JPH08101038A - Motion identification device, motion identification method, shape estimation device, and shape measurement method - Google Patents

Abstract

(57) [Summary] (Correction) [Purpose] To measure the position, orientation, motion parameters, etc. of an object even when there is no information about the shape. An image input device for inputting image data of a moving object, an image processing unit for extracting an object from the image data, a position calculation unit for measuring a position to a feature point with respect to an image plane, and a feature point. The motion identifying unit 5 for determining the posture of the object and the motion parameter of the object based on the time-series position data is provided, and the determined position, posture, and motion parameter are output. In addition, the shape of the object is estimated along with the motion.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a motion identification device and a motion identification method for obtaining the posture, position, and motion parameters (translational velocity, rotation axis, angular velocity, etc.) of an object that is moving at a constant velocity or at a constant velocity. The present invention relates to a shape estimating device and a shape measuring method for obtaining the shape of the object.

[0002] In space development and the like, there are cases where it is necessary to capture an artificial satellite or the like floating in outer space.
In order to perform such work automatically, first, it is necessary to approach the artificial satellite to be captured, so it is necessary to measure its position, attitude, and motion parameters (translational velocity, angular velocity). In addition, it is necessary to capture dust not only in artificial satellites but also in outer space, and in such cases it is necessary to measure its position and movement. When capturing artificial satellites and space debris, it is also necessary to measure their three-dimensional shape (surface shape, three-dimensional shape, etc.).

[0003]

2. Description of the Related Art Conventionally developed technology is a measurement method based on a model of an artificial satellite, in which the size and shape of the artificial satellite are known, or the artificial satellite is preliminarily marked with a distance measuring mark. Originally, the position, attitude, and motion parameters of satellites were measured. Also,
The shape of the object was estimated only based on the image data for those with known motion parameters (translational velocity, angular velocity, etc.).

[0004]

In the prior art, when there is no model of the shape of the artificial satellite, or when the model is actually deformed and the model cannot be used, the shape of the satellite is like dust in space. If there was no information about the, it was not possible to measure the position, posture, motion parameters, etc. of the object.

The present invention provides a motion identifying apparatus and a motion identifying method, a shape estimating apparatus and a shape measuring method capable of measuring the position, posture, motion parameter and the like of an object even when there is no information about the shape. The purpose is to

[0006]

[Means for Solving the Problem] Positions, postures, and motion parameters of an object are measured by time-series image data and distances of the characteristic points by measuring the distances of the characteristic points by using the corners and patterns of the object as characteristic points. (Translational velocity, rotation axis, angular velocity, etc.) can be measured. For example, image data of an object is obtained by an image input device such as an image pickup device, and feature points are extracted. Further, at that time, if the feature points of the object can be extracted from a plurality of viewpoints (positions of the image input device), the distance and direction from the image input device to the feature points of the object can be calculated by the principle of triangulation. Alternatively, a distance image of the object can be generated by irradiating the object with a laser or the like, and the angle (feature point) of the object can be extracted from the distance image to measure the distance and direction of the feature point.

If the motion parameter of the object is obtained,
As for the shape of an object, the surface shape can be measured by time-series data of the contour (silhouette) on the image plane and the posture, position, and motion parameters.

FIG. 1 shows the basic configuration of the present invention. In FIG. 1, reference numeral 1 denotes an object to be measured, which has a uniform velocity motion and a uniform angular velocity motion.

F is a motion identification device. An image input device 2 is an image pickup device having a plurality of viewpoints, or a laser device for measuring a distance between an image pickup device having one viewpoint and an object. The image input device also includes an interface for inputting stored data from the image data storage device 10 such as a magnetic disk device for storing captured image data.

K is a motion identification means. K'is a shape estimation means. An image processing unit 3 extracts the feature points of the object based on the left and right image data of the object.

A position calculation unit 4 associates the left and right characteristic points obtained from a plurality of (left and right) viewpoints,
The distance and direction between the feature point and the image plane are calculated to obtain the position of the feature point. Alternatively, the position and the position may be obtained by measuring the distance and direction of the characteristic point with an image pickup device for one viewpoint and a distance measuring device such as a laser device.

