JP2001074412A - Method for correcting lens position and three- dimensional input device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for positioning as designed irrespective of a magnitude of the error of an unfixed quantity such as variations of parts, an assembling error in lens moving. SOLUTION: In this method for correcting a lens position in an image pickup device 2, capable of varying a magnification, comprising a lens moving in the optical axis direction and a controller 58A moving a lens based on control data indicating a lens moving amount in correspondence with each of a plurality of setting magnifications, the control data are stored in a memory 61 writable at least once, and an actual magnification is measured when moving the lens, based on the control data stored in the memory 61A, and the control data are corrected in correspondence with an error between the measured magnification and the setting magnification in correspondence with the control data, and the corrected control data are stored in the memory 61A instead of the control data prior to correction.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for correcting a lens position in zooming and an optical three-dimensional input device having a zooming function.

[0002]

2. Description of the Related Art An optical three-dimensional input device (range finder) is used for data input to a CG system or a CAD system, body measurement, visual recognition of a robot, and the like. Methods for optically measuring the shape of an object without contact include a slit light projection method (light cutting method), a pattern light projection method, a stereoscopic method, and an interference fringe method. For example, in the slit light projection method, a triangle is obtained from the projection angle of the slit light, the incident angle of the slit light reflected by the object to be measured and returned, and the base line length (the distance between the projection start point and the light receiving principal point). The distance between the objectives is calculated by a calculation to which the survey method is applied. By calculating the inter-object distance for a plurality of points on the object, three-dimensional data indicating the shape of the object can be obtained.

In a range finder having a zooming function for changing a measurement range, optical parameters such as a principal point position of light reception and an image distance are related to calculation. The values of these optical parameters change depending on the position of the movable part (zooming state). In particular, the influence of the movement of the variator lens is great. Therefore, in three-dimensional input using this type of range finder, lens information indicating optical parameter values at a plurality of appropriate set positions within the lens movement range is prepared in advance. Optical parameter values at lens positions other than these set positions can be calculated by interpolation based on lens information. In the case of using the slit light projection method, it is desirable to optimize the projection angle range in accordance with the change in the angle of view of the received light, so that lens information indicating the relationship between the lens position and the angle of view is required.

[0004] Such lens information can be created by calculation based on the design values of the zoom lens system.
However, in practice, there are variations in components and assembly errors in the zoom lens system, so that the optical parameter values do not always match the design values. That is, in mass production of the range finder, a certain degree of variation occurs in the optical parameter value between products.

On the other hand, in the related art, a DC motor is used for moving a movable portion in a zoom lens system, and feedback type movement control is performed in which the amount of movement is adjusted while detecting the lens position with an encoder (for example, a metal brush type). I was Then, in order to enhance the reproducibility of the movement stop position, one-sided feeding with a minute return amount that cancels backlash of the gear has been performed.

[0006]

In order to perform an accurate three-dimensional input, it is necessary to calibrate the lens information by measuring the optical parameter value at each of the above-described set positions. In this calibration, it is essential that the positioning state when the lens is disposed at each set position has reproducibility. If the positioning state differs between the state at the time of calibration and the time of actual measurement thereafter, an error occurs in the three-dimensional data. Further, it is desirable that the set position and the actual lens arrangement position are made to coincide with each other. The greater the displacement between these positions, the lower the accuracy of calibration.

Conventionally, there are variations in parts related to lens control, assembly errors, and a time lag from detection by the encoder to stop of movement, so that it is difficult to accurately arrange the lens at the set position, and it is difficult to adjust the imaging magnification. A displacement of about 5 to 10%, which is represented by an error, occurred. The lens posture when the movement was stopped was unstable and the reproducibility was low. In addition, there is a problem in that when there is a variation in parts related to lens control and an assembly error, the influence appears as a displacement of positioning.

An object of the present invention is to realize positioning as designed irrespective of the magnitude of an indeterminate error such as component variation and assembly error in moving a lens. Another object is to enable a three-dimensional input without variable error effects.

[0009]

According to a first aspect of the present invention, there is provided a method according to the first aspect, wherein the lens is movable based on control data indicating a lens movement amount corresponding to each of a plurality of set magnifications. A method for correcting a lens position in a variable magnification imaging apparatus comprising a controller for moving the control data, wherein the control data is stored in a memory that can be rewritten at least once, and the control data stored in the memory is stored in the memory. The actual magnification when the lens is moved based on the control data is measured, and the control data is corrected according to an error between the measured magnification and a set magnification corresponding to the control data. The data is stored in the memory in place of the control data before correction.

According to a second aspect of the present invention, there is provided an apparatus having variable magnification imaging means having a lens movable in the direction of the optical axis, wherein the magnification is variable according to the object image obtained by the imaging means and the magnification of the imaging. A three-dimensional input device that outputs data obtained from a reference position determined by an arrangement state of the origin switch for detecting the position of the lens and an arrangement of the origin switch, to a lens position corresponding to each of a plurality of set magnifications. Control data indicating the amount of lens movement of at least one
A nonvolatile memory that can be rewritten a number of times; and a controller that moves the lens based on the control data stored in the memory.

In a three-dimensional input device according to a third aspect of the present invention, the movement of the lens is performed by a pulse motor.
The control data is a numerical value per rotation step of the pulse motor.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS [System Configuration and Operation] FIG. 1 is a configuration diagram of a measurement system according to the present invention.

