JP2001041745A - Photogrammetry method and photogrammetry image processing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a high-precision survey drawing by a simple work. SOLUTION: First, second and third images are obtained by photographing from first, second and third camera positions M1, M2, M3. A target 30 is shifted into the three images. The third camera position M3 is nearer to the target 30 than the first and second camera positions M1, M2. A third open angle σ3 between a second optical axis O2 and a third optical axis O3 is larger than a first open angle σ1 between a first optical axis O1 and the second optical axis O2, and a second open angle σ2 between the first optical axis O1 and the third optical axis O3 is larger than the first open angle σ1 and the third open angle σ3. The camera positions and inclinations of the optical axes are calculated by using the three images, and three-dimensional coordinates of an optional object point are calculated based on the first and second images.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a photographing process and an image process in photogrammetry.

[0002]

2. Description of the Related Art Photogrammetry is widely used for creating maps,
It is also used as an extremely effective means for recording local situations such as on-site inspections of traffic accidents. Conventional photogrammetry uses a stereo camera in which two cameras are fixed while being separated from each other, and three-dimensional coordinates of each survey point are calculated from two images taken by the two cameras. Here, a stereo camera is a heavy and large equipment, and it is necessary to record detailed information such as camera position information, tilt angle, and actual measurement length of the subject in calculating three-dimensional coordinates. Is cumbersome and hard work. Also, there are many obstacles around the traffic accident site,
In many cases, a sufficiently wide shooting environment is not secured, and it is often difficult to perform on-site verification using a stereo camera.

[0003] The applicant of the present application has proposed a photogrammetry method using a monocular camera (Japanese Patent Application Laid-Open Nos. Hei 10-293026 and Hei 10-221072), and a pair of images (hereinafter referred to as "photographs") for improving the accuracy of the photogrammetry. , "Pair image"
(See JP-A-10-30702)
No. 5, JP-A-10-293026, JP-A-1
0-185563, JP-A-10-185562, JP-A-10-170263, JP-A-10-
No. 141951) has been proposed to realize efficient photogrammetry using simple equipment.

In such a photogrammetry method, a paired image of the same target and the object to be surveyed taken from arbitrary different directions is obtained. Based on the coordinates of survey points (hereinafter, referred to as “object points”) commonly captured in the images, the relative position of the camera and the inclination of the optical axis when each image was obtained (hereinafter, “camera parameters”) )
The three-dimensional coordinates of the object point are obtained. Then, a survey map is drawn using the three-dimensional coordinates of many object points.

[0005]

However, when there are two images including the same target, sufficient accuracy is not obtained with respect to camera parameters calculated based on the coordinates of an object point specified manually. As a result, there arises a problem that a highly accurate survey map cannot be obtained.

Conventionally, the surveying accuracy has been improved by increasing the number of images. However, not only is it difficult to find and specify the corresponding object point in all images, but also the surveying accuracy is taken into consideration. Choosing points is an extremely cumbersome task, even for skilled operators. In addition, since the number of objects to be processed increases, three
The calculation process for calculating the dimensional coordinates becomes complicated and requires a lot of time.

The present invention has been made in view of such a point, and an object of the present invention is to obtain a highly accurate survey map by a simple operation.

[0008]

SUMMARY OF THE INVENTION In a photogrammetry method according to the present invention, a target having a plurality of reference points is provided at a predetermined position, and first, second and third cameras which are respectively separated from the predetermined position by a predetermined distance. An image acquisition step of taking an image from a position to obtain first, second, and third images; and a first image acquisition step based on the photographic coordinates of the reference point in the first, second, and third images and the size and shape of the target. And a camera parameter calculation step of calculating camera parameters including the coordinates of the second camera position and the inclinations of the first and second optical axes at the first and second camera positions, and the first and second images. An object point specifying step of specifying an included object point and obtaining its photo coordinates; and, based on the camera parameters and the photo coordinates of the object point, converting the three-dimensional coordinates of the object point in a predetermined coordinate system. And the three-dimensional coordinates calculating step of leaving, is characterized in that it comprises a survey plan creation step of creating a survey map based on the three-dimensional coordinates of the object point. In such a photogrammetry method, the second opening angle formed by the first and third optical axes is larger than the first opening angle formed by the first and second optical axes, and the second opening angle is larger than the first opening angle. , The third opening angle formed by the second and third optical axes is large. By this photogrammetry method, the accuracy of the camera parameters is improved, and a highly accurate survey map can be obtained.