Reference numeral 5 denotes a motion identification unit, which obtains the position, orientation, and motion parameters of the object based on the time-series position data of the characteristic points. 6 is a shape estimation unit,
The shape of the object is estimated based on the time-series data of the contour of the image obtained by the image input device 2 and the position, posture, motion parameter and the like.

Reference numeral 7 is an output unit,
It outputs motion parameters, shapes, etc. An image data storage device 10 stores image data obtained by an imaging device or the like.

[0014]

The operation of the basic configuration of the present invention shown in FIG. 1 will be described. The case where the image input device has a plurality of viewpoints (for example, there are two left and right imaging devices) will be described as an example.

The image input device 2 inputs the left and right image data of the object 1. The image processing unit 3 extracts the characteristic points (the corner points of the image, etc.) of the left and right image data. The position calculation unit 4 associates the characteristic points of the left and right image data with each other, calculates the distance and direction between each characteristic point and the imaging device, and obtains the position of the characteristic point.

The motion identifying section 5 determines the posture and motion parameters of the object based on the time series data of the positions of the characteristic points. For example, the distance is calculated for the three characteristic points A, B, and C. Then, with the point A as the origin, the vector AB is x
If the cross product of the axes, the vector AB and the vector AC, is the z axis, and the cross product of the x axis and the z axis is the y axis, the Cartesian coordinate system (center of gravity coordinate system) of the object is determined. The origin of the barycentric coordinates is, for example, the average of the coordinates of the feature points. It is possible to identify the motion by calculating the direction vector and the position of the origin of the orthogonal coordinates of the coordinate axes of the barycentric coordinate system with respect to the camera coordinates by the coordinates of each feature point, and determining the pose and motion parameters of the object from the time series data. .

The output unit 7 outputs the position, posture, and motion parameters obtained by the motion identification unit 5 to a posture control device for an object or the like. Next, the shape measuring method of the shape estimating apparatus of the present invention will be described with reference to FIG.

FIG. 2 (a) is a sectional view of a plane perpendicular to the rotation axis, and FIG. 2 (b) shows an image plane parallel to the rotation axis. FIG. 2 shows a state in which the viewpoint A, the viewpoint B, and the viewpoint C rotate around the object 20 (in the actual device, the viewpoint is fixed and the object 20 rotates, but as shown in FIG. This is relatively the same as when the object 20 is fixed and the object 20 rotates).

In FIGS. 2A and 2B, 20 is an object. Reference numeral 21 is a rotation axis of the object.

Reference numeral 22 denotes a photographing range, which is a range of the object 20 photographed at the viewpoint A (24). An image plane 23 is an image plane at time t ₁ . A viewpoint A is a viewpoint at time t ₁ .

A viewpoint B is a viewpoint B at time t ₂ . Reference numeral 26 is a viewpoint C, which is a viewpoint at time t ₃ (time is assumed to pass in the order of t ₁ , t ₂ , and t ₃ ).

Reference numeral 30 denotes an image plane, which is parallel to the rotation axis viewed from the direction of the viewpoint A (24). Reference numeral 31 is an image, which is an image of the object 20 reflected on the image surface 30. 32
Is the contour and is the contour of the image 31.

The time-series data of the contours of the image 31 such as the contour of the image 31 at time t ₁ , the contour of the image at time t ₂ , the contour of the image at time t ₃ , and the posture, position, and motion parameters of the object 20 The shape can be estimated.

The shape estimation unit 6 inputs the image data, extracts the contour data at each time, and estimates the shape of the object based on the posture, position, and motion parameters of the object obtained by the motion identification unit.

In the above, an image pickup device having one viewpoint and a laser device for measuring the position of an object are provided, feature points are extracted from image data, and the feature points are irradiated with laser light to generate reflected light. You may make it receive light and measure a position.

The image data taken by the image pickup device may be stored in the image data storage device 10 such as a magnetic disk device, and the image data of the image data storage device 10 may be extracted and input.

According to the present invention, it is possible to obtain motion parameters such as the position, orientation, translational speed, rotation axis, and angular velocity of an object whose shape is not known, which is moving at a constant velocity or at a constant angular velocity. In addition, shape estimation can be performed on an object whose motion parameters such as shape, posture, position, translational velocity, and angular velocity are unknown.

