JPH0658209B2 - Coordinate measuring method and device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial application] The present invention relates to a coordinate measuring method and apparatus, and more particularly to three-dimensional coordinates of a measuring point on an object or three-dimensional measuring points arbitrarily provided in a stereo image. A method and apparatus for contactlessly measuring coordinates.

[Prior art]

As an example of the conventional non-contact type three-dimensional coordinate measuring method,
A method is known in which corresponding measurement points are extracted from both the pair of stereo photographs, and three-dimensional coordinate values are calculated from the coordinate values of the measurement points.

[Problems to be solved by the invention]

In the above-mentioned conventional three-dimensional coordinate measuring method, it takes a lot of skill and time to extract the measuring points, and especially when the shape of the object to be measured is complicated and the number of measuring points is large, the fatigue is great and it is extremely inefficient. Work was needed.

The present invention has been made in view of the above problems of the conventional three-dimensional coordinate measuring method, in which the designation of the measurement point is input to only one of the stereo image data, and the corresponding point in the other stereo image data is automatically processed by the correlation processing. It is an object of the present invention to provide a coordinate measuring method and a device which can be easily measured without any special skill.

The present invention is also, when the extraction of the corresponding point to the measurement point is not performed accurately, the measurement point is observed floating up or sinking from the object to be measured, the extraction of an erroneous corresponding point (hereinafter referred to as "mismatching"). It is an object of the present invention to provide a coordinate measuring method and device capable of easily checking whether or not

The present invention further provides a window having a different width in the first image data to detect the corresponding points, and performs a correlation process on the second image data based on the window so that the mismatching can be performed regardless of the situation of the image data. An object of the present invention is to provide a coordinate measuring method and a device that can extract corresponding points with low frequency and high accuracy.

[Structure of Invention]

A first invention of the present application is a storage unit that stores first image data and second image data that form a stereo image, a setting unit that sets a measurement point for the first image data, and means other than stereoscopic vision. A corresponding point detecting unit that obtains a corresponding point in the second image data with respect to the measuring point, and a measuring point mark data is added to the first image data based on the data about the measuring point from the setting unit, and the corresponding point detecting unit A marker part for adding corresponding point mark data to the second image data based on the corresponding point data and outputting the same, and a first measuring point mark for receiving the output of the marker part.
An image forming unit that forms a stereo image based on the image and the second image to which the corresponding point mark is added, three-dimensional data is obtained by the measurement point data and the corresponding point data obtained by the corresponding point detection unit. A coordinate measuring apparatus having a calculating unit for obtaining.

A second invention of the present application is that a first step of storing first image data and second image data forming a stereo image, a second step of setting a measurement point in the first image data, and means other than stereoscopic vision. A third step of obtaining a corresponding point in the second image data corresponding to the measuring point, a measuring point mark is added to the first image data based on the measuring point, and the second image data is handled based on the corresponding point. A fourth step of providing a point mark, a fifth step of forming an image based on the first image data having the measurement point mark and the second image data having the corresponding point mark, the measurement point data and the above And a sixth step of calculating three-dimensional data of the measurement point based on the corresponding point data.

[Effects of the Invention] Since the present invention is configured to extract corresponding points by the correlation processing as described above, coordinate measurement is automatically and highly accurately performed by inputting measurement points to one stereo image data. It has the advantage that it can. Further, the present invention, when the extraction of the corresponding points is performed inaccurately, the measurement point image can be observed with respect to the measurement target object image by floating or sinking, so the extraction accuracy of the corresponding points, that is, the measurement It has an advantage that the accuracy can be visually confirmed easily. Further, the present invention has an advantage that the frequency of mismatching can be reduced and the corresponding points can be accurately extracted by providing windows of different widths for extracting the corresponding points.

〔Example〕

The embodiments of the present invention will be described below separately for the data reading system and the processing output system.

(Data Reading System) As shown in FIG. 1, the data reading system 100 includes an optical system for obtaining an optical image of the object 3 to be measured, and a detection unit 10 including means for photoelectrically detecting the optical image. , Detection unit 10
And a storage unit 30 for storing the data obtained by the control unit 20.

First, considering the target object 3, as shown in FIG. 1, x,
Determine y, z coordinates. The detection unit 10 arranges two objective lenses 105 and 106 above the target object 3 side by side in the x-axis direction, and further, a linear CCD (charge coupled device) at the imaging position of the target object 3 by the objective lenses 105 and 106. 1
03 and 104 are arranged so that the detection direction is parallel to the x-axis direction. CCD 103, 104, objective lens 10
Reference numerals 5 and 106 respectively form a detection unit (A) 101 and a detection unit (B) 102, and the detection devices 101 and 102 can be integrally moved by a pulse motor 107 in the y-axis direction.

