JPH0658212B2 - Three-dimensional coordinate measuring device - Google Patents

Description

The present invention relates to an improvement of a three-dimensional coordinate measuring device.

(Prior Art) Conventionally, there are various methods for measuring the three-dimensional coordinates of a measurement object having a three-dimensional spread, and one of them is a focus position changing method. In this method, the measurement target is imaged many times while changing the focus position of the imaging device, and the in-focus image for each part of the measurement target is determined from among the obtained many images, and The distance from the state of the optical system of the image pickup device to each part of the measurement object, that is, the three-dimensional coordinates for each part is obtained.

In order to determine the focused image described above, the magnitude of image contrast is usually used.

When a measurement target having a three-dimensional spread is imaged by using an imaging device having an optical system having a predetermined depth of focus, the in-focus portion of the image has the highest contrast, and conversely Since the displaced part causes blurring,
The contrast decreases. That is, when the focus position is continuously changed from the front to the back, the contrast of the edge portion in the image is initially small due to the blur, becomes larger as the focus comes in, and becomes the largest when the focus comes in. , And then it gets smaller again due to the blur.

Therefore, the contrast of the edge portion, that is, the distance corresponding to the in-focus position of the optical system when the image with the largest differential component can be regarded as the distance to the characteristic point,
The distance can be measured by finding the one having the largest differential component in the edge portion among a large number of images having different in-focus positions.

FIG. 2 shows an example of a three-dimensional coordinate measuring device using the above method. The video camera 1 forms an image of the measuring object (subject) A on the image sensor 3 via the lens 2,
Convert to video signal. The video signal is converted from an analog signal to a digital signal by an analog / digital (A / D) converter 4 and transferred to the processing device 5. The processing device 5 performs a differential calculation process on the digital signal to detect the contrast of the image of the subject A, and further controls the lens controller 6 based on this result to control the position of the lens 2, that is, the focus position of the video camera 1. To change. Further, the image of the subject A after changing the in-focus position is converted into a video signal by the video camera 1 in the same manner as described above, transferred to the processing device 5 via the A / D converter 4, and processed. By repeating such a series of processes, the contrast in a large number of images with different in-focus positions is detected, and the object A
The distance to each part of the subject A is obtained by selecting the image having the largest contrast for each part.

(Problems to be Solved by the Invention) By the way, the measurement accuracy in the above-described device depends on the contrast detection accuracy.

When the object point is imaged on the conjugate image plane by the optical system, if the image plane is moved back and forth, it passes through the exit pupil of the optical system,
The bundle of rays that converges on the image point produces a blurred image on the image plane. The size of this blur is proportional to the angle of opening of the bundle of rays that converges toward the image point and the amount of deviation from the correct position on the image plane. If the size of the blur is smaller than the resolution ε of the image sensor, It is considered that the image is practically clear, that is, has a high contrast. When the position of the image plane is fixed, the range of the object that produces a clear image on this is called the object (field of field) depth.

FIG. 3 shows the relationship between the image on the image plane and the object depth. In the figure, Q is the image plane, O is the object plane conjugate to the image plane Q, E is the optical system, and here the focal length f Is the center plane of the lens 7,
The distance between the image plane Q and the center plane E is a, and the distance between the center plane E and the object plane O is b.

The object point on the object plane O forms a clear image on the image plane Q via the lens 7. Further, the object points on the planes O1 and O2 existing at the distances Δ1 and Δ2 from the object surface O before and after the object surface O form a circular image of a predetermined size on the image surface Q via the lens 7. However, if the diameter of the circular image is equal to or less than the resolution ε, it can be considered that a clear image is formed on the image plane Q.

When the diameter of the circular image on the image plane Q corresponding to the object points on the planes O1 and O2 is equal to the resolution ε, the planes O1 and O
The interval between the two, that is, Δ = Δ2-Δ1 (1) is called the object depth.