[0028]

EXAMPLE FIG. 3 shows an example of the motion identifying apparatus of the present invention. In FIG. 3, reference numeral 50 denotes a floating object such as an artificial satellite, which has a translational velocity (dv) and an angular velocity (ω) about a rotation axis (kv).
Is. dv is a velocity vector and kv is a rotation vector.

Reference numeral 51 is a motion identification device. Image input devices 52 and 53 are left and right image pickup devices. 54,5
Reference numeral 5 is a left and right image processing unit for obtaining characteristic points.

Reference numeral 56 denotes a position calculation unit, which correlates the left and right characteristic points, calculates the distance and direction between the characteristic points and the image input device, and obtains the position. Reference numeral 57 is a motion identification unit that obtains the position, orientation, and motion parameters (translational velocity (dv), rotation axis (kv), angular velocity (ω)) of the floating object 50.

An output unit 58 is an interface for outputting the result obtained by the motion identifying unit 57. Reference numeral 59 denotes a magnetic disk device, which stores the result obtained by the motion identification unit 57, the image data of the image input device 52, and the like.

Reference numeral 60 is a display. 61 'is a printer. The output result is also output to an external device such as a posture control device for the floating object 50.

The operation of the configuration shown in FIG. 3 will be described. The left and right image input devices 52 and 53 capture the floating object 50. The left and right imaging data are respectively processed by the image processing unit 54 and the image processing unit 55
And the feature points are extracted.

The position calculation unit 56 associates the characteristic points of the left and right image data with each other, and the characteristic points and the image input devices 52 and 53.
Calculate the distance and direction to and find the position. The motion identification unit 57 calculates the posture and motion parameters of the floating object 50 based on the time-series data of the positions of the characteristic points. The calculation result of the motion identification unit 57 is output to the display 60, the printer 61 ′, etc. as necessary. In addition, it is stored and saved in the magnetic disk device 59.

FIG. 4 shows an embodiment of the system configuration of the motion identification device of the present invention. In FIG. 4, parts common to those in FIG. 3 are indicated by common numbers. In FIG. 4, 51 is a motion identification device.

An image input device 52 is an image pickup device or the like. Reference numeral 58 is an output unit. Reference numeral 59 is a magnetic disk device.

Reference numeral 60 is a display. 61 'is a printer. Reference numeral 65 is a CPU.

Reference numeral 66 is a memory. In the memory 66, K is a motion identification means. In the memory 66, 54 is an image processing unit, which is a program for extracting features from image data.

A position calculation unit 56 is a program for associating feature points of left and right image data with each other, calculating the distance and direction between the image input device and the object, and obtaining the position. A motion identification unit 57 is a program that performs coordinate system conversion, posture calculation, motion identification calculation, and the like.

The operation of FIG. 4 is similar to that of FIG. Details of the operation of the motion identification means will be described later. FIG. 5 is an explanatory diagram of the motion identification calculation method of the present invention.

FIG. 5 (a) shows the coordinate system of the first embodiment.
(b) is the coordinate system of the second embodiment. [Calculation Method for Motion Identification of Embodiment 1] In FIG.
Reference numeral 30 denotes an image plane of the image input device, x-axis, y-axis and z-axis are orthogonal coordinates of the camera coordinate system, x-axis and y-axis are on the image plane 30, and z-axis is perpendicular to the image plane 30 ( Optical axis direction).

Reference numeral 50 is an object. At the object 50,
A, B, and C are feature points, and G is the center of gravity (average of position coordinates). The object 50 is a floating object such as an artificial satellite.

The position vector of the feature point A at time f is set to Av _f
And Av _f = (A _x , A _y , A _z ). Similarly, the position vector of the feature point B at time f is Bv _f , and the position vector of the feature point C at time f is Cv _f .

Tv _f is a vector representing the position of the center of gravity G, and the components in the x, y, z coordinate system are (t _x , t _y , t _z )
Is. nv is a direction vector in the x-axis direction of the barycentric coordinate system. mv is a direction vector in the y-axis direction of the barycentric coordinate system. ov is a direction vector in the z-axis direction of the barycentric coordinate system. nv is taken in the direction of the AB vector. The position that is the origin of the vector representing the vector kv of the rotation axis is also determined in advance.

Nv = (n _x = i _x , n _y = i _y , n _z = i _z ) av = (a _x = j _x , a _y = j _y , a _z = j _z ) ov = (o _x , O _y , o _z ).

Nv, av and ov at time f are respectively nv
Letting _f , av _f , and ov _f , nv _f = Bv _f −Av _f av _f = (Bv _f −Av _f ) × (Cv _f −Av _f ) ov _f = nv _f × av _f . (“×” is an outer product) Rv _f represents the posture of the object 50, and is represented by the following matrix having nv, av, and ov as components.

[0047]

[Equation 1]

Is as follows. kv is the rotation vector of angular velocity ω,
dv is a translation vector of the object 50. mv is a vector representing the position of the center of gravity G with respect to the rotation center coordinate system.

Let tv and Rv at time f be tv _f and R, respectively.
_Let v _f be the rotation conversion matrix of the vector kv (angular velocity ω) be Rω _{, then} Rv _{f + 1} = RωRv _{f, and} thus Rω = Rv _{f + 1} Rv _f ⁻¹ .

At time f (f frame), the coordinates (P _fx , P _fy , P _fz ) of the P-th feature point in the camera coordinate system are P
_When expressed as _fx = x _fp , P _fy = y _fp and P _fz = z _fp , the barycentric coordinates of each feature at time f are as follows.

[0051]

[Equation 2]

The position of the feature point (camera coordinate system) viewed from the center of gravity is Og-XgYgZg in the center of gravity coordinate system.

[0053]

(Equation 3)

It is Next, according to the coordinates of each time in the barycentric coordinate system,

[0055]

[Equation 4]

When a matrix called

[0057]

(Equation 5)

It becomes Here, (n _xf , _nyf , _nzf ) is a direction vector with respect to the camera coordinate system on the Xg axis at time f.

(O _xf , o _yf , o _zf ) are direction vectors with respect to the camera coordinate system on the Yg axis at time f. (S _xp , S
_yp , S _zp ) is the position of the feature point in the barycentric coordinate system. The pose of the object is

[0060]

(Equation 6)

Since Rv _{f + 1} = RωRv _f at times f and f + 1, the rotation conversion matrix Rω is calculated by Rω = Rv _{f + 1} Rv _f ^-1 .

If the position of the center of gravity at time f is tv _f ,

[0063]

(Equation 7)

Is as follows. From the above relationship, the direction (kv) of the rotation axis with respect to the camera coordinates, the position (gv), the translational velocity (dv), and the angular velocity ω are obtained. kv and ω are Rω = Rv
It can be obtained from _{f + 1} Rv _f ^-1 .

Since gv _{f + 1} = gv _f + dv tv _f = gv _f + Rv _f · mv, tv _f = gv ₀ + dv · f + Rv _f · mv.

Therefore, t _x = gv _0x + f · dv _x + n _x · mv _x + o _x · mv
_{y + a x · mv z t} y = gv 0y + f · dv y + n y · mv x + o y · mv
_{y + a y · mv z t} z = gv 0z + f · dv z + n z · mv x + o z · mv
By the relation of _y + a _z · mv _z or more, tv _f and Rv _f at three times are measured,
Solving the simultaneous equations obtained from the measured values, gv ₀ ,
It is possible to obtain dv and mv.

FIG. 6 is a flowchart (1) of the system configuration embodiment of the present invention. S1 and S1 'Input left and right image data from the image input device. The S2, S2 'image processing unit 54 extracts respective features from the left and right image data.

The S3 position calculator 56 associates the left and right feature points. S4 The position calculation unit 56 calculates the distance and direction of the feature point to obtain the position. The S5 motion identification unit 57 converts the coordinates (camera coordinate system) of the position of each feature point into the barycentric coordinate system.

S6 The matrix of the coordinates of each feature point is singularly decomposed. S7 The motion identification unit 57 obtains the posture and rotation components. S8 The motion identifying unit 57 calculates the translational component.

S9 The motion identification result is output. S10 Request input of image data at the next time. [Calculation Method of Motion Identification Method of Second Embodiment] In the above description, motion parameters are obtained by performing singular value decomposition, but motion identification can be performed without singular value decomposition.

With reference to FIG. 5B, a method of calculating the posture and motion parameters of the object according to the second embodiment will be described.
In FIG. 5 (b), 30 is the image plane of the image input device, and the x-axis, y-axis, and z-axis are the orthogonal coordinates of the camera coordinate system (origin O, and the x-axis and y-axis are on the image plane 30). , Is parallel to each side of the image plane, and the z axis is perpendicular to the xy plane (optical axis direction).

Reference numeral 50 is an object. At the object 50,
A, B and C are characteristic points. The position vector of the feature point A at time f is Av _f , and the coordinates in the camera coordinate system are Av _f =
(A _x , A _y , A _z ).

Similarly, the position vector of the characteristic point B at time f is Bv _f , and the position vector of the characteristic point C at time f is Cv _f . The coordinate system of the object is the barycentric coordinate system, and the vector of its origin (centroid) G with respect to the origin O of the camera coordinate system is tv.
_{Let f} . The origin of the barycentric coordinate system is set to the characteristic point A.

Tvf = Avf. The component of tvf is (t _x , t _y , t _z ). nv is a unit direction vector of the x-axis of the barycentric coordinate system. mv is a unit direction vector of the y-axis of the barycentric coordinate system. ov is a unit direction vector of the z-axis of the barycentric coordinate system. nv, av, represent ov the camera coordinate system and _{nv = (n x, n y} , n z) av = (a x, a y, a z) ov = (o x, o y, o z) and .

Nv, av, and ov at time f are respectively nv
_{If f} , av _f , and ov _f are set, av _f = (Bv _f −Av _f ) × (Cv _f −Av _f ) ov _f = nv _f × av _f nv _f = Bv _f −Av _f (av _f , ov _f and nv _f are unit vectors, which are actually standardized calculation results on the right side, but are represented by the above formulas for convenience).

Rv _f represents the posture of the object 50, and nv, a
It is represented by the following matrix having v and ov as components.

[0077]

(Equation 8)

It is something. kv is a rotation vector of the angular velocity ω, and dv is a translation vector of the object 50. mv is a vector representing the position of the center of gravity G with respect to the rotation center coordinate system.

Let tv and Rv at time f be tv _f and R, respectively.
_Let v _f be the rotation conversion matrix of the vector kv (angular velocity ω) be Rω _{, then} Rv _{f + 1} = RωRv _{f, and} thus Rω = Rv _{f + 1} Rv _f ⁻¹ .

From the above relationship, the direction (kv) of the rotation axis with respect to the camera coordinates, the position (gv), the translational velocity (dv),
Calculate the angular velocity (ω). kv and ω are Rω = Rv _{f + 1} R
It can be obtained from v _f ^-1 .

Since gv _{f + 1} = gv _f + dv tv _f = gv _f + Rv _f · mv, tv _f = gv ₀ + dv · f + Rv _f · mv.

Therefore, t _x = gv _0x + f · dv _x + n _x · mv _x + o _x · mv
_{y + a x · mv z t} y = gv 0y + f · dv y + n y · mv x + o y · mv
_{y + a y · mv z t} z = gv 0z + f · dv z + n z · mv x + o z · mv
By the relation of _y + a _z · mv _z or more, tv _f and Rv _f at three times are measured,
Solving the simultaneous equations obtained from the measured values, gv ₀ ,
dv and mv are obtained.

The calculation method for motion identification is described in Japanese Patent Application No.
No. 145633. FIG. 7 is a flowchart (2) of the embodiment of the system configuration of the present invention, which is a flowchart in the case of performing motion identification by the above-mentioned motion identification method (2).

The S1, S1 'image data input section 61 inputs image data from the left and right image input devices at each time. The S2, S2 ′ image processing unit 54 extracts the characteristic points A, B, C of the left and right images at each time.

The S3 position calculator 56 associates the left and right feature points at each time. The S4 position calculation unit 56 calculates the distance and direction of the feature point at each time to obtain the position. That is, the position vector Av _f = (A _x , A _y , of the feature points A, B, C at time f
_{A z), Bv f = (} B x, B y, B z), Cv f = (C
_x , C _y , C _z ) is obtained.