Between the detection devices 101 and 102, the objective lens 109,
An illumination unit 108 including a pattern film 110 and a light source 111 is arranged. The illumination unit 108 has a grid pattern on the surface of the target object 3 when there is no pattern on the surface.
It is used to project a stripe pattern or a random pattern relating to density or period.

On the other hand, the control unit 20 and the storage unit 30 are configured as follows. That is, first, the control unit 20 determines the detection unit (A) 1
01 and A / D converter (A) 121 and A / D converter (B) 122 for converting the signals from the detection unit (B) 102 into digital signals, and RAMs 123 and 124, which will be described later, convert digital signals for one scan. A counter 125 that manages the addresses of the RAMs 123 and 124 when storing, control of the maximum value detection unit (A) 127 and maximum value detection unit (B) 126 that detects the maximum signal during one scan, control of the CCDs 103 and 104, and the timing of each unit. And a timing pulse generator 128 that generates a timing pulse that takes

The storage unit 30 stores an RA for storing digital signals for one scan.
It is composed of an M (A) 123, a RAM (B) 124, and a disk 2 that stores data for all scans.

The timing pulse generator 128 includes the A / D converter 12
A pulse signal for starting the A / D conversion is output to the first and second 122 through the signal line AS, while the A / D converters 121 and 122 are output.
From the A / D conversion end signal is received through the signal lines AEA and AEB. When the timing pulse generator 128 receives the A / D conversion end signals from both the A / D converters 121 and 122, it outputs the clock signal T1 (S) to the CCDs 103 and 104 through the signal line T1 (see FIG. 3). ) Is converted from “0” to “1” or “1” to “0”,
The CCDs 103 and 104 are respectively shifted by 1 bit, and the pulse signal for starting the A / D conversion is output again to the A / D converters 121 and 122. That is, an oscillation loop is formed by the timing pulse generator 128 and the A / D converters 121 and 122, and the operation of the oscillation loop causes C
The scanning of the CDs 103 and 104 is executed, and the output signals thereof are converted into digital signals by the A / D converters 121 and 122, and are input to the RAMs 123 and 124.

The timing pulse generator 128 also receives periodic data from the main controller 1 through the data line TC. Inside the timing pulse generator 128, the digital signal output through the signal line T1 is counted as a clock signal T1 (S) (see FIG. 3), and 1 is calculated based on the cycle data.
A cycle counter is included, and one pulse of the timing pulse signal T2 (S) (see FIG. 3) is output to the CCDs 103 and 104 through the signal line T2 every one cycle. By the way, the CCDs 103 and 104 are of the accumulation effect type, and the magnitude of the output signal can be changed by controlling the accumulation time.
Timing pulse signal T2 output to 03 and 104
(S) is for controlling this accumulation time.

The timing pulse generator 128 further includes a main controller 1
From the signal line S to receive the read start signal S (S) (see FIG. 3). Timing pulse generator 128
Receives the read start signal S (S) and outputs the timing pulse signal T2 (S) to the CCDs 103 and 104, the write signal T3 (S) (see FIG. 3) synchronized with the clock signal T1 (S) is signaled. The pulse signal T3 (S) is output through the line T3 to the counter 125, while the generation of the pulse signal T3 (S) is stopped by receiving the reading end signal E (S) (see FIG. 3) which will be described later and is input through the signal line E from the counter 125. To be

The counter 125 is reset by receiving the read start signal S (S) from the main controller 1 through the signal line S, and then counts the write signal T3 (S). This count value by the counter 125 indicates the address value of the CCD 103, 104, and RA through the address data line ADR.
It is output to M123 and M124. When the count value reaches a predetermined number, for example, the number of bits of the CCDs 103 and 104, the counter 125 sends a read end signal E (S) through the signal line E.
To the timing pulse generator 128 and the main controller 1. The RAMs 123 and 124 have write signals T3 (S).
The output data from the A / D converters 121 and 122 is written at the address position according to the count value of the counter 125.

The maximum value detectors 126 and 127 are each composed of a comparator and a latch, and the latch is reset by the read start signal S (S), and then the A / D converters 121 and 122.
If the data from the latch is larger than the data in the latch, the contents of the latch are replaced by the write signal T3 (S) with the data output from the A / D converters 121 and 122. That is, the latch has the maximum value (Amax) of the read data at the time when the reading of one scan is completed,
(Bmax) is held. This maximum value (Amax), (Bma
x) receives the control signal output from the main control unit 1 through the signal lines ME1 and ME2, and thereby the data line DA
It is output to the main control unit 1 through T.

