JP2012137933A - Position specifying method of planimetric features to be photographed, program thereof, display map, photographic position acquiring method, program thereof and photographic position acquiring device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method grasping a position of planimetric features, e.g., a disaster-stricken position, and a shape of the planimetric features, e.g., a disaster scale, without causing a staff such as a surveying engineer to enter a disaster-stricken area and a program thereof, and further to provide a method capable of specifying a photographic position by a photographed photo image, a device and a program thereof.SOLUTION: A position specifying method of planimetric features to be photographed according to the present invention comprises the steps of: selecting four or more feature points from a shape of planimetric features to be photographed while providing the feature points with coordinates; selecting four or more points corresponding to the feature points from a model shape, on which a land surface model is projected, while providing the points corresponding thereto with coordinates; associating any coordinate system in a plane of a photo image with any coordinate system in the plane representing the model shape by causing the feature points of the planimetric features to be photographed to correspond to the points corresponding to the feature points of the model shape; providing the planimetric features to be photographed with space information based on a numerical altitude model; and specifying a position of the planimetric features to be photographed or/and the shape thereof.

Description

  The present invention relates to a method and program for specifying the position of an object to be photographed in a photograph taken, a display map, a method for identifying a spot where a photograph is taken, a program and an apparatus therefor.

  Japan is known as a country where earthquakes occur frequently. In recent years, large earthquakes such as the Hyogoken-Nanbu Earthquake and the Niigata Chuetsu Earthquake have occurred, and each time they are suffering enormous damage. In addition, damage caused by typhoons occurs almost every year, and recently, various types of damage can be caused by sudden heavy rain. These natural disasters often cause slope failures and in some cases can trigger landslides.

  Slope failures and landslides (hereinafter referred to as “slope failures, etc.”) damage villages, roads, rivers, and sabo facilities directly below them. It has an impact on the environment. When villages or roads are damaged by disasters such as slope failures, measures are taken promptly to restore these functions.

  In order to implement countermeasures planning and design, it is necessary to ascertain as accurately as possible the location and scale of damage (such as collapsed shape and size) due to slope failure. On the other hand, slope failures, etc. may also be a secondary disaster, and it is not possible to approach the disaster area immediately after the disaster. Conventionally, after waiting for safety confirmation, a survey engineer enters a disaster-stricken area with a surveying device such as a total station, and measures the location and scale of the disaster. In some cases, such as when the scale of the disaster is large or the social impact of the damage is large, the location of the disaster and the scale of the damage may be ascertained using stereo photography by aircraft, but the flight for photography depends on the weather. In some cases, it lacks mobility and costs and labor.

  As described above, it is necessary to take quick recovery measures for disasters such as slope failures, and for that purpose, the location and scale of damage must be as quickly as possible. However, the measurement method using surveying equipment such as a total station requires a certain period of time for measurement to confirm the safety of the disaster-stricken area, lacks speed, and takes the harsh labor to transport the equipment and the risk of entering the disaster-stricken area. Requested a surveyor. In addition, the method using stereo photography by aircraft is not very mobile, such as being affected by the weather, is laborious and expensive, and requires measurement points around the disaster-stricken area. It takes a period for confirmation and lacks quickness.

  As described above, the conventional methods for determining the location and scale of the disaster have various problems, so surveying engineers and other people do not enter the disaster-stricken area, and promptly after a disaster such as a slope failure occurs, There was a long-awaited approach to grasp the scale. The present invention has been made to meet this expectation and uses a photograph taken with a digital camera or the like and a digital elevation model.

  Although no technology has been proposed to identify the location of a disaster-stricken area using a photograph and a digital elevation model, it has been partially used in the present invention. Several technologies have been proposed.

International Publication WO2006 / 134962 JP 05-46080

  The projection map generation system disclosed in Patent Document 1 is used for car navigation and the like, displays a two-dimensional projection plane when a three-dimensional terrain model is viewed from a certain viewpoint, and hides a destination road by a shielding object. In some cases, the shield is removed and displayed.

  The perspective view creation apparatus of Patent Document 2 displays a projection plane obtained by viewing three-dimensional digital position data from an arbitrary viewpoint as a perspective view, and the viewpoint position moves at a high speed when mounted on a sailing airplane. However, processing corresponding to the speed is performed and a perspective view is displayed in real time.

As in Patent Document 1 and Patent Document 2, a projection plane (or a perspective plane, hereinafter referred to as “projection plane”) obtained by viewing a digital elevation model (three-dimensional terrain model) from an arbitrary viewpoint is created. The technology using the surface has been conventionally used in various fields. The present invention also utilizes the projection plane from this arbitrary viewpoint, and is an invention made by paying attention to the fact that the position of the feature in the photograph can be specified by comparing with the photographed photograph.

  Further, by collating the projection surface with the feature in the photograph, the photographing position that is the viewpoint position of the projection surface can be specified. If you want to check where a photograph taken in the past was taken, or if you want to know the current position in an unfamiliar place such as a sightseeing spot, specify the shooting position by taking a picture or by taking a picture No technology that could be proposed has been proposed before.

An object of the present invention is a method for grasping a feature position (for example, a damaged position) and a feature shape (for example, a damaged scale) without a person such as a surveying engineer entering the damaged area, a program for the method, and a map displaying the position And a method and apparatus capable of specifying a photo shooting position based on a photograph taken, and a program therefor.

  The object specifying method of the present invention is a method for specifying the position or / and the shape of an object photographed in a photograph based on a number of elevation models, Four or more feature points are selected from the shape, and coordinates are assigned to these feature points in an arbitrary coordinate system in the photographic plane, and the feature points are derived from a model shape obtained by projecting a ground surface model based on the digital elevation model. 4 or more points are selected, and the coordinates corresponding to the feature points are given coordinates in an arbitrary coordinate system in the plane representing the model shape, By associating the points corresponding to the feature points of the model shape, the arbitrary coordinate system in the photographic plane is associated with the arbitrary coordinate system in the plane representing the model shape. Assign spatial information to the feature and A method of locating and / or shape of the feature.

  The method for specifying the position of the object to be photographed according to the present invention is a method for specifying the position or / and shape of the object to be photographed in the photograph based on the digital elevation model, wherein A two-dimensional model shape is created by creating a two-dimensional reference shape based on a part or all of the shape, assigning coordinates to the reference shape using an arbitrary coordinate system in the photograph plane, and projecting a ground surface model based on the digital elevation model And a coordinate is given to the model shape by an arbitrary coordinate system in the plane representing the model shape, and the component of the reference shape and the component of the model shape are made to correspond to each other in the photographic plane. An arbitrary coordinate system is associated with an arbitrary coordinate system in a plane representing a model shape, spatial information is given to the object based on the digital elevation model, and the position of the object or / and Those who specify the shape It can also be a.

  The method for specifying the position of the object to be photographed according to the present invention is a method for specifying the position or / and shape of the object to be photographed in the photograph based on the digital elevation model, wherein A ground surface model based on a digital elevation model to which a two-dimensional reference shape is created based on a part or all of an image, and coordinates are assigned to pixels constituting the reference shape by an arbitrary coordinate system in a photographic plane. A two-dimensional model shape is created from the projection image obtained by projecting the image, and coordinates are given to the pixels constituting the model shape by an arbitrary coordinate system in a plane representing the model shape, and the reference shape pixel and the model By associating the shape pixels with each other, the arbitrary coordinate system in the photographic plane is associated with the arbitrary coordinate system in the plane representing the model shape, and spatial information is assigned to the object to be imaged based on the digital elevation model. Grant, It may be a method of locating and / or shape of 該被 copy feature.

  The object specifying method of the present invention is the object specifying method according to claim 2 or 3, wherein a plurality of ground surfaces are changed by changing a viewpoint and a viewing angle for projecting the ground model. A shape obtained by projecting a model is created as a model shape candidate, and by comparing the model shape candidate with a reference shape, a model shape corresponding to the reference shape can be selected from a plurality of model shape candidates.

  The display map of the object according to the present invention is a map on which the position or / and shape of the object captured in the photograph is displayed based on the digital elevation model. The object in which the position or / and the shape is specified by the described object specifying method is displayed.

  The object location specifying program of the present invention is a program for specifying the position or / and shape of the object to be photographed on a photograph based on a digital elevation model. An electronic computer (hereinafter referred to as “computer”) executes the method for specifying the position of the object to be described.

  The shooting position acquisition method of the present invention is a method for acquiring a point where a photographed object was photographed based on a photographed object and a digital elevation model, A viewpoint and a viewing angle for creating a two-dimensional reference shape based on a part or all of the shape, assigning coordinates to the reference shape by an arbitrary coordinate system in a photographic plane, and projecting a ground surface model based on the digital elevation model To create a two-dimensional shape projecting a plurality of ground models as model shape candidates, and assign coordinates to the model shape candidates by an arbitrary coordinate system in a plane representing these model shape candidates. By collating the candidate with the reference shape, a model shape corresponding to the reference shape is extracted from among a plurality of model shape candidates and corresponds to the extracted model shape. The point is a method for the shooting location of the copy feature.

  The shooting position acquisition method of the present invention is a method for acquiring a point where a photographed object was photographed based on a photographed object and a digital elevation model, In order to create a two-dimensional reference shape based on a part or all of the image, assign coordinates to pixels constituting the reference shape by an arbitrary coordinate system in the photographic plane, and project a ground surface model based on the digital elevation model A plurality of two-dimensional shapes are created as model shape candidates from projection images obtained by projecting a plurality of ground models by changing the viewpoint and viewing angle of the model, and the model shape is determined by an arbitrary coordinate system in the plane representing these model shape candidates. A coordinate corresponding to the reference shape is selected from among a plurality of model shape candidates by assigning coordinates to the pixels constituting the model and collating the model shape candidate with the reference shape. Extracted, a viewpoint corresponding to the extracted model shape it may be a method for the imaging location of the object scene feature.

  The shooting position acquisition program according to the present invention is a program for acquiring position information of a point where a photographed object is photographed based on a photographed object and a digital elevation model. According to another aspect of the present invention, there is provided a computer for executing the photographing position acquisition method.

  The subject position photographing position acquisition device of the present invention is a device for obtaining position information of a point where a subject feature is photographed based on a subject feature photographed in a photograph and a digital elevation model. The means for photographing the object to be photographed and acquiring the image data, the means for creating the reference shape according to claim 7 or 8 from the image data, and the plurality according to claim 7 or claim 8 Means for generating a model shape candidate, means for collating the model shape with a reference shape to extract a model shape corresponding to the reference shape from a plurality of model shapes, and the extracted model shape And a means for setting the viewpoint corresponding to the shooting point of the object to be photographed.

The method for specifying the position of the object to be imaged according to the present invention, the program thereof, and the display map have the following effects.
(1) Only the shooting of the object to be performed is performed locally, and the position and shape of the object to be captured can be grasped extremely easily.
(2) For example, when grasping the position and shape of slope failure, etc., it is safe because people do not enter the disaster area, and it is possible to grasp the position and shape immediately after the disaster, so it is necessary to implement countermeasures promptly it can.

The imaging position acquisition method, the program thereof, and the imaging position acquisition apparatus of the present invention have the following effects.
(1) Since the shooting location of a photo taken in the past can be specified, the information value of the photo increases.
(2) Since the current position can be specified only by taking a picture, it is possible to accurately move to a destination even in an unguided place such as a sightseeing spot.

Explanatory drawing explaining aerospace laser measurement. Explanatory drawing for demonstrating the condition which has image | photographed the collapse of a slope. (A) is the model figure which looked at "the situation which is photographing the to-be-photographed object" from the side, (b) is explanatory drawing which shows the image projected on the camera image surface. (A) is explanatory drawing which shows the condition which projected the digital elevation model on the model image surface, (b) is explanatory drawing which shows the image projected on the model image surface. The execution flowchart in the case of making a reference | standard shape and a model shape respond | correspond by two-dimensional projective transformation. (A) is a model diagram for explaining a reference shape extracted from an object, (b) is a model diagram for explaining a model shape created based on a digital elevation model, and (c) is a reference shape and model shape. The model figure explaining the state which collated. Explanatory drawing which shows the concept which produces a model shape. Explanatory drawing which shows the concept which aims at a response | compatibility with the composing point of a reference | standard shape, and the composing point of a model shape. The execution flowchart in the case of giving spatial information with respect to a reference | standard shape, without performing collation judgment. Explanatory drawing which shows arrangement | positioning of the apparatus which comprises the "positioning system of a to-be-photographed object". The implementation flowchart in the case of giving spatial information with respect to a reference | standard shape after performing collation judgment. The implementation flowchart in the case of making a reference | standard shape and a model shape based on an image, and accompanying a collation judgment. Explanatory drawing which shows the condition which is acquiring the position which image | photographed using the imaging position acquisition apparatus. The implementation flowchart in the case of acquiring the imaging | photography position information of a photograph. (A) shows the function which comprises a "photographing position acquisition apparatus", and is explanatory drawing at the time of integrating all the functions, (b) shows the function which comprises a "photographing position acquisition apparatus". Explanatory drawing when function is distributed.

  An example of an embodiment of the object location specifying method and program thereof, a display map, a shooting position acquisition method and program thereof, and a shooting position acquisition apparatus according to the present invention will be described with reference to the drawings.

  Recently, with the advent of aviation laser measurement and the like, measurement techniques for acquiring a large amount of terrain information have been remarkably advanced, and the technology for handling terrain information has progressed dramatically with the evolution of computers. As shown in FIG. 1, the aviation laser measurement is performed by flying the aircraft P over the terrain Gr to be measured and receiving the reflection of the laser L irradiated to the terrain Gr during the flight. According to this aerial laser measurement, countless three-dimensional point cloud data (X, Y, Z) can be obtained densely as information representing the topography, and if this point cloud data is processed by a computer, the topography Can be modeled three-dimensionally.

(Digital elevation model M)
A model constituted by the three-dimensional point cloud data is called “numerical elevation model M” (FIGS. 4A and 4B), and representative examples include DEM (Digital Elevation Model) and DSM (Digital Surface Model). ). Of course, the present invention is not limited to this, and various digital elevation models M conventionally used can be used. In carrying out the present invention, it is desirable to use an existing digital elevation model M, but if there is no existing one, a new digital elevation model M can be created and used.

  Here, a DEM that is a typical digital elevation model M will be briefly described. The DEM is a numerical model of the topography, which is the shape of the ground surface, and is generally a lattice model. A DEM is formed based on so-called point cloud data, which is a set of points having a plane coordinate (X, Y) and an elevation value (Z) on the ground surface. The denser the point cloud data, the more accurately the original terrain is reproduced. can do. This point cloud data is generally acquired by aviation laser measurement. Of course, in addition to the aerial laser measurement, point cloud data with three-dimensional spatial information may be generated based on stereo aerial photographs and satellite photographs, or the point of having direct three-dimensional spatial information by surveying the site. Group data may be acquired. When acquiring point cloud data by aerial laser measurement, it is common to perform so-called filtering processing that removes measurement data that is not on the ground surface, such as the top of a tree, to obtain point cloud data on the ground surface.

  The point cloud data acquired by the aviation laser measurement is merely a set of laser measurement points measured at random, and the DEM is created by the following procedure. That is, a plurality of grids (for example, square lattices) arranged at predetermined intervals (for example, 2 m) are placed on the laser measurement points. By being divided by this square lattice, lattice points are generated, and a large number of quadrangles (mesh) are formed. One representative point is provided in the mesh, and the position of the representative point is set at the center of the mesh, or a lattice point at the upper right corner of the mesh, and is appropriately set according to the situation.

  Based on the three-dimensional coordinates (X, Y, Z) of the laser measurement points, the plane coordinates (X, Y) and elevation values (Z) of the mesh representative points are calculated, and the DEM is completed. This calculation method includes an interpolation method based on TIN (Triangulated Irregular Network) that obtains the height from an irregular triangle network from a laser measurement point, a nearest neighbor method that employs the nearest laser measurement point (Nearest Neibor), and an inverse distance weighting method. Various methods such as (IWD), Kriging method, and average method are adopted.

  As for the digital elevation model M, development of the digital elevation model M in the jurisdiction area is being promoted in each local government including the country. Some private businesses have a digital elevation model M in the form of a library. In recent years, the digital elevation model M has become familiar and easy to use. The present invention is based on the fact that the digital elevation model M has been accumulated throughout the country as described above, and is described as “the shape of an object photographed in a photograph and a model calculated from the digital elevation model M”. The necessary spatial information is acquired by collating with the shape (two-dimensional) ”. Various embodiments described below have this technical feature in common, but for the sake of convenience, the implementation of the “positioning method of the object and its program, and the display map of the object” will be described. The embodiment and the “shooting position acquisition method, program thereof, and shooting position acquisition apparatus” will be described separately.

[Embodiment of Method for Specifying Location of Object and its Program, and Display Map of Object)
The present embodiment will be described with respect to “the method for specifying the position of the object to be photographed, its program, and the display map of the object to be photographed” of the present invention. In addition, although "the thing reflected in the photograph" is demonstrated here as features, such as the ridgeline A of a mountain (henceforth "skyline A") and the collapse part B (FIG. 2), this is only description. It is needless to say that the present invention can be implemented even if the “object photographed” is any other object (feature).

FIG. 2 is an explanatory diagram for explaining a situation where a slope collapse is imaged. As shown in this figure, the slope of this mountain has been damaged due to collapse, and the collapsed portion B can be visually confirmed. A photograph 2 taken by a photographer standing at a position slightly away from the disaster area (site where the slope collapsed) was photograph 2, and photograph 2 contains features including skyline A and collapsed part B. ing. Here, the skyline A captured in the photograph 2 is referred to as the “photographed skyline 3a”, and the collapsed portion B captured in the photograph 2 is referred to as the “photographed collapsed portion 3b”. The generic term of the features copied in the photograph 2 such as “the subject collapsed portion 3b” is referred to as “the subject feature 3”. The photograph 2 is not limited to a photograph taken by a person from the ground, but may be a photograph taken while moving with a moving body such as an aircraft.

(Embodiment 1)
Here, the first embodiment of “the method for specifying the position of the object to be imaged, its program, and the display map of the object to be imaged” will be described. In this embodiment, a reference shape is created from the object 3 in the photographic image, and a model shape is created based on the digital elevation model M separately. A method of extracting the points and matching the reference shape with the model shape by associating the feature points with the constituent points of the model shape will be described. In other words, the feature of the present invention is that coordinates can be given to the feature by associating one photo 2 (single photo) with one model shape 5, and the coordinates are given using a set of two photos. This point is significantly different from the conventional method (stereo matching method).

(Standard shape)
FIG. 3A is a model view of the “situation where the object 3 is being photographed” shown in FIG. 2 as viewed from the side, and shows a situation where the camera 1 has copied an image from the front of the mountain. It is explanatory drawing. Further, as shown in FIG. 2, assuming that the coordinate axes of the real space are the X axis, the Y axis, and the Z axis, FIG. 3A is a cross section taken along the YZ plane. As shown in FIG. 3A, a mountain that is an object is photographed by the camera 1, and the photographed image is projected on the camera image plane Ph. The coordinates of the center position of the camera 1 (hereinafter referred to as “viewpoint”) are represented by (X ₀ , Y _0, Z ₀ ) in the real space coordinate system, and the tilt of the camera 1 (hereinafter referred to as “photographing posture”). )), The inclination around the X axis is ω (pitch), the inclination around the Y axis is φ (roll), and the inclination around the Z axis is κ (yaw). Note that (X ₀ , Y _0, Z ₀ ) and (ω, φ, κ) that determine the spatial arrangement of the camera 1 are called external orientation elements.

  FIG. 3B is an image shown on the camera image plane Ph shown in FIG. The reference shape 4 is extracted from the features (things) in the image. In this case, the reference shape 4 may be created based on all the object features 3 stored in the photograph 2, but in order to correspond to the model shape 5 (FIG. 6B) described later, It is also possible to create the reference shape 4 by extracting features that are easily associated with the model shape 5 (that is, can be expressed by the digital elevation model M). In the present embodiment, the photographed skyline 3 a is extracted as the reference shape 4 from the photographed object 3 photographed in the photograph 2.

The reference shape 4 can be drawn to trace the imaged skyline 3a using a drawing application such as CAD. Alternatively, the image data of Photo 2 is used, the skyline A is automatically generated by a computer as an edge from the difference in brightness (or hue, saturation, and brightness) of the sky and mountains in the image data, and this edge is used. A reference shape 4 can also be created. The reference shape 4 created in this way is a two-dimensional shape, and geometric elements (points, line segments, surfaces, etc.) constituting this shape are arbitrary in the camera image plane Ph (Photo 2). It can be represented by the coordinates of the coordinate system (x _a , y _a ). Hereinafter, the arbitrary coordinate system in the camera image plane Ph is referred to as a photographic coordinate system. Further, four or more constituent points characterizing the shape are selected from the points constituting the reference shape 4 and set as feature points. In FIG. 3B, five points of feature point A ₁ to feature point A ₅ are extracted.

(Model shape)
FIG. 4A is an explanatory diagram illustrating a situation in which the digital elevation model M is projected onto the model image plane G. This figure shows a cross-sectional view of a three-dimensional terrain model created based on the digital elevation model M. Like FIG. 3 (a), FIG. It was cut out on the Z plane. A plane for projecting the three-dimensional terrain model based on the digital elevation model M in two dimensions is a model image plane G shown in FIG. In this figure, for convenience, the model image plane G is arranged parallel to the Z axis (that is, a vertical plane) and parallel to the X axis. However, the present invention is not limited to this, and the model image plane G is inclined in an arbitrary direction. It doesn't matter.

FIG. 4B is an image projected on the model image plane G shown in FIG. The model shape 5 is extracted from the features (things) in the image. The model shape 5 is a two-dimensional shape, and geometric elements (points, line segments, surfaces, etc.) constituting this shape include arbitrary coordinate systems (x _b , y _b ) in the model image plane G. It can be expressed in coordinates. Hereinafter, the arbitrary coordinate system in the model image plane G is referred to as a model coordinate system.

When the model shape 5 is extracted, the model shape 5 is extracted so that at least four points corresponding to the feature points selected from the constituent points of the reference shape 4 are included (A ′ _{1 to} A ′ _{5 in} FIG. 4B). . This is because the feature points of the reference shape 4 and the feature points of the model shape 5 need to correspond to each other as will be described later. Therefore, it is desirable that the viewpoint (X ₀ , Y _0, Z ₀ ) of the camera 1 and the inclination κ (that is, the azimuth) around the Z axis of the camera 1 are known. Usually, the digital elevation model M is often created in a wide range, and it takes a lot of time and labor to extract the model shape 5 including the feature points of the reference shape 4 without any clues. It is. In order to select a point corresponding to the feature point of the reference shape 4, it can be determined by visual observation or the like.

(Correspondence between standard shape and model shape)
As shown in FIG. _4B , the two-dimensional coordinates are given to the components (component points, etc.) of the model shape 5 in the model image plane G by the model coordinate system (x _b , y _b ). As can be seen from a), each component of the model shape 5 corresponds to the spatial coordinates of the digital elevation model M. That is, the constituent elements of the model shape 5 are associated with the spatial coordinates of the digital elevation model M. In other words, the constituent elements of the model shape 5 have three-dimensional coordinates in the real space.

On the other hand, the reference shape 4 in the camera image plane Ph does not correspond to a known altitude model M or the like having a known spatial coordinate. It cannot be expressed in dimensional coordinates. Therefore, by making the feature points A _{1 to} A _{5 of} the reference shape 4 correspond to the points A ′ _{1 to} A ′ ₅ corresponding to the feature points of the model shape 5, the components of the reference shape 4 are three-dimensional in real space. Give coordinates.

Since the reference shape 4 is a two-dimensional shape represented on the same plane, and the model shape 5 is also a two-dimensional shape represented on the same plane, a two-dimensional projective transformation formula can be used. . That is, if projective transformation of the feature point of the reference shape 4 is a point corresponding to the feature point of the model shape 5, both coordinate systems, that is, “photo coordinate system (x _a , y _a )” and “model” The coordinate system (x _b , y _b ) ”can be spatially related. The two-dimensional projective transformation formula is given by the following formula.

ここで、
x_a : 写真座標系における特徴点のx座標
y_a : 写真座標系における特徴点のy座標
x_b : モデル座標系における特徴点のx座標
y_b : モデル座標系における特徴点のy座標

here,
x _a : x coordinate of the feature point in the photographic coordinate system
y _a : y coordinate of the feature point in the photographic coordinate system
x _b : x coordinate of the feature point in the model coordinate system
y _b : y coordinate of the feature point in the model coordinate system

There are eight unknowns b _{1 to} b ₈ in these two formulas, and a total of eight formulas are required to obtain solutions of these unknowns. That is, four feature points of the reference shape 4 and four points corresponding to the feature points of the model shape 5 must be known. The reason why four or more feature points (points corresponding to feature points) are required for each of the reference shape 4 and the model shape 5 is to obtain an unknown solution of the two-dimensional projective transformation equation. When there are exactly 4 known points, one solution can be obtained by solving the 8-ary linear equation, but when 5 or more points are known, a solution can be obtained by the least-order method.

When the eight unknowns b _{1 to} b ₈ are obtained in this way, the “photo coordinate system (x _a , y _a )” and the “model coordinate system (x _b , y _b )” are spatially related. That is, if the above-described two-dimensional projective transformation formula is used, an arbitrary point arranged in the camera image plane Ph (that is, an arbitrary point to which coordinates are assigned in the photographic coordinate system) is arranged in the model image plane G. Can do. This means that an arbitrary point in the camera image plane Ph is to obtain a two-dimensional coordinate of the model coordinate system, that is, is associated with a spatial coordinate (three-dimensional coordinate) of the digital elevation model M. In other words, by expressing the reference shape 4 represented in the photographic coordinate system in the model coordinate system, the three-dimensional coordinates of the real space can be given to the reference shape 4. It should be noted that the shape in which the reference shape 4 based on the camera image plane Ph is represented by the model image theoretically matches the model shape 5 (broken line shape in FIG. 4B) represented in the model image plane G. Although it should be, it does not necessarily correspond for reasons, such as the creation accuracy of the digital elevation model M.

Similarly to the reference shape 4, three-dimensional spatial coordinates can be given also to the subject collapsing portion 3 b shown in FIG. The procedure is shown below. The object collapsing part 3b displayed in the camera image plane Ph is created as the output shape 6 by the same method as the reference shape 4. The components (component points, etc.) of the output shape 6 are given two-dimensional coordinates by _a photographic coordinate system (x _a , y _a ), and _a model coordinate system (x _b , y is obtained by using a two-dimensional projective transformation formula. _b ). In this way, the components of the output shape 6 are associated with the coordinates of the model coordinate system, and as a result, the three-dimensional coordinates of the real space can be given. That is, the position, shape, and size of the collapsed portion B that actually exists can be specified.

(Create map)
If the position and shape of the collapsing part B can be specified, it can be displayed and displayed on an existing map, and a new map in which information on the collapsing part B is recorded can be created. Of course, it is not limited to displaying on an existing map, but other spatial information can be added to create a new map describing the position and shape of the collapsed portion B. Thus, since it can plot on a map only by image | photographing sufficiently away from the collapse part B, the map which displayed the collapse part B can be obtained easily and safely.

(Implementation flow)
An example of a procedure for carrying out this embodiment will be described with reference to the implementation flow shown in FIG.

  An image (photo 2) including a feature (here, the collapsed portion B) whose position and shape are to be specified is acquired by a camera 1 such as a digital camera (501). At this time, shooting information (viewpoint coordinates, shooting direction) when a photograph is taken is measured using a measuring instrument such as a GPS or an electronic compass.

The object feature 3 (here, the object skyline 3a) having a characteristic part is extracted from the object feature 3 stored in the photograph 2 (that is, the camera image plane Ph). The boundary (edge) of the captured skyline 3a is automatically generated by the computer from the difference in brightness between the mountains and the sky in the image data of the photograph 2, and the reference shape 4 is created using this edge (502). Next, four or more feature points characterizing the shape are selected from the points constituting the reference shape 4, and the coordinates of the photographic coordinate system (x _a , y _a ) are given to these (503).

  Shooting information (viewpoint coordinates, shooting direction) for creating the model shape 5 is input (504). Based on this input value and the existing digital elevation model M, a model shape 5 is created by projecting it onto the model image plane G (505). Next, four or more points corresponding to the feature points of the reference shape 4 are extracted from the model shape 5 (506). If four or more points corresponding to feature points cannot be extracted, the model shape 5 is created again by changing the creation range.

Based on the feature points (4 points or more) of the reference shape 4 and the feature point equivalent points (4 points or more) of the model shape 5, a two-dimensional projective transformation is performed to obtain a “photo coordinate system (x _a , y _a )” “Model coordinate system (x _b , y _b )” is associated (507). As a result, the reference shape 4 (the subject skyline 3a) and the subject collapsed portion 3b can be converted into the model coordinate system (x _b , y _b ), and can be associated with the spatial coordinates of the digital elevation model M. (508). If the collapsing part B can be identified in the real space, this is described on the existing map, and the map which described the information of the collapsing part B is created (509).

(Embodiment 2)
Here, a second embodiment of “the method for specifying the position of the object to be imaged, its program, and the display map of the object to be imaged” will be described. In the present embodiment, a case where the correspondence between the reference shape 4 and the model shape 5 is directly achieved in place of the projective transformation based on the feature point extraction of the first embodiment will be described. The explanation about the contents is omitted.

(Standard shape)
6A is a model diagram for explaining the reference shape 4 extracted from the object 3 to be photographed. FIG. 6B is a model diagram for explaining the model shape 5 created based on the digital elevation model M. FIG. FIG. 6C is a model diagram for explaining a state in which the reference shape 4 and the model shape 5 are collated. The object 3 stored in the photograph 2 photographed in FIG. 2 can be given spatial information by directly corresponding to the model shape 5 created based on the digital elevation model M. In other words, the feature of the present invention is that coordinates can be given to the feature by associating one photo 2 (single photo) with one model shape 5, and the coordinates are given using a set of two photos. This point is significantly different from the conventional method (stereo matching method).

  When the object feature 3 and the model shape 5 are made to correspond to each other, instead of directly (for example, by visual observation), the object feature 3 is made to correspond to the model shape 5 after making the object feature 3 a reference shape 4 that can be processed by a computer. It is desirable. In addition, the reference shape 4 corresponding to the model shape 5 may be created based on all the object features 3 contained in the photograph 2, but it is easy to correspond to the model shape 5 (that is, the digital elevation model). It is preferable that the feature (which can be expressed by M) is extracted from the object feature 3 and created. In the present embodiment, the subject skyline 3a is extracted from the subject features 3 photographed in the photograph 2.

The reference shape 4 can be drawn to trace the imaged skyline 3a using a drawing application such as CAD. Alternatively, the image data of Photo 2 is used, the skyline A is automatically generated by a computer as an edge from the difference in brightness (or hue, saturation, and brightness) of the sky and mountains in the image data, and this edge is used. A reference shape 4 can also be created. The reference shape 4 created in this way is a two-dimensional shape, and the geometric elements (points, line segments, surfaces, etc.) constituting this shape are represented by an arbitrary coordinate system (x _a , y in the photograph 2). _a ) can be expressed in coordinates. Hereinafter, the arbitrary coordinate system in the photograph 2 is referred to as a photograph coordinate system. In addition, a part characterizing the shape (hereinafter referred to as “characteristic part”) can be selected from the points constituting the reference shape 4. As this characteristic part, not only a characteristic point (for example, a peak of a mountain or a change point of a valley) but also a part having a local characteristic such as a curvature of a line segment can be selected. Note that in FIG. 6A, three locations of the ridge portion F1, the valley portion F2, and the ridge portion F3 are selected as characteristic portions.

(Model shape)
FIG. 7 is an explanatory diagram showing the concept of creating the model shape 5. The model shape 5 is a two-dimensional shape created by performing spatial calculation based on the digital elevation model M, and corresponds to the reference shape 4. As described above, since the digital elevation model M has three-dimensional spatial coordinates, it is possible to create a three-dimensional terrain model, as shown in FIG. "Ground model S") can be reproduced. A projected landform (two-dimensional shape) created by projecting the surface model S from an arbitrary point is a model shape 5. That is, if the surface model S is photographed from an arbitrary point by a camera, the topography that should be obtained is the model shape 5.

  Specifically, when the “elevation” information and the “viewing angle” information are given to the digital elevation model M and the spatial calculation processing is performed, a projected landform, that is, a model shape 5 is created. In FIG. 7, three different model shapes 5 are created by projecting from three different points. The “viewpoint” here is a three-dimensional coordinate (X, Y, Z) in the real space to be projected, and is a concept corresponding to the camera center position. The “viewing angle” refers to a projection direction and a viewing angle, and this projection direction is a shooting posture when shooting with a camera (tilt ω (pitch) around the X axis, tilt φ (roll) around the Y axis, Z This is a concept corresponding to a tilt around the axis (yaw), and the viewing angle corresponds to a camera photographing range defined by a screen distance or the like.

If the viewing angle (projection direction and viewing angle) and viewpoint are determined and spatial calculation is performed based on the digital elevation model M, the range on the ground model S to be projected is specified, and the terrain within the specified range is the same. It can be expressed as a two-dimensional shape on a plane (hereinafter, this plane is referred to as “model image plane G”). The two-dimensional shape expressed here is “model shape 5”, and the geometric elements (points, line segments, surfaces, etc.) constituting this shape are arbitrary coordinate systems (x _b , y _b ). Hereinafter, the arbitrary coordinate system in the model image plane G is referred to as a model coordinate system. In addition, such a technique is conventionally used as introduced in Patent Document 1 and Patent Document 2 described above, and an application program for executing this technique is also commercially available.

(Correspondence between standard shape and model shape)
When the reference shape 4 shown in FIG. 6 (a) is created and the model shape 5 shown in FIG. 6 (b) is created, the reference shape 4 and the model shape 5 are compared with each other as shown in FIG. 6 (c). Try to correspond between each component.

  Actual shooting condition values (camera center position coordinates, shooting posture, shooting range) when the viewpoint and viewing angle (projection direction and viewing angle) input when creating the model shape 5 are shot to create the reference shape 4 ) Theoretically, the reference shape 4 and the model shape 5 completely coincide. That is, the coordinates (X, Y, Z) of the camera center position are set as the “viewpoint”, the shooting posture (ω, φ, κ) is set as the projection direction of the “viewing angle”, and the shot shooting range is set as the “viewing angle”. The model shape 5 created as the viewing angle is completely coincident with the reference shape 4. In order to measure the coordinates of the camera center position and the camera shooting posture, a total station or direction meter (magnet, electronic compass, etc.) can be used. Especially when shooting in flight, GPS and IMU are used. be able to. Further, in order to obtain a photographed photographing range, it can be obtained from the specifications (view angle, screen distance, etc.) of the camera 1.

  If the reference shape 4 and the model shape 5 completely coincide with each other, it is easy to compare them with each other by comparing them. However, actually, the reference shape 4 and the model shape 5 may not completely match due to the influence of the measurement accuracy of the photographing condition value and the position accuracy of the digital elevation model M (FIG. 6C). In such a case, the reference shape 4 and the model shape 5 are arranged so as to overlap as much as possible (hereinafter, this state is referred to as “closest arrangement”), and both adjacent components are made to correspond to each other.

  For example, the reference shape 4 and the model shape 5 are arranged in the same coordinate system (photographic coordinate system or model coordinate system), and further, an error between the constituent points of the reference shape 4 and the constituent points of the model shape 5 by the least square method or the like ( Both shapes can be arranged closest to each other so that (absolute value) is minimized. Alternatively, paying attention to the feature portions (ridge portion F1, valley portion F2, ridge portion F3 in FIG. 6A) reflected in the reference shape 4, a characteristic shape of the model shape 5 corresponding to this feature portion is extracted. Then, it can be arranged closest to the feature part of the reference shape 4 so as to match.

  The following method can be exemplified as a method of associating the constituent elements of the reference shape 4 and the model shape 5 in the state of closest approach. That is, in the state of the closest arrangement, a method of associating the constituent points of the reference shape 4 and the constituent points of the model shape 5 that are closest to each other, the reference shape 4 and the model shape 5 with the X coordinate of a predetermined pitch. For example, a method of calculating the respective component points to correspond the component points having the same X coordinate (FIG. 8), or a method of replacing the reference shape 4 with the shape (component) of the model shape 5 is exemplified.

  The concept of making the constituent elements of the reference shape 4 and the model shape 5 correspond can be shown, for example, as shown in FIG. In this figure, the configuration point 5_01, the configuration point 5_02, the configuration point 5_03,..., The configuration point 5_n of the model shape 5 are the configuration point 4_01, the configuration point 4_02, the configuration point 4_03,. Corresponds to 4_n. As described above, if the constituent points of the reference shape 4 and the constituent points of the model shape 5 can be associated with each other, the spatial information on the constituent points 4_01, 4_02, 4_03,. Can be given. That is, since the configuration point 5_01, the configuration point 5_02, the configuration point 5_03,..., The configuration point 5_n of the model shape 5 are originally created based on the digital elevation model M, each of the configuration points of the model shape 5 Is associated with (that is, has) spatial information including three-dimensional coordinates in real space (that is, attribute information may be included). If given to the constituent points of the shape 4, the constituent points of the reference shape 4 can obtain spatial information. Here, the constituent points are made to correspond as constituent elements. However, the constituent points are not limited to this, and may be made to correspond by using a line segment, a combination of line segments, or a surface. In short, the reference shape 4 and the model shape 5 are configured. What is necessary is just to make the elements to perform correspond. As described above, the components of both the reference shape 4 and the model shape 5 may be associated with each other, and the spatial information of the components of the model shape 5 may be given to the components of the associated reference shape 4, Instead of this, the reference shape 4 may be replaced with the model shape 5 to give spatial information to the reference shape 4 (the spatial information of the model shape 5 is used as it is).

  When the constituent points of the reference shape 4 obtain the spatial information, the spatial information can be given to the object 3 that is not the target of the reference shape 4 in the object 3. That is, in the present embodiment, the “photographed skyline 3a” of the subject features 3 is the target of the reference shape 4, and the subject collapse part 3b is not the target of the reference shape 4, but this subject collapse Spatial information can also be given to the part 3b.

A two-dimensional shape (hereinafter referred to as “output shape 6”) is created in the same manner as the reference shape 4 is created for the object to be given spatial information (here, the object collapsed portion 3b). create. Since the output shape 6 can be arranged in the photographic coordinate system (x _a , y _a ), it can be represented by the coordinates of the photographic coordinate system. The state in which the reference shape 4 and the model shape 5 was closest arrangement, since both are arranged on the same plane, hidden (deviation between x _b-axis of x _a-axis of the photo coordinate system and the model coordinate system ), hidden and y _b axis y _a-axis of the photo coordinate system and the model coordinate system (shift) can be seen. Therefore, an arbitrary point arranged in the photographic coordinate system (x _a , y _a ) can be converted into the model coordinate system (x _b , y _b ), and the output shape 6 is also expressed by the coordinates of the model coordinate system. As a result, the three-dimensional coordinates of the real space can be given. That is, the position, shape, and size of the collapsed part B that actually exists can be specified.

(Create map)
If the position and shape of the collapsing part B can be specified, it can be displayed and displayed on an existing map, and a new map in which information on the collapsing part B is recorded can be created. Of course, it is not limited to displaying on an existing map, but other spatial information can be added to create a new map describing the position and shape of the collapsed portion B. Thus, since it can plot on a map only by image | photographing sufficiently away from the collapse part B, the map which displayed the collapse part B can be obtained easily and safely.

(Implementation flow)
An example of a procedure for carrying out the present embodiment will be described with reference to the implementation flow shown in FIG.

  An image (photo 2) including an object to be identified for its position and shape (here, the object collapsing part 3b) is acquired by a camera 1 such as a digital camera (901). At this time, the photographing condition values (the coordinates of the camera center position, the photographing posture, and the photographing range) when the photograph is taken are measured using a measuring instrument such as GPS or IMU.

  The object feature 3 (here, the object skyline 3a) that is easily associated with the model shape 5 is extracted from the object features 3 stored in the photograph 2 (902). The boundary (edge) of the captured skyline 3a is automatically generated by the computer from the difference in brightness between the mountains and the sky in the image data of Photo 2, and the coordinates of the photographic coordinate system are assigned to the constituent points constituting this edge. A reference shape 4 is set (903). At this time, the ridge part F1, the valley part F2, and the ridge part F3 can be reflected in the reference shape 4 as characteristic parts (904).

  From the existing digital elevation model M, an area including the object feature 3 is selected (905). This is because, when the digital elevation model M covers a wide range, it is possible to reduce the load on the arithmetic processing by limiting the range to some extent. If the range of the digital elevation model M is relatively limited, the region is not necessarily limited. There is no need to select.

  The photographing condition values (the coordinates of the camera center position, the photographing posture, and the photographing range) are input as a viewpoint and a viewing angle (projection direction and viewing angle) for creating the model shape 5 (906). Among these, the values measured at the time of shooting are used for the coordinates of the camera center position and the shooting posture, and the values obtained from the specifications of the camera 1 (focal length, screen distance, etc.) are used for the shooting range. Spatial calculation processing is performed based on the viewpoint and viewing angle input here and the digital elevation model M selected in the region, and a model shape 5 which is a two-dimensional shape is created (907).

  The reference shape 4 and the model shape 5 are arranged in the same coordinate system (two-dimensional), and both shapes are placed in the closest arrangement so that the reference shape 4 and the model shape 5 match as much as possible. At this time, the reference shape 4 and the model shape 5 may be illuminated while paying attention to the feature part reflected in the reference shape 4. Further, the closest arrangement may be made such that the error (absolute value) between the constituent point of the reference shape 4 and the constituent point of the model shape 5 is minimized by the least square method or the like. In this closest arrangement state, the constituent points closest to each other in the constituent points of the reference shape 4 and the model shape 5 (or those having the same X coordinate (FIG. 8), etc.), both constituent points Make an association. Alternatively, the reference shape 4 is set as the shape of the model shape 5 (908).

  When the constituent points of the reference shape 4 and the constituent points of the model shape 5 can be associated (or replaced with the shape of the model shape 5), the three-dimensional coordinates of the constituent points of the model shape 5 are converted into the reference shapes related thereto. 4 is assigned to the component number 4 (909). An output shape 6 is created for the subject collapsing portion 3b using a drawing application such as CAD. The output shape 6 is represented by coordinates in the photographic coordinate system, and the coordinates of the output shape 6 in the model coordinate system are obtained from the spatial relationship between the photographic coordinate system and the model coordinate system. Further, based on the digital elevation model M, spatial information is given to the output shape 6, and the position and shape of the subject collapsing part 3b are specified (910). If necessary, the position and shape of the collapsed portion B are described and displayed on an existing map, and a map in which information of the collapsed portion B is written is created.

(System configuration)
FIG. 10 is an explanatory diagram showing the arrangement of the devices constituting the “target object position specifying system”. Based on this figure, the device configuration of the “target object position specifying system” in the present embodiment. An example will be described.

  As shown in FIG. 10, a host computer 7, a terminal computer 8, and a digital elevation model database 9 are connected by a wired or wireless network. The terminal computer 8 reads the image data of the photograph 2 from the camera 1 (digital camera), and stores the photographing condition value when actually photographing. Furthermore, the reference shape 4 is created by automatically generating the boundary (edge) of the object 3 based on the difference in luminance in the image data and assigning two-dimensional coordinates on the same plane to the constituent points of the edge. To do. It is also possible to input a characteristic part and reflect it in the reference shape 4.

  The host computer 7 receives the reference shape 4 and the photographing condition value from the terminal computer 8 through communication, and receives the digital elevation model M from the digital elevation model database 9. At this time, it is possible to specify a necessary area and receive the digital elevation model M limited to the range from the digital elevation model database 9. Further, the host computer 7 executes steps 907 to 910 shown in the execution flow of FIG. 9 (or steps 1107 to 1113 of the execution flow of FIG. 11 described in the third embodiment).

  In FIG. 10, the host computer 7, the terminal computer 8, and the digital elevation model database 9 are connected via a network. However, the present invention is not limited to this, and the host computer 7 and the terminal computer 8 are integrated, or the host computer 7 and the numerical value model 9 are connected. The elevation model database 9 can be integrated, or the host computer 7, the terminal computer 8, and the numerical elevation model database 9 can all be integrated.

(Embodiment 3)
Here, a third embodiment of “the method for specifying the position of the object to be imaged, its program, and the display map of the object to be imaged” will be described. The present embodiment describes a case where any or all of the viewpoint and the viewing angle for creating the model shape 5 in the second embodiment are unknown, and technical contents common to the first and second embodiments. The description regarding is omitted.

  When the object 3 is photographed, only the coordinates of the camera center position are measured, and the photographing posture may not be measured. Alternatively, it may be possible to measure only the shooting posture or not to measure both the coordinates of the camera center position and the shooting posture. In such a case, the model shape 5 may be created by performing a collation determination according to the following procedure. In other words, unknown viewpoints and viewing angles (projection direction and viewing angle) are treated as variables, and these values are gradually changed from the initial values (arbitrary values). It is created as a model shape candidate 5 ′ ”. For example, when the coordinates of the camera center position are known and the shooting posture is unknown, the viewpoint (coordinate value) is fixed and the viewing angle is changed from the initial value to create a model shape candidate 5 '. When the coordinates of the camera center position and the shooting posture are unknown, the model shape candidate 5 'is created by changing the viewpoint (coordinate value) and the viewing angle from the initial values as variables. The created model shape candidate 5 'is collated with the reference shape 4, and collation judgment is performed based on a predetermined approximate condition. The calculation of the model shape candidate 5 ′ and the matching determination with the reference shape 4 are repeatedly performed until it is determined that matching is performed, and the model shape candidate 5 ′ determined to be verified is extracted as the model shape 5.

  The collation judgment between the model shape candidate 5 ′ and the reference shape 4 is for judging the coincidence of both shapes by superimposing them, and of course, not limited to the case where both shapes are completely coincident (congruent). Similarly, when both shapes are approximated to such an extent that they can be determined to be coincident, it can be determined that they are identical, that is, collated. The following method can be exemplified when determining the consistency (matching determination). That is, the model shape candidate 5 'and the reference shape 4 are collated when the predetermined condition (approximate condition) is satisfied in the closest arrangement (similar to the description in the second embodiment). to decide. As this approximation condition, for example, the condition that the total error (absolute value) in the closest approach is less than the threshold, or the number of component points that are outside the allowable error (preset) in the closest approach is less than the threshold. Conditions, etc. Alternatively, focusing on only the characteristic part (similar to the description of the second embodiment, for example, the ridge part F1, the valley part F2, and the ridge part F3 in FIG. 6A), the characteristic part is an approximate condition as described above. The reference shape 4 and the model shape candidate 5 ′ may be verified at the time of the closest approach without providing an approximation condition (unconditionally). it can.

  When it is determined that the model shape candidate 5 ′ is matched with the reference shape 4, and the model shape candidate 5 ′ is extracted as the model shape 5, the correspondence between the reference shape 4 and the components of the model shape 5 is the same as in the second embodiment. Thus, the spatial information of the real space by the digital elevation model M is given to the constituent elements of the reference shape 4. In addition, since the constituent points of the reference shape 4 have obtained spatial information, the object feature 3 (for example, the object collapsing part 3b) that is not the object of the reference shape 4 among the object features 3 is detected. Since spatial information can also be given, the position and shape can be described and displayed on an existing map as in the second embodiment.

(Implementation flow)
An example of a procedure for carrying out the present embodiment will be described with reference to the execution flow shown in FIG.

  An image (photo 2) including the object to be identified for its position and shape (here, the object collapsing part 3b) is acquired by the camera 1 such as a digital camera (1101). At this time, the photographing condition value (the coordinates of the camera center position) when the photograph is taken is measured using a measuring instrument such as a total station.

  The object feature 3 (here, the object skyline 3a) that is easily associated with the model shape 5 is extracted from the object features 3 stored in the photograph 2 (1102). The boundary (edge) of the skyline 3a is automatically generated by the computer from the difference in brightness between the mountains and the sky in the image data of Photo 2, and two-dimensional coordinates on the same plane are given to the constituent points that constitute this edge. Thus, the reference shape 4 is created (1103). At this time, the ridge part F1, the valley part F2, the ridge part F3 (FIG. 6A) can be reflected in the reference shape 4 as a characteristic part (1104).

  Of the existing digital elevation model M, an area including the object feature 3 is selected (1105). This is because, when the digital elevation model M covers a wide range, it is possible to reduce the load on the arithmetic processing by limiting the range to some extent. If the range of the digital elevation model M is relatively limited, the region is not necessarily limited. There is no need to select.

  A viewpoint and a viewing angle (projection direction and viewing angle) for creating the model shape 5 are input (1106). Here, assuming that the coordinates of the camera center position are known among the photographing condition values, this is input as a viewpoint, and an arbitrary initial value is input for the viewing angle.

  Spatial calculation processing is performed based on the digital elevation model M selected in the region and the input viewpoint and viewing angle, and a model shape 5 candidate that is a two-dimensional shape (that is, model shape candidate 5 ′) is created. (1107). The reference shape 4 and the model shape candidate 5 ′ are arranged in the same coordinate system (two-dimensional), and the error (absolute value) between the constituent point of the reference shape 4 and the constituent point of the model shape 5 is minimized by the least square method or the like. Both shapes are brought into a state of closest approach so as to become (1108). In this state, collation determination is performed based on the approximate condition (1109), and if the approximate condition is satisfied, that is, if it can be determined that collation has been performed, the process proceeds to the next step (1110). When the approximation condition is not satisfied, that is, when it is determined not to collate, the viewpoint / viewing angle is input again (1106). However, at this time, the viewpoint value is left as it is, and the previous viewing angle input value is changed and input. In this way, the visual angle input (1106) to the collation determination (1109) are repeatedly performed until the reference shape 4 and the model shape candidate 5 'are collated.

  When it is determined that the reference shape 4 and the model shape candidate 5 ′ are collated, the model shape candidate 5 ′ is extracted as the model shape 5 (1110), and the constituent points of the reference shape 4 and the constituent points of the model shape 5 are associated with each other. (Or replacement of the model shape 5 with the shape) is performed (1111). When the constituent points of the reference shape 4 and the constituent points of the model shape 5 can be associated (or replaced with the shape of the model shape 5), the three-dimensional coordinates of the constituent points of the model shape 5 are converted into the reference shapes related thereto. 4 is assigned to the constituent points (1112). Next, an output shape 6 is created for the subject collapsing portion 3b using a drawing application such as CAD. The output shape 6 is represented by coordinates in the photographic coordinate system, and the coordinates of the output shape 6 in the model coordinate system are obtained from the spatial relationship between the photographic coordinate system and the model coordinate system. Further, based on the digital elevation model M, spatial information is given to the output shape 6 to identify the position and shape of the subject collapsing portion 3b (1113). If necessary, the position and shape of the collapsed portion B are described and displayed on an existing map, and a map in which information of the collapsed portion B is written is created.

(System configuration)
Similarly to the first embodiment, the arrangement of the devices constituting the “target object position specifying system” in the present embodiment can also be the device configuration shown in FIG.

(Embodiment 4)
Here, a description will be given of a fourth embodiment of “the method for specifying the position of the subject feature, its program, and the display map of the subject feature”. In the present embodiment, the case where the reference shape 4 and the model shape 5 are created based on the images in the second to third embodiments will be described, and the description regarding the technical contents common to the first to third embodiments will be omitted.

(Standard shape)
In order to make the object feature 3 correspond to the model shape 5, it is desirable to make the object feature 3 correspond to the model shape 5 after making it the reference shape 4 that can be processed by a computer. The reference shape 4 is created. As in the first to third embodiments, the reference shape 4 may be created based on all the object features 3 contained in the photograph 2, but can be easily associated with the model shape 5 (that is, the digital elevation model M). It is preferable to extract and create the features that can be expressed by In the present embodiment, the object mountain 3c (the shape of the mountain shown in the photograph 2, FIG. 2) is extracted from the object 3 photographed in the photograph 2.

  As the reference shape 4, an image (a set of pixels) that constitutes the subject mountain portion 3 c in the photograph 2 is extracted, and this image is set as the reference shape 4. In this case, the reference shape 4 can be created by cutting out the image of the mountain portion 3c while the photograph 2 is displayed on the monitor and visually confirmed. Or, based on the difference in brightness (or saturation or lightness) in the image data of Photo 2, the boundary of the feature (edge) is automatically recognized by the computer, and the set of pixels constituting this edge is used as a reference. It can also be created as shape 4.

(Model shape)
The model shape 5 is created based on the digital elevation model M and the texture assigned thereto. If stereo matching is performed using a pair of stereo aerial photographs and satellite photographs, an image (a set of pixels) corresponding to the topography, that is, a texture can be given to the digital elevation model M. Or, calculate the topographic amount (elevation value, inclination angle, etc.) of each mesh in the digital elevation model M, and give the topographical amount map (hue, saturation, brightness) according to the inclination amount to each mesh, It can also be given to the digital elevation model M as a texture. As the topographic map, there are various types such as a shadow map (shaded and shaded) created by giving a light source and a water system map. In this way, if a space calculation is performed based on the digital elevation model M to which the texture (aerial photograph, topographic map, etc.) is given, and the viewpoint and viewing angle (projection direction and viewing angle) described in the second and third embodiments. The range on the ground surface model S projected by the viewpoint and the viewing angle is specified, and an image within the specified range can be expressed.

(Correspondence between standard shape and model shape)
When the reference shape 4 based on the image and the model shape 5 based on the image are created, these are collated, and each pixel is associated with each other. Specifically, based on the respective pixel arrays of the reference shape 4 and the model shape 5, the pixels of the reference shape 4 and the pixels of the model shape 5 are associated with each other. At this time, the method described in the second and third embodiments (a method using the outer shape and the characteristic part of the object 3) can be combined. Spatial information can be given to the pixels of the reference shape 4 after comparing the corresponding pixels and judging the matching (matching judgment). When the model shape 5 can be created in a state that approximates the reference shape 4 to some extent, such as when the coordinates of the center position, the shooting posture, and the shooting range are adopted, the pixel of the reference shape 4 is omitted as it is without matching judgment. Spatial information can also be given to.

  The collation determination in this embodiment is performed based on the attributes of corresponding pixels. The attribute in this case is a color attribute, but the color is originally recognized by human vision and is accompanied by individual differences. However, in recent years, a technique for quantifying color attributes has been established so that this color can be handled by a computer. There are various methods for quantifying the color attributes, and RGB, Cyan, Magenta, and Yellow, whose basic colors are three colors of red, green, and blue. CMYK having four basic colors of black (Keycolor), NCS and Ostwald color system having six basic colors of yellow, red, blue, green, black, and white are known. In the present embodiment, a case will be described below in which color attributes are digitized in RGB, but other techniques may be employed to implement the present invention.

Here, RGB will be briefly described. Colors have three attributes consisting of hue, saturation, and lightness. RGB expresses three attributes of various colors by additive color mixing that mixes these three primary colors with red, green, and blue as basic colors. It is. Specifically, RGB expresses red, green, and blue with the respective lightness values. If red lightness is r, green lightness is g, and blue lightness is b, RGB is (r, g B), and the three attributes of color can be expressed by a combination of the values of r, g, and b.

The collation determination in the present embodiment will be specifically described as follows. That is, with respect to the reference shape 4 and the model shape 5, RGB is compared for each corresponding pixel, and the presence or absence of collation is determined according to an approximate condition set in advance. As this approximate condition, for example, a condition in which the ratio of the number of pixels in which both RGB differences are within an allowable value (a ratio to the number of previous pixels) is equal to or greater than a threshold value, or a condition in which the maximum value of RGB differences is equal to or less than the threshold value. , Etc. Under such an approximate condition, the reference shape 4 and the model shape 5 are collated, and if this condition is satisfied, it is determined to be verified, and if not satisfied, it is determined not to be verified.

  When the reference shape 4 and the model shape 5 are collated and determined, the viewpoint and viewing angle values are gradually changed from initial values (for example, photographing condition values), and the calculation is repeated. That is, the shape calculated at each viewpoint and viewing angle is set as a “model shape candidate 5 ′”, and this is compared with the reference shape 4 to perform collation determination based on a predetermined approximate condition. This collation determination is repeated until it is determined that the reference shape 4 and the model shape candidate 5 ′ have been collated, and the model shape candidate 5 ′ that has been determined to be collated is extracted as the model shape 5.

  Spatial information is given to the pixel of the reference shape 4 corresponding to the pixel of the model shape 5 (including the case where the collation is not determined) extracted in this way. Alternatively, as described in the second and third embodiments, the image of the reference shape 4 (a set of pixels) is replaced with the image of the model shape 5 (a set of pixels). Since the pixels constituting the model shape 5 are related to the digital elevation model M, each pixel has spatial information (including attribute information in some cases) including three-dimensional coordinates. If the spatial information which the pixel of the model shape 5 has is provided to the corresponding pixel of the reference shape 4, the pixel of the reference shape 4 can obtain the spatial information.

  Furthermore, when the pixels of the reference shape 4 obtain the spatial information, the spatial information can be given to the object 3 that is not the target of the reference shape 4 among the objects 3 to be imaged. In the present embodiment, the “photographed mountain portion 3 c” of the subject feature 3 is the target of the reference shape 4, and the subject collapse portion 3 b is not the target of the reference shape 4. Spatial information can be given also to 3b.

  A two-dimensional shape (hereinafter referred to as “output shape 6”) is created based on the image in the same manner as the reference shape 4 is created for the subject collapsing portion 3b to which spatial information is to be added. As in the second and third embodiments, the relationship between the subject collapsing portion 3b and the model shape 5 is determined based on the arrangement relationship between the subject collapsing portion 3b and the subject mountain portion 3c. As a result, the spatial information of the real space by the digital elevation model M can be given to the subject collapsing part 3b. That is, spatial information is given to the subject collapsing part 3b, and the position and shape of the collapsing part B can be specified.

(Implementation flow)
The procedure for implementing this embodiment is the same as that of Embodiments 2 and 3 except for the method of creating the reference shape 4 and the model shape 5. Here, based on FIG. 12, the implementation flow in the case of collation judgment is demonstrated. Note that the execution flow when the collation determination is not made is the same as in FIG.

  An image (photo 2) including the subject collapse part 3b whose position and shape are desired to be specified is acquired by a camera 1 such as a digital camera (1201). At this time, the photographing condition values (the coordinates of the camera center position, the photographing posture, and the photographing range) when the photograph is taken are measured using a measuring instrument such as GPS or IMU.

  From the object 3 stored in the photograph 2, the object mountain portion 3c that is easily associated with the model shape 5 is extracted (1202). The boundary (edge) of the captured skyline 3a is automatically recognized by the computer from the difference in brightness between the mountains and the sky in the image data of the photograph 2, and a set of pixels constituting the edge is created as the reference shape 4. (1203). At this time, the ridge part F1, the valley part F2, the ridge part F3 (FIG. 6A) can be reflected as a characteristic part in the reference shape 4 (1204).

  From the existing digital elevation model M, an area including the object 3 is selected (1205). This is because, when the digital elevation model M covers a wide range, it is possible to reduce the load on the arithmetic processing by limiting the range to some extent. If the range of the digital elevation model M is relatively limited, the region is not necessarily limited. There is no need to select.

  The viewpoint and viewing angle (projection direction and viewing angle) for creating the model shape 5 are input as shooting condition values (camera center position coordinates, shooting posture, shooting range) (1206). Among these, the values measured at the time of shooting are used for the coordinates of the camera center position and the shooting posture, and the values obtained from the specifications of the camera 1 (focal length, screen distance, etc.) are used for the shooting range. Spatial calculation processing is performed based on the viewpoint and viewing angle input here, and the digital elevation model M to which texture is added, and the range on the ground model S that is projected by the viewpoint and viewing angle is specified, and the identification is performed. A candidate for model shape 5 (that is, model shape candidate 5 ′) is created from the image within the range (1207).

  Based on the pixel arrangement of each of the reference shape 4 and the model shape candidate 5 ′, the pixel of the reference shape 4 and the pixel of the model shape candidate 5 ′ are made to correspond (1208), and collation determination is performed based on a predetermined approximate condition ( 1209) When the approximation condition is satisfied, that is, when it can be determined that the collation has been performed, the process proceeds to the next step (1210). When the approximation condition is not satisfied, that is, when it is determined not to collate, the viewpoint / viewing angle is input again (1206). However, in this case, the viewpoint / viewing angle is input in place of the previous value. In this manner, the viewpoint / viewing angle input (1206) to the collation determination (1209) are repeatedly performed until the reference shape 4 and the model shape candidate 5 'are collated.

  Spatial information is given to the pixel of the reference shape 4 corresponding to the pixel of the model shape 5 (1210). Similarly to the creation of the reference shape (1206), the output shape 6 is also created for the subject collapsing portion 3b. The disintegration part 3b and the model shape 5 are associated with each other based on the arrangement relationship between the object collapse part 3b and the object mountain part 3c, and the arrangement relation between the object mountain part 3c and the model shape 5. Spatial information of the real space by the digital elevation model M can be given to the photo collapsing part 3b. That is, spatial information is given to the subject collapsing part 3b, and the position and shape of the collapsing part B can be specified (1211). If necessary, the position and shape of the collapsed portion B are described and displayed on an existing map, and a map in which information of the collapsed portion B is written is created.

[Embodiment of Shooting Position Acquisition Method and Program, and Shooting Position Acquisition Device]
The present embodiment describes the “shooting position acquisition method, program thereof, and shooting position acquisition apparatus” of the present invention. In addition, this embodiment is the spatial information to the to-be-photographed object 3 implemented in [Embodiment of the position specifying method of the to-be-grounded object and its program, and the display map of the to-be-photographed object] (Embodiment 1-4). A case where the viewpoint and the viewing angle when creating the model shape 5 are obtained instead of giving (specifying the position and shape) will be described. Therefore, in addition to those relating to the provision of spatial information to the object 3, that is, Embodiments 1 to 4 of [Embodiment of the method for specifying the position of the object to be imaged and its program, and display map of the object to be imaged] The present embodiment relates to the photographing of the single photograph 2, the extraction of the object 3 to be photographed, the creation of the reference shape 4, the creation of the model shape 5, the collation between the reference shape 4 and the model shape 5 (including collation judgment). However, the same applies, and a description thereof is omitted here. The creation of the reference shape 4 and the creation of the model shape 5 in the present embodiment are the first to third embodiments of [Method for specifying position of object to be imaged and its program, and embodiment of display map of object to be imaged]. The method described in the above (the method based on the outer shape) or the method described in the fourth embodiment (the method based on the image) may be used.

  FIG. 13 is an explanatory diagram showing a situation in which the current shooting position is acquired using the shooting position acquisition device 10 of the present invention. The photographing position acquisition device 10 includes image acquisition means such as a digital camera, and a photographer shown in the figure takes a picture of a cityscape. The photograph 2 taken here contains a hotel D and an apartment E, and the object to be photographed 3 is acquired as the photographed hotel 3d and the photographed apartment 3e, respectively.

  The photographed hotel 3d and the photographed apartment 3e are extracted from the photographed features 3 of the photograph 2 shown in FIG. 13, and a reference shape 4 is created for the photographed hotel 3d and the photographed apartment 3e. In this case, as described above, the reference shape 4 can be created by the method described in the first to third embodiments of [the position specifying method of the object to be detected and its program, and the embodiment of the display map of the object to be detected]. The method described in the fourth embodiment may be used.

  In the city area including the hotel D and the apartment E, a digital elevation model M (or a digital elevation model M provided with a texture) has already been created, and the digital elevation model M (or a digital elevation model M provided with a texture). ) To create a model shape 5. Since the viewpoint and viewing angle at this time are unknown, a predetermined initial value (default value) is adopted. Model shape candidates 5 ′ are calculated while gradually changing the viewpoint and the viewing angle from the initial values, and are compared with the reference shape 4. The calculation of the model shape candidate 5 ′ and the matching with the reference shape 4 are repeated until the approximate condition is satisfied, and the model shape candidate 5 ′ determined to be matched is extracted as the model shape 5. In this case as well, the model shape 5 can be created by the method described in the second and third embodiments of [the method for specifying the position of the target object and its program, and the display map of the target object]. The method described in 4 may be used.

  The viewpoint and viewing angle adopted when calculating the model shape 5 extracted in this way are determined to be the photographing condition values when the photograph 2 is photographed, that is, (camera center position coordinates, photographing posture, photographing range). can do.

  Knowing the coordinates taken when shooting (that is, in real time) is extremely convenient for tourists and other people who are unfamiliar with the land, and can reach the destination without getting lost. Alternatively, in conjunction with the shooting position information acquired by the shooting position acquisition device 10, it is possible to generate a sound of sightseeing guidance (for example, explanation of the hotel D that was shot).

  The present embodiment is not limited to acquiring shooting position information in real time when shooting, but can also be used to acquire shooting position information of a photograph 2 taken in the past. FIG. 14 is an execution flow in the case where the shooting position information of the photograph 2 is acquired.

  There is a photograph 2 (1401) taken in the past. Although this hotel 2 contains a hotel D and an apartment E, it is currently unknown where the photograph was taken. The photographed hotel 3d and the photographed apartment 3e are extracted from the photographed objects 3 contained in the photograph 2 (1402). The boundary (edge) between the hotel 3d and the apartment 3e is automatically generated by the computer based on the difference in brightness between the building (hotel D and apartment E) and the sky in the image data of Photo 2. The reference shape 4 is created by giving two-dimensional coordinates on the same plane (1403). At this time, the characteristic part of the hotel D or the apartment E can be reflected in the reference shape 4 (1404).

  When the general area where the photograph 2 is taken is known, an area including the object 3 is selected from the existing digital elevation model M (1405). This is because, when the digital elevation model M covers a wide range, it is possible to reduce the load on the arithmetic processing by limiting the range to some extent. There is no need to select.

  Since the viewpoint and viewing angle for creating the model shape 5 are unknown, a predetermined initial value (default value) is input (1406). Spatial calculation processing is performed based on the digital elevation model M selected in the region and the viewpoint and viewing angle to which initial values are input, and a candidate for the model shape 5 that is a two-dimensional shape (that is, model shape candidate 5 ′). Is created (1407). The reference shape 4 and the model shape candidate 5 ′ are arranged in the same coordinate system (two-dimensional), and the error (absolute value) between the constituent point of the reference shape 4 and the constituent point of the model shape 5 is minimized by the least square method or the like. Both shapes are brought into a state of closest approach so as to become (1408). In this state, collation determination is performed based on the approximate condition (1409), and if the approximate condition is satisfied, that is, if it can be determined that collation has been performed, the process proceeds to the next step (1410). When the approximation condition is not satisfied, that is, when it is determined not to collate, the viewpoint / viewing angle is input again (1406). However, in this case, the viewpoint / viewing angle is input in place of the previous value. In this manner, viewpoint / viewing angle input (1406) to collation determination (1409) are repeatedly performed until the reference shape 4 and the model shape candidate 5 'are collated.