Further, assuming that the lateral magnification of the lens 7 is β and the diameter of the exit pupil by the diaphragm 8 is D, if | β | D >> ε, Δ2≈−Δ1
Therefore, the object depth Δ is Δ = Δ2-Δ1≈2ε · b / | β | D (2)

As described above, since a clear image can be obtained on the image plane Q, that is, the range of the object plane where a large contrast can be obtained has a certain width, it is difficult to measure the distance with high accuracy by the device. In particular, the distance to the subject, that is, the distance b in equation (2)
There is a problem that the larger the value is, the larger the object depth Δ is, and the measurement accuracy is deteriorated.

SUMMARY OF THE INVENTION It is an object of the present invention to solve the above problems and provide a three-dimensional coordinate measuring device which has high measurement accuracy and can measure a distant subject with high accuracy.

(Means for Solving the Problems) FIG. 1 shows an outline of a three-dimensional coordinate measuring apparatus of the present invention. In the figure, 9 is an image pickup apparatus having an optical system 9a whose magnification can be changed, and 10 is an image pickup apparatus. The measurement target A is imaged many times while changing the in-focus position of 9, and the in-focus image for each part of the measurement target A is determined from the obtained many images, and the imaging device at the time of imaging each image The coordinate calculation means for obtaining three-dimensional coordinates for each part of the measurement object A from the state of the optical system 9a of the reference numeral 9, and 11 is the magnification of the optical system 9a of the imaging device 9 from the lowest magnification to the highest magnification (or the highest magnification). From the lowest to the lowest magnification), and with each change,
It is a magnification changing unit that restarts the coordinate calculating unit 10.

Further, FIG. 4 shows an outline of another three-dimensional coordinate measuring apparatus of the present invention. In the figure, 12 and 13 are a pair of left and right image pickup devices arranged at a predetermined interval, and 14 and 15 are image pickup devices. Corresponding to 12 and 13, respectively, the measurement object A is imaged many times while changing the in-focus position of the imaging devices 12 and 13, and each part of the measurement object A is focused from the obtained many images. 3) of the left image and the right image for each part of the measuring object A from the states of the optical systems 12a and 13a of the imaging devices 12 and 13 at the time of capturing each image.
First and second coordinate calculating means for obtaining dimensional coordinates, 16 are provided for the left image and the right image based on the three-dimensional coordinates of the left image and the right image obtained by the first and second coordinate calculating means 14, 15. A search range limiting means for dividing an image into a plurality of ranges having similar distances, and 17 searches for corresponding points in the ranges in the left image and the right image captured by the image capturing devices 12 and 13, and their three-dimensional It is a third coordinate calculating means for obtaining coordinates.

Further, FIG. 5 shows an outline of still another three-dimensional coordinate measuring apparatus of the present invention. In the figure, 14 and 15 are first and second coordinate calculating means 16 similar to those of the apparatus of FIG. Is a search range limiting means similar to that of the apparatus of FIG. 4, 17 is third coordinate calculating means similar to that of the apparatus of FIG. 4, and 18 and 19 are a pair of left and right having optical systems 18a and 19a whose magnification can be changed, respectively. Image pickup devices, 20 and 21 are a pair of left and right image pickup devices 1.
The magnifications of the optical systems 18a and 19a of 8 and 19 are sequentially changed from the lowest magnification to the highest magnification (or from the highest magnification to the lowest magnification), and each time the change is performed, the first magnification is changed.
And first restarting the second coordinate calculating means 14 and 15
And second magnification changing means.

(Operation) According to the apparatus of FIG. 1, the magnification changing means 11 can change the magnification of the optical system 9a of the image pickup apparatus 9, for example, when the magnification is increased, the object depth of the optical system 9a of the image pickup apparatus 9 is also changed.
That is, it becomes smaller.

FIG. 6 shows the relationship between the lateral magnification of the optical system and the object depth. If the lateral magnification of the lens 22 is larger than that of the lens 23 in the optical system, for example, the lenses 22 and 23, that is, the focal length is longer, The object depth Δ01 of the lens 22 is smaller than the object depth Δ02 of the lens 23.

Therefore, in the above-mentioned apparatus, when the coordinate changing means 11 is restarted by the magnification changing means 11, the in-focus image is decided out of the many images picked up by the image pickup device 9 in the state where the object depth is small. From this, the three-dimensional coordinates of the measuring object A can be obtained with high accuracy.

