JPH06105166B2 - Beam center position detector - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention is a third-order method in which a slit light is projected onto a three-dimensional object and a bright line generated on the object surface is imaged by a television camera to measure the distance to the object surface. The present invention relates to an original object measuring device, and more particularly to a beam center position detecting device that can determine the center position of a slit light even when the slit light is split by the measuring object.

[Conventional technology]

Generally, as a visual function used in an inspection process or an assembly robot in a factory, three-dimensional recognition ability is required, but in reality, the conventional method of recognizing an object from luminance image information is not sufficient. Is. for that reason,
Attempts to facilitate recognition and extraction of an object by obtaining not only luminance image information but also distance information to the object have become popular.

FIG. 15 (a) is a diagram showing a configuration of such a conventional three-dimensional object measuring device, and FIG. 15 (b) is a diagram showing bright lines. In the figure, 1 is a television camera, 3 is a projector, and 5 is The measurement surface, 7 is slit light.

In the figure, it is assumed that the television camera 1 and the light projector 3 are arranged at a predetermined position with an interval l, and the light projector 3 projects the slit light 7 over the visual field range P1 to P2 by changing the angle of the measurement surface by a unit angle. . The television camera 1 is capable of capturing an image of the field of view, and captures the bright line of the slit light and displays it on a monitor (not shown).

In such a configuration, first, the television camera scans the object plane to obtain normal luminance image information. Next, the projector 3 is scanned by a predetermined number of steps, and as a result,
The bright line image as shown in FIG. Position of measuring surface 5 now
Suppose that the bright line at Px is observed. Since the angle of the projector 3 in a unit step is known, the angle α of the projector 3 can be known from this and the number of steps, and the incident angle β of the bright line to the television camera can be known from the position of the observed bright line. The distance from the TV camera to the measurement position can be calculated by the trigonometric method with the distance 1 between the TV cameras.

In this way, it is possible to obtain distance information as well as luminance image information.

By the way, when observing a bright line with a television camera, since the slit light has a width, it is necessary to accurately determine the center position of the bright line. Method for obtaining the center position of the binarized image of B, and the method for obtaining the left and right edges of the binarized image of the bright line and estimating the center value by preliminarily measuring and assuming the width of the binarized image of the bright line And so on.

These methods will be described with reference to FIG.

FIG. 16 (a) is a diagram showing a one-dimensional profile of an image taken by a TV camera in a state where a plane where the bright lines are not split is irradiated with laser light, in which the horizontal axis represents the number of pixels and the vertical axis represents luminance. It was taken.

Fig. 16 (b) shows the method of obtaining the center position from the brightness distribution of the bright line. When the brightness distribution of the bright line as shown in Fig. 16 (a) is obtained, a distribution function such as Gaussian distribution is assumed. The central pixel is calculated from the maximum value.

Figure 16 (c) shows the image of the bright line of the laser beam binarized into 2
This figure shows the method of obtaining the center position of the binarized image. The center position of the one-dimensional profile image of the binarized image by setting a certain binarization level for the luminance distribution as shown in Fig. 16 (a). The position where the slit light is projected is obtained by obtaining

FIG. 16 (d) shows a method of preliminarily measuring the width of a binarized image of a bright line and presuming it, and obtaining the left and right edges of the binarized image of the bright line to roughly estimate the center value. The width of the binarized image is obtained in advance, and the central pixel is obtained by adding or subtracting 1/2 of this width as a correction amount from the position of either the left or right edge.

[Problems to be solved by the invention]

However, the observed bright line is divided into 11 and 12 in the case of a three-dimensional object having a shape as shown in Fig. 17 (a), and the bright line image is shown in Fig. 17 (b) (black in the figure). The part will be observed as the bright line. In this case, if the center position of the beam is the position of Pc in Fig. 17 (a), it is necessary to obtain this position accurately.
As shown in Fig. 17 (c), the brightness distribution is obtained, and an appropriate binarization level is set for the brightness distribution, and the brightness distribution as shown in Fig. 17 (d) is set.
A digitized image will be obtained. The center of the beam is 2 each
When the width of the binarized image is A and B, it is calculated as (A + B) / 2.

However, even if the profile of such a binarized image is obtained, in the method of obtaining the brightness distribution described above, the method of obtaining the center value of the binarized image, and the method of obtaining the edge of the bright line and roughly estimating the center value, The position cannot be determined accurately.

In order to solve the above-mentioned problems, the present invention applies the above-described method of obtaining the center value of the binarized image. Therefore, even if the bright line is divided, the center of the projected slit light is accurately detected, and as a result, the object is detected. An object of the present invention is to provide a beam center position detecting device that enables highly accurate measurement of a shape.

[Means for solving problems]

Therefore, the beam center position detection device of the present invention projects the slit light to the measurement object while scanning in steps, and measures the shape of the measurement object by imaging the bright line on the measurement object by the slit light of each scanning step. In the three-dimensional object measuring device, means for obtaining the total bright line width of all the bright lines by integrating the period clocks in which the imaging data of all the bright lines generated by the slit light for each scanning step are obtained,
Means for halving the integrated value, and an address corresponding to ½ of the integrated value with respect to the total bright line width is used as the address onto which the slit light is projected.

[Action]

The beam center position detection device of the present invention projects slit light to a measurement object, obtains a bright line width by counting a period clock during which bright line imaging data by the slit light is obtained, and corresponds to 1/2 of the coefficient value. The address is the address on which the slit light is projected.

〔Example〕

 Embodiments will be described below with reference to the drawings.

FIG. 1 is a block diagram showing a configuration of a beam center position detecting device of the present invention, FIG. 2 is a timing chart, and FIG. 3 is a diagram showing division of a bright line by a measuring object. In the figure, 101 is a binarization circuit, 102 is a clock generator, 103 is an AND circuit, 105
Is a width counter, 107 is a 1/2 circuit, 111 is a selector, 113 is RA
M and 117 are X address counters.

As shown in FIG. 3 (a), the spread of the slit light from the projector 3 divides it into two bright lines on the measurement object, and this is divided into two bright lines by the television camera 1 as shown in FIG. 3 (b). If it is imaged, the brightness distribution is shown in Fig. 3 (c).
As shown in FIG. 3, by setting a predetermined binarization level to this, a binarized image as shown in FIG. 3D can be obtained.

In this way, the binarization circuit 101 generates a binarized signal as shown in FIG. Since this signal and the clock from the clock generator 102 (FIG. 2 (A)) are applied to the AND circuit 103, the width counter 105 counts the count values corresponding to the widths of the two bright lines. That is, as shown in FIG. 2C, for example, when the clock is counted up to n + 3 on the left bright line, the counting starts from n + 4 on the right bright line, and the total value becomes 2n, for example. Then, the count value of the X address counter 117 that counts the clock using the value of the width counter 105 as an address is read as data into the RAM 113, and the value when the count of the width counter 105 is finished is 1/2 in the 1/2 circuit 107. Do 2 If this value n is used as an address and the value of the data read into RAM is taken out, the second
The address value Xn corresponding to the center of the beam can be obtained as shown in FIG.

In the above description, an example in which the beam is divided into two has been described, but the beam center can be similarly obtained regardless of how the beam is divided.

The present invention can be applied to the conventional three-dimensional object measuring device shown in FIG. 15, but can also be suitably applied to the following three-dimensional object measuring device of the present invention.

In the conventional three-dimensional object measuring apparatus shown in FIG. 15 described above, even if the luminance image is obtained instantaneously, in the case of the object having the shape shown in FIG. Therefore, the slit light is first projected onto the TV camera from one side and then scanned, and then the slit light is projected from the opposite side to the TV camera and similarly scanned, so that a total of two scannings are performed. Without it, complete measurement cannot be performed. Since it takes 1/30 seconds, which is the scanning time for one screen, to image one bright line, if this is scanned in the horizontal direction, for example, 512 pixels, it takes about 17 seconds (= 512 × 512 × 1 scan). 1/30) is also required, 3 times in 2 times
It takes 4 seconds.