The measurement system 1 is a three-dimensional camera (range finder) for performing three-dimensional measurement by a slit light projection method.
2 and a host 3 for processing output data of the three-dimensional camera 2
It is composed of

The three-dimensional camera 2 outputs a color image of the object Q and data necessary for calibration, together with measurement data corresponding to the three-dimensional position of a sampling point on the object Q as a measurement target. The host 3 is in charge of calculating the coordinates of the sampling points using the triangulation method.

The host 3 includes a CPU 3a, a display 3
b, a keyboard 3c, a mouse 3d, and the like. Software for processing measurement data is incorporated in the CPU 3a. Two forms of data transfer between the host 3 and the three-dimensional camera 2 are possible: online by cable or infrared communication, and offline by a portable recording medium 4. Examples of the recording medium 4 include a magneto-optical disk (MO), a mini disk (MD), and a memory card.

FIG. 2 is a diagram showing the appearance of the three-dimensional camera. A light emitting window 20a and a light receiving window 2 are provided on the front of the housing 20.
0b is provided. The light projecting window 20a is located above the light receiving window 20b. The slit light U having a predetermined width w emitted from the internal optical unit OU passes through the light projecting window 20a toward the object (subject) to be measured. The radiation angle φ in the length direction M1 of the slit light U is fixed. A part of the slit light U reflected on the surface of the object enters the optical unit OU through the light receiving window 20b. The optical unit OU is
A two-axis adjustment mechanism is provided for optimizing the relative relationship between the light emitting axis and the light receiving axis.

On the upper surface of the housing 20, zoom buttons 25a and 25b, a manual focusing button 26
a, 26b, an auto-focusing button 26c, and a shutter button 27 are provided. As shown in FIG. 2B, the liquid crystal display 2
1. cursor button 22, select button 23, cancel button 24, analog signal terminal 32, SCSI terminal 33 as a digital signal terminal, and recording medium 4.
Are provided.

A liquid crystal display (LCD) 21 is used as an operation screen display means and an electronic finder.
The user (photographer) can set the photographing mode by using the buttons 21 to 24 on the back. From the analog signal terminal 32, a color image signal is output, for example, in the NTSC format.

FIG. 3 is a block diagram showing a functional configuration of the three-dimensional camera 2. In the figure, solid arrows indicate the flow of electric signals, and broken arrows indicate the flow of light. 3D camera 2
Has two optical systems 40 and 50 on the light projecting side and the light receiving side that constitute the optical unit OU described above. Optical system 40
, The wavelength 6 emitted by the semiconductor laser (LD) 41
The 85 nm laser beam becomes slit light U by passing through the light projecting lens system 42 and is deflected by a galvanomirror (scanning means) 43. Semiconductor laser 4
1 driver 44 and drive system 46 of galvanometer mirror 43
Is controlled by the system controller 61. Data required for control is recorded in a rewritable nonvolatile control data memory 61A unique to the present invention. The drive system 45 of the light projecting lens system 42 includes the lens controller 5
8. The lens controller 58 has a RAM 58A used as a work area for control.

In the optical system 50 on the light receiving side, the light condensed by the zoom unit 51 is transmitted to the beam splitter 5.
2 is split. Light in the oscillation wavelength band of the semiconductor laser 41 is incident on the image sensor 53 for measurement. The light in the visible band enters the monitor color image sensor 54. Both the image sensor 53 and the color image sensor 54 are CCD imaging devices. To perform auto focusing (AF), the zoom unit 51
AF module 57 is arranged in the vicinity of. The lens controller 58 controls the focusing drive system 59 based on the distance data (distance measurement result) output from the AF module 57. The zooming drive system 60 is provided for electric zooming.

The flow of object information in the three-dimensional camera 2 is as follows. First, shooting information from the image sensor 53 is transferred to the signal processing circuit 62 in synchronization with a clock from the driver 55. The signal processing circuit 62 includes an amplifier for amplifying the photoelectric conversion signal of each pixel output from the image sensor 53, and an AD converter for converting the photoelectric conversion signal into 8-bit light reception data. Signal processing circuit 62
Are temporarily stored in the memory 63 and then sent to the center-of-gravity calculation circuit 73. The address designation at this time is performed by the memory controller 65. Center of gravity calculation circuit 7
Reference numeral 3 calculates data (hereinafter referred to as a “center of gravity ip”) serving as a basis for calculating a three-dimensional position based on the input light receiving data, and outputs it to the output memory 64. The center of gravity ip is also sent to the display memory 74 as monitor information.
The LCD 21 is used for displaying a distance image. The system controller 61 gives a character generator (not shown) an instruction to display appropriate characters and symbols on the screen of the LCD 21.

On the other hand, imaging information from the color image sensor 54 is transferred to the color processing circuit 67 in synchronization with a clock from the driver 56. The color-processed imaging information is supplied to the NTSC conversion circuit 70 and the analog output terminal 3
2 or on-line output or digitizing circuit 6
8 and stored in the color image memory 69.
Thereafter, the color image data is transferred from the color image memory 69 to the SCSI controller 66, and is output online or stored in the recording medium 4.

It should be noted that the color image is output from the image sensor 53.
Is an image having the same angle of view as that of the distance image, and is used as reference information in application processing in the host 3. Examples of the process using a color image include a process of generating a shape model by combining a plurality of sets of measurement data having different camera viewpoints and a process of thinning out unnecessary vertices of the shape model.

FIG. 4 is a diagram showing the construction of the light projecting lens system. FIG.
(A) is a front view and FIG. 4 (b) is a side view. The light projection lens system 42 includes three lenses: a collimator lens 421, a variator lens 422, and an expander lens 423. The laser beam emitted from the semiconductor laser 41 is subjected to an optical process for obtaining an appropriate slit light U in the following order. First, the collimator lens 421 converts the beam into substantially parallel light slightly inward. Next, the beam diameter of the laser beam is adjusted by the variator lens 422. Finally, the beam is expanded in the slit length direction M1 by the expander lens 423.

The variator lens 422 is provided to allow the slit light U having a width corresponding to a plurality of pixels to enter the image sensor 53 regardless of the photographing distance and the angle of view. The drive system 45 moves the variator lens 422 so as to keep the width w of the slit light U on the image sensor 53 constant according to an instruction from the lens controller 58. The variator lens 422 works with the zoom lens on the light receiving side.

By increasing the slit length before deflection by the galvanomirror 43, distortion of the slit light U can be reduced as compared with the case where deflection is performed after deflection. By arranging the expander lens 423 at the last stage of the light projecting lens system 42, that is, by bringing it closer to the galvanometer mirror 43, the size of the galvanometer mirror 43 can be reduced.

FIG. 5 is a block diagram of the zoom unit. The zoom unit 51 includes a front fixed lens 515, a variator unit 514, a focusing unit 511, and a rear fixed lens 512. The variator unit 514 and the focusing unit 511 can move independently of each other along the optical axis.

The movement of the focusing unit 511 is carried out by the focusing drive system 59. Focusing drive system 59
Has a pulse motor 59A for driving the lens and an origin switch 59B for positioning. Focusing unit 511 moves by a distance corresponding to the amount of rotation of pulse motor 59A with reference to the position where origin switch 59B is turned on. The movement of the variator unit 514 is performed by the zooming drive system 60. The zooming drive system 60 also includes a pulse motor 60A and an origin switch 60B. The variator unit 514 moves by a distance corresponding to the rotation amount of the pulse motor 60A based on the position where the origin switch 60B is turned on.

FIG. 6 is a schematic diagram showing an example of the configuration of the beam splitter 52. The beam splitter 52 includes a color separation film (a dichroic mirror) 521, two prisms 522 and 523 sandwiching the color separation film 521, and an infrared cut filter 52 provided on an emission surface 522 b of the prism 522.
4, a visible cut filter 525 provided on the front side of the image sensor 53, an infrared cut filter 526 provided on the exit surface 523b of the prism 523, and low-pass filters 527 and 528.

Light UC incident from the zoom unit 51
Enter the color separation film 521 through the low-pass filter 527 and the prism 522. Light U0 in the oscillation band of the semiconductor laser 41 is reflected by the color separation film 521, and is reflected by the prism 522.
After being reflected by the incident surface 522a, the light exits from the exit surface 522b toward the sensor 53. Of the light U0 emitted from the prism 522, the light transmitted through the infrared cut filter 524 and the visible cut filter 525 becomes the image sensor 5
3 is received. On the other hand, the light C0 transmitted through the color separation film 521 passes through the prism 523 and exits from the exit surface 523b toward the color image sensor 54. Of the light C0 emitted from the prism 523, the light transmitted through the infrared cut filter 526 and the low-pass filter 528 is received by the color image sensor 54.

The color separation film 521 has a wavelength of the slit light (6
85 nm) is reflected. However, since the infrared cut filter 524 and the visible cut filter 525 are provided, the light finally incident on the image sensor 53 is light in a narrow range including 685 nm. As a result, measurement with a small influence of ambient light, that is, a large optical SN ratio can be realized. On the other hand, since the infrared cut filter 528 blocks light in the infrared band transmitted through the color separation film 521, only visible light enters the color image sensor 54. Thereby, the color reproducibility of the monitor image is improved.

FIG. 7 is a diagram showing a read range of the image sensor. The reading of one frame by the image sensor 53 is performed not on the entire imaging surface S2 but only on the effective light receiving area (band image) Ae that is a part of the imaging surface S2 in order to increase the speed. The effective light receiving area Ae is
This is an area where a bright line indicating the undulation of the object within the measurable distance range forms an image, and shifts by one pixel for each frame with the deflection of the slit light U. In the present embodiment, the number of pixels in the shift direction of the effective light receiving area Ae is fixed to 32, and the number of pixels in the length direction (horizontal direction) is selected to be, for example, 200. A method of reading out only a part of the captured image of the CCD area sensor is disclosed in Japanese Patent Application Laid-Open No. 7-174536.

FIG. 8 is a principle diagram for calculating a three-dimensional position in the measurement system. In the figure, for easy understanding,
Only five samples of the received light amount of each pixel g are shown.

The three-dimensional camera 2 outputs a relatively wide slit light U whose slit width on the light receiving surface (imaging surface) S2 of the image sensor 53 is a plurality of pixels g arranged at a pitch pv to the object Q. To project. Specifically, the width of the slit light U is set to about 3 to 5 pixels. The slit light U is deflected around the starting point A at a constant angular velocity in the vertical direction in the figure. The slit light U reflected by the object Q passes through the light receiving optical system and enters the light receiving surface S2 of the image sensor 53. By periodically sampling the amount of light received by each pixel g on the light receiving surface S2 during the projection of the slit light U, the object Q (strictly, a virtual surface orthogonal to the depth direction) is scanned. One frame of photoelectric conversion signal is output from the image sensor 53 every sampling period.