Further, according to the apparatus of FIG. 4, the measurement object A is selected from a large number of left and right images imaged by the first and second coordinate calculating means 14 and 15 while changing the in-focus positions of the imaging devices 12 and 13. The in-focus left image and right image are determined for each part of
The three-dimensional coordinates of the left image and the right image for each part of the measurement object A are obtained from the states of the optical systems 12a and 13a of No. 3. Further, from the three-dimensional coordinates of the left image and the right image, the search range limiting means 16 divides the left image and the right image into a plurality of ranges having the same distance, and the third coordinate calculating means 17 further causes the image capturing device 12 to be divided. And 13 are searched for the corresponding points from the ranges in the left and right images, that is, within the range where the corresponding points are expected to exist from the measurement results by the first and second coordinate calculating means 14 and 15. , Its three-dimensional coordinates are obtained.

According to the apparatus of FIG. 5, the object depth is changed by the first and second magnification changing means 20 and 21 corresponding to the first and second coordinate calculating means 14 and 15, for example, in a small state. The three-dimensional coordinates of the left image and the right image for each part of the measurement object A are obtained based on a large number of left and right images captured by the image capturing devices 18 and 19, and the left image and the right image are obtained in the same manner as described above. The image pickup devices 18 and 19 are divided into a plurality of ranges having similar distances.
Corresponding points are searched for in the ranges in the left and right images captured by, that is, the range in which the presence of corresponding points is predicted with a higher probability, and the three-dimensional coordinates thereof are obtained.

(Embodiment) FIG. 7 shows a first embodiment of the three-dimensional coordinate measuring apparatus of the present invention, and here shows an example corresponding to the apparatus of FIG. 1 described above. In the figure, 31 is a video camera, 32 is a camera mount, 33 is an A / D converter, 34 is a camera controller, 35 is a lens controller, and 36 is a processing device.

The video camera 31 is provided with an optical system whose magnification can be changed, here, a zoom lens 31a, and an image of a measurement object (subject) A is formed on an image sensor (not shown) through the zoom lens 31a. Convert to signal. The camera mount 32 is configured to support the video camera 31 so that it can be swung freely in any of horizontal and vertical directions.

The camera controller 34 drives the camera mount 32 based on the control command from the processing device 36, and changes the imaging direction of the video camera 31. Lens controller 3
Reference numeral 5 drives the zoom lens 31a of the video camera 31 on the basis of a control command from the processing device 36 to change the magnification and the focus position.

The processing device 36 comprises a well-known computer and the like, controls the camera controller 34 and the lens controller 35 in accordance with a predetermined program, and receives the image pickup signal of the subject A converted into a digital signal via the A / D converter 33. Then, the three-dimensional coordinates of the subject A are calculated by performing a predetermined calculation process.

FIG. 8 is a flow chart corresponding to the program of the processing device 36, and the program and the processing device 36 realize the coordinate calculating means 10 and the magnification changing means 11 in FIG.

Next, the operation will be described.

First, the processing device 36 operates the camera mount 32 and the zoom lens 31a via the camera controller 34 and the lens controller 35, and sets the magnification and focus position of the zoom lens 31a in a predetermined initial state (for example, the state where the magnification is the lowest). ) And the video camera 31
The imaging field of view is set to a predetermined imaging start position (step s
1) Input an image (step s2). At this time, if the subject A is not captured (step s3), the camera mount 32 is operated via the camera controller 34 to move the field of view of the video camera 31 in a predetermined scanning direction (step s4), and image input is performed again. To do.

Next, the processing device 36 receives an input from the video camera 31
The video signal converted into the digital signal by the / D converter 33 is subjected to the same differential operation as in the conventional example to detect the contrast (image blur) of the image of the subject A (step s5). Further, the processing device 36 changes the focus position of the zoom lens 31a based on the detection result (steps s6, s7), repeats the image input (step s8) and the image blur detection (step s5), and the maximum value of the contrast is The obtained focus position is determined (step s9).

After that, the processing device 36 moves the imaging visual field of the video camera 31, and steps s2 to s2 described above for the entire visual field.
The process of step s9 is repeated (steps s10 and s11).