S5 The motion identifying section 57 performs coordinate conversion into the barycentric coordinate system and calculates the posture. That is, the direction vectors nv, av, ov of the orthogonal axes of the barycentric coordinates at time f are respectively n
v _f, av _f, when the ov _f, obtaining the _{av f = (Bv f -Av f} ) × (Cv f -Av f) ov f = nv f × av f nv f = Bv f -Av f.

S6 The motion identification unit 57 also calculates the distances of the feature points.
Tv for 3 consecutive hours_f, Rv _f(F = 0, 1,
2) is asked. And the rotation matrix Rω = Rv_{f + 1}Rv_f ^-1
Ask for.

Furthermore, gv _{f + 1} = gv _f + dv (dv
Is a translational velocity vector) tv _f = gv ₀ + dv · f + Rv _f · mv (f = 0,
1, 2), the simultaneous equations obtained from the measured values of tv _f and Rv _f at three times are solved to obtain gv ₀ , dv, mv.

The S7 output section outputs the motion identification result. S8 Request image data for the next three consecutive times. FIG. 8 shows an embodiment of the shape estimation device of the present invention.

FIG. 8A shows the device configuration. FIG. 8 (b) is a flow chart of the apparatus of FIG. 8 (a). In FIG. 8 (a), 50 is a floating object, which is a translational velocity (kv), an angular velocity (ω), and a rotation axis (kv).

Reference numeral 51 'is a shape estimating device for estimating the shape of the floating object 50. Reference numerals 52 and 53 are left and right image input devices. Reference numerals 54 and 55 are left and right image processing units for extracting feature points of left and right image data.

Reference numeral 56 is a position calculation unit, which correlates the left and right characteristic points and calculates the positions of the characteristic points.
An output unit 58 is an interface for outputting the result of shape estimation.

Reference numeral 70 denotes a shape estimation unit, which is a floating object 50.
Shape estimation is performed. In the shape estimating unit 70, 57 is a motion identifying unit, which is the posture of the floating object 50,
This is to obtain a motion parameter.

Reference numeral 65 is a contour point coordinate holding unit, which extracts and holds contour data of an image. Reference numeral 66 'is a shape estimation calculation unit, which estimates the shape from the contour data of the image and the motion parameters.

Reference numeral 60 denotes a display, which displays the estimated shape on the screen. A magnetic disk device 59 stores the estimated shape data.

Reference numeral 61 'is a printer for printing out the data of the shape estimation result. The output unit also outputs the position, posture, motion parameter, and shape of the object to an external device such as a posture control device.

The description of FIG. 8B will be given later. FIG. 9 shows an embodiment of the system configuration of the shape estimation device of the present invention. In FIG. 9, 51 'is a shape estimation device.

Reference numeral 52 is a left and right image input device. An image data input unit 61 is an interface for inputting image data.

Reference numeral 65 is a CPU. 66 is a memory. K'is a shape estimation means. An output unit 58 is an interface that outputs shape estimation data to an external device.

A magnetic disk device 59 stores image data, shape estimation data and the like. 6
Reference numeral 0 denotes a display, which displays the estimated shape on the screen.

Reference numeral 61 'is a printer. In the memory 66, the left and right image processing units 54 are programs that perform processing such as extracting feature points based on image data.

A position calculation unit 56 is a program for associating left and right characteristic points with each other and for obtaining the position of each characteristic point. 57 is a motion identification unit, which is a program for motion identification.

Reference numeral 70 is a shape estimation unit. Shape estimating unit 70
In the above, reference numeral 64 is a contour extraction unit, which is a program for extracting the contour of an image based on image data.

Reference numeral 65 denotes a contour point coordinate holding unit, which is an area for holding the coordinates of the contour. Reference numeral 66 'is a shape estimation calculation unit, which is a program for calculating shape estimation. 59
Is a magnetic disk device.

Reference numeral 60 is a display. 61 'is a printer. The operation of the shape estimation device of the present invention will be described with reference to FIG.

The process from image input to motion identification of the shape estimation device of the present invention is the same as that of the motion identification device described above. S1 Images are input from the left and right image input devices 52 and 53.

S2 The left and right image processing units 54 and 55 extract feature points from the image data. The S3 position calculation unit 56 associates the left and right feature points. The S4 position calculation unit 56 calculates the distance between the feature points.

S5 The motion identification unit 57 obtains the posture and motion parameters from the time series data of the positions of the characteristic points of the floating object 50. S6 The contour extraction unit 64 extracts a contour from the image data.

S7 The shape estimating unit 70 estimates the shape from the contour data at a plurality of consecutive time points and the posture, position and motion parameters. S8 The position, posture, motion parameter, and estimated shape are output.

S9 Input request for the next image data is made.

[0111]

According to the motion identifying apparatus of the present invention, the position, orientation, and motion parameters of an object can be obtained without a model of the object or a mark for distance measurement. Further, according to the shape measuring apparatus of the present invention, it is possible to obtain the position, orientation, and motion parameters of the object without using the object model or the mark for distance measurement, and to obtain the shape of the object using the measurement result. .

[Brief description of drawings]

FIG. 1 is a diagram showing a basic configuration of the present invention.

FIG. 2 is an explanatory diagram of a shape measuring method of the shape estimating device of the present invention.

FIG. 3 is a diagram showing an embodiment of the motion identification device of the present invention.

FIG. 4 is a diagram showing a system configuration example of the motion identification device of the present invention.

FIG. 5 is an explanatory diagram of calculation methods (1) and (2) for motion identification according to the present invention.

FIG. 6 is a flowchart of a system configuration example of the present invention.
It is a figure which shows (1).

FIG. 7 is a flowchart of a system configuration example of the present invention.
It is a figure which shows (2).

FIG. 8 is a diagram showing an embodiment of a shape estimation device of the present invention.

FIG. 9 is a diagram showing a system configuration example of a shape estimation device of the present invention.

[Explanation of symbols]

 F: Motion identification device 1: Object to be measured 2: Image input device (shape estimation device) K: Motion identification means K ': Shape estimation means 3: Image processing part 4: Position calculation part 5: Motion identification part 6: Shape Estimating unit 7: Output unit 10: Image data storage device

Continuation of front page (51) Int.Cl. ⁶ Identification code Office reference number FI Technical display location G01P 3/36 C G06T 7/00 7/60

Claims (4)

[Claims] 1. An image input device for inputting image data of a moving object, an image processing unit for extracting a feature point of the object from the image data, and a position calculation for measuring a position to the feature point with respect to an image plane. Section, a motion identifying section for obtaining the posture of the object and motion parameters of the object based on the time-series position data of the feature points, and an output section for outputting the obtained position, posture, and motion parameters Motion identification device characterized by. 2. An image input device, a motion identifying means for determining the position and orientation of a moving object, and a motion parameter, and an output part, wherein the motion identifying means uses the image data input from the image input device to detect the object. Extract feature points, calculate the position of the feature points with respect to the image plane, determine the pose and motion parameters of the object based on the time-series position data of the feature points, and output the calculated position, pose, and motion parameters A method for identifying movement, which comprises: 3. An image input device for inputting image data of a moving object, an image processing unit for extracting a feature point of the object from the image data, and a position calculation for calculating a position of the feature point with respect to an image plane. Section, a motion identification section that obtains the pose and motion parameters of the object based on time-series position data of the feature points, and the object based on the contour, position, pose, and motion parameters of the object on the image plane. A shape estimation apparatus comprising: a shape estimation unit that estimates the shape of the object; and an output unit that outputs the obtained position, orientation, motion parameter, and shape. 4. An image input device, a shape estimating means for obtaining the position and orientation of a moving object, and a motion parameter, and an output part. The shape estimating means uses the image data input from the image input device to detect the object. The feature points are extracted, the positions of the feature points with respect to the image plane are calculated, the posture and motion parameters of the object are obtained based on the time-series data of the positions, the contour data of the time-series image is obtained, and the contours of the time-series are obtained. A shape measuring method characterized by estimating the shape of an object based on data, a position, a posture, and a motion parameter, and outputting the obtained position, posture, motion parameter, and shape.