The main control unit 1 formed by a microcomputer or a personal computer, when measurement starts,
First, the cycle data is sent to the timing pulse generator 128 through the data line TC, and then the read start signal S
(S) is the same as the signal line S and the timing pulse generator 12
Send to 8. As described above, the timing pulse generator 128 controls the storage time of the CCDs 103 and 104 by the above input, and the output signals for one scan from the CCDs 103 and 104 are written in the RAMs 123 and 124, and the output for one scan is output. The maximum values (Amax) and (Bmax) of the signal are detected by the maximum value detection units 126 and 127.

The main controller 1 further includes the counter 125 as described above.
When the reading end signal E (S) is received from the maximum value detecting units 126 and 12 through the signal lines ME1 and ME2.
7 and outputs the maximum value (Ama) through the data line DAT.
x) and (Bmax) are input to the main controller 1. Here, if the data having the larger maximum value (Amax) or (Bmax) is not included in the predetermined range, that is, the CCD 1
If the outputs of 03 and 104 are too low, the cycle data is adjusted and the read start signal S (S) is output again to the timing pulse generator 128 to read the read data again. On the other hand, when the data having the larger maximum value (Amax) or (Bmax) is included in the predetermined range, the main controller 1 controls the pair of one scans stored in the RAMs 123 and 124. The image data (A) and (B) are transferred to the control unit 1 through the data line DAT by controlling the RAMs 123 and 124 through the address data line ADR and the read signal lines RR1 and RR2 and input to the disk 2. After that, the main controller 1 drives the pulse motor 107 through the signal line MT to move the detectors 101 and 102 in the y-axis direction, and scanning is performed again in the x-axis direction. In this way, the image data (A) and (B) regarding the x and y axes of the target object 3 are stored in the disc 2.

(Processing Output System) The processing output system 200, as shown in FIG.
Images of the target object 3 are alternately displayed on the monitor TV 254 from the pair of image data (A) and (B) stored in
The target object image is stereoscopically observed by observing this with the separation glasses 259 having shutters that open and close alternately to the left and right, and the coordinates of the x, y, and z axes of an arbitrary measurement point on the target object 3 are calculated and displayed. To do.

As shown in FIG. 2, the processing output system 200 includes a storage unit 210 that stores image data (A) and (B) that form a stereo image, and a measurement point setting that sets measurement points in the image data (A). A unit 220, a correlation unit 230 that obtains corresponding points in the image data (B) with respect to the measurement points by a correlation process, and adds measurement point mark data to the image data (A) based on the measurement points set by the setting unit 220. And a marker unit 240 that adds the corresponding point mark data to the image data (B) based on the corresponding points from the correlation unit 230 and outputs the image data (A) including the measurement point mark data output from the marker unit 240. ) And the image data (B) including the corresponding point mark data, an image forming unit 250 that forms an image, the measurement point mark data, and the corresponding point mark data. Display unit 260 for displaying et measuring point
Composed of and. The processing output system 200 having the above configuration is controlled by the main controller, and the main controller 1 is connected to the input device 4 so as to perform desired control.

The video signal control unit 252 of the image forming unit 250 outputs the horizontal synchronizing signal H, the vertical synchronizing signal V, and the blanking signal B to the monitor TV 254 by using the output signal of the oscillator 251 including a counter or the like as a clock signal, and stores the same. The image data in the video memory VRAM (A) 213 and the video memory VRAM (B) 214 of the unit 210 are monitored by the monitor TV 254.
The address data for displaying at the upper predetermined position is output to the VRAMs 213 and 214 through the address data line ADV. Then, the video signal control unit 252 includes the main control unit 1
When the inhibit signal output from the signal line VCE is input from the VRAM (A) 214, (B)
It stops outputting to 214, but at other times, it always outputs the address data to VRAM (A) 213, (B) 2
It is repeatedly output to 14. The main control unit 1 uses the input device 4 to display the image data of a desired area among the image data on the disk 2 in the VRAM (A) 213,
(B) Input into 214.

The output data of the VRAMs 213 and 214 is the marker section 24.
It is input to the switch 256 of the image forming unit 250 via the marker (A) 242 and the marker (B) 244 of 0. The switch 256 receives a signal of a frequency obtained by dividing the vertical synchronizing signal V of the video signal controller 252 by half through the flip-flop 258, and based on this, a marker (A) 242,
(B) The output from 244 is alternately switched and output to the blanking 257. Therefore, the monitor TV 254 receives the image data (A) by the switching operation of the switch 256.
The target object image of (B) is displayed alternately.