Next, the processing device 36 changes the focal length of the video camera 31, that is, the magnification, and repeats the processing of steps s2 to s11 described above for all the magnifications (steps s12, s).
13) Based on these results, the distance to each part of the subject A, that is, the three-dimensional coordinates is calculated (step s14).

According to the present embodiment, an image of the subject A is obtained by the zoom lens 31a whose magnification is changed, that is, the object depth is changed, and the focus position of the video camera 31 in the state where the object depth is small is about each part of the subject A. Since these are detected and the distances from these to each part of the subject A are calculated, highly accurate three-dimensional coordinates can be obtained.

Further, in the present embodiment, the magnification is changed after scanning the entire visual field, but the magnification may be sequentially changed in a certain visual field range, and then the visual field may be changed. The entire field of view means the movable range of the camera mount 32, or the video camera 31 in the state where the magnification is the lowest.
Is the angle of view.

FIG. 9 shows a second embodiment of the three-dimensional coordinate measuring apparatus of the present invention, and here shows an example corresponding to the apparatus of FIG. 4 described above. In the figure, 37 and 38 are video cameras, 39 and 40 are camera mounts, 41 and 42 are A / D converters, 43 and 44 are camera controllers, and 45 and 46.
Is a lens controller, and 47 is a processing device.

The video cameras 37 and 38 are standard lenses 37a, respectively.
And 38a, which are arranged at predetermined intervals, form a left image and a right image of the subject A on an image sensor (not shown) via the standard lenses 37a and 38a,
Convert to video signal. The camera mounts 39 and 40 are designed to support the video cameras 37 and 38 so that they can be swung freely in horizontal and vertical directions.

The camera controllers 43 and 44 drive the camera mounts 39 and 40 based on the control command from the processing device 47, and change the imaging directions of the video cameras 37 and 38. The lens controllers 45 and 46 are controlled by the video cameras 37 and 38 based on the control command from the processing unit 47.
The standard lenses 37a and 38a are driven to change their in-focus positions.

The processing device 47 comprises a well-known computer and the like, and the camera controllers 43 and 44 are operated according to a predetermined program.
In addition to controlling the lens controllers 45 and 46, the video signals corresponding to the left and right images of the subject A converted into digital signals via the A / D converters 41 and 42 are received, and predetermined arithmetic processing is performed. Subject A going
The three-dimensional coordinate of is calculated.

10 (a) and 10 (b) are flow charts corresponding to the program of the processing device 47. The program and the processing device 47 allow the first and second coordinate calculating means 14 and 15 and the search range limiting means 16 in FIG. In addition, the third coordinate calculating means 17 is realized.

Next, the operation will be described.

First, the processing device 47 detects the distances to the respective parts of the left image and the right image of the subject A, that is, the three-dimensional coordinates, based on the focusing position change method using the video cameras 37 and 38, respectively (step sp1, sp2).

The operation in steps sp1 and sp2 is the tenth.
The description is omitted because it is the same as the first embodiment except that the magnification is not changed in the video camera as shown in FIG.

Next, the processing device 47 divides the left image and the right image of the subject A into a plurality of ranges having the same distance based on the value of the distance to each part of the left image and the right image of the subject A described above (step sp3).

Further, the processing device 47 performs a corresponding point search based on the well-known binocular stereoscopic vision for each of the ranges (search range) in the left image and the right image of the subject A input via the video cameras 37 and 38. (Step sp4), and the distance to each part of the subject A, that is, the three-dimensional coordinates is calculated based on this (step sp5).

According to the present embodiment, prior to the detection of the three-dimensional coordinates by the binocular stereoscopic method, the same is done from the three-dimensional coordinates of the left image and the right image by the focus position changing method using the two video cameras 37 and 38. Since a range having a large distance, that is, a range in which a corresponding point is expected to exist, is obtained, the range of the corresponding point search can be narrowed, the calculation amount associated with the corresponding point search can be reduced, and the accuracy thereof can be improved. And more accurate three-dimensional coordinates can be obtained.

FIG. 11 shows a third embodiment of the three-dimensional coordinate measuring apparatus of the present invention, and here shows an example corresponding to the apparatus of FIG. 5 described above. In the figure, 39 and 40 are camera mounts, 4
1 and 42 are A / D converters, 43 and 44 are camera controllers, 48 and 49 are video cameras, and 50 and 51.
Is a lens controller, and 52 is a processing device.