FIG. 4 is a diagram showing an embodiment of a three-dimensional object measuring device according to the present invention that solves this point, FIG. 5 is a diagram showing a scanning method, and FIG. 6 is an example of measuring a hemispherical measuring object. It is a figure which shows, 1 is a TV camera, 3a, 3b is a left and right floodlight,
S is a reference plane, 5 is a measurement plane, 7a and 7b are slit lights, 11 and 13
Is a step motor, 15 is a preprocessing unit, 17 is a position detection unit, 19
Is a distance calculation unit, 21 is a distance data recording unit, 23 is a distance image memory, 25 is a table correction unit, 27 and 29 are control tables, 31
Is a left projection control unit, and 33 is a right projection angle control unit.

In the figure, projectors 3a and 3b are arranged on both sides of the TV camera 1, and the slit lights 7a and 7b are projected toward the same position on the reference plane S at a constant distance d from the TV camera 1,
As shown in the figure, scanning is performed for each pixel. Floodlight
The projection angles of 3a and 3b are controlled by controlling the step motors 11 and 13 by the left projection angle control unit 31 and the right projection angle control unit 33. In this case, the rotation of the step motor is decelerated and transmitted to the projector by using a gear or the like. Therefore, it is added to the step motor due to a gear pitch error, a bearing diameter error, or a mechanical error such as backlash. Since the relationship between the number of pulses and the rotation angle of the projector is not perfectly linear, the relationship between the pulse and rotation angle is measured in advance, and the number of pulses corresponding to each pixel on the reference plane is set as the control tables 27 and 29. The values are created in advance, and the values in the control table are read out to control the projection angle so that scanning is performed for each pixel. In order to improve the accuracy, it is better to create the control table in advance as described above. Before the measurement, the reference plane is scanned pixel by pixel and the measurement is performed. The number of pulses is converted from the deviation between the original position and the desired position, and the table correction unit 25 corrects the contents of the control tables 27 and 29.

As shown in FIG. 5, the measurement surface 5 is on the front side of the reference surface S by a predetermined distance. By doing this, the position of the bright line on the measurement surface is always separated into the right and left with respect to the target pixel on the reference surface,
By measuring the bright line on the left side by the left slit light and the right bright line by the right slit light respectively, it is possible to perform measurement by eliminating the shadow portion by one scanning.

For example, when a hemispherical object is placed on the measurement surface as shown in FIG. 6 (a), a brightness image is obtained as shown in FIG. 6 (b). If the bright lines at the positions A, B, and C on the reference plane S are as shown in FIG. 6C, when the slit light is projected at the position A on the reference plane, the left slit light is projected on the plane. Therefore, the bright line becomes linear, and the right side slit light projects on the spherical surface, so that the bright line becomes arcuate (FIG. 6 (d)). Further, when light is projected from the position B on the reference plane, both the left slit light and the right slit light are projected on the spherical surface, so that the base line has an arc shape (Fig. 6 (e)), and C on the reference surface is C. Similarly, when the light is projected at the position of, the bright line becomes an arc shape, but since the right side slit light projects the end of the spherical surface,
It becomes a crushed arc shape (Fig. 6 (f)). The broken line indicates the position of the bright line on the reference plane.

The brightness image and the bright line signal thus obtained are used in the television camera 1.
Are taken into the preprocessing unit 15. The pre-processing unit 15 has a frame memory for a luminance image, a circuit for binarizing and extracting the bright line, and a circuit for taking a difference between the luminance image and the bright line image for facilitating the recognition of the bright line in a bright environment. ing.

The position detection unit 17 detects the position of the slit light based on the data from the pre-processing unit, and the distance calculation unit 19 based on this detects the position of the emission line, the position of the television camera, the position of the projector, and the projection angle. Similarly to, the distance calculation is performed by the triangle method.
The calculated distance data is recorded in the distance data recording unit 21, and is written in the address on the image whose position is detected in the distance image memory unit 23.

By the way, if the slit light beam has a width, the beams are not separated in the region where the beams of the left and right projectors overlap, and as a result, the captured bright lines are not separated into the left and right, and the measurement accuracy is degraded.

FIG. 7 is a diagram showing this situation, and shows the case where the width of the beam is 5 pixels. Since the left and right beams overlap in the region from the reference plane to R, the measurement plane should be in front of R. There is a need to. The size of the overlapping area D is the beam width,
Since it changes according to the angle of incidence on the reference plane, it is necessary to set the measurement plane in consideration of these. When the pixel at address x is projected by the left beam and the right beam as shown in the figure, the distance conversion value r1 (in the case of left projection) and the distance conversion value r2 (right projection) of the beam projecting one pixel width are shown. It can be seen that (in the case of light) becomes the distance resolution, and these depend on the incident angle of the beam.

FIG. 8 is a diagram for explaining the distance resolution.
Is an overall configuration diagram, (b) is an enlarged view of the central portion P of the visual field, and (c) is a diagram showing the relationship of distance resolution with respect to the distance between the television camera and the projector.

As shown in FIG. 8A, the incident angle of the laser beam is θ,
θ ′, the distance between the television camera 1 and the light projector 3 is 1, the measurement distance is d, the visual field of the television camera is V, and the visual field of one pixel in the central portion of the visual field is ΔV as shown in FIG. 8B. Then the distance resolution is defined as the resolution Δd for ΔV. This .DELTA.d depends on the incident angle .theta. ', And .theta.' Changes depending on l. Therefore, the relationship between .DELTA.d and l is as shown in FIG. As can be seen from FIG. 8C, the distance resolution can be increased by increasing the distance between the television camera and the projector.

FIG. 9 is a diagram showing an example of a specific configuration of the three-dimensional measuring apparatus of FIG. 4, in which 41 is a He-Ne laser, 43 is a half mirror, 45 and 47 are movable mirrors, and D and E are It is the bright line of the slit light.

In this embodiment, the laser light from the He-Ne laser 41 is divided into two by the half mirror 43, and the movable mirrors 45 and 47 respectively generate left and right slit light. The movable mirror vibrates in the Y direction, and as described above, the stepper motor drives the X mirror.
It is composed of a mirror that is driven in the direction, and slit light is generated by vibrating in the Y direction, and scanning is performed by driving in the X direction. D and E in the figure show the bright lines of the slit light emitted to the object by the left and right projectors.

It should be noted that the number of laser light sources is not one as in this embodiment, but it may be provided for each of the left and right projectors.

FIG. 10 is a diagram showing an example of measuring a three-dimensional object according to the present invention. For example, a hard object having the same reflectance is shown in FIG.
In the case of the arrangement as shown in Fig. 5, the luminance image alone is observed as shown in Fig. 9B, and the relative positional relationship of the corner rigidities cannot be actually discriminated. Signals are obtained as shown in (c) in the figure, and bright lines due to the edges of each coin are obtained as shown in (d) in the figure. Even if they have the same reflectance, their relative positional relationship can be clearly discriminated. It becomes possible to do.

FIG. 11 is a diagram showing a configuration when the beam center position detecting device shown in FIG. 1 is applied to the three-dimensional object measuring device of FIG. 4,
FIG. 12 is a diagram showing a timing chart, and FIG. 13 is a diagram for explaining the operation. In the figure, 101 is a binarization circuit, 103 is an AND circuit, 105 is a width counter, 107 is a 1/2 circuit, 109 is a delay circuit,
111 is a selector, 113 is RAM, 115 is a delay circuit, 117 is an X address counter, 119 is a comparison circuit, 121 is a binary counter,
123 is an X counter.