The blanking unit 257 uses the blanking signal B output from the video signal control unit 252 to monitor the TV 254.
The output of the image data is stopped at the time of the return of the line, and at other times, the output of the switch 256 is output to the D / A converter 258. D
The A / A converter 258 converts the digital image data signal from the blanking unit 257 into an analog signal and monitors the TV.
Output to 254.

The separation glasses 259 used when the measurer observes the monitor TV 254 are provided with an optical shutter made of liquid crystal or the like for the left and right eyeglass lens frames, and the optical shutter is provided by the output of the flip-flop 258.
By alternately opening and closing the right and left in synchronization with the switching image on 254, the target object image based on the image data (A) and (B) is observed by the left and right eyes of the measurer, respectively, and the stereoscopic view of the target object 3 is enabled. To do.

The measurement point setting unit 220 is for inputting the coordinates x _P , y _P for setting the measurement point on the monitor TV 254, and the output of the tablet or key 221 for inputting the coordinates of the measurement point is input to the setting unit 222. Is entered. Setting unit 22
2 stores the input measurement point coordinates x _P , y _P and outputs them to a comparator described later, and adds a window constant ± ω input in advance for the coordinate x _P to obtain (x
_{P + ω), (x P} -ω) each comparator (x _P ± ω) 2
To 24.

In the correlator described later, a data string similar to the data string in the vicinity of the measurement point in the image data (A) is extracted from the image data (B), so the window constant (±
ω) regulates the data string length at this time.

The comparator (x _P ± ω) 224 receives the address data from the video signal control unit 252 and the window constant signals (x _P + ω) and (x _P −ω) from the setting unit 222, and receives the address in the x-axis direction. The output signal is generated when the data is between (x _P + ω) and (x _P −ω). Comparator (y _P ) 22
6 receives the address data from the video signal control unit 252 and the coordinate y _P from the setting unit 222, and generates an output signal when the address data in the y-axis direction matches y _P.

The AND circuit 232 of the correlator 230 is a comparator (x _P ± ω)
When it receives both the output signal of 224 and the comparator y _P 226, it produces an output. The correlator 234 receives the output signal of the oscillator 251 as a clock signal via the terminal CL, and when the output signal of the AND circuit 232 is input to the terminal A, the output data from the VRAM (A) 213 is input via the terminal DA. On the other hand, when the output signal of the comparator y _P 226 is input to the terminal B, the output data from the VRAM (B) 214 is input via the terminal DB. That is, the correlator 234 stores VRAM as correlation data.
(A) “x _P ±” in the row of coordinate y _P stored in 213
The image data in the range of ω ″ and the image data for one scan of the row of the coordinate y _P stored in the VRAM (B) 214 are input. The correlator 234 is for one scan input from the terminal DB. Image data of (x
A portion of the image data similar to the image data of ( _P ± ω) is detected, and in other words, the position signal relating to the position of this portion is output, and the coordinate x _P ′ of the corresponding point corresponding to the coordinate x _O is output. Details of the configuration of the correlator 234 will be described later.

The comparator (x _P ) 238 of the correlator 230 has a setting unit 222.
Data x _P signal input from the video signal controller 2
An address data signal in the x-axis direction is received from 52, an output is generated when the two coincide with each other, and this is output by the marker unit 240.
To the marker (A) 242. Marker (A) 242
When output signals are input from both the comparator (y _P ) 226 and the comparator (x _P ) 238, the output data from the VRAM (A) 213 is marked data, for example, the maximum value of the image data is hexadecimal. When it is smaller than "FF", the mark data is replaced with "FF" and output. That is, VRAM
The image data (A) from the (A) 213 becomes the image data (A) containing the mark data indicating the measurement point and is output from the marker (A) 242.

The comparator (x _P ′) 236 receives the signal of the coordinate x _P ′ of the corresponding point corresponding to the coordinate x _P output from the correlator 234 and the address data signal in the x-axis direction from the video signal control unit 252, When they match, an output is generated and output to the marker (B) 244. The marker (B) 244 is connected to the comparator (y _P ) 226 and the comparator (x _P ′) 236, and when the outputs of both comparators 226 and 236 are input, V
The output data from the RAM (B) 214 is replaced with mark data and output. That is, the image data (B) from the VRAM (B) 214 is output from the marker (B) 244 as the image data (B) containing the mark data indicating the corresponding points.

The image data (A) and (B) including the mark data output from the markers (A) 243 and (B) 244 are switched by the switch 2
56, the blanking 257, and the D / A converter 258 are input to the monitor TV 254 and displayed as a target object image on the monitor TV 254. The target object image is stereoscopically observed through the separation glasses 259. At this time, if the above-mentioned correlation processing result is not correct, the mark at the measurement point will be observed as floating or sinking from the surface of the target object image. Can be determined.