  If it is determined that the reference shape 4 and the model shape candidate 5 'are collated, the model shape candidate 5' is extracted as the model shape 5 (1410). The viewpoint and viewing angle corresponding to the model shape 5 extracted here, that is, the viewpoint and viewing angle adopted when the model shape 5 is calculated are the shooting condition values (the coordinates of the camera center position, the shooting posture) , The shooting range), and the position where the shot was taken can be specified (1411).

(System configuration)
FIG. 15 is an explanatory diagram showing the functions constituting the “imaging position acquisition device”. FIG. 15A is an explanatory diagram of the imaging position acquisition device when all the functions are integrated, and FIG. It is explanatory drawing of the imaging | photography position acquisition apparatus at the time of making it into a distributed type.

  As shown in FIG. 15A, the photographing position acquisition device 10 stores a main calculation function 11, a reference shape calculation function 12, and a digital elevation model database 13, and further acquires and stores a photograph 2. It also has a camera function. The reference shape calculation function 12 reads the image data of the photograph 2 acquired by the camera function, automatically generates a boundary (edge) of the object 3 based on a difference in luminance in the image data, The reference shape 4 is created by giving two-dimensional coordinates on the same plane. It is also possible to input a characteristic part and reflect it in the reference shape 4.

  The main calculation function 11 receives the reference shape 4 calculated by the reference shape calculation function 12 and also receives the digital elevation model M from the digital elevation model database 13. At this time, it is also possible to specify a necessary area and receive the digital elevation model M limited to the range from the digital elevation model database 13. Further, the main calculation function 11 has a function of executing steps 1406 to 1411 shown in the implementation flow of FIG. 14, and identifies the shooting point of the photograph 2 by executing these steps.

  As shown in FIG. 15B, the photographing position acquisition apparatus 10 has a camera function and a communication function, and uses a wireless (or wired) communication means to provide a host computer 7, a terminal computer 8, and a digital elevation model database 9 It can also be set as the structure connected with. The host computer 7, the terminal computer 8, and the digital elevation model database 9 are also connected by a wired or wireless network, and these can be installed at a location away from the shooting site.

  The terminal computer 8 receives the image data of the photograph 2 from the photographing position acquisition device 10 via the communication means, and executes the reference shape calculation function 12 to create the reference shape 4. The host computer 7 receives the reference shape 4 from the terminal computer 8 via the network, and further receives the digital elevation model M from the digital elevation model database 9. At this time, it is possible to specify a necessary area and receive the digital elevation model M limited to the range from the digital elevation model database 9. Further, the host computer 7 has a main calculation function 11, that is, a function of executing steps 1406 to 1411 shown in the execution flow of FIG. 14, and by specifying these, the shooting point of the photograph 2 is specified.

  The method for specifying the position of the object to be photographed, the method for obtaining the photographing position, the device for obtaining the photographing position, the program for identifying the position of the object to be photographed, and the program for obtaining the photographing position of the object to be photographed Since the position and shape of the photographed feature can be grasped, it is more effective particularly when used for a feature such as a stricken area where a person cannot enter. In addition, since it is possible to take quick measures against disasters, it is an invention that can quickly recover the functions of villages, roads, etc., and can be used not only industrially but also make a great social contribution.

DESCRIPTION OF SYMBOLS 1 Camera 2 Photograph 3 Object to be photographed 3a Photographed skyline 3b Photograph collapsing part 3c Photographed mountain part 3d Photographed hotel 3e Photographed apartment 3e
4 Reference Shape 4_n Reference Shape Component 5 Model Shape 5 ′ Model Shape Candidate 5_n Model Shape Component 6 Output Shape 7 Host Computer 8 Terminal Computer 9 Digital Elevation Model Database 10 Imaging Position Acquisition Device 11 Main Calculation Function 12 Reference Shape Calculation Function 13 Digital elevation model database A Mountain ridgeline (skyline)
B Collapse part D Hotel E Mansion F1 Characteristic part (ridge part)
F2 feature (Tanibe)
F3 feature (ridge)
G Model image plane Gr Topography L Laser M Digital elevation model P Aircraft Ph Camera image plane S Ground model X X-axis in real space Y Y-axis in real space Z Z-axis in real space x _a x-axis in photo coordinate system y _a photo y-axis in the x-axis y _a model coordinate system in the y-axis x _a model coordinate system in the coordinate system

Claims (10)

A method for identifying the position or / and shape of an object to be photographed in a photograph based on a digital elevation model,
Four or more feature points are selected from the shape of the object to be featured, and coordinates based on an arbitrary coordinate system in the photograph plane are given to these feature points,
Four or more points corresponding to the feature points are selected from a model shape obtained by projecting a ground surface model based on the digital elevation model, and an arbitrary coordinate system in a plane representing the model shape is selected for the points corresponding to the feature points. Give the coordinates of
By associating the feature point of the object to be photographed with the point corresponding to the feature point of the model shape, the arbitrary coordinate system in the photographic plane is associated with the arbitrary coordinate system in the plane representing the model shape,
A method for specifying a position of an object to be photographed, wherein spatial information is given to the object to be imaged based on the digital elevation model, and the position or / and shape of the object to be imaged is specified. A method for identifying the position or / and shape of an object to be photographed in a photograph based on a digital elevation model,
A two-dimensional reference shape is created based on a part or all of the shape of the object, and coordinates are given to the reference shape by an arbitrary coordinate system in the photographic plane,
Creating a two-dimensional model shape by projecting a surface model based on the digital elevation model, and giving coordinates to the model shape by an arbitrary coordinate system in a plane representing the model shape;
By associating the component of the reference shape with the component of the model shape, the arbitrary coordinate system in the photographic plane is associated with the arbitrary coordinate system in the plane representing the model shape,
A method for specifying a position of an object to be photographed, wherein spatial information is given to the object to be imaged based on the digital elevation model, and the position or / and shape of the object to be imaged is specified. A method for identifying the position or / and shape of an object to be photographed in a photograph based on a digital elevation model,
Creating a two-dimensional reference shape based on a part or all of the image of the object, and giving coordinates to pixels constituting the reference shape by an arbitrary coordinate system in a photographic plane;
A two-dimensional model shape is created from a projection image obtained by projecting a ground surface model based on a digital elevation model to which an image is assigned, and coordinates are set to pixels constituting the model shape by an arbitrary coordinate system in a plane representing the model shape. Grant,
By associating the pixel of the reference shape with the pixel of the model shape, the arbitrary coordinate system in the photographic plane is associated with the arbitrary coordinate system in the plane representing the model shape,
A method for specifying a position of an object to be photographed, wherein spatial information is given to the object to be imaged based on the digital elevation model, and the position or / and shape of the object to be imaged is specified. In the method for specifying the position of the object to be photographed according to claim 2 or 3,
By changing the viewpoint and viewing angle for projecting the ground model, create a shape that projects multiple ground models as model shape candidates,
A method for specifying a position of an object to be imaged, wherein a model shape corresponding to a reference shape is selected from a plurality of model shape candidates by collating the model shape candidate with a reference shape. A map displaying the position or / and shape of the object in the photograph based on the digital elevation model,
5. A display map of a feature to be displayed, wherein the subject feature whose position or / and shape is specified by the method for specifying the position of the feature to be recorded according to claim 1 is displayed. A program for specifying the position or / and shape of an object featured in a photograph based on a digital elevation model,
A computer-executable object position specifying program that causes a computer to execute the object specifying object position specifying method according to any one of claims 1 to 4. A method for acquiring a point where a photographed feature is photographed based on a photographed feature and a digital elevation model,
A two-dimensional reference shape is created based on a part or all of the shape of the object, and coordinates are given to the reference shape by an arbitrary coordinate system in the photographic plane,
By changing a viewpoint and a viewing angle for projecting a ground surface model based on the digital elevation model, a two-dimensional shape projecting a plurality of ground models is created as a model shape candidate, and in a plane representing these model shape candidates Coordinates are given to model shape candidates by an arbitrary coordinate system,
By collating the model shape candidate with a reference shape, a model shape corresponding to the reference shape is extracted from among a plurality of model shape candidates,
A shooting position acquisition method, wherein a viewpoint corresponding to the extracted model shape is set as a shooting point of a subject. A method for acquiring a point where a photographed feature is photographed based on a photographed feature and a digital elevation model,
Creating a two-dimensional reference shape based on a part or all of the image of the object, and giving coordinates to pixels constituting the reference shape by an arbitrary coordinate system in a photographic plane;
A plurality of two-dimensional shapes are created as model shape candidates from projection images obtained by projecting a plurality of ground models by changing the viewpoint and viewing angle for projecting the ground surface model based on the digital elevation model, and these model shapes Give coordinates to the pixels that make up the model shape by an arbitrary coordinate system in the plane representing the candidate,
By collating the model shape candidate with a reference shape, a model shape corresponding to the reference shape is extracted from among a plurality of model shape candidates,
A shooting position acquisition method, wherein a viewpoint corresponding to the extracted model shape is set as a shooting point of a subject. A program for obtaining position information of a point where a photographed feature is photographed based on a photographed feature and a digital elevation model,
A shooting position acquisition program that causes a computer to execute the shooting position acquisition method according to claim 7 or 8. A device that acquires position information of a point where a photographed feature is photographed based on a photographed feature and a digital elevation model,
Means for photographing the object and acquiring this image data;
Means for creating a reference shape according to claim 7 or 8, from the image data;
Means for creating a plurality of model shape candidates according to claim 7 or claim 8;
Means for extracting a model shape corresponding to a reference shape from a plurality of model shapes by collating the model shape with a reference shape;
Means for taking a viewpoint corresponding to the extracted model shape as a shooting point of the object to be photographed, and a photographing position acquisition device for the object to be photographed.

Images