In the figure, the image clock of the television camera has a cycle of about 80 ns (about 12.5 MHz), is counted by the X address counter 117, and is reset every horizontal scanning (HD). Therefore, the contents of the X address counter 117 represent the X address in each scan line. Since the measurement range must be entirely on the TV screen, the effective measurement area is shorter than one horizontal scan (Fig. 12 (c)). Further, the vertical deflection (VD) signal is input to the binary counter 121, divided into 1/2 and counted by the X counter 123. Therefore, the X counter 123 is counted up every odd field or even field, that is, every one screen update,
This shows the number of scanning steps of the slit light, that is, the current light projection position on the reference plane shown by the broken line in FIG. Therefore, since the scanning position can be seen to the left or right of the projected pixel position on the reference surface by the size of the X address counter and the X counter, the X address counter and the X counter contents when the bright line is detected are compared to determine the X address. The counter <X counter → left bright line X address counter> X counter → right bright line can be determined, and if the output of the comparison circuit 119 is “1”, it is detected as the bright line by the left projector, and if “0”, it is detected as the bright line by the right projector. can do. Although the output of the X address counter is shown as a 9-bit configuration,
This may be appropriately changed depending on the image clock frequency.

On the other hand, the video signal is applied to the binarization circuit 101 and compared with the threshold level to detect the slit projection image, that is, the bright line signal. This bright line signal and the image clock
Since it is added to the AND circuit 103, the width counter 105 counts clocks only while the bright line image is being obtained.

Therefore, for example, the beam is divided into two, and the beam is divided into two as shown in FIG.
In the case where the binarized bright line signal as shown in (1) is obtained, the width counter counts the total width of the divided beams as shown in FIG. When this is completed, the center of the beam can be obtained by halving this value by the 1/2 circuit 107. The 1/2 circuit may be composed of, for example, a shift register, and 1/2 operation may be performed by shifting 1 bit.

Therefore, the value of the X address counter (FIG. 12 (d)) is read as data into the RAM 113 as the address of the value of the width counter, the output of the 1/2 circuit is read as the address, and the output of the comparison circuit 119 at this time is read. If is "1", the first
As shown in FIG. 2 (g), for example, 55 indicates the center position of the beam from the projector from the left side.

When this state is shown in FIG. 13, the clock count number 55 from the scanning start position indicates the center position of the left bright line. Then, the width counter 105 is cleared by the output of the comparison circuit 119 when the contents of the X address counter and the contents of the X counter match. That is, the counter is cleared when the scanning position reaches the light projecting position on the reference surface shown by the broken line in FIG. In the same way, the center position of the right side bright line is detected, and at the same time as the horizontal deflection (HD) blanking is started, the X address is read with the center value of the width as the address.
The position of the beam can be detected by the right projector as the number of counts up to the HD blanking start position (value shown as m in FIG. 13).

After that, the width counter 105 and the 1/2 circuit 107 are reset by HD, and the same detection is performed for each horizontal scan to detect the positions of the left bright line and the right bright line.

The selector 111 is necessary when the RAM 113 has one port, and is not necessary if random access is possible with dual ports. The delay circuits 109 and 115 are for adjusting the timing of each circuit device.

Next, referring to FIG. 14, description will be made on the comparison result of the projection position detection accuracy between the present invention and the conventional method for the state of a typical bright line.

FIG. 14 shows that (A) the object is a flat surface, (B) the object has a reduced width, and (C) the object has an edge width of 2.
The luminance distribution of the bright line in the case where the width is wide and the left bright line is divided, (D) is divided into two by the object edge and the right bright line is wide, and (E) the width is changed due to the change in the reflectance of the object. , Binary image of bright line, center pixel to be obtained, center pixel obtained from luminance distribution, center pixel obtained from binary image, center pixel obtained from left and right edges of binary image, center obtained by the present invention Each pixel is obtained. The width of the bright line is 10 pixels, and the correction amount of the method is 5 pixels.

When the object is a plane (A), the central pixel obtained by each of the methods (1) to (3) coincides with the true central pixel of (3), and no deviation occurs by either method.

The width of the luminance distribution changed due to the edge of the object (B)
In the case of (in the example of the figure, the width changes to half and the center pixel to be obtained is the leftmost pixel), in the method, a Gaussian distribution function etc. is fitted to the binarized image of the bright line and the center position is calculated from the maximum value of the function. In order to obtain, the center position of the binarized image by the method of shifts to the right from the true center pixel shifts by 2 pixels to the right of the true center pixel, and by the method, 5 pixels are added to the left edge. Since 5 pixels are subtracted from the right edge, the center of the left edge becomes the right edge, the center of the right edge becomes the left edge, and the shift amount becomes 0 or 5 pixels. Further, in the present invention, the position is the same as the method.