On the other hand, the display unit 260 for calculating the coordinates (X, Y, Z) of the measurement point includes a calculation unit (Y) 262 and a calculation unit (X,
Z) 264. The calculation unit (Y) 262 is the setting unit 2
The signal of the coordinate y _P from 22 is received, and Y = αy _P + Y _O is calculated. Here, α represents a movement pitch in the Y-axis direction by the pulse motor 107, and Y _O is a reference address of the VRAM (A) 213 with respect to an origin arbitrarily set on the table on which the target object 3 is placed (for example, x = 0, y = 0) Indicates the position in the Y-axis direction.

The calculation unit (X, Z) 264 receives the coordinates x _P from the setting unit 222.
And the signal of coordinate x _P ′ from the correlator 232, And outputs the X and Z values. Where β is C
It is the element spacing of the CDs 103 and 104, and as shown in FIG. 4, L ₁ and L ₂ are from the left end of the CCDs 103 and 104 to the lens 1
Distances to the centers O ₁ and O _{2 of} 05 and 106, L
Is the distance between the lens centers O ₁ and O ₂ , and f is the lens 105, 10
The distance between the CCD 6 and the CCDs 103 and 104, X ₀ and Z _0, are the X and Z coordinates of O ₁ with respect to the origin of the coordinate system X, Y and Z.

Equations (1) and (2) are obtained from the following relational equations (3) to (8).

The calculation of the coordinates (X, Y, Z) is performed by using the data y _P , x _P ,
Alternatively, x _P ′ may be input to a microcomputer or a personal computer forming the main control unit 1 to be operated.

(Correlator) Next, the above correlator 234 will be described. The principle of the correlator includes an absolute difference method and a correlation coefficient method. The absolute difference method is In this method, x _P ′ is sequentially changed to find x _P ′ that minimizes G ₁ . On the other hand, the correlation coefficient method In this method, x _P ′ is sequentially changed to find x _P ′ when G ₂ is closest to “1”. Where Ax _P , B
x _P ′ means the average value.

The configuration of the correlator 234 based on the absolute difference method will be described below with reference to FIG. The terminal A to which the output of the AND circuit 232 is input is the AND circuit 301 and the switch 3
04, and the terminal B to which the output of the comparator (y _P ) 226 is input is the reset terminal RS of the counter (1) 307.
T, AND circuit 308, RAM 305, and inverter 3
06 is connected. The terminal DA to which the output of the VRAM (A) 213 is input is connected to the switch 304 and
(B) The terminal DB to which the output of 214 is input is the RAM 305.
Connected to. Terminal CL to which the output of the oscillator 251 is input
Is a clock terminal CL of the counter (1) 307, AND circuits 301, 302, RAM 305, cumulative adders (1), (2) 3
14 and 315.

The output terminal of the counter (1) 307 is the counter (2) 309,
It is connected to the cumulative adder (1) 314, the minimum value detector 317, the cumulative adder (2) 315, and the adder 310 and the comparator 320. The output terminal of the AND circuit 308 is a counter
(2) Connected to 309. The output terminal of the inverter 306 is connected to the AND circuit 302 and the delay element 322,
The outputs of the ND circuits 301 and 302 are shift registers 303.
Entered in. The output of the switch 304 is the shift register 3
03, and the output of the shift register 303 is input to the switch 304 and the subtractor 312.

The output of the counter (2) 309 is the adder (1) 310 and the adder
(2) Input to 323, output of adder (1) 310 is RAM
305 and the comparator (END) 311. The output of the comparator (END) 311 is the counter (2) 309 and A.
The output of the delay element 322 is also input to the AND circuit 321.

On the other hand, the output of the RAM 305 is passed through a subtractor 312 and an absolute value calculator 313 to a cumulative adder (1) 314 and a cumulative adder.
(2) Input to 315. The outputs of the cumulative adder (1) 314 and the cumulative adder (2) 315 are input to the combiner 316, and the output of the combiner 316 is input to the minimum value detector 317.
The output of the AND circuit 321 is input to the latch 319 and the minimum value detector 317. Further, the output of the adder (2) is input to the latch 319 via the address latch 318,
The output of the minimum value detector 317 is input to the address latch 318. The output of the latch 319 becomes the output of the correlator 234.

In the correlator 234 having the above structure, when a signal is input to the terminal A from the AND circuit 232, the clock signal input from the terminal CL is input to the shift register 303 through the AