JP3423346B2 - Measuring device for moving objects - Google Patents

Description

Detailed Description of the Invention

[0001]

The present invention is used for measuring a moving object. In particular, it relates to a configuration for obtaining the coordinates of an imaged moving object at high speed.

[0002]

2. Description of the Related Art Conventionally, as a method of measuring a moving object, the coordinates are obtained by imaging the moving object and processing a video signal thereof. In such a measuring method, the moving object needs to be constantly imaged in the screen. Therefore, conventionally, (1) when the moving object moves in a wide range, the wide-angle lens is used to measure the moving object so that the moving object is imaged within the angle of view of the image pickup device (2) the movement of the moving object (3) Fix the imaging device to the platform with an angle encoder and track the moving object while manually moving the platform. , The angle of the platform at that time is recorded together with the image, and the coordinate of the moving object is checked for each frame by a method such as digitizing from the recorded image and the angle of the platform. .

In addition, two cameras simultaneously capture two moving objects.
It is also known to measure dimensional coordinates and calculate three-dimensional coordinates of a moving object from these two two-dimensional data.

[0004]

However, in the method (1), when a wide-angle lens is used to capture an image of a moving object, the moving object itself is also imaged in a small size, making it difficult to separate the moving object from the background. was there. Further, since the image processing apparatus that inputs the video signal and calculates the geometric center of gravity (centroid) of the moving object processes in pixel units, it is impossible to measure the fine movement of the moving object.

Further, in the method (2), since the moving range of the moving object is limited, there is a limit to what can be measured.

Such a problem similarly occurs when two cameras are used at the same time. Moreover, if a wide-angle lens is used, the three-dimensional coordinate measurement accuracy of the moving object will be reduced. For this reason, such a measuring method was exclusively used for the measurement in the laboratory. Further, the image processing in this case is based on the measurement of the two-dimensional coordinates, and does not utilize the property of the movement of the moving object in the three-dimensional measurement space. Therefore, an invalid measurement result may be output due to noise included in the image or noise when the moving object is separated from the background.

The method (3) can be used to measure the movements and actions of the player outdoors or in the stadium. However, since one frame of a normal video signal is composed of 1/60 seconds, 432,000 frames of images must be digitized in order to check the recording for 2 hours, for example. Even if it takes seconds, it takes a long time of 600 hours in total.

The present invention solves such a problem, and even when a moving object moves in a wide range, the moving object is imaged without using a wide-angle lens, and the coordinates of the imaged moving object can be obtained at high speed. An object of the present invention is to provide a measuring device for a moving object that can be used.

[0009]

The measuring device for a moving object according to the present invention is obtained by an image pickup means for picking up an image of a moving object, a driving means for setting an image pickup direction and an image pickup range of the image pickup means, and an image pickup means. Image processing means for calculating the position in the image of the moving object included in the video signal, information about the position obtained by the image processing means and the imaging direction and imaging range of the imaging means when the video signal is obtained, In a measuring device for a moving object, which comprises one or more sets of position calculating means for calculating the coordinates in the real space of the moving object by means of three-dimensional real space coordinates, there is a possibility that the moving object to be imaged may move. Three-dimensional area setting means set in the system, and area mapping means for mapping the area of the three-dimensional real space coordinate system into a two-dimensional area corresponding to the imaging direction and imaging range of the imaging means. , And a two-dimensional area setting means for setting the corresponding image processing means for mapping the two-dimensional regions,
The image processing means is characterized by including area limiting means for limiting the target area of the calculation to the two-dimensional area set by the two-dimensional area setting means.

Three-dimensional vector setting means for setting the moving speed vector of the moving object to be imaged in the three-dimensional real space coordinate system, and the moving speed vector of this three-dimensional real space coordinate system in the imaging direction and the imaging range of the imaging means. Further comprising vector mapping means for mapping into a two-dimensional moving velocity vector, and vector setting means for setting the mapped two-dimensional moving velocity vector for each corresponding imaging means,
The image processing means includes means for detecting a positional change of the moving object between the previous screen and the current screen by a predetermined number within the two-dimensional area limited by the area limiting means, and the detected position It is possible to provide means for activating the position calculation means on condition that the change is within a predetermined range with respect to the two-dimensional moving velocity vector set by the vector setting means.

The image pickup means includes a platform which is rotatable around at least one axis, and an image pickup device which is fixedly mounted on the platform, and the driving means includes respective axes of the platform. It is preferable to include an angle setting means for setting a rotation angle of the surroundings and a zoom lens means attached to the image pickup apparatus.

When the accuracy of the angle setting means and the zoom lens means is so high that the difference between the set target value and the actual value hardly occurs, the target value itself is used as the information regarding the image pickup direction and the image pickup range of the image pickup means. It can also be used. However, practically, it is preferable to perform feedback control so that the difference between the actual value and the target value is reduced, and use the actual value as information regarding the image capturing direction and the image capturing range of the image capturing unit. That is, the angle setting means includes, for at least one axis, a motor that rotates the platform, an angle encoder that detects an actual rotation angle, a detected angle, and a target azimuth angle set by the host device. The zoom lens means includes a zoom drive motor that sets a zoom position, a zoom encoder that detects an actual zoom position, and a detected zoom position. And zoom control means for controlling the zoom drive motor so that the difference between the target position set by the host device and the target position set by the host device is small, and the output of the angle encoder and the output of the zoom encoder are the image pickup direction and the image pickup range of the image pickup means. It is preferable to be supplied to the position calculation means as the information regarding.