The video cameras 48 and 49 are zoom lenses 48, respectively.
a and 49a, which are arranged at a predetermined interval, and the subject A through the zoom lenses 48a and 49a.
The left image and the right image of are imaged on an image sensor (not shown) and converted into video signals.

The lens controllers 50 and 51 drive the zoom lenses 48a and 49a of the video cameras 48 and 49 based on the control command from the processing unit 52, and change the magnification and focus position thereof.

The processing device 52 comprises a well-known computer and the like, and the camera controllers 43 and 44 according to a predetermined program.
In addition to controlling the lens controllers 50 and 51, the video signals corresponding to the left and right images of the subject A converted into digital signals via the A / D converters 41 and 42 are received, and predetermined arithmetic processing is performed. Subject A going
The three-dimensional coordinate of is calculated. The other structure is the same as that of the second embodiment.

FIGS. 12 (a) and 12 (b) are flowcharts corresponding to the program of the processing device 52. The program and the processing device 52 allow the first and second coordinate calculating means 14 and 15 and the search range limiting means 16 in FIG. , Third coordinate calculating means 17 and first and second magnification changing means 20 and 21 are realized.

Next, the operation will be described.

First, the processing device 52 detects the distances to the left and right images of the subject A, that is, the three-dimensional coordinates, based on the focusing position change method involving magnification change using the video cameras 48 and 49, respectively. (Steps sp6, sp7).

The operations in steps sp6 and sp7 are the twelfth.
Since it is the same as the first embodiment as shown in FIG. 6B, its explanation is omitted.

Hereinafter, as in the case of the second embodiment, the processing device 52 has the same distance between the left image and the right image of the subject A from the value of the distance to each part of the left image and the right image of the subject A. Each of the ranges (search range) in the left image and the right image of the subject A, which are divided into a plurality of ranges (step sp3) and input via the video cameras 48 and 49, is based on the well-known binocular stereoscopic method. Corresponding point search is performed (step sp4), and the distance to each part of the subject A, that is, three-dimensional coordinates is calculated based on this (step sp5).

According to the present embodiment, the corresponding point search range obtained prior to the detection of the three-dimensional coordinates by the binocular stereoscopic method is the video cameras 48 and 4 equipped with the zoom lenses 48a and 49a.
Since it is obtained more accurately based on the image input from 9, it is possible to further reduce the amount of calculation involved in the corresponding point search as compared with the second embodiment, and it is possible to further improve its accuracy. Further accurate three-dimensional coordinates can be obtained.

In the embodiments, a zoom lens that can continuously change the focal length is used as the optical system whose magnification can be changed, but a lens that can change the focal length stepwise may be used.

(Effects of the Invention) As described above, according to the present invention, by changing the magnification of the optical system of the image pickup apparatus, a focused image of a subject in a state where the object depth is reduced can be obtained, and based on this, Since the three-dimensional coordinates are calculated, the three-dimensional coordinates with less error can be obtained, and particularly, the accurate three-dimensional coordinates can be obtained even when the distance to the subject is large.

Further, according to the present invention, focused images are respectively obtained from a pair of left and right imaging devices, three-dimensional coordinates of a left image and a right image of a subject are calculated, and based on these, a left image and a right image of the subject are calculated. Limits the range with a similar distance of
Since three-dimensional coordinates are calculated by performing the corresponding point search by the well-known binocular stereoscopic method within the range, the amount of calculation for obtaining the corresponding points from the left and right images can be reduced, and the accuracy can be reduced. Corresponding points can be obtained well, and therefore highly accurate three-dimensional coordinates can be obtained.

Furthermore, according to the present invention, focused images in a state where the object depth is reduced are obtained from a pair of left and right image pickup devices, from which the three-dimensional coordinates of the left and right images of the subject are calculated, and based on these, Since the range having the same distance in the left and right images of the subject is limited and the corresponding points are searched by the well-known binocular stereoscopic method within the range, the three-dimensional coordinates are calculated. It is possible to remarkably reduce the amount of calculation for obtaining the corresponding points from the image, and it is possible to obtain the corresponding points with extremely high accuracy, and therefore, it is possible to obtain more highly accurate three-dimensional coordinates.