In the case of (C) in which the left side bright line is 7 pixels and the right side bright line is 3 pixels, the right side bright line is ignored in the method of (2), so the center position is 1 to 2 to the left of the true center position. In the pixel shift method, 5 pixels are added to and subtracted from the left edge of the left bright line and the right edge of the right bright line, so the center of the left edge is 5 pixels from the left end of the left bright line, and the center of the right edge is 5 pixels from the left end of the right bright line. Therefore, the deviation shown in the figure occurs. In the present invention, since the center of the sum of the widths of the two divided pixels is obtained as described above, no deviation occurs.

In the case of (D) in which the left side bright line is 3 pixels and the right side bright line is 7 pixels, the right side bright line is ignored in the method, so that the center position is largely shifted to the left side from the true center position. In this method, 5 pixels are added to and subtracted from the left edge of the left bright line and the right edge of the right bright line, so the center of the left edge is 5 pixels from the left end of the left bright line, and the center of the right edge is 5 pixels from the right end of the right bright line. The center is displaced due to. In the present invention, since the sum of the widths of the divided pixels is calculated as described above, no deviation occurs.

In the case of (E), in which the width of the bright line is changed to 5 pixels, which is half due to the change of the reflectance of the object, there is no deviation in ,,
Then, the center of the left edge is the right edge, the center of the right edge is the left edge, and the shift amount is ± 2 pixels.

〔The invention's effect〕

As described above, according to the present invention, the center position of the beam can be accurately obtained even when the beam is divided, so that the position of the luminance can be detected with high accuracy. It is possible to obtain a highly accurate range image by applying to.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a beam center position detecting device of the present invention, FIG. 2 is a timing chart, and FIG. 3 is a diagram showing division of bright lines. In the drawings, FIG. 4 is a diagram showing an embodiment of a three-dimensional object measuring device according to the present invention, FIG. 5 is a diagram showing a scanning method, FIG. 6 is a diagram showing a measurement example of a hemispherical measuring object, and FIG. FIG. 8 is a diagram for explaining an overlapping region of beams, FIG. 8 is a diagram showing a relationship of distance resolution with respect to a distance between a television camera and a projector, and FIG. 9 is a diagram showing a specific example of the three-dimensional measuring apparatus of the present invention. FIG. 10 is a diagram showing a measurement result according to the present invention, FIG. 11 is a diagram showing a configuration when the beam center position detecting device of FIG. 1 is applied to the three-dimensional object measuring device of FIG. 4, and FIG. FIG. 13 is a diagram showing a timing chart, FIG. 13 is a diagram for explaining the operation, FIG. 14 is a diagram showing a comparison result of projection position detection by the present invention and a conventional method, and FIG. 15 is a conventional three-dimensional The figure which shows the structure of a measuring device, FIG. 16 is a figure for demonstrating the conventional method of calculating | requiring the light projection position, FIG. The figure is a diagram showing an example of division of a bright line by a three-dimensional object. 101 ... Binarization circuit, 102 ... Clock generator, 103 ... A
ND circuit, 105 …… width counter, 107 …… 1/2 circuit, 111 ……
Selector, 113 …… RAM, 117 …… X address counter.

Claims (1)

[Claims] 1. A three-dimensional object measuring device for measuring the shape of a measuring object by projecting slit light onto a measuring object while scanning stepwise and imaging a bright line on the measuring object by the slit light of each scanning step. , Means for obtaining a total bright line width of all the bright lines by integrating clocks during which image data of all bright lines generated by the slit light for each scanning step are obtained, and means for halving the integrated value Equipped with
A beam center position detecting device, wherein an address corresponding to 1/2 of an integrated value with respect to the total bright line width is used as an address on which slit light is projected.