[0013]

A region of a three-dimensional real space coordinate system in which a moving object to be imaged may move is mapped to a two-dimensional region,
The coordinates of the moving object are obtained only in that area. For example, when the moving course of a moving object is determined to some extent, such as skiing, skating, athletics, car racing, horse racing, bicycle racing, and boat racing, it is not necessary to measure other areas. Therefore, the processing for such an area is omitted. As a result, it is possible to reduce the number of video signals to be processed and obtain the coordinates of the moving object in real time.

In some cases, the speed of a moving object can be predicted in advance. In that case, the moving speed vector of the moving object is set in the three-dimensional real space coordinate system, and it is mapped to the two-dimensional moving speed vector, and the moving object is detected and designated within the set two-dimensional area. The coordinates are measured when the velocity vector satisfying a certain condition can be measured for the obtained velocity vector. As a result, even if an object other than the target moving object moves within the imaging range, it is unlikely to measure it by mistake.

According to the present invention, a plurality of image pickup means (for example, a TV camera) are arranged on a moving course of a moving object to be imaged, and a drive means, an image processing means and a position calculation means are provided respectively, and these are used as a common higher-order calculation It is suitable for the usage form that is integratedly managed by the device. In this case, the host computing device comprehensively manages the coordinates of the moving object imaged by each imaging means and the state of the imaging means, and measures the measurement information from a plurality of angles and the movement of the moving object registered in advance. From the above feature, it is possible to effectively and accurately measure the position of a moving object that moves in a wide area at high speed. Further, the direction and the imaging range of the image pickup means are controlled in accordance with the movement of the moving object, the comparison data is displayed on the image obtained by each image pickup means, the measurement accuracy is displayed, and the information of other image pickup means is displayed. And so on.

[0016]

1 is a block diagram showing a measuring apparatus for a moving object according to an embodiment of the present invention.

This measuring device is equipped with a TV camera 21 as an image pickup means for picking up an image of a moving object.
An electric zoom lens 22, an electric camera platform 23, and a drive control device 25 are provided as driving means for setting the image pickup direction and the image pickup range of the position of the moving object included in the image signal obtained by the TV camera 21 in the image ( In this example, an image processing device 24 is provided as an image processing means for calculating the geometrical center of gravity, and the position obtained by this image processing device 24 and the image pickup direction and image pickup range of the TV camera 21 when the video signal thereof is obtained are related. A position calculation device 26 is provided as position calculation means for calculating the coordinates of the moving object in the real space from the information.

A TV camera 21, an electric zoom lens 22,
The electric platform 23 and a part of the drive control device 25 form a camera head, and these, the image processing device 24, the drive control device 25, and the position calculation device 26 form a camera head / calculation device 12. A plurality of camera heads / arithmetic units 12 are provided, and each is connected to the host arithmetic unit 11 via a data link 13. Camera head / calculator 1
2, a video synthesizing device 27 for superimposing additional information on the video signal output from the TV camera 21,
An input / output device 28 for inputting / outputting data via the data link 13 is provided.

Each camera head / arithmetic unit 12 outputs coordinate data to the data link 13 for each measurement cycle, and the camera head / arithmetic unit 1 other than the one outputting coordinate data.
2 can capture that data. The upper-level arithmetic device 11 integrally manages the plurality of camera heads / arithmetic devices 12, and outputs the coordinate measurement result to a screen display or other output device. When each camera head / arithmetic unit 12 finishes outputting coordinate data, the upper arithmetic unit 11
A command can be individually output to each camera head / arithmetic unit 12 according to the measurement situation.

The number of camera heads / arithmetic units 12 that can be connected to the data link 13 is limited by the data transfer speed of the data link 13 and the arithmetic speed of the host arithmetic unit 11. The camera head / arithmetic unit 12 can calculate three-dimensional coordinates, but the higher-level arithmetic unit 11 can calculate three-dimensional coordinates from a plurality of two-dimensional coordinate data.

The TV camera 21 is fixed to an electric pan head 23 together with an electric zoom lens zoom 22, and an image processing device 24 is provided.
And a video signal is output to the video synthesizer 27. The image processing device 24 obtains the geometric center of gravity of the moving object included in the video signal input from the TV camera 21, and outputs it to the position calculation device 26 as centroid data. Drive controller 25
Outputs a zoom control signal to the electric zoom lens 22 to capture a zoom position signal. The drive control device 25 also outputs a tripod head control signal to the electric tripod platform 23 and takes in the tripod head direction signal. The drive control device 25 further generates zoom position / pan head direction data from the captured zoom position signal and pan head direction signal, and outputs it to the position calculation device 26. The position calculation device 26 controls the zoom position and the tripod head direction, and therefore uses these control data as the drive control device 2.
Output to 5. The position calculator 26 also calculates the coordinates of the moving object from the centroid data from the image processor 24 and the zoom position / head direction data from the drive controller 25,
The obtained coordinate data in the real space is sent to the data link 13 via the input / output device 28.

The position calculation device 26 includes a data link 13
And another camera head / via the input / output device 28.
The coordinate data output by the arithmetic unit 12 and the host arithmetic unit 11
The command output by is input. In response to this, the position calculation device 26 outputs the superimposed image data to the video synthesizing device 27, and the video synthesizing device 27 superimposes this data on the video signal from the TV camera 21 and outputs the synthetic video signal. . As a result, various information can be output by superimposing various information on the video used for measurement.

The feature of this embodiment is that the upper arithmetic unit 11 includes a three-dimensional area setting means for setting an area where a moving object to be imaged may move in a three-dimensional real space coordinate system. Is provided as an internal control program, and the area of this three-dimensional real space coordinate system is set in each TV camera
Area mapping means for mapping to a two-dimensional area corresponding to the image pickup direction and the image pickup area of is provided as a control program in the position calculation device 26, and the position calculation device 26 sets the mapped two-dimensional region in the image processing device 24. The output is connected as described above, and the image processing device 24 includes the area limiting means for limiting the target area of the calculation to the two-dimensional area set by the two-dimensional area setting means. This is shown in FIG. 2 and FIG.
Will be described with reference to.

FIG. 2 shows the higher-level arithmetic unit 11 and the position arithmetic unit 2.
6 is a flowchart showing a part of the processing with 6; The upper arithmetic device 11 integrally manages a plurality of camera heads / arithmetic devices 12. Only the setting of the imaging area and the velocity vector and the measurement of the three-dimensional coordinates will be described here.

Each higher-order arithmetic unit 11 has a TV camera 21
Before the start of imaging, the area in which the moving object to be imaged may move and its moving velocity vector are input in the three-dimensional real space coordinate system. Upper arithmetic device 11
Indicates the area and moving velocity vector of the position calculation device 2
Notify 6. When the image pickup is started, the position calculation device 26 takes in the image pickup direction and the image pickup range of the TV camera 21 from the drive control device 25, and correspondingly, the region of the three-dimensional real space coordinate system and the moving velocity vector are two-dimensionally obtained. Map to the region and velocity vector of. The mapped area and velocity vector data are output to the image processing device 24.

FIG. 3 is a block diagram showing the details of the image processing device 24.

The video signal output from the TV camera 21 shown in FIG.
It is input to the 1- and binary-coding circuit 304. Further, the two-dimensional area setting data 320 from the position calculation device 26 is input to the timing generation circuit 302, and the velocity vector setting data 321 is input to the central processing unit 310.

The sync separation circuit 301 extracts the horizontal sync signal and the vertical sync signal included in the input video signal 311 and inputs them to the timing generation circuit 302. The timing generation circuit 302 receives the vertical timing signal 31 from the clock signal output from the clock generation circuit 303.
3, horizontal timing signal 314 and effective range signal 31
5 is generated. The vertical timing signal 313 is a signal having the same frequency as the vertical synchronizing signal, and has a pulse width narrowed to match the timing of the horizontal synchronizing signal so as to be convenient for use inside the image processing apparatus. . The horizontal timing signal 314 is a signal having the same frequency as the horizontal synchronizing signal and a narrow pulse width. The effective range signal 315 is a signal for limiting the calculation target area to the area set by the two-dimensional area setting data 320, and based on the vertical timing signal 313 and the horizontal timing signal 314,
It is output only within the effective range for measurement in the video signal.

The clock signal output from the clock generation circuit 303 is, in addition to the timing generation circuit 302, a clock input to the horizontal address generation circuit 306, a clock input to the vertical accumulation circuit 307, and a clock input to the horizontal accumulation circuit 308. , Are input to the clock input of the area counting circuit 309. The clock signal output from the clock generation circuit 303 is adjusted to have an integral multiple frequency of the horizontal timing signal 314. By adjusting in this way, a moving object signal can be created at a stable timing with respect to the video signal 311.

The vertical timing signal 313 output from the timing generation circuit 302 is the vertical address generation circuit 305.
Reset input and the input port of the central processing unit (CPU) 310. The horizontal timing signal 314 output from the timing generation circuit 302 is the clock input to the vertical address generation circuit 305 and the horizontal address generation circuit 3
06 reset input.

The vertical address generation circuit 305 is a counter that is incremented each time a horizontal timing signal is input and is reset by the vertical timing signal.
The line number (vertical position) of the currently input video signal is output. The horizontal address generation circuit 306 is a counter that is incremented each time a clock signal is input and reset by a horizontal timing signal, and outputs the horizontal position of the currently input video signal.

The effective range signal 315 output from the timing generation circuit 302 is the enable input E2 of the vertical accumulation circuit 307, the enable input E2 of the horizontal accumulation circuit 308, and the enable input E2 of the area counting circuit 309.
Entered in.

The binarization circuit 304 separates the luminance indicating the moving object and the luminance indicating the background, which are included in the input video signal 311, and outputs a moving object signal 312. The moving object signal 312 is obtained by enabling input E1 of the vertical accumulation circuit 307 and enabling input E of the horizontal accumulation circuit 308.
1 and the enable input E1 of the area counting circuit 309.

The vertical address generation circuit 305 outputs the current vertical address of the input video signal, and the vertical address signal is input to the vertical direction accumulation circuit 307. The horizontal address generation circuit 306 outputs the current horizontal address of the input video signal, and the horizontal address signal is input to the horizontal accumulation circuit 308.

The vertical accumulation circuit 307 accumulates the vertical address signal being input only when signals are simultaneously input to the enable inputs E1 and E2, and the reset signal is input to the reset input. Reset the accumulator counter. That is, the moving object signal 3
Only when 12 indicates a moving object and the effective range signal 315 indicates an effective range, the vertical address signal is accumulated at the timing of the clock signal. Therefore, the output becomes the vertical first moment of the moving object included in the video signal.

The horizontal accumulation circuit 308 accumulates the horizontal address signal being inputted only when the signals are inputted to the enable inputs E1 and E2 at the same time, and the reset signal is inputted to the reset input. Reset the accumulation counter. That is, the moving object signal 31
Only when 2 indicates a moving object and the effective range signal 315 indicates an effective range, the horizontal address signal is accumulated at the timing of the clock signal. Therefore, the output becomes the horizontal first moment of the moving object included in the video signal.

The area counting circuit 309 has an enable input E.
The output is incremented only when signals are simultaneously input to 1 and E2, and the accumulator counter is reset when a reset signal is input to the reset input. That is, the output is incremented at the timing of the clock signal only when the moving object signal 312 indicates the moving object and the effective range signal 315 indicates the effective range.
Therefore, the output becomes the area (0th moment) of the moving object included in the video signal.

The cumulative calculation force of the vertical accumulation circuit 307 is input to the input port SY of the central processing unit 310, and the cumulative calculation force of the horizontal accumulation circuit 308 is input to the input port SX of the central processing unit 310 to determine the area. The area output of the counting circuit 309 is input to the input port N of the central processing unit 310.

A counter initialization signal 316 is output from the central processing unit 310, and is input to the reset input of the vertical accumulation circuit 307, the reset input of the horizontal accumulation circuit 308, and the reset input of the area counting circuit 309. . The central processing unit 310 also outputs vertical centroid data 317 and horizontal centroid data 318.

Further, in this embodiment, the central processing unit 310
A storage device 319 is connected to the storage device 319, and a predetermined number of geometric centroids in the previous screen can be stored. The central processing unit 310 detects a position change between the geometric center of gravity stored in the storage device 319 and the geometric center of gravity of the subtraction screen, and the position change is within a predetermined range for the velocity vector setting data 321. Then, the position calculation activation signal 322 is output.

The purpose of this image processing apparatus is to detect a moving object signal included in an input video signal in a limited two-dimensional area and output centroid data thereof. This operation will be described in detail.

When the video signal 311 is input, the timing generation circuit 302 generates a vertical timing signal 313, which is input to the central processing unit 310. The central processing unit 310 outputs a counter initialization signal 316 to indicate that the video signal has started. This resets the accumulating circuits 307 and 308 and the area counting circuit 309. The vertical timing signal 313 also resets the vertical address generation circuit 305.

When the vertical timing signal 313 ends,
The timing generation circuit 302 uses the horizontal timing signal 31
4. Output the effective range signal 315, and the binarization circuit 3
04 outputs the moving object signal 312, the accumulation circuit 3
07 and 308 output the vertical first moment and the horizontal first moment of the moving object, respectively. At the same time, the area counting circuit 309 outputs the area of the moving object.

One field of the video signal 311 is completed,
The timing generation circuit 302 uses the vertical timing signal 313.
Is output again, the central processing unit 310 reads these first moment and area (0th moment),
[1st moment in vertical direction ÷ area] and 1 in horizontal direction
Next moment ÷ area] is calculated. The results of this calculation are the vertical and horizontal centroids of the moving object, respectively. The centroid data is output to the output port, and then the counter initialization signal 316 is output. After that, this sequence is repeated.

When the centroid data is obtained, the central processing unit 310 also compares the centroid data with the centroid data of a few frames before stored in the storage device 319 and detects the position change. Further, when the detected position change is within the range predetermined for the velocity vector setting data 321, the position calculation activation signal 322 is output to the position calculation device.

FIG. 4 is a trihedral view showing the structure of the camera head in a simplified manner. FIG. 4A is a plan view, FIG.
(C) is a side view. Here, the platform has 2 degrees of freedom,
That is, an example of the configuration when there are two rotatable axes is shown.

The platform is rotatable about a vertical axis and a horizontal axis, and the TV camera 21 and the zoom lens 22 are fixed to the platform. A vertical axis drive motor / angle encoder 41 is provided on the vertical axis of the platform to output a pulse at a constant angle. The TV camera 21, the zoom lens 22, and the vertical axis drive motor / angle encoder 41 are rotatable around a horizontal axis of the platform, and the horizontal axis is provided with a horizontal axis drive motor / angle encoder 42, and a fixed angle. A pulse is output every time.

FIG. 5 is a block diagram showing an example of the drive controller 25.