Further, any of the devices of the present invention can measure by natural light, can measure the three-dimensional coordinates of an object, a landscape, etc. in the natural light, and can accurately measure three-dimensional information for an image database and CAD. It has the advantage of being applicable to topographical analysis by aerial photography, automatic driving of automobiles, visual control of robots, etc.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a functional block diagram showing an outline of a three-dimensional coordinate measuring device of the present invention, FIG. 2 is a configuration diagram showing an example of a conventional three-dimensional coordinate measuring device, and FIG. 3 is an image. FIG. 4 is an explanatory diagram showing a relationship between an image on a surface and an object depth, FIG. 4 is a functional block diagram showing an outline of another three-dimensional coordinate measuring device of the present invention, and FIG. 5 is still another three-dimensional coordinate of the present invention. 6 is a functional block diagram showing the outline of the measuring device, FIG. 6 is an explanatory diagram showing the relationship between the lateral magnification of the optical system and the object depth, and FIG. 7 shows the first embodiment of the three-dimensional coordinate measuring device of the present invention. Configuration diagram, FIG. 8 is a flow chart corresponding to the program of the processing device of FIG. 7, FIG. 9 is a configuration diagram showing a second embodiment of the three-dimensional coordinate measuring device of the present invention, and FIG. 10 (a) ( b)
Is a flow chart corresponding to the program of the processing apparatus of FIG. 9, FIG. 11 is a configuration diagram showing a third embodiment of the three-dimensional coordinate measuring apparatus of the present invention, and FIGS. 12 (a) and (b) are FIG. 4 is a flowchart corresponding to the program of the processing device of FIG. 9, 12, 13, 18, 19 ... Imaging device, 9a, 12
a, 13a, 18a, 19a ... Optical system, 10 ... Coordinate calculating means, 11 ... Magnification changing means, 14 ... First coordinate calculating means, 15 ... Second coordinate calculating means, 16 ... Search range limiting means, 17 ... Third coordinate calculating means, 20 ... First magnification changing means, 21 ... Second magnification changing means.

Claims (3)

[Claims] 1. An image pickup device and a measurement target imaged a number of times while changing an in-focus position of the image pickup device, and an in-focus image is determined for each part of the measurement target object from the obtained many images. Then, in a three-dimensional coordinate measuring device having coordinate calculating means for obtaining three-dimensional coordinates for each part of the measurement object from the state of the optical system of the image capturing device at the time of capturing each image, the three-dimensional coordinate measuring device has an optical system whose magnification can be changed. Using the imaging device, the magnification of the optical system of the imaging device is sequentially changed from the lowest magnification to the highest magnification (or from the highest magnification to the lowest magnification), and the coordinate calculation means is restarted each time the magnification is changed. A three-dimensional coordinate measuring device characterized in that a changing means is provided. 2. A pair of left and right image pickup devices arranged at a predetermined interval, and a pair of left and right image pickup devices respectively corresponding to a pair of left and right image pickup devices. An in-focus image is determined for each part of the measurement object from among the obtained many images, and the left image and the right image for each part of the measurement object are determined from the state of the optical system of the imaging device at the time of imaging each image. Based on the three-dimensional coordinates of the left image and the right image obtained by the first and second coordinate calculating means, the left image and the right image are similar to each other. Search range limiting means for dividing into a plurality of ranges having a distance, and searching for corresponding points in the ranges in the left and right images picked up by a pair of left and right image pickup devices, and obtaining three-dimensional coordinates thereof. With the coordinate calculation means of And a three-dimensional coordinate measuring device. 3. A pair of left and right image pickup devices each having an optical system capable of changing the magnification is used, and the magnification of the optical system of the pair of left and right image pickup devices is changed from the lowest magnification to the highest magnification (or from the highest magnification to the lowest magnification). The first and second magnification changing means for sequentially changing the magnifications and restarting the first and second coordinate calculating means each time the change is made are provided. Three-dimensional coordinate measuring device.