Focusing on one pixel g on the light receiving surface S2,
In the present embodiment, 32 samples of light reception data are obtained by 32 samplings performed during scanning. A barycenter (time barycenter) ip is obtained by a barycenter calculation for these 32 received light data. The center of gravity ip is such that the center of a range ag of the object surface where the pixel of interest g is glazed is a slit light U
Is the point at which the optical axis passes.

Assuming that the surface of the object Q is flat and there is no noise due to the characteristics of the optical system, the amount of light received by the target pixel g is
As shown in FIG. 8 (B), the number increases in a period during which the image of the slit light U passes, and usually changes almost like a normal distribution curve. In the example of FIG. 8B, the amount of received light is maximum between the _{n- th} sampling time T _n and the immediately preceding (n−1) -th sampling time T _n−1 , It substantially matches the center of gravity ip of the calculation result. Here, the line-of-sight direction of each pixel g is uniquely determined from the positional relationship between the principal point O and each pixel g on the light receiving surface S2.

The position (coordinates) of the object Q is calculated based on the relationship between the direction of irradiation of the slit light at the obtained center of gravity ip and the direction of the line of sight of the target pixel g. This enables measurement with a higher resolution than the resolution defined by the pixel pitch pv of the light receiving surface. Note that the amount of light received by the target pixel g is
Depends on the reflectance of However, the relative ratio of each received light amount of sampling is constant regardless of the absolute amount of received light. That is, the shading of the object color does not affect the measurement accuracy.

FIG. 9 is a diagram showing the relationship between lines and frames on the light receiving surface of the image sensor, and FIG. 10 is a diagram showing the storage state of the received light data of each frame. The first frame 1 of the light receiving surface S2 includes light receiving data for 32 lines from the line 1 to the line 32, which are the first lines. Frame 2 is shifted by one line per frame, such as from line 2 to line 33, frame 3 from line 3 to line 34, and so on. The frame 32 is 32 lines from a line 32 to a line 63.

The received light data from frame 1 to frame 32 are sequentially transferred to the memory 63 via the signal processing circuit 62 and stored. That is, in the memory 63,
The light receiving data is stored in the order of frames 1, 2, 3,.
The data of the line 32 which is the first line of the sampling range is the 32nd line of the frame 1 and the data of the frame 2
1 for each frame, such as the 31st line
The data is shifted upward by lines and stored. Frame 1
Is stored in the memory 63, the barycenter ip is calculated for each pixel on the line 32. While the calculation for the line 32 is being performed, the received light data of the frame 33 is transferred to the memory 63 and stored. The light receiving data of the frame 33 is stored in the memory 63 at the next address of the frame 32. When the data of the frame 33 is stored in the memory 63, the lines 3 to 3
The calculation of the centroid ip is performed for each pixel of No. 3.

FIG. 11 is a diagram showing the concept of the center of gravity. The barycenter ip is a time barycenter for 32 time-series received light data obtained by 32 samplings. Sampling numbers 1 to 32 are assigned to 32 pieces of received light data for each pixel. The i-th received light data is represented by xi. i is an integer of 1 to 32. At this time, i indicates the number of frames for one pixel after the pixel enters the effective light receiving area.

The center of gravity ip for the 1st to 32nd received light data x1 to x32 is i ·
It is obtained by dividing the sum Σi · xi of xi by the sum Σxi of xi. That is,

[0042]

(Equation 1)

Is as follows. The center-of-gravity calculating circuit 73 calculates the center of gravity ip by performing the calculation of the equation (1). The value of the center of gravity ip is large when the position of the surface of the object Q is close to the three-dimensional camera 2, and is small when it is far from the three-dimensional camera 2. Therefore, by displaying a grayscale image using the barycenter ip of each pixel on the light receiving surface S2 as density data, the distance distribution as a measurement result can be visualized. The host 3 calculates the position (coordinates) of the object Q by applying a triangulation method based on the relationship between the irradiation direction of the slit light U at the calculated center of gravity ip and the line of sight of the target pixel g.

In the measurement by the three-dimensional camera 2, the user determines the camera position and the direction while setting the angle of view while watching the color monitor image displayed on the LCD 21. At that time, one or both of a zooming operation and a focusing operation are performed as necessary. The variator unit 514 and the focusing unit 5 in the zoom unit 51 according to the operation
11 moves. When the user turns on the shutter button 27, the angle of view is set. In response to this, the system controller 61 sets the deflection conditions (scan start angle, scan end angle, deflection angular velocity) of the slit light U.

FIG. 12 is a schematic diagram of a positional relationship between a main point of the optical system and an object. The basic contents of setting the deflection conditions are as follows. Assuming that the reference plane Ss exists at the position of the inter-object distance D measured by the AF module 57,
From the imaging magnification of the zoom unit 51, an area of the reference surface Ss which is projected (projected) on the light receiving surface S2 is determined. The scanning start angle and the scanning end angle are determined so that the slit light U is projected from the lower end side over the entire area, and the deflection angular velocity is determined so that the required scanning time becomes a constant value.

In the actual setting, the Z-direction offset Doff between the front principal point H of the light receiving system, which is the distance measurement reference point of the inter-object distance D, and the projection start point A is considered. Also, in order to secure the same measurable distance range d at the end in the scanning direction as at the center, an overscan of a predetermined amount (for example, for 16 pixels) is performed. The scan start angle th1, the scan end angle th2, and the deflection angular velocity ω are represented by the following equations. th1 = tan ⁻¹ [β × pv (np / 2 + 16) + L) / (D + Doff)] × 180 / π (2) th2 = tan ⁻¹ [−β × pv (np / 2 + 16) + L) / (D + Doff) × 180 / π (3) ω = (th1−th2) / np (4) β: magnification (= D / effective focal length freal) pv: pixel pitch np: number of effective pixels in the Y direction of the light receiving surface S2 L: Base line length The center of gravity i obtained by scanning under the conditions calculated as described above.
p is the host 3 or the recording medium 4 as a measurement result by the SCSI controller 66 together with the sum xi of xi.
Is output to At the same time, device information including deflection conditions and specifications of the image sensor 53 is also output. Table 1 is 3
The main data sent by the two-dimensional camera 2 to the host 3 is summarized.

[0047]

[Table 1]

FIG. 13 is a flowchart showing the processing procedure of the three-dimensional position calculation in the host. First, it is determined whether or not the sum ｉxi of xi sent from the three-dimensional camera 2 exceeds a predetermined value (# 11). When xi is small, that is, when the sum ス リ ッ ト xi of the slit light components does not satisfy the predetermined criterion, a large error is included, and the calculation of the three-dimensional position is not executed for the pixel. Then, data indicating "error" is set and stored for the pixel (# 17). If Σxi exceeds a predetermined value, a sufficient accuracy is obtained, so that the calculation of the three-dimensional position is executed.

Prior to the calculation of the three-dimensional position, the passage timing nop of the slit light U is calculated (# 12). The passage timing nop is calculated by adding the line number to the center of gravity ip. That is, the calculated center of gravity i
Since p is the timing in the 32 frames at which the output of the pixel is obtained, the pixel is converted into a passage timing nop from the start of scanning by adding a line number. Specifically, the line number is the line 32 calculated first.
The pixel of "32" is "32", and the next line 33 is "33". Each time the line of the target pixel g advances by one, the line number increases by one. However, these values can be other suitable values.

Then, a three-dimensional position is calculated (# 13).
The calculated three-dimensional position is stored in the memory area corresponding to the pixel (# 14), and the same processing is performed for the next pixel (# 16). When the processing for all the pixels is completed, the process exits from the routine (Yes in # 15).

The method of calculating the three-dimensional position is as follows. The camera line-of-sight equations are the following equations (5) and (6). (U-u0) = (xp) = (b / pu) × [X / (Z-FH)] (5) (v−v0) = (yp) = (b / pv) × [Y / (Z −FH)] (6) b: Image distance FH: Front principal point position pu: Horizontal pixel pitch on the imaging surface pv: Vertical pixel pitch on the imaging surface u: Horizontal pixel position on the imaging surface u0: Center pixel position in the horizontal direction on the imaging plane v: Pixel position in the vertical direction on the imaging plane v0: Center pixel position in the vertical direction on the imaging plane The slit plane equation is given by equation (7).

[0052]

(Equation 2)

The geometric aberration depends on the angle of view. Distortion occurs in the target substantially at the center pixel. Therefore, the amount of distortion is represented by a function of the distance from the center pixel. Here, the distance is approximated by a cubic function. The secondary correction coefficient is d1, and the tertiary correction coefficient is d2. The corrected pixel positions u ′ and v ′ are given by equations (8) and (9).

[0054] u '= u + d1 × t2 2 × (u-u0) / t2 + d2 × t2 3 × (u-u0) / t2 ... (8) v' = v + d1 × t2 2 × (v-v0) / t2 + d2 × t2 ³ × (v−v0) / t2 (9) t2 = (t1) ⁻² t1 = (u−u0) ² + (v−v0) ^{2 In the} above equations (5) and (6), u By substituting u ′ instead of, and substituting v ′ instead of v, it is possible to obtain a three-dimensional position in consideration of distortion. For calibration, IEICE Technical Report, PRU91-113 [Geometric correction of images without camera positioning] Onodera and Kanaya, IEICE Transactions D-II vol. J74-D-II No.9 pp.1227-1235, '91 / 9 [High-precision calibration method of range finder based on three-dimensional model of optical system] Ueshiba, Yoshimi, Oshima, etc. have detailed disclosure. [Lens Position Correction and Lens Control] Hereinafter, a lens position correction method to which the present invention is applied will be described. In the specification of the three-dimensional camera 2, zooming is performed in a stepwise manner, and the lens position of each zoom stage is defined by the origin switch 60B. That is, the origin switch 60B is connected to the variator 5
The position of the variator 514 at the time of detecting and turning on the variator 14 is set as a reference position (origin), and zooming for moving the variator 514 by a certain distance from the origin is performed. The movement amount is set for each zoom stage, and the number of pulses corresponding to the movement amount is recorded in the control data memory 61A as a control amount for the pulse motor 60A.

FIG. 14 is a diagram showing the relationship between the origin and the lens stop position. In the example, the number m of zoom stages is 5, and the stop positions of the variator unit 514 are a total of five positions of stage 1, stage 2, stage 3, stage 4, and stage 5. However, at the time of design, a total of stage 0.5, stage 1, stage 1.5,... Stage 4.5, stage 5, and stage 5.5 in which the five stop positions and the positions sandwiching each of the five stop positions are combined. Optical parameter values such as magnification are calculated for 11 (= 2m + 1) positions.

If the origin P0 is a regular position as designed as shown in FIG. 14 (A), the variator 514 can be arranged at a regular stop position by moving by the number of design pulses. However, actually, the origin P0 'may be slightly deviated from the design position (P0) as shown in FIG. 14B due to a dimensional error or a mounting error of the origin switch 60B or the support. In this case, the variator unit 514 cannot be arranged at a regular stop position by moving the number of design pulses. That is, a deviation Δz occurs in the relative position between the beam splitter 52 and the image sensor 53, and the optical parameter values such as the magnification are different from the design values. Since the design value of the optical parameter value according to the zoom stage is used for the calculation of the three-dimensional position, the accuracy of the three-dimensional input is reduced due to the shift of the origin P0 '.

Therefore, it is necessary to correct the lens position in order to realize a highly accurate measurement. In the three-dimensional camera 2, the number of pulses defining the lens movement amount is corrected so that the optical parameter value becomes as designed at each zoom stage, and the corrected pulse number is replaced with the design pulse number and stored in the control data memory 61A. Write.

FIG. 15 is a diagram showing design values of optical parameters, and FIG. 16 is a diagram showing design pulse numbers. Front principal point,
The back principal point, the back focus, and the magnification (lateral magnification) are expressed as a function of the distance between the objects. The measurable distance range in the example specification is 0.5 m to 2.5 m, and the design value of each zoom stage is calculated for each of the ten distances selected as representatives of this range. In the actual three-dimensional measurement, when the inter-object distance is other than the ten representative distances, the optical parameters at the inter-object distance are calculated by interpolation based on the optical parameters at the representative distance. The number of design pulses indicates the distance (in mm) between the lens stop position and the reference position by the pulse motors 59A and 60A.
It is a numerical value converted into the number of steps by applying the mechanical conditions such as the step angle and the gear ratio of A.

The pulse number correction as a lens position correction operation is performed before shipment from the factory, and a chart is used for the operation. FIG. 17 is a diagram showing an example of the chart, and FIG. 18 is a schematic diagram of the arrangement of the chart.

The chart CT is composed of a stripe pattern RP for focus adjustment, which extends radially from the center CO, and a stripe pattern RP for magnification measurement.
4 is a test image plate having a size of about 1 m square having two black points BP1 to BP4. With the center CO aligned with the optical axis of the zoom unit 51, the chart CT is arranged at a position corresponding to one of the ten representative distances. The magnification is calculated by measuring the dimensions of a specific portion in the photographed image of the chart CT.

Prior to the actual measurement of the magnification by the chart CT, a graph for quantitatively grasping the deviation between the measured value and the design value is prepared. Specifically, for each of the five zoom stages in total, the relationship between the three stages (positions) of the two adjacent stages and the design value of the optical parameter of each stage is approximated by, for example, a quadratic function, Draw an approximation curve. A similar operation is performed for each of the 10 representative distances. For example,
For the zoom stage 3, an approximate curve of the position range from the stage 2.5 to the stage 3.5 is obtained in advance.

FIG. 19 is a flowchart showing the procedure of the correction operation. With respect to the five zoom stages, the appropriateness of the number of design pulses is checked in order from the zoom stage 1, and if improper, the number of correction pulses is determined for the zoom stage.

The operator selects any one of the ten representative distances as the inter-object distance, and arranges the chart CT at the position of the inter-object distance. Then, in a state where the variator unit 514 returns to the origin, for example, the zooming button 25a,
Designate zoom stage 1 by operating 25b (# 21, # 2
2). In response to this, the lens is moved based on the corresponding design pulse numbers for zooming and focusing (# 23, # 24). The actual movement type at this time is a single feed. For example, in a configuration in which the origin of the variator unit 154 is selected on the WIDE side, when sending data from the WIDE side to the TELE side, the T
After sending extra to the ELE side, it returns to the target position (see FIG. 14). When sending from the TELE side to the WIDE side, stop without sending extra. In the configuration in which the origin of the focusing unit 511 is selected on the infinity side, when sending from the long distance side to the short distance side, the extra distance is once sent from the target position to the short distance side and then returned to the target position. When sending from the short distance side to the long distance side, stop without sending extra. The purpose of such single feed is not to cancel the backlash of the driving system, but to stabilize the posture when the lens is stopped, to prevent the optical axis from shifting, and to improve the reproducibility of optical characteristics related to three-dimensional input. . Therefore, the number of single-feed pulses defining the extra feed amount is selected to be a sufficiently large value. More specifically, while the amount of movement between zoom stages is 400 to 500, the number of single-feed pulses is about 100. The number of one-sided pulses is also stored by the control data memory 61A.

A monitor image obtained by photographing the chart CT with the lens positioned based on the number of design pulses is displayed on the LCD.
21 is displayed (# 25). The monitor image can be displayed on another display or printed. In the zoom unit 51 of the present embodiment, focusing is performed by changing the position of the focusing unit 511 in accordance with the position of the variator unit 514. However, at the stage of adjusting the magnification, the stop position of the focusing unit 511 has not been determined. This is because the position of the variator unit 514 has not been determined, and the arrangement and operating points of the origin switch 59B vary. Therefore, the contrast of the fringe pattern image is obtained by image calculation, and the number of focusing pulses is corrected and focused so that the contrast of the fringe pattern is maximized (# 26,
# 27).

Subsequently, the magnification is obtained from the in-focus monitor image and compared with the design value. The normal error range caused by the displacement of the operating point and the arrangement of the origin switch 60B described above is about ± 5%. If the error of the magnification is ± 0.5% of the design value, the magnification can be regarded as designed. If the magnification error is outside the allowable range, the number of zooming pulses is corrected to bring the magnification closer to the design value (# 28, # 29). Then, the number of pulses when the magnification error falls within the allowable range is recorded as the number of corrected pulses (# 30). The number of correction pulses becomes control data thereafter.

The same operation is performed for the other four zoom stages (# 31, # 22). Since the magnification error is mainly caused by a variation in the relative position between the origin switch 60B and the variator unit 514, it is sufficient to adjust the magnification for any of the ten representative distances.

In this way, the moving amounts of the variator 514 and the focusing unit 511 are determined, and then the front principal point position and the image distance are calibrated in detail. In zooming, the variator unit 514 is always arranged at a position close to the design value (for example, a position where the magnification error becomes + 0.5%), and the reproducibility of control is high. If it is found by the measurement of the magnification that the position is shifted by, for example, + 0.5%, the design value of the principal point position and the image distance at the position shifted by + 0.5% can be obtained by interpolating from the approximate curve. + 5% without magnification correction
Compared to the case where interpolation is performed at a shifted position, higher approximation accuracy is obtained in the interpolation near the design value. As a result, the accuracy of calibration is improved, and accurate three-dimensional input can be performed.

Then, in the calibration of the optical system 50, the projection is performed by using the accurate optical parameter values (the front principal point position and the magnification for each of the five zoom stages for each of the five zoom stages) obtained by the interpolation approximation. The calibration of the optical system 40 on the light side is performed. [Lens Control in Three-Dimensional Measurement] Hereinafter, the operation of the lens controller 58 will be described in detail.

FIG. 20 is a main flowchart showing an outline of lens control. When the power is turned on, the lens controller 58 initializes the internal RAM 58A (# 4).
1). On the other hand, the system controller 61 reads data related to the control of the zoom unit 51 and the projection lens system 42 from the control data memory 61A and sends the data to the lens controller 58.

The lens controller 58 stores the data received from the system controller 61 in the RAM 58A (# 42), and causes the drive systems 60, 59, and 45 to be controlled to start the initialization movement (# 43). As a result, the variator 514, the focusing unit 511, and the variator lens 422 of the light projecting lens system 42 are arranged at a predetermined initial position in the above-described single-feed format. Thereafter, the lens controller 58 waits for an instruction from the system controller 61, and executes a process according to the received command.

When the system controller 61 detects the operation of the auto-focusing button 26c or receives an AF command from the host 3 via the SCSI controller 66, the AF command is given to the lens controller 58. AF processing is performed in response to the reception of the AF command (# 44, # 45).

The system controller 61 detects the operation of the zoom buttons 25a and 25b, or
When the zoom command is received from the lens controller 58, the zoom command is given to the lens controller 58. Zoom processing is performed in response to the reception of the zoom command (# 46, # 4)
7).

When the system controller 61 detects the operation of the shutter button 27 or receives a measurement command (three-dimensional input instruction) from the host 3, the lens controller 58 is given the measurement command. In response to the reception of the measurement command, high-precision AF processing for active AF that projects slit light is performed (# 4).
8, # 49).

FIG. 21 is a flowchart of the AF processing subroutine. The lens controller 58 receives the distance measurement result (subject distance information) from the AF module 57,
The target position of the focusing unit 511 according to the current zoom stage and the distance measurement result is obtained with reference to the control data of the RAM 58A (# 451, # 452). The moving direction and the number of pulses of the focusing unit 511 are calculated from the obtained target position and the current position of the focusing unit 511 (# 45).
3). Then, the focusing drive system 59 is operated to move the focusing unit 511 to the target position (# 4).
54). In this AF operation, since the user feels uncomfortable when performing the single-feed movement, a simple movement without extra feeding is performed.

FIG. 22 is a flowchart of a zoom processing subroutine. The lens controller 58 receives data indicating the target zoom stage from the system controller 61, and outputs the data to the variator unit 514 and the variator lens 422.
The number of pulses corresponding to the target zoom stage and the number of single-feed pulses are read out from the RAM 58A (# 471, # 4).
72). In addition, a target position of the focusing unit 511 that becomes the best focus according to the target zoom stage and the distance measurement result is obtained (# 473).

For the variator section 514 and the variator lens 422, the number of pulses for moving from the current position to the target zoom stage is determined (# 474, # 475). To avoid a focus shift due to zooming, a target position corresponding to the target zoom stage and the distance measurement result is obtained with reference to the control data of the RAM 58A for the focusing unit 511, and a pulse for moving from the current position to the target position is obtained. The number is obtained (# 476). Then, the variator unit 514, the focusing unit 511, and the variator lens 422 are moved to respective target positions (# 477). The movement of the variator unit 514 and the focusing unit 511 at this time is of a single feed type.