The drive control device 25 includes a vertical axis angle encoder 501, a vertical axis drive motor 502 and a vertical axis control circuit 503, a horizontal axis angle encoder 511, a horizontal axis drive motor 512 and a horizontal axis control circuit 513, and a zoom / zoom. An encoder 521, a zoom drive motor 522, and a zoom control circuit 523 are provided. The vertical axis angle encoder 501 and the vertical axis drive motor 502 constitute the vertical axis drive motor / angle encoder 41 in FIG. 4, and the horizontal axis angle encoder 511 and the vertical axis drive motor 512.
Constitutes a horizontal axis drive motor / angle encoder 42.

The drive motor and encoder for each axis and zoom are mechanically coupled to each other, and when the rotation angle of the drive motor changes, this is mechanically transmitted and an encoder pulse containing information about the rotation angle is transmitted from the encoder. Is output. The driving methods for zooming the vertical axis, the horizontal axis, and the electric zoom lens are essentially the same, and the configurations and operations of the vertical axis control circuit 503, the horizontal axis control circuit 513, and the zoom control circuit 523 are substantially the same. is there.
The operation regarding the vertical axis will be described below.

The vertical axis control circuit 503 includes a pulse / angle conversion circuit 504, a difference circuit 505, a gain adjustment circuit 506, and a motor drive circuit 507. The encoder pulse output from the vertical angle encoder 501 is input to the pulse / angle conversion circuit 504. The pulse / angle conversion circuit 504 converts this pulse into a signal indicating the angle (voltage, digital data, etc.), supplies it to one input of the difference circuit 505, and outputs it as an angle signal to the position calculation device.

The target angle signal is input from the position calculation device to the other input of the difference circuit 505. As a result, the difference circuit 505 calculates the difference between the target angle signal and the angle signal and outputs the difference angle. This difference angle is determined by the gain adjustment circuit 5
It is input to 06. The gain adjustment circuit 506 adjusts the drive amount of the vertical axis drive motor 502 according to the input difference angle, that is, the residual difference from the target angle. At this time, the gain adjustment circuit 506 adjusts the drive amount of the vertical axis drive motor 502 so that the overshoot and the offset error are eliminated. The motor drive control signal output from the gain adjustment circuit 506 is input to the motor drive circuit 507. The motor drive circuit 507 power-amplifies the input signal and outputs a motor drive signal to drive the vertical axis drive motor 502. In this way, the drive motor is driven so that there is no difference between the target angle and the actual angle.

Although the case where the encoder outputs the angle information as a pulse signal has been described here, a potentiometer may be used as the encoder. In that case, since the angle information is output as a voltage signal and it is not necessary to convert the information so that the difference circuit 505 can perform the calculation, the pulse / angle conversion circuit 504 becomes unnecessary.

FIG. 6 shows the relationship between horizontal and vertical addresses.
The effective range 62 is limitedly set with respect to the television screen range 61 (the figure is an example not so limited), and the moving object 63 therein is detected.

Next, a position calculation method in the position calculation device 26 will be described with reference to FIGS. 7 and 8. FIG. 7 is a diagram showing the position of the camera head in the measurement space coordinate system.
FIG. 4 is a diagram showing a relationship between a measurement space coordinate system and a coordinate system of an image processing device.

First, a method of calculating the position of one moving object by using two camera heads will be described here.

The coordinate system (X, Y, Z) representing the coordinates of the camera head in the measurement space is a coordinate system fixed to the measurement space, that is, the ground or floor, and is constant regardless of the direction of the camera head. The coordinates of the camera head are camera head 1: M ₁ (X ₁ , Y ₁ , Z ₁ ) and camera head 2: M ₂ (X ₂ , Y ₂ , Z ₂ ).

The direction of the camera head, that is, the central direction (angle) of the image observed from the camera head is defined as camera head 1: (θ ₁ , ψ ₁ ) camera head 2: (θ ₂ , ψ ₂ ). This angle is an azimuth angle and an elevation angle (or depression angle) based on a vector fixed in the coordinate system (X, Y, Z), for example, (1, 0, 0), and is the drive control device shown in FIG. It is the direction of the pan head output from 25.

The imaging range (angle) of the camera head is camera head 1: (α ₁ , β ₁ ) camera head 2: (α ₂ , β ₂ ). Generally, the imaging range has a horizontal to vertical ratio of 4: 3.
So between the alpha and beta each camera _{head, tan (α 1/2)} : tan (β 1/2) = 4: 3 tan (α 2/2): tan (β 2/2) = 4: There are 3 relationships. Further, there is a unique correspondence between the imaging range of the camera head and the zoom position (focal length of the lens). That is, it can be calculated from the zoom position data output by the drive control device 25 in FIG.

As shown in FIG. 8, the barycentric coordinates of the moving object included in the image observed from the camera head are as follows: camera head 1: (u ₁ , v ₁ ) camera head 2: (u ₂ , v ₂ ). To do. These coordinates are the coordinate system of the image on the screen, that is, FIG.
Is a coordinate system output by the image processing device 24 in FIG. The origin of this coordinate system is the center of the screen, and the horizontal size (horizontal pixel count) and vertical size (vertical pixel count) within the effective range are 2 × u _max and 2 × v _max , respectively. Therefore, it matches the imaging range. This is a constant common to both camera heads.

From the coordinates of the camera head and the direction of the moving object, an equation of a straight line passing over the moving object is obtained for each camera head, and camera head 1: p ₁ = (X ₁ , Y ₁ , Z ₁ ) + A ₁ ×
t Camera head 2: p ₂ = (X ₂ , Y ₂ , Z ₂ ) + A ₂ ×
t p ₁ and p ₁ are points on a straight line, t is a parameter, A ₁ and A ₂ are vectors, and A ₁ = (cos Ψ ₁ × cos Θ ₁ , cos Ψ ₁ × sin Θ ₁ , si
n Ψ ₁ ) A ₂ = (cos Ψ ₂ × cos Θ ₂ , cos Ψ ₂ × sin Θ ₂ , si
n Ψ ₂ ). Where Θ and Ψ are the azimuth angle and elevation angle (or depression angle) from the camera head to the moving object, and Θ ₁ = θ ₁ + tan ⁻¹ ((u ₁ / u _max ) × tan (α ₁ /
2)) Ψ ₁ = ψ ₁ − tan ⁻¹ ((v ₁ / v _max ) × tan (β ₁ /
2)) Θ ₂ = θ ₂ + tan ^-1 ((u ₂ / u _max ) × tan (α ₂ /
2)) Ψ ₂ = ψ ₂ − tan ⁻¹ ((v ₂ / v _max ) × tan (β ₂ /
2)). The intersection of these two straight lines is the coordinate of the moving object. Actually, due to the error of each part, the intersection coordinates may not be obtained from the two straight lines formed from the measured values. In this case, the closest point coordinates to the two straight lines may be obtained.
That is, the coordinates of the midpoint of the distance between the two straight lines may be the coordinates of the moving object.

Next, a method of measuring three-dimensional coordinates with one camera head in consideration of the constraint condition of the moving object will be described. The restraint conditions in this case may be the movement conditions of each player on the stadium, and for example, considering that the vertical movement of each player is sufficiently smaller than the movement on the plane,
A constraint condition is obtained that each player moves on a plane.

At this time, the coordinates of the camera head are M ₁ (X ₁ , Y ₁ , Z ₁ ). Further, the direction of the camera head, that is, the central direction (angle) of the image observed from the camera head is (θ ₁ , ψ ₁ ). This angle is the platform direction output by the drive control device 25 in FIG. Further, the imaging range (angle) of the camera head is (α ₁ , β ₁ ). Generally, the imaging range has a horizontal to vertical ratio of 4: 3.
_{So, tan (α 1/2)} : tan (β 1/2) = 4: 3 The alpha ₁ and beta _2, the drive control device in FIG. 1 25
Can be calculated from the zoom position data output by.

It is assumed that the barycentric coordinates of the moving object included in the image observed from the camera head are (u ₁ , v ₁ ) as shown in FIG. The coordinates are the coordinate system of the in-screen image, that is, the coordinate system output by the image processing device 24 in FIG. The origin of this coordinate system is the center of the screen, and the horizontal size (horizontal pixel count) and vertical size (vertical pixel count) within the effective range are 2 × u _max and 2 × v _max , respectively. Yes, it matches the imaging range.

From the coordinates of the camera head and the direction of the moving object, an equation for a straight line passing on the moving object is obtained by p ₁ = (X ₁ , Y ₁ , Z ₁ ) + A ₁ × t. p ₁ is a point on a straight line, t is a parameter, A ₁
A ₁ = (cos Ψ ₁ × cos Θ ₁ , cos Ψ ₁ × sin Θ ₁ , si
n Ψ ₁ ). Where Θ ₁ and Ψ ₂ are the azimuth angle and elevation angle (or depression angle) from the camera head to the moving object, and Θ ₁ = θ ₁ + tan ⁻¹ ((u ₁ / u _max ) × tan (α ₁ /
2)) Ψ ₁ = ψ ₁ −tan ⁻¹ ((v ₁ / v _max ) × tan (β ₁ /
2)).

On the other hand, the moving object is on the next plane due to the constraint condition.

(P−b) · B = 0 p is a point on the constraining plane, b is a vector on the constraining plane (height of the plane, etc.), B is a normal vector of the constraining plane, and is an inner product of the vectors. is there. The coordinates of the intersection of this straight line and the plane are the coordinates of the moving object.

As described above, the camera head /
The computing device 12 measures the centroid position of each moving object,
By transferring the two-dimensional data to all the other camera heads / arithmetic units 12 through the data link 13, each camera head / arithmetic unit 12 can know the three-dimensional coordinates of the moving object as a result. Further, the upper arithmetic device 11 may be configured to obtain three-dimensional coordinates.

In the above description, an example in which the three-dimensional coordinates of one moving object is measured by two camera heads / arithmetic units has been described, but it is necessary to connect a plurality of processing units to one camera head. Thus, it is possible to simultaneously measure the centroid positions of a plurality of moving objects on the screen. At that time, it is better to change the generation timing of the effective range signal according to the position of the moving object in the screen measured by each image processing apparatus. With such a configuration, it is possible to measure the three-dimensional coordinates of a plurality of moving objects with two camera heads / arithmetic devices. Further, even in a configuration having a plurality of camera heads / arithmetic devices, it is possible to measure the three-dimensional coordinates of a plurality of moving objects by appropriately combining them.

Next, the position computing device 26 shown in FIG. 1 receives the three-dimensional region set in the higher-level computing device 11 and maps the three-dimensional region to the two-dimensional region in accordance with the direction of the camera head and the imaging range. Will be described.

This three-dimensional area represents the characteristics of the motion of the moving object. That is, it is an area that may or may not move, a moving direction, and a moving speed. The user of the measuring device may know the outline of the three-dimensional movement path of the moving object. In this case, the user will specify the three-dimensional area for the host computer. For example, in the case of coordinate measurement of an aircraft approaching the runway, since the approach route is specified, there will be no moving objects in an area that deviates significantly from this route, and the moving speed on the approach route will also change extremely. There is nothing to do.

In some cases, the outline of the three-dimensional movement path of the moving object may not be known at all. In this case, the user specifies the movement path of the moving object for the image captured by each camera head. . That is, the two-dimensional area from the installation position of each camera head is designated. The host computer integrates the two-dimensional areas from the installation positions of the camera heads into the three-dimensional motion characteristic area of the moving object. For example, when a general-purpose coordinate measuring device is carried to an athletic stadium to measure each motion along a running route of a track and field athlete, it is difficult to know the running route coordinates in advance. However, shooting with the camera head during practice before the game of track and field athletes,
It is possible to know the travel route. If the travel route is imaged using a plurality of camera heads, the coordinate measuring device can know the outline of the three-dimensional movement route of the moving object even if the travel route has undulations.

In either case, it is preferable to specify the three-dimensional area based on the spherical area. By designating a plurality of spheres at overlapping positions, almost any three-dimensional area can be represented. Moreover, to represent a sphere,
Specify the center and diameter of the sphere. A large number of spheres are required to represent a cubic solid such as a rectangular parallelepiped using only spheres, but this is rarely the case for the characteristics of the motion of a moving object. Moreover, since the sphere is a sphere seen from any angle, the amount of calculation when mapping to a two-dimensional area observed from the installation coordinates of each camera head is reduced. The upper-level computing device has a three-dimensional region represented by a set of spheres for each feature such as a region in which a moving object may move or a region in which it does not move, and transfers this to each camera head / arithmetic device through a data link. . Each camera head / arithmetic device maps a three-dimensional area to a two-dimensional area viewed from the imaging direction using the center coordinates and diameter of the sphere, and measures it by collating it with centroid data output from the image processing apparatus. The validity of the coordinates can be determined.

The speed of the moving object measured from each camera head differs depending on the moving direction of the moving object viewed from each camera head. For example, the moving speed of a moving object moving toward a certain camera head is measured at a low speed, but the speed measured at this time does not correspond to the moving speed of the moving object. That is, according to the conventional coordinate measuring device, it is impossible to output the three-dimensional coordinate data of the moving object when the moving speed of the moving object is compared with a preset speed and a certain condition is satisfied. .
Since each camera head of the coordinate measuring apparatus according to the present invention grasps the three-dimensional coordinates of the moving object in the real space, the three-dimensional velocity vector of the moving object can be known by examining the difference between the three-dimensional coordinates. By comparing this with a preset three-dimensional velocity vector, the three-dimensional coordinate data of the moving object can be output when this and a certain condition are satisfied. For example, it is possible to output the three-dimensional coordinate data of the moving object from the time when the moving object moves in a specific direction at a specific position at a moving speed equal to or higher than a specific speed. If the measurement data is output in response to such a specific motion of the moving object, it becomes unnecessary to record and analyze meaningless and useless measurement data. This is exactly the trigger function that starts recording when the phenomenon you want to observe with the oscilloscope occurs.

In the above embodiments, the centroid position is used as the position of the moving object measured from the video signal output from the image pickup device, but the position where the shape matching degree by pattern matching is the highest is the position of the moving object. However, the present invention can be similarly implemented.

Further, by using an infrared TV camera as the image pickup device, it is possible to reduce the noise contained in the image when measuring a distant measurement target. At this time, the influence of other light sources can be reduced by attaching an infrared light emitting element to the moving object. By using a TV camera in which an optical filter is attached to the image pickup device, a wavelength band including only a specific color can be selected, and a moving object can be easily separated from the background.

The diaphragm of the image pickup device may be controlled so that the moving object can be easily separated from the background. At this time, the diaphragm can be controlled based on the luminance histogram of the video signal including the moving object and the data indicating the measurement state such as the area of the moving object. The focus of the electric zoom lens may be controlled based on the coordinate data of the moving object calculated by the arithmetic device. When one camera head measures the positions of a plurality of moving objects, the diaphragm, the imaging range, and the focus can be controlled so that all the moving objects are within the depth of field of the zoom lens.

[0078]

As described above, the moving object measuring apparatus of the present invention sets a region of a three-dimensional real space coordinate system corresponding to a moving object to be imaged and maps it to a two-dimensional region. As a result, the processing for obtaining the position of the moving object can be significantly reduced, and the coordinates of the moving object can be obtained in real time. Therefore, it is possible to automatically control the imaging direction and the imaging range by tracking a moving object, and it is possible to measure the three-dimensional coordinates of a measurement target that moves in a wide area at high speed with high accuracy.

[Brief description of drawings]

FIG. 1 is a block configuration diagram showing a moving object measuring apparatus according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a part of processing performed by a higher-order arithmetic device and a position arithmetic device.

FIG. 3 is a block configuration diagram showing details of an image processing apparatus.

FIG. 4 is a three-sided view showing a simplified structure of a camera head, in which (a) is a plan view, (b) is a front view, and (c) is a side view.

FIG. 5 is a block configuration diagram showing an example of a drive control device.

FIG. 6 is a diagram showing a relationship between horizontal and vertical addresses.

FIG. 7 is a diagram showing a position of a camera head in a measurement space coordinate system.

FIG. 8 is a diagram showing a relationship between a measurement space coordinate system and a coordinate system of the image processing apparatus.

[Explanation of symbols]

11 Upper-level computing device 12 Camera head / Calculator 13 Data link 21 TV camera 22 Electric zoom lens 23 Electric pan head 24 Image processing device 25 Drive controller 26 Position calculation device 27 Video synthesizer 28 I / O device 301 Sync separation circuit 302 Timing generation circuit 303 clock generation circuit 304 Binarization circuit 305 Vertical address generation circuit 306 Horizontal address generation circuit 307 Vertical accumulation circuit 308 Horizontal direction accumulation circuit 309 Area counting circuit 310 Central Processing Unit 311 video signal 312 Moving object signal 313 Vertical timing signal 314 Horizontal timing signal 315 effective range signal 316 Counter initialization signal 317 Vertical Centroid Data 318 Horizontal Centroid Data 319 storage device 320 Two-dimensional area setting data 321 Speed vector setting data 322 Position calculation start signal 41 Vertical axis drive motor / angle encoder 42 Horizontal axis drive motor / angle encoder 501 vertical axis angle encoder 502 Vertical axis drive motor 503 Vertical axis control circuit 504 pulse / angle conversion circuit 505 Difference circuit 506 Gain adjustment circuit 507 Motor drive circuit 511 Horizontal axis angle encoder 512 Horizontal axis drive motor 513 Horizontal axis control circuit 521 Zoom encoder 522 Zoom drive motor 523 Zoom control circuit 61 TV image range 62 Effective range 63 Moving object

Front page continued (58) Fields surveyed (Int.Cl. ⁷ , DB name) G01B 11/00-11/30 H04N 7/18

Claims (4)

(57) [Claims] 1. An image pickup means (21) for picking up an image of a moving object.
When, Drive for setting the imaging direction and imaging range of this imaging means
Means (22, 23, 25), Moving object included in the video signal obtained by the image pickup means
Image processing means for calculating the position of the body in the image
(24), The position obtained by this image processing means and its video signal are obtained.
Information about the imaging direction and imaging range of the imaging means when
Calculate the coordinates of the moving object in real space from the information
One or more pairs of position calculation means (26), The image processing means is a moving object within the screen for the calculation target area.
Area limiting means to limit the area where the body may move
including In a measuring device for moving objects, Area where moving object to be imaged may move
Three-dimensional region setting that sets the three-dimensional real space coordinate system
Means and The method of imaging the area of the three-dimensional real space coordinate system by the imaging means
Area that maps to a two-dimensional area corresponding to the orientation and the imaging range
Mapping means, The mapped two-dimensional areaLimited by the area limiting means
As the target area of the calculationTwo-dimensional area setting means to set
AndGotA measuring device for a moving object, characterized in that 2. A three-dimensional vector setting means for setting a moving speed vector of a moving object to be imaged in a three-dimensional real space coordinate system, and a moving speed vector of the three-dimensional real space coordinate system in an imaging direction of the imaging means. And a vector mapping means for mapping a two-dimensional moving speed vector corresponding to the imaging range, and a vector setting means for setting the mapped two-dimensional moving speed vector for each corresponding imaging means. , A means for detecting a positional change of a moving object between a previous screen and a current screen by a predetermined number in the two-dimensional area limited by the area limiting means, wherein the detected positional change is The position calculation means on condition that it is within a predetermined range with respect to the two-dimensional moving velocity vector set by the vector setting means. Start moving object measuring apparatus according to claim 1, further comprising means for. 3. The imaging means includes a platform that is rotatable around at least one axis, and an imaging device fixedly mounted on the platform, and the driving means includes the platform. 3. An apparatus for measuring a moving object according to claim 1, further comprising angle setting means for setting a rotation angle around each of the axes, and zoom lens means attached to the image pickup apparatus. 4. The angle setting means is the at least one.
For each axis, a motor that rotates the platform, an angle encoder that detects the actual rotation angle, and the difference between the detected angle and the target azimuth angle set by the host device are reduced. The zoom lens means includes an angle control means for controlling a motor, and the zoom lens means is configured by a zoom drive motor for setting a zoom position, a zoom encoder for detecting an actual zoom position, a detected zoom position and a host device. A zoom control means for controlling the zoom drive motor so that the difference between the target position and the target position is reduced, and the output of the angle encoder and the output of the zoom encoder are information relating to the imaging direction and the imaging range of the imaging means. The measuring device for a moving object according to claim 3, wherein the measuring device is supplied to the position calculating means as the above.