The photogrammetry method may further include an auxiliary point specifying step of specifying an auxiliary point that is commonly included in the first, second, and third images and that is different from the reference point and obtains the photo coordinates thereof. In this case, in the camera parameter calculation step, the photograph coordinates of the auxiliary points in the first, second, and third images are taken into account. Further, the auxiliary point may be located between the predetermined position where the target is placed and the third camera position. As a result, the camera parameters are corrected, and a more accurate survey map can be obtained.

In the photogrammetry method, preferably, the third
The magnification of the image of the image is higher than the magnification of the images of the first image and the second image. Thereby, the accuracy of the photographic coordinates of the reference point in the third image is improved, and a more accurate survey map can be obtained.

In the photogrammetry method, the direction of the optical axis at the time of shooting at the third camera position may be substantially parallel to a straight line passing through the first and second camera positions in a horizontal plane. Thus, the effective third camera position can be easily recognized at the time of shooting.

Further, the photogrammetry image processing apparatus according to the present invention provides a target having a plurality of reference points at a predetermined position, and separates the target from the first, second and third camera positions separated by a predetermined distance from the predetermined position. Image reading means for reading the first, second, and third images obtained by performing the photographing; and, based on the photographic coordinates of the reference points in the first, second, and third images and the size and shape of the target, Camera parameter calculating means for calculating camera parameters including coordinates of the first and second camera positions and inclinations of the first and second optical axes at the first and second camera positions;
And an object point designating means for designating an object point included in the second image to obtain the photograph coordinates thereof, and a three-dimensional object point in a predetermined coordinate system based on the camera parameters and the photograph coordinates of the object point. It is characterized by comprising three-dimensional coordinate calculating means for calculating coordinates and surveying map creating means for creating a surveying map based on the three-dimensional coordinates of the object point. Further, the first and second photogrammetric image processing apparatuses used in such a photogrammetric image processing apparatus.
And the third image, the second opening angle formed by the first and third optical axes is larger than the first opening angle formed by the first and second optical axes, and the second opening angle is larger than the first opening angle.
The third opening angle formed by the second and third optical axes is characterized by being large. With this photogrammetric image processing apparatus, the accuracy of camera parameters is improved, and a highly accurate survey map can be obtained.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a photogrammetry method and a photogrammetry image processing apparatus according to the present invention will be described below with reference to the accompanying drawings.

FIG. 1 is a diagram showing a photographing situation in an embodiment of a photogrammetry method, and is a horizontal plan view as viewed from vertically above. A target 30 is placed at a predetermined position on the road. The camera 10 captures images from the first camera position M1, the second camera position M2, and the third camera position M3, and the target 30 is imprinted on each image.

In the drawing, arrows extending from the camera positions M1 to M3 indicate optical axes O1, O2 and O of the photographing optical system of the camera 10 at the time of photographing at the camera positions M1, M2 and M3.
3 shows the direction. The camera 10 is a CCD (not shown)
And converts the optical image into an electronic image, that is, digital pixel data, and stores it in an image storage medium such as a memory card.

The angle formed by the optical axis O1 and the optical axis O2 is the first opening angle δ1, and the angle formed by the optical axis O1 and the optical axis O3 is the second opening angle δ.
2. If the angle formed by the optical axis O2 and the optical axis O3 is defined as a third opening angle δ3, the second opening angle δ2 is larger than the first opening angle δ1, and the third opening angle δ3 is larger than the first opening angle δ1. It is largely determined. In particular, the optical axis O3 is substantially parallel to the base line BL connecting the first and second camera positions as shown in FIG.

In preparing the survey map, a first image IM1 (FIG. 5) obtained at the first camera position M1 and a second image IM2 (FIG. 6) obtained at the second camera position M2 are used. Based on the target 30 projected on the first image IM1 and the second image IM2, each camera position M
1 and M2 are calculated. However, since the first image IM1 and the second image IM2 are electronic images by the CCD, the distance from the camera position along the optical axis direction of the camera 10 is increased. The error increases. Therefore, if the target 30 is separated from the camera positions M1 and M2, the positions of the camera positions M1 and M2 cannot be obtained with high accuracy, and a high-accuracy survey drawing cannot be obtained.

Here, as described above, the direction of the third optical axis O3 is set substantially parallel to the base line BL, that is, substantially perpendicular to the depth direction, and when obtaining the camera positions M1 and M2,
The accuracy of the camera positions M1 and M2 is improved using the third image IM3 obtained at the third camera position M3. As a result, a highly accurate survey chart can be obtained.

Further, the third camera position M3 is located at a position closer to the target 30 than the first camera position M1 and the second camera position M2.
, And the target 30 is projected on the third image IM3 at a magnification of an image larger than the first image IM1 and the second image IM2. Therefore, the depth error of the reference point is small, and the camera position and the like can be obtained with higher accuracy. The magnification of the image is a ratio between the size of the image formed by the photographing optical system (not shown) of the camera 10 and the actual size of the subject.

FIG. 2 is a perspective view of the target 30. The target 30 includes two black columnar members 31 and 33 connected to each other at one end, and has an L-shape.
Reference points 32 and 34 are provided on the upper surfaces of the ends of these columnar members.
And 36 are provided. The reference points 32, 34, and 36 are, for example, white point members, and black disk members 35, 37, and 39 are provided therearound. Thereby, the white reference points 32, 34, and 36 are emphasized, and the identification on the image is facilitated.

The target 30 is mounted on a road surface, and a plane including the reference points 32, 34 and 36 is substantially parallel to the road surface. The distance between the reference points 32 and 34 and the reference point 3
The distance between 4 and 36 is both LT (LT is known). By connecting the three reference points 32, 34 and 36, a right-angled isosceles triangle having the reference point 34 as a corner (90 degrees) is formed.

First and second camera positions M1 and M2
Are obtained, the two-dimensional coordinates of the three reference points 32, 34, and 36 in the images IM1 to IM3 are obtained. These two-dimensional coordinates and reference points 32, 34 and 36
The first camera position M1 and the second camera position M2 are calculated based on the actual size of the right-angled isosceles triangle formed by

FIG. 3 is a diagram showing a format of surveying photograph data stored in the image storage medium by the camera 10, and a plurality of photograph data (in FIG. 3, the (n-2) th to (n + 1) th photograph data are shown. ) Are sequentially stored. 1
Photo data (n-th) are header H and image data IM
D, a spare space SP is provided after the image data IMD to separate it from the adjacent photo data. The image data IMD is a digital pixel data sequence.

The header H includes an image name H1, an identification number H2, a photographing date / photographing condition H3, a rotation angle / movement amount H4, and the like. The shooting dates of the image names H1 and H3 are manually input in the camera. The identification number H2 is, for example, 1
It includes a photographing point number that is incremented by one and a target position number that is incremented by one when the target 30 moves, and is used for identification of an image in a photogrammetric image processing apparatus described later.

The photographing conditions of H3 are inputted from the camera at the time of photographing, and include the focal length f of the camera, the angles of view Δh and Δv in the horizontal and vertical directions, the resolution rp of the CCD, and the like. H
The rotation angle of 4 includes the rotation angle and azimuth of the target 30 on the horizontal plane at the time of shooting, and the movement amount is the movement amount from the first target position.

Next, with reference to FIG. 4, an overall configuration of a photogrammetric image processing apparatus that reads an image from an image storage medium and creates a survey map will be described. FIG. 1 is a block diagram showing the entire configuration of the photogrammetric image processing apparatus.

The photogrammetric image processing device includes a display device 10,
It has an input device 12, such as a keyboard and a mouse, an image storage medium 13, and a CPU 14, which are directly or indirectly connected to a bus 15.

The CPU 14 is provided with an input state management section 41, a display state management section 42, a calculation control section 43, and a data management section 44, where necessary management, calculation and processing are executed. The input device 12 is connected to an input device control device 17 connected to the bus 15, whereby the input device 1
2 is transferred to the bus 15 and the input device 12
Is set. The image storage medium 13 is inserted into a storage medium control device 18 such as a memory card reader,
As a result, the photograph data (FIG. 3) stored in the image storage medium 13 is appropriately read.

Further, a working memory 19 and a display memory 20 are connected to the bus 15, and the working memory 19 is used as a cache memory or the like in calculation and processing of the CPU 14, and the display memory 20 stores contents to be displayed on the display device 10. Will be retained. The display device 10 is connected to a display device control device 16 connected to the bus 15, and converts digital data in the display memory 20 into analog RGB signals for the display device 10.

An input state management unit 41 of the CPU 14 manages settings of the input device 12 and converts input information such as coordinates of a mouse pointer moving on the screen with a mouse, characters input from a keyboard, and the like into a predetermined digital form. Convert to data. The display state management unit 42 manages the content to be displayed on the display device 10, and changes the display content when a setting related to the display is changed. The arithmetic control unit 43 is used for calculation processing of a camera position and the like described later. The data management unit 44 manages data contents read from the image storage medium 13, and manages image data, various coordinate data obtained by calculation, data of a plotted survey map, and the like.

In the photogrammetric image processing apparatus shown in FIG. 4, first, the photographic data of the images IM1 to IM3 are read from the image storage medium 13, and based on these, the first, second and third camera positions M1 to M1, respectively. 3 and the inclinations of the first, second and third optical axes O1 to O3 are calculated.

FIG. 5 is a conceptual diagram showing a first image IM1 obtained at the camera position M1, and FIG.
2 shows the second image IM2 obtained at the camera position M3, and FIG. 7 shows the third image IM3 obtained at the camera position M3. The images IM1 to IM3 have the reference points 32, 3
Only 4 and 36 are shown, reference points 32, 34 and 3
The right-angled isosceles triangle formed by 6 is indicated by hatching.

In the first image IM1 of FIG. 5, a two-dimensional orthogonal coordinate system (Xp, Yp) having the origin at the imaging center c is set, and the coordinates of the image point of the reference point 32 (hereinafter referred to as photographic coordinates). Is called p11 (px11, py11). Similarly, the photograph coordinates of the reference point 34 are p21 (px21, py2
1) The photograph coordinates of the reference point 36 are p31 (px31, py
31).

Similarly, in the second image IM2 of FIG.
The photograph coordinates of the reference points 32, 34 and 36 are p12 (px
12, py12), p22 (px22, py22) and p32 (px32, py32), and in the third image IM3 of FIG. 7, the reference points 32, 34 and 36
Are photograph coordinates of p13 (px13, py13), p23
(Px23, py23) and p33 (px33, py23)
33).

Actually, the image IM is stored in the image storage medium 13.
When 1 is read by the CPU 14, image processing such as binarization processing is performed, and the reference points 32, 34, and 36 are automatically extracted, and their photograph coordinates pi (xpi, ypi) (i = 1 to
3) is obtained and stored in the working memory 19.

FIG. 8 is a diagram conceptually showing the relationship between the camera positions M1 to M3, the three images IM1 to IM3, and the target 30. Here, the imaging planes of the camera 10 that substantially match the images IM1 to IM3 are set on the screens SIM1 to SIM3.

The two-dimensional coordinates (hereinafter referred to as screen coordinates) of the image points of the reference points 32, 34 and 36 projected on the first screen SIM1 are s11 (sx11,
sy11), s21 (sx21, sy21), s31
(Sx31, sy31). This screen coordinate s
It is preferable that 11, s21, and s31 match the photograph coordinates p11, p21, and p31 specified in the actual image, but strictly, there is an error between the two.

The screen coordinates of the reference points 32, 34 and 36 on the second screen SIM2 are s12 (sx12, sy12) and s22 (sx22,
sy32) and s32 (sx32, sy32), and the screen coordinates of the reference points 32, 34, and 36 on the third screen SIM3 are s13 (sx13, sx32, sy32), respectively.
sy13), s23 (sx23, sy23), s33
(Sx33, sy33).

With reference to FIG. 8, a method of calculating the camera position using the reference points 32, 34 and 36 and the inclination of the optical axis thereof will be described. Here, i is a parameter indicating the reference point (i = 1 to 3), and j is a parameter corresponding to the camera position (j = 1 to 3).

In the state where the reference points 32, 34 and 36 are imaged on the j-th screen SIMj, the j-th optical axis Oj
Passes through the imaging center C of the j-th camera position Mj and the j-th screen SIMj, and the image points sij of the reference points 32, 34, and 36 projected on the j-th screen SIMj become the j-th camera position Mj and the respective reference points 32, 34 and 36, respectively.

The j-th screen set in the j-th screen SIMj
The origin of the screen coordinate system (Xs, Ys) is the photographing center C
Matches. In FIG. 8, the j-th camera position Mj
Is set as the camera coordinate system (Xc, Yc, Zc) with the origin being (0, 0, 0), and the Xc axis and the Yc axis are respectively the Xs axis and the Ys axis of the j-th screen coordinate system. , And the Zc axis coincides with the optical axis Oj.

Reference points 32, 34, in the camera coordinate system
The three-dimensional coordinates of 36 are represented by Pcij (Pcxij, Pcyi
(j, Pczij), the relationship between the screen coordinates sij (xsij, ysij), which are image points, and the camera coordinates (Pcxij, Pcyij, Pczij) of the reference point is represented by Expressions (1) and (2). In Expressions (1) and (2), f is the focal length of the camera 50.

(Equation 1)

Further, in FIG. 8, the three-dimensional coordinates having the reference point 34 of the target 30 as the origin are represented by a reference coordinate system (X
b, Yb, Zb). Xb of reference coordinate system
The axis and the Zb axis are along the sides of the reference points 34 and 32 and the reference points 34 and 36, respectively, and the Yb axis is perpendicular to the Xb axis and the Zb axis. When the target 30 is placed on an inclined surface, the inclination of the Xb axis and the Zb axis is corrected based on the rotation angles of the Xb axis and the Zb axis with respect to the horizontal plane. Therefore, the Yb axis is made to coincide with the vertical direction, and X
The b-Zb plane is made substantially coincident with the horizontal plane.

Here, if the coordinates of the j-th camera position Mj in the reference coordinate system are defined as (ΔXj, ΔYj, ΔZj) and the inclination of the j-th optical axis Oj as (αj, βj, γj), the camera coordinates Pcij (Pcxij) , Pcyij, Pczi
j) and reference coordinates Pbij (Pbxij, Pbyij, P
bzij) is represented by equation (3).

(Equation 2)

R in equation (3) is a rotation matrix, and as shown in equation (4), j-th optical axis Oj (Zc
The direction cosine of (axis) is represented by cosαj, cosβj, and cosγj. In Expression (3), Δ is the coordinate origin moving amount, and the reference coordinates (ΔXj, ΔXj) of the j-th camera position Mj.
Yj, ΔZj).

Reference points 32, 34, 3 in the reference coordinate system
6, the reference coordinates are Pb1 (−LT, 0, 0),
Pb2 (0,0,0) and Pb3 (0,0, LT), and these values are substituted into equations (1) to (4) to obtain screen coordinates s11 (xs11, ys11) and s21 (xs
21, ys21) and s31 (xs31, ys31) are obtained.

The screen coordinates s11, s21 and s31 obtained by these equations and the photograph coordinates pi1 (xpi1, ypi1) in the first image IM1 (FIG. 5).
In order to minimize the error Φ (Equation (5)) from (i = 1 to 3), the first camera position M1 (Δ
X1, ΔY1, ΔZ1) and the inclination (α) of the first optical axis O1.
1, β1, γ1) are calculated. Since the successive approximation method is known, it will not be described in detail.

(Equation 3)

At the j-th camera position, six parameters consisting of the three-dimensional coordinates Mj (ΔXj, ΔYj, ΔZj) and the inclination (αj, βj, γj) of the j-th optical axis Oj are called camera parameters.

As described above, when obtaining the camera parameters at the first and second camera positions, three images IM
1-3 are used. As a result, it is possible to obtain camera parameters with higher accuracy than calculation using only two images.

Although only three reference points are used in the calculation of the camera parameters described above, in addition to the reference points, one or more auxiliary points commonly printed on three images are actually used in addition to the reference points. Points are added. The auxiliary points are reference points 32, 3
Any point located outside the right-angled isosceles triangle formed by 4 and 36 may be sufficient, and it is more preferably located between the third camera position M3 and the target 30. Since the auxiliary point is closer to the camera position M3 in the optical axis direction than the reference points 32, 34, and 36, the depth error on the third image IM3 is further reduced. By adding such an auxiliary point having a small depth error to the calculation, the accuracy of the camera parameters can be further improved.

FIGS. 9 to 11 are flowcharts showing an object point coordinate calculation routine for obtaining three-dimensional coordinates of an object point using the above-described camera parameter calculation method. In this object point coordinate calculation routine, photograph data of three images IM1 to IM3 are read by the arithmetic control unit 43 of the CPU 14,
This is executed after the photograph coordinates of the reference point are stored in the work memory 19.

First, in step S100, a camera parameter calculation subroutine is executed, and the first and second camera parameter calculation subroutines are executed.
Camera parameters at the camera positions M1 and M2 are obtained. Then, in step S204, the image IM
In 1 and IM2, image points corresponding to one object point are manually specified, and the photographic coordinates of the two image points are determined. In this manual designation, two images IM are displayed on the screen of the display device 10 (FIG. 4).
1 and IM2 are displayed in parallel, and when an image point on the screen is designated by clicking with the input device 12 such as a mouse, the photograph coordinates are determined based on the mouse coordinates of the designated position.

Next, step S206 is executed, and the three-dimensional coordinates of the object point are calculated based on the two camera parameters obtained in step S100 and the photograph coordinates of the corresponding two image points obtained in step S204. Desired. The calculation of the three-dimensional coordinates is obtained using the equations (1) to (5) used for calculating the camera parameters described above.

When the calculation of the three-dimensional coordinates of the object point is completed in step S206, the process returns to step S202, and it is determined whether or not the surveying has been completed, that is, whether or not the object point designation has been completed. For example, when the end button is clicked on the operation screen (not shown) of the display device 10, it is determined to be the end, and the object point coordinate calculation routine ends.

FIGS. 10 and 11 are flowcharts showing details of the camera parameter calculation subroutine (step S100) of FIG.

First, in step S102, the reference coordinates of the reference points 32, 34, and 36 in the reference coordinate system are Pb1 (-LT, 0, 0) and Pb2 (0, 0,
0) and Pb3 (0, 0, LT). Next, 3
The photograph coordinates pij (pxij, pyij) of the reference points 32, 34, 36 in the images IM1 to IM3 (i = 1 to 3,
j = 1 to 3) are read from the working memory 19. Then, in step S106, the three images IM1 to IM3
An appropriate initial value is given to each of the three camera parameters corresponding to.

Subsequently, in step S108, im
The sum of the number of reference points (= 3) and the number n of auxiliary points is substituted for ax, and 4 is substituted for parameter i assuming that designation of reference points (i = 1 to 3) has been completed. Note that the number n of auxiliary points is determined in advance.

In step S112, the three-dimensional coordinates (Pbx4, Pby) of the auxiliary point Pb4 in the reference coordinate system
4, Pbz4) is given an initial value, and step S114 is performed.
, The photographic coordinates p of the auxiliary points in the images IM1 to IM3
4j (px4j, py4j) (j = 1 to 3) are obtained.

More specifically, three images IM1 to IM3 are displayed side by side on the screen of the display device 10, and the image points of the auxiliary points which are commonly captured in each of the images IM are respectively click-designated. Thus, the photograph coordinates p4j (j = 1 to 3) of the three image points are determined.

Subsequently, in step S116, the parameter i is incremented by one, and in step S11
Return to 0. In step S110, the parameter i is set to ima
x or less, it is determined whether the parameter i is ima
If it exceeds x, the process moves to step S118. That is, steps S110 to S116 are performed n times, whereby the reference coordinates of the n auxiliary points are set to the initial values, and the photograph coordinates of the n auxiliary points are obtained.

In step S118, the screen coordinates sij (sxij, syi) for imax points
j) (i = 1 to imax, j = 1 to 3) are calculated, and in step S120, these screen coordinates sij
(Sxij, syij), steps S104 and S104
114, the photograph coordinates pij (pxij, py)
ij) (i = 1 to imax, j = 1 to 3) and the camera parameter that minimizes the difference is calculated. Step S12
When 0 ends, the camera parameter calculation subroutine ends. The camera parameters are obtained corresponding to the images IM1 to IM3, but the camera parameters used in step S206 are only those corresponding to the images IM1 and IM2.

As described above, in the present embodiment, the two images IM1 and IM2 are used for creating the survey map, but when the camera parameters are calculated, the third image IM3 is further added. The calculation time required to obtain the three-dimensional coordinates of the object point can be reduced with high accuracy, and a coordinate value with high accuracy can be obtained.

Further, by adding points other than the reference points 32, 34, and 36 as auxiliary points to the camera parameter calculation, the accuracy of the camera parameters can be improved, and a highly accurate survey map can be obtained.

[0064]

According to the present invention, a highly accurate survey drawing can be obtained by a simple operation.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating an embodiment of a photogrammetry method according to the present invention, and is a diagram illustrating a shooting situation.

FIG. 2 is an enlarged perspective view showing a target shown in FIG. 1;

FIG. 3 is a conceptual diagram showing a format of photographic data processed by the photogrammetric image processing apparatus.

FIG. 4 is a block diagram showing an overall configuration of a photogrammetric image processing apparatus according to the present invention.

FIG. 5 is a conceptual diagram showing an image at a first camera position.

FIG. 6 is a conceptual diagram showing an image at a second camera position.

FIG. 7 is a conceptual diagram showing an image at a third camera position.

FIG. 8 is a schematic diagram illustrating a relationship between a camera position, a target, and a screen.

9 is a flowchart showing an object point coordinate calculation routine executed by the CPU shown in FIG. 4;

FIG. 10 is a flowchart illustrating a first half of a camera parameter calculation subroutine illustrated in FIG. 9;

11 is a flowchart showing the latter half of the camera parameter calculation subroutine shown in FIG.

[Explanation of symbols]

 10 Camera 30 Target 32, 34, 36 Reference point M1, M2, M3 Camera position O1, O2, O3 Optical axis δ1 First opening angle δ2 Second opening angle δ3 Third opening angle

Claims (6)

[Claims] 1. A target having a plurality of reference points is provided at a predetermined position, and images are taken from first, second, and third camera positions that are respectively separated by a predetermined distance from the predetermined position. And an image acquisition step of obtaining a third image; and coordinates of the first and second camera positions based on photographic coordinates of the reference point in the first, second, and third images and dimensions and shape of the target. Calculating camera parameters including first and second tilts of the optical axis at the first and second camera positions; and specifying an object point commonly included in the first and second images. And calculating the three-dimensional coordinates of the object point in a predetermined coordinate system based on the camera parameters and the photo coordinates of the object point. Three-dimensional coordinate calculation step, and a survey map creation step of creating a survey map based on the three-dimensional coordinates of the object point, wherein the first opening angle formed by the first and second optical axes is A photograph wherein a second opening angle formed by the first and third optical axes is large, and a third opening angle formed by the second and third optical axes is larger than the first opening angle. Survey method. 2. An auxiliary point specifying step of specifying an auxiliary point different from the reference point and obtaining a photograph coordinate of the auxiliary point, the auxiliary point specifying step being included in the first, second and third images, and the camera parameter calculating step 2. The photogrammetry method according to claim 1, wherein photocoordinates of the auxiliary points in the first, second, and third images are taken into account. 3. The method according to claim 2, wherein the auxiliary point is located between the predetermined position and the third position.
3. The camera according to claim 2, wherein the camera is located between the camera and the camera.
The photogrammetric method described in 1. 4. The photogrammetry method according to claim 1, wherein the magnification of the image of the third image is higher than the magnification of the images of the first image and the second image. 5. The apparatus according to claim 1, wherein, in a horizontal plane, a direction of an optical axis at the time of shooting at the third camera position is substantially parallel to a straight line passing through the first and second camera positions. The photogrammetric method described. 6. A first target obtained by providing a target having a plurality of reference points at a predetermined position and performing photographing from first, second, and third camera positions separated by a predetermined distance from the predetermined position. Image reading means for reading the second and third images; and the first and second camera positions based on the photograph coordinates of the reference point in the first, second and third images and the size and shape of the target. And camera parameter calculation means for calculating camera parameters including the coordinates of the first and second camera axes at the first and second camera positions, and an object commonly included in the first and second images. Object point designating means for designating a point and obtaining the photograph coordinates thereof; and three-dimensional coordinates of the object point in a predetermined coordinate system based on the camera parameters and the photograph coordinates of the object point. A three-dimensional coordinate calculating means for calculating; and a surveying map creating means for creating a surveying map based on the three-dimensional coordinates of the object point, wherein a first opening angle formed by the first and second optical axes is provided. A second opening angle formed by the first and third optical axes is larger, and a third opening angle formed by the second and third optical axes is larger than the first opening angle. Photogrammetric image processing device.