FIG. 23 is a flowchart of a high-precision AF processing subroutine. The lens controller 58 notifies the system controller 61 of the current lens position and the result of distance measurement by the AF module 57 (# 491). In response to this, the system controller 61 projects the slit light in a predetermined direction, and based on the light receiving condition (incident direction and intensity) of the slit light reflected by the object, an accurate inter-object distance and an optimum laser emission intensity. Is calculated. Then, the target position of the focusing unit 511 that becomes the best focus according to the accurate inter-object distance obtained by the active AF / AE operation and the current zoom stage is stored in the control data memory 61.
The lens controller 58 is obtained by referring to the data of A and notified.

Upon receiving the target position of the focusing unit 511, the lens controller 58 calculates the moving direction and the number of pulses for the movement based on the current position and the target position of the focusing unit 511 (# 493). Then, the focusing drive system 59 is operated to dispose the focusing unit 511 at the target position in a single-feed format (#
494).

In the embodiment described above, instead of distance measurement by the AF module 57, preliminary measurement for projecting the slit light U at a predetermined angle to detect the incident angle is performed, and the distance between the objects is determined by triangulation. Active distance measurement for calculation may be performed. Alternatively, lens control may be performed according to a preset inter-object distance or an inter-object distance input by an operator without performing distance measurement.

[0080]

According to the first to third aspects of the present invention,
In the movement of the lens, it is possible to realize the positioning as designed, regardless of the magnitude of an indefinite amount of error such as variation of parts and assembly error.

According to the second or third aspect of the present invention, three-dimensional input can be performed without being affected by an indeterminate error.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a measurement system according to the present invention.

FIG. 2 is a diagram illustrating an appearance of a three-dimensional camera.

FIG. 3 is a block diagram illustrating a functional configuration of a three-dimensional camera.

FIG. 4 is a configuration diagram of a light projecting lens system.

FIG. 5 is a configuration diagram of a zoom unit.

FIG. 6 is a schematic diagram illustrating an example of a configuration of a beam splitter.

FIG. 7 is a diagram illustrating a reading range of the image sensor.

FIG. 8 is a principle diagram of calculation of a three-dimensional position in the measurement system.

FIG. 9 is a diagram illustrating a relationship between a line and a frame on a light receiving surface of the image sensor.

FIG. 10 is a diagram showing a storage state of received light data of each frame.

FIG. 11 is a diagram illustrating a concept of a center of gravity.

FIG. 12 is a schematic diagram of a positional relationship between a main point of an optical system and an object.

FIG. 13 is a flowchart illustrating a processing procedure of a three-dimensional position calculation in the host.

FIG. 14 is a diagram illustrating a relationship between an origin and a lens stop position.

FIG. 15 is a diagram showing design values of optical parameters.

FIG. 16 is a diagram showing the number of design pulses.

FIG. 17 is a diagram showing an example of a chart.

FIG. 18 is a schematic diagram of an arrangement of a chart.

FIG. 19 is a flowchart illustrating a procedure of a correction operation.

FIG. 20 is a main flowchart showing an outline of lens control.

FIG. 21 is a flowchart of an AF processing subroutine.

FIG. 22 is a flowchart of a zoom processing subroutine.

FIG. 23 is a flowchart of a high-precision AF processing subroutine.

[Explanation of symbols]

 2 3D camera (imaging device, 3D input device) 50 Optical system (imaging means) 58A Lens controller (controller) 60A Pulse motor 60B Origin switch 61A Control data memory (memory) 514 Variator (lens) P0, P0 'Origin position (reference position)

 ────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Tadashi Fukumoto 2-13-13 Azuchicho, Chuo-ku, Osaka-shi, Osaka F-term in Osaka International Building Minolta Co., Ltd. 2F065 AA04 AA53 FF04 FF09 FF16 FF67 GG06 HH05 JJ02 JJ25 LL06 LL09 LL10 LL15 LL20 LL26 MM16 PP22 QQ31 RR06 2H087 KA03 MA15 NA09 SA23 SA27 SA29 SA32 SA63 SA65 SA72 SA74

Claims (3)

[Claims] 1. A variable magnification imaging apparatus comprising: a lens movable in an optical axis direction; and a controller for moving the lens based on control data indicating a lens movement amount corresponding to each of a plurality of set magnifications. Wherein the control data is stored in a rewritable memory at least once, and the lens is moved based on the control data stored in the memory. The actual magnification is measured, and the control data is corrected in accordance with an error between the measured magnification and a set magnification corresponding to the control data, and the corrected control data is stored in the memory instead of the uncorrected control data. A method of correcting a lens position, which is stored. And a variable magnification imaging means having a lens movable in an optical axis direction, and outputs data corresponding to an object image obtained by said imaging means and a magnification of the imaging.
A dimension input device, comprising: an origin switch for detecting the position of the lens; and a lens movement amount from a reference position determined by an arrangement state of the origin switch to a lens position corresponding to each of a plurality of set magnifications. A three-dimensional memory comprising: a nonvolatile memory that stores control data and is rewritable at least once; and a controller that moves the lens based on the control data stored by the memory. Input device. 3. The three-dimensional input device according to claim 2, wherein the movement of the lens is performed by a pulse motor, and the control data is a numerical value in rotation step units of the pulse motor.