensional model Gki in the memory 57. If there are, the same technique as the virtual overhead view Fi generated at the time of photographing depression angle determination is used. Thus, a virtual overhead view FKi including the three-dimensional shape (polygon of the building) of the surrounding features included in the angle of view of the photograph to be judged this time is generated in the memory 57.

  On this virtual bird's-eye view FKi, the area of the collimable surface where the target feature is not hidden by other features such as buildings and terrain is obtained, and the collimation is possible when there is no feature other than the target feature The ratio with the area of the surface is calculated as the occlusion rate Qi.

  Using this value as an easy-to-see index by occlusion determination, a photograph with a small number of portions where the target feature is hidden behind other features is determined.

  This will be further described with reference to FIG.

FIG. 27 is an explanatory diagram for explaining the calculation of the visibility index by the occlusion determination using the collimable surface of the three-dimensional shape data of the feature.

  After obtaining the total area Ac of the collimable surfaces on the virtual overhead view Fi generated based on the three-dimensional shape data, elevation data, and imaging information Bip of the target feature Hi, the surrounding area existing within the same angle of view. Similarly, the virtual overhead view Fki including the three-dimensional shape data of the object is generated in the memory 57. Then, the total area Ac ′ of the collimable surfaces of the target feature that is not hidden behind other features on the virtual overhead view is obtained.

  Specifically, the surrounding polygon and the polygon of the target feature are overlapped, and the surface of the target feature that is not hidden by other features is extracted and used as a collimable surface A′ci.

  Occlusion with the ratio of the total area Ac and Ac 'of the collimable surface Ac' / Ac as the occlusion rate when only the 3D shape data of this target feature and other features within the angle of view are included It becomes an easy-to-read index by judgment. The closer this index is to 1, it can be determined that the photograph is less affected by the surrounding features.

  For example, in FIG. 27, the total area of collimable surfaces on the virtual overhead view including the peripheral features in photo 1 is Ac′1 = Ac′11 + Ac′12. Ac'2 = Ac'21 + Ac'22. Here, the area of Ac′11 in the virtual overhead view of Photo 1 is smaller than that of Ac11 in FIG. 23 by the amount hidden behind the feature in front.

  On the other hand, in the virtual overhead view of Photo 2, there is no change in the area of the collimable surface as compared with FIG. Therefore, the occlusion rate Ac '/ Ac is (Ac'1 / Ac1) <(Ac'2 / Ac2) = 1, and in Photo 2, the collimable surface of the target feature is hidden by the surrounding features. Judge that it is not easy to see.

  Next, a description will be given using the flowchart of FIG.

  After obtaining the total area Ac of the collimable surfaces on the virtual overhead view Fi generated based on the three-dimensional shape data, altitude data, and imaging information Bip of the target feature Hi, the third optimum photograph providing unit The shooting information Bip including the target feature Hi is read into the memory 56 (S81).

  Next, the three-dimensional model Gki is generated in the memory 57 using the three-dimensional information of the shooting range of the read shooting information Bip (S82: see FIG. 29).

  Next, a virtual overhead view Fki when the stereoscopic model Gk is photographed with the camera of the photographing information Bip that has been read is generated (S83).

  Next, the surface including the collimation boundary line Li of the generated virtual overhead view Fki is compared with the surface of the three-dimensional shape HBp of another feature in the virtual overhead view Fki to obtain the occlusion rate Qi, This is stored in the memory 56 in association with the photographing information Bip (S83).

  After determining the total area Ac of the collimable surface on the virtual overhead view Fi generated based on the imaging range of the three-dimensional shape data, the elevation data, and the imaging information Bip of the target feature, Obtain the total area Ac 'of the collimable surface of the target feature that is not hidden behind other features on the virtual bird's-eye view Fki including the three-dimensional shape data of the surrounding features existing in the same angle of view. (The memory 56 is linked to the shooting information Bip and stored).

  Specifically, the surrounding polygon and the polygon of the target feature are overlapped, and the surface of the target feature that is not hidden by other features is extracted and used as a collimable surface A′ci.

  The ratio Ac '/ Ac of the total area of the collimable surface Ac and Ac' when only the 3D shape data HBp of this target feature and other features within the angle of view are included as the occlusion rate Qi The visibility index is based on occlusion determination.

  Next, it is determined whether or not other shooting information Bip exists in the memory 27. If it does not exist, it is updated to the next shooting information Bip and the process returns to step S89 (S85).

  That is, as shown in FIG. 29, the virtual overhead view Fki stored in the memory 57 and the three-dimensional shape of Hi in the shooting information Bip stored in the memory 56 (the shooting information Bip and the determination boundary line Lj and this shooting The comparison result between the three-dimensional shape of Hi in the information Bip, the total sum Aci of collimable surfaces, and the occupancy rate Qi is obtained as an occlusion rate, and this is stored in the memory 56 . In FIG. 29, the three-dimensional shape (building polygon) of the target feature Hi in the shooting range of the shooting information is described as HBp.

  Next, the photo information Dip of the shooting information having the occlusion rate Qi closest to “1” among these occlusion rates Qi is determined as the optimum photo information and stored in the memory 52 (S87).

  Then, the polygon of the target feature Hi of the virtual overhead view Fki corresponding to the optimum photograph information is superimposed, emphasized (S88), and transmitted to the user terminal by the providing unit 55 (S89).

  FIG. 30 is an explanatory diagram for explaining the enhancement processing.

  A region on the virtual overhead view Fki of the three-dimensional polygon of the target feature that is not hidden behind the surrounding feature within the angle of view is obtained using the same method as the calculation of the visibility index by the occlusion determination. This region includes both the collimable surface and the non-collimable surface.

  Next, overlay the generated virtual overhead view and photographic data, and for the target feature area, for example, by adding image processing such as coloring of the wire frame, filling of the area, translucent, etc. Generate an image that highlights the visible part above. In the present system, the image processing technique related to this highlighting is not particularly limited. In FIG. 30, the target feature region on the photograph is emphasized by being painted by semi-transparent processing.

  Therefore, in the present embodiment, the processing shown in the flowchart in FIG. 31 can be performed.

  The flowchart of FIG. 31 shows the concept of processing in time series.

  When the user designates a feature on the geographic information user terminal 1 (S110), the corresponding feature information Ai and geographic coordinates are retrieved from the database by the geographic information retrieval unit of the center (S111, S112). It is also possible to input geographical coordinate information from a system different from this system (S100).

  Next, the photographing information database 12 is searched for photographing information Bip of a photograph in which the geographical coordinates described in the feature information of the input feature are included in the photographing range Bci (S113).

  Next, the two-dimensional map information included in the shooting range of the shooting information extracted by the search of the shooting information and the three-dimensional information included in the angle of view of the photograph are searched from the feature information database 10 (S114). ).

  In this information search, for example, when the image is taken horizontally with respect to the ground surface, the depth direction coordinates of the shooting range are infinite, and the amount of information to be extracted becomes enormous. By ignoring data only in the farther range, it is possible to reduce the load according to the hardware capability.

  Next, using the shooting information Bip and the extracted two-dimensional map information, a legibility index based on the shooting direction is calculated (S115).

  Next, an easy-to-see index based on the shooting depression angle is calculated using the shooting information Bip and the extracted three-dimensional information (S116).

  Next, the visibility index based on the occlusion rate Qi is calculated using the shooting information Bip and the extracted three-dimensional information (S117).

  Next, when there are a plurality of photographs of the target feature, the most suitable target feature in accordance with the threshold value and the extraction condition that are arbitrarily set from the photographing direction, the depression angle, the occlusion determination result, and the information of the photographing information database 12 Determine the optimal photo that is easy to see.

  If there is no photo that meets the conditions, a message that there is no corresponding photo is displayed on the user terminal 1 (S118).

  Next, the optimum photograph and the corresponding virtual overhead view are superimposed, and the target feature on the photograph is emphasized by changing the color tone of the range of the target feature on the virtual overhead view (S119). The feature emphasized image is displayed on the user terminal 1 (photo display unit).

  As a highlighting method, for example, when the 3D shape information of the underlying feature is provided in polygon format, the wire frame connecting the vertices of the polygon can be displayed in any color (specifically, any color). Can you imagine a method to superimpose colored images by adding a colored or semi-transparent process within the range? .

  Then, the photograph closest to the search condition is transmitted to the user terminal 1 and displayed (S120).

  One or more photographs may be displayed. There is also a method of arranging and displaying the calculated indices in an order that is easy to see.

<Embodiment 4>
That is, as shown in the flowchart in FIG. 32, it is possible to obtain any one of the optimum photographs based on the shooting direction, the depression angle, the occlusion rate, or a combination thereof.

  This example will be described as a fourth embodiment.

  As shown in FIG. 32, when there are a plurality of photos, the optimum photo determination unit narrows down the shooting information by the shooting date and time (S131). Then, a shooting direction determination process is performed using these shooting information, and a photograph of the shooting direction determination process is transmitted (S133).

  Further, a shooting depression angle determination process is performed using shooting information narrowed down by shooting date and time (S134), and these are transmitted (S135). Also, the occlusion determination process is performed using the shooting information narrowed down by the shooting date and time (S136), and these are transmitted (S137).

  These photos are preferably sent on a set of thumbnail screens.

  Then, the user looks at the screen, determines the size that is easy to see, the angle that is easy to see, whether it is not hidden, and the like, and selects an optimal photo from these (S138 to S141).

  In steps S138 to S141, thumbnails are displayed on the screen of a set of a number of optimum photos for each shooting direction, a number of sets of optimum photos according to depression angles, and a number of sets of optimum photos based on an occlusion rate. It is preferable that the image is made the optimum photograph by the user.

<Embodiment 5>
In the fifth embodiment, a screen in which a photograph and a three-dimensional shape are superimposed is provided on a user terminal, and a photograph obtained as a result of shooting direction, depression angle determination, and occlusion processing from the image is provided.

  FIG. 33 is a flowchart for explaining the fifth embodiment.

  If the user wants to change the feature to be displayed or if the feature to be searched has not been determined, the optimal photo of any feature on the photo from the state in which the photo is first displayed in the photo display area The procedure for displaying is described.

  The user selects a desired feature from the image on the screen (S210). The corresponding shooting information is searched from the shooting information database 12 (S211).

  Next, regarding the shooting information Bip extracted by the shooting information search, the two-dimensional map information within the shooting range and the three-dimensional information included within the angle of view of the photograph are extracted from the feature information database 10 (S212).

  The two-dimensional map information 1 related to the photographing range of the selected photo can be displayed on the user terminal (two-dimensional map display unit).

  Next, from the shooting information Bip and the extracted three-dimensional information, a three-dimensional image of the city in the three-dimensional virtual space based on the elevation data of the ground surface and the three-dimensional shape data of the feature being photographed within the angle of view. A model is generated, and a two-dimensional virtual overhead view is generated as if the three-dimensional model was photographed under the same conditions as the orientation information of the photographing information.

  In this virtual overhead view, it is desirable to distinguish a portion of the feature within the angle of view into a two-dimensional area, and to associate the feature on the virtual overhead view with the feature information for each feature. Then, the virtual overhead view is superposed on the selected photo, and the overlay image is displayed on the user terminal 1 (photo display unit) (S213).

Next, when the user selects an arbitrary feature on the displayed overlay image with a pointing device such as a mouse (S215),
As can be imagined from the above-described generation process of the overlay image, the feature information associated with the coordinates on the image is searched from the feature information database 10 (S215). The searched feature information Aip can be displayed on the user terminal 1 (feature attribute information display section) (S216).

  Next, a new optimum photograph is retrieved along the optimum photograph retrieval procedure shown on the basis of the geographical coordinates of the feature information (S217).

  The search for the optimum photograph is performed by superimposing the optimum photograph and the corresponding virtual overhead view, and highlighting the target feature on the photograph by changing the color tone of the range of the target feature on the virtual overhead view. The feature emphasized image is displayed on the user terminal 1 (photo display unit) (S218).

  As a highlighting method, as shown in FIG. 34, for example, when the three-dimensional shape information of a base feature is provided in a polygon format, a wire frame connecting the vertices of the polygon is displayed in an arbitrary color ( I can imagine a method of superimposing colored images by adding a color or other processing such as semi-transparency, etc. There is no particular problem with this system.

  If a 2D map related to the shooting range of the initially selected photo is displayed (2D map display unit), the map is replaced with a map centered on the selected feature on the overlay image. It is also possible.

  The display of the optimum photo is to display the photo closest to the search condition (photo display unit). One or more photographs may be displayed. There is also a method of arranging and displaying the calculated indices in an order that is easy to see.

  However, when the number of photographs related to the target feature is one or less, the process of calculating the visibility index may be omitted.

  FIG. 34 schematically shows that the two-dimensional region segmented on the photograph in S213 and the feature information are related.

  If the photo used to search for the feature in the photo is the photo selected by the optimal photo search result, each feature on the photo obtained from the overlay result with the virtual overhead view including the peripheral features generated during the search Using the two-dimensional area information, a position on the photograph displayed on the photograph display unit, for example, by specifying coordinates with a pointing device such as a mouse and selecting an area including the coordinates, By highlighting each area in order by key operation and selecting the area, the feature information associated with the area is called, for example, displaying information on the feature attribute information display unit, You can easily imagine using it for processing.

However, in the case of an arbitrarily selected photograph, it is necessary to newly generate a virtual overhead view including the surrounding features and similarly to obtain two-dimensional area information of each feature on the photograph.

  If it is a newly selected feature from the feature information corresponding to the selected two-dimensional region information, unless it is already the target feature of the optimal photo in the previous operation, select that feature. It is also possible to execute the optimum photo search process using the set feature as the target feature and newly display the optimum photo on the photo display unit.

  In the above embodiment, the user terminal and the service center are connected to the network to provide the optimum photograph, but may be configured by a single personal computer. The functions of the servers and terminals shown in the drawings for explaining the embodiments of the present invention and the arrangement of data to be held are merely examples, and are not limited to realizing the functions of the present invention. Therefore, it goes without saying that various combinations of arrangements of processing means are conceivable within a range in which the functions of the present invention can be realized.

It is a schematic block diagram of the optimal photograph provision system of this Embodiment 1. It is explanatory drawing of the screen displayed on the user terminal of this Embodiment 1. FIG. It is explanatory drawing of the screen displayed on the user terminal of this Embodiment 1. FIG. It is explanatory drawing explaining an example of 2D map information and 3D information. It is explanatory drawing of the imaging | photography method of a diagonal photograph. It is explanatory drawing of an imaging | photography range. Photographs, photographing information, three-dimensional information, and two-dimensional map information are explanatory diagrams for linking. It is a sequence diagram until it obtains the optimal photograph of this Embodiment. It is explanatory drawing explaining the input information of a user terminal. It is explanatory drawing of search conditions. It is explanatory drawing of the map information displayed on a user terminal. It is a flowchart explaining operation | movement of the photographic information acquisition part including a feature. It is explanatory drawing of a common map. 4 is an explanatory diagram illustrating information stored in a memory 27. FIG. 4 is a flowchart for explaining the operation of a first optimum photograph determination processing unit 22. 4 is a flowchart for explaining the operation of a first optimum photograph determination processing unit 22. It is explanatory drawing of the method of obtaining the optimal photograph according to imaging | photography direction. It is explanatory drawing explaining the processing operation | movement of a 1st optimal photograph determination part in association with a memory. It is explanatory drawing explaining a superimposition process. It is explanatory drawing explaining the production | generation method of the determination boundary line in case the division shape of a target feature is only a point coordinate. FIG. 5 is a schematic configuration diagram of an optimum photograph providing system according to a second embodiment. It is explanatory drawing which illustrates the method of calculating | requiring a collimable surface typically. It is a flowchart of the method of calculating | requiring a collimable surface. It is explanatory drawing explaining the parameter | index of the visibility by the imaging depression angle using the collimable surface. It is a flowchart which calculates | requires the parameter | index of the visibility by the imaging depression angle using a collimable surface. 3 is an explanatory diagram for explaining information stored in a memory 36 and a memory 37. FIG. FIG. 6 is a schematic configuration diagram of an optimum photograph providing system according to a third embodiment. It is explanatory drawing explaining the calculation 7 of the visibility parameter | index by the occlusion determination using the collimable surface of the three-dimensional shape data of a feature. It is a flowchart for calculating | requiring an occlusion ratio. 4 is an explanatory diagram for explaining information in a memory 56 and a memory 57. FIG. It is explanatory drawing explaining an emphasis process. It is a schematic flowchart explaining a 3rd search service. 10 is a flowchart for explaining the fourth embodiment. 10 is a flowchart for explaining the fifth embodiment. FIG. 10 is an explanatory diagram of building emphasis according to the fifth embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 User terminal 2 Service provision center 14 User determination / condition input screen provision part 16 Map provision / designation | designated feature reading part 18 Feature including photographic information acquisition part 20 Common 2D map information creation part 22 1st optimal photograph determination part 24 Provision Department

Claims (14)

The shooting point in which the target target feature is photographed in contact with the designated feature among a plurality of oblique photographs in which the feature is photographed from the oblique direction with the camera for each photographing point (Pi) with different flight paths. (Pi) An optimal oblique photograph providing method for providing an oblique photograph for each as an optimum oblique photograph,
First storage means for storing a planar map in which a planar shape of a feature including the target feature is defined, and attribute information for the feature in the planar map;
Wherein for each shot point (Pi), shooting position and the shooting point (Pi) each of the planar imaging range and their photographic point of the camera in the map camera focal length and the orientation and the camera of the CCD of the (Pi) of Second storage means for storing shooting information including an angle of view;
A third storage means for storing an oblique photograph of the camera for each shooting point (Pi) in association with the shooting information of the shooting point (Pi);
Computer
Reading the attribute information of the target feature and extracting all of the shooting information of the shooting point (Pi) having the shooting range including the target feature of the attribute information from the second storage means;
Defining a predetermined range in which the target feature is shown in common based on each of the shooting ranges included in the extracted shooting information as a common map in a fourth storage unit;
Defining the shooting position and the angle of view included in each of the extracted shooting information in the common map;
Obtaining, as a judgment boundary line (EFGH), a line where the designated feature and the target feature existing on the common map are in contact with each photographing position of the photographing point (Pi);
For each shooting position of the shooting point (Pi), a straight line is drawn from the shooting point (Pi) to the target feature, and it never passes through the inside of the target feature. Obtaining a set of points intersecting the object as collimable lines (EF) on the determination boundary line (EFGH) facing the photographing position of the photographing point (Pi);
For each shooting position of the shooting point (Pi), the width of the angle of view of the shooting point (Pi) is a projection plane on which the target feature is projected on the CCD, and the collimable line (EF) is Projecting onto the projection plane and determining the projected line as a collimable line projection plane width (wcl);
For each imaging position of the imaging point (Pi) , the value of the ratio of the collimable line projection plane width (wcl) to the projection plane width (WLl) is set as a collimable line projection ratio Wi (Wi = Wcl / ( WLl)) as a step
For each shooting position of the shooting point (Pi), the value of the ratio of the length of the collimable line (EF) of the shooting position of the shooting point (Pi) to the length of the determination boundary line (EFGH) is determined as a determination boundary. Obtaining the collimable line ratio (Li);
Comparing each of the determination boundary collimable line ratios ( Li ) for each photographing position of the photographing point (Pi ) to obtain a determination boundary collimable line ratio ( Li ′ ) having the largest value;
Comparing the collimable line projection ratios (Wi) to obtain the largest collimable line projection ratio (Wi ′);
The imaging information including the imaging position of the imaging point (Pi) where the collimable line projection ratio (Wi ′) and the determination boundary collimable line ratio (Li ′) having the largest values are obtained . Read from the storage means of the second, the oblique photograph of the third storage means associated with the shooting position of the shooting point (Pi) included in this shooting information from the shooting direction of the target feature of interest And providing an optimum oblique photograph. The method of providing an optimum oblique photograph. Said first memory hand stage,
The attribute information for the feature is stored in association with the three-dimensional shape data of the feature,
The computer is
The three-dimensional shape data corresponding to the target feature of the optimum oblique photograph is read based on the attribute information associated with the target feature, and three-dimensional shape data with contour enhancement is generated, 2. The method for providing an optimum oblique photograph according to claim 1, wherein the three-dimensional shape data is superimposed on the photograph data of the target feature in the optimum oblique photograph and output. 3. The optimal diagonal according to claim 2, further comprising a step of outputting an image obtained by combining a collection of three-dimensional shape data other than the target feature in the optimal diagonal photograph and the photograph data of the target feature. Photo offer method. The computer is
When the target feature does not exist in the common map, a circle in a predetermined range starting from the geographic coordinates included in the attribute information of the target feature is defined in the common map, and the designation in the circle The optimal oblique photograph providing method according to claim 1, wherein a line of a feature is used as the determination boundary line. Preparing a virtual overhead view for each shooting position of the shooting point (Pi), a fifth storage means for generating a three-dimensional model;
The computer is
Said common the photographing points defined in the map (Pi) three-dimensional contained in the attribute information associated with the feature of each of the shooting information and the common map associated with the shooting position location of The stereo model is generated based on the coordinates and the three-dimensional shape data included in the attribute information of each of these features, and a virtual overhead view for each shooting position of the shooting point (Pi) based on the stereo model. Generating a diagram in the fifth storage means;
Reading three-dimensional shape data of the target feature in the three-dimensional model, and determining the determination boundary line (EFGH) of the target feature with respect to the designated feature based on the three-dimensional shape data;
Each time the determination boundary line (EFGH) of the three-dimensional shape data is obtained, a predetermined position that faces the determination boundary line (EFGH) and is outside the target feature is obtained, and is used as a virtual collimation point. Defining the solid model as (Mi);
Every time the virtual collimating points (Mi) is defined, three-dimensional shape data of said object the said target area product foreseeable three-dimensional shape data of a feature in the three-dimensional model from the said virtual collimating points (Mi) Are defined as the first collimable surface (Rpi) by the photographing depression angle at the photographing position of the photographing point (Pi) , and the total area (Sc) of these first collimable surfaces (Rpi) is defined as the first collimable surface (Rpi). And obtaining a total area (Sn) by setting a surface other than the first collimable surface (Rpi) as a first collimable surface (Rqi) based on a shooting depression angle;
For each shooting position of the shooting point (Pi), a second collimable surface (Rpi ′) and a second collimation failure of the target feature in the virtual overhead view of the shooting position of the shooting point (Pi). Determining a possible surface (Rqi ′);
For each photographing position of the photographing point (Pi) , the total area (Ac) of the second collimable surface (Rpi ′) and the second non-collimable surface (Rqi ′) in the virtual overhead view. Determining a total area (An);
The total area (Sc) of the first collimable surface (Rpi) of the three-dimensional shape data of the target feature in the three-dimensional model and the second collimation failure of the target feature in the virtual overhead view. The first product (Sc · An) with the total area (An) of the possible surface (Rqi ′) is obtained, and the first collimation of the three-dimensional shape data of the target feature in the three-dimensional model is not possible. Second product (Sn · Ac) of the total area (Sn) of the surface (Rqi) and the total area (Ac) of the second collimable surface (Rpi ′) of the target feature in the virtual overhead view A step of seeking
For each shooting position of the shooting point (Pi) , a ratio value {(Sc) of the first product (Sc · An) of the shooting position of the shooting point (Pi) to the second product (Sn · Ac). · an) / (Sn · Ac ) seek}, the virtual overhead view of the imaging position location of the shooting point shows the smallest value the value of the ratio (Pi), from the most easy-to-see the imaging point (Pi) Outputting as a virtual overhead view based on an oblique photograph;
The method for providing an optimal oblique photograph according to claim 1, wherein: Computer
In the virtual bird's-eye view for each shooting position of the shooting point (Pi) , for each virtual bird's-eye view, the second collimation of the target feature at a portion where the target feature is not hidden by another feature is possible. Sequentially calculating the total area (Ac ′) of the area of the surface (Rpi ′);
For each virtual overhead view, the total area (Ac ′) of the total area of the second collimable surface (Rpi ′) and the second collimable surface (Rpi ′) when there is no feature other than the target feature. ) For the ratio (Ac ′ / Ac) of the ratio of the total area (Ac) to the total sum (Ac),
According to claim 5, characterized in that the step of outputting the virtual overhead view of the imaging position of the imaging point the occlusion rate was obtained a larger value (Pi) as the optimal virtual overhead view by occlusion rate Optimal oblique photo providing method. The computer is
2. The method for providing an optimum oblique photograph according to claim 1, wherein the oblique photograph for each photographing position of the photographing point (Pi) is output as a thumbnail. A user terminal and a server of the optimum photo providing service center are connected by a communication network, and the server has a plurality of oblique images obtained by photographing the feature from the oblique direction with a camera at each photographing point (Pi) with different flight paths. An optimum oblique photograph providing system for transmitting an oblique photograph at each photographing point (Pi) where a target target feature is photographed in contact with a specified feature among the photographs as an optimum photograph to the user terminal. ,
The server
First storage means for storing a planar map in which a planar shape of a feature including the target feature is defined, and attribute information for the feature in the planar map;
Wherein for each shot point (Pi), shooting position and the shooting point (Pi) each of the planar imaging range and their photographic point of the camera in the map camera focal length and the orientation and the camera of the CCD of the (Pi) of Second storage means for storing shooting information including an angle of view;
Third storage means for storing an oblique photograph of the camera for each photographing point (Pi) in association with the photographing information of the photographing point (Pi);
Means for storing user information from the user terminal and the search condition for the oblique photograph in a user information search condition storage means;
Means for transmitting, via the communication network, input screen information for allowing the user terminal to input search conditions for the user information and the oblique photograph;
When the attribute information of the target feature is input from the user terminal, the attribute information of the target feature is read, and the shooting point (Pi having the shooting range including the target feature of the attribute information) Means for extracting all of the shooting information from the second storage means;
Means for defining, in a fourth storage means, a predetermined range in which the target feature is shown in common based on each of the shooting ranges included in the extracted shooting information;
Means for defining the shooting position of the shooting point (Pi) included in each of the extracted shooting information and the width of the angle of view in the common map;
Means for obtaining, as a determination boundary line (EFGH), a line where the designated feature and the target feature existing on the common map are in contact with each photographing position of the photographing point (Pi);
For each shooting position of the shooting point (Pi) , a straight line is drawn with respect to the target feature from the shooting point (Pi) of the shooting point (Pi), and never passes through the inside of the target feature. Means for first obtaining a set of points intersecting the target feature as collimable lines (EF) on the determination boundary line (EFGH) facing the photographing position of the photographing point (Pi);
For each shooting position of the shooting point (Pi) , the width of the angle of view at the shooting position of the shooting point (Pi) is a projection plane on which the target feature is projected on the CCD, and the collimable line ( EF) is projected onto the projection plane, and the projected line is obtained as a collimable line projection plane width (wcl);
For each imaging position of the imaging point (Pi) , the value of the ratio of the collimable line projection plane width (wcl) to the projection plane width (WLl) is set as a collimable line projection ratio Wi (Wi = Wcl / ( WLl)) means to obtain
For each photographing position of the photographing point (Pi), a value of a ratio of the length of the collimable line (EF) of the photographing point (Pi) to the length of the judgment boundary line (EFGH) is determined as a judgment boundary collimable line. Means for obtaining the ratio (Li);
Means for comparing each of the determination boundary collimable line ratios ( Li ) for each photographing position of the photographing point (Pi ) to obtain a determination boundary collimable line ratio ( Li ′ ) having the largest value ;
Means for comparing the collimable line projection ratios (Wi) to obtain the largest collimable line projection ratio (Wi ′);
The imaging information having the imaging position of the imaging point (Pi) where the collimable line projection ratio (Wi ′) and the determination boundary collimable line ratio (L1 ′) having the largest values are obtained . Read from the storage means of the second, the oblique photograph of the third storage means associated with the shooting position of the shooting point (Pi) included in this shooting information from the shooting direction of the target feature of interest Means to make an optimal oblique photo;
Means for transmitting the optimum oblique photograph to the user terminal via the communication network,
The user terminal is
Means for receiving the input screen information from the server, and transmitting the attribute information of the target target feature and the designated feature input to the input screen information to the server as the search condition;
An optimum oblique photograph providing system comprising: means for receiving an optimum oblique photograph from the server and displaying the optimum oblique photograph on a screen. The first storage means is
The attribute information for the feature is stored in association with the three-dimensional shape data of the feature,
The server
The three-dimensional shape data corresponding to the target feature of the optimum oblique photograph is read based on the attribute information associated with the target feature, and three-dimensional shape data with contour enhancement is generated, 9. The optimum oblique photograph providing system according to claim 8, further comprising means for superimposing and outputting the three-dimensional shape data on photograph data of the target feature in the optimum oblique photograph. The server
10. The optimal oblique according to claim 9, further comprising means for outputting an image obtained by combining a collection of three-dimensional shape data other than the target feature in the optimal oblique photograph and the photograph data of the target feature. Photo offer system. The server
When the target feature does not exist in the common map, a circle in a predetermined range starting from the geographic coordinates included in the attribute information of the target feature is defined in the common map, and the designation in the circle 9. The optimum oblique photograph providing system according to claim 8, further comprising means for setting a feature line as the determination boundary line. The server
A virtual overhead view for each shooting position of the shooting point (Pi), and a fifth storage means for generating a three-dimensional model,
The server
Said common the photographing points defined in the map (Pi) three-dimensional contained in the attribute information associated with the feature of each of the shooting information and the common map associated with the shooting position location of coordinates as well as to generate the three-dimensional model based on three-dimensional shape data included in the attribute information of each of these features, the virtual shooting position 置毎 of the photographing point based on the three-dimensional model (Pi) Means for generating an overhead view in the fifth storage means;
Means for reading three-dimensional shape data of the target feature in the three-dimensional model, and determining the determination boundary line (EFGH) of the target feature with respect to the designated feature based on the three-dimensional shape data;
Each time the determination boundary line (EFGH) of the three-dimensional shape data is obtained, a predetermined position that faces the determination boundary line (EFGH) and is outside the target feature is obtained, and is used as a virtual collimation point. Means for defining the solid model as (Mi);
Every time the virtual collimating points (Mi) is defined, three-dimensional shape data of said object the said target area product foreseeable three-dimensional shape data of a feature in the three-dimensional model from the said virtual collimating points (Mi) Are defined as the first collimable surface (Rpi) by the photographing depression angle at the photographing position of the photographing point (Pi) , and the total area (Sc) of these first collimable surfaces (Rpi) is defined as the first collimable surface (Rpi). Means for determining the total area (Sn) by determining a surface other than the first collimable surface (Rpi) as a first collimable surface (Rqi) by an imaging depression angle;
For each shooting position of the shooting point (Pi), a second collimable surface (Rpi ′) and a second collimation failure of the target feature in the virtual overhead view of the shooting position of the shooting point (Pi) . Means for determining the possible surface (Rqi ′);
For each photographing position of the photographing point (Pi) , the total area (Ac) of the second collimable surface (Rpi ′) and the second non-collimable surface (Rqi ′) in the virtual overhead view. Means for determining the total area (An);
The total area (Sc) of the first collimable surface (Rpi) of the three-dimensional shape data of the target feature in the three-dimensional model and the second collimation failure of the target feature in the virtual overhead view. The first product (Sc · An) with the total area (An) of the possible surface (Rqi ′) is obtained, and the first collimation of the three-dimensional shape data of the target feature in the three-dimensional model is not possible. Second product (Sn · Ac) of the total area (Sn) of the surface (Rqi) and the total area (Ac) of the second collimable surface (Rpi ′) of the target feature in the virtual overhead view A means of seeking
For each shooting position of the shooting point (Pi) , a ratio value {(Sc) of the first product (Sc · An) of the shooting position of the shooting point (Pi) to the second product (Sn · Ac). · an) / (Sn · Ac ) seek}, the virtual overhead view of the imaging position location of the shooting point shows the smallest value the value of the ratio (Pi), from the most easy-to-see the imaging point (Pi) The optimal oblique photograph providing system according to any one of claims 8 to 11, further comprising means for transmitting to the user terminal via the communication network as a virtual overhead view based on the oblique photograph. The server
In virtual overhead view of each imaging position of the imaging point (Pi), each said virtual overhead view, the second collimation before Symbol target feature portion being the target feature is not hidden another feature Means for sequentially obtaining the sum (Ac ′) of the areas of the possible surfaces (Rpi ′);
For each virtual overhead view, the total area (Ac ′) of the total area of the second collimable surface (Rpi ′) and the second collimable surface (Rpi ′) when there is no feature other than the target feature. ) With respect to the total sum (Ac) of the areas (Ac ′ / Ac), respectively, as means for determining the occlusion rate,
Means for transmitting the virtual overhead view of the photographing position of the photographing point (Pi) having a large value of the occlusion rate to the user terminal as the optimal virtual overhead view based on the occlusion rate via the communication network. The optimal oblique photograph providing system according to claim 12 . The server
9. The optimum oblique photograph providing system according to claim 8, further comprising means for converting an oblique photograph of each photographing position of the photographing point (Pi) into a thumbnail and transmitting the thumbnail to the user terminal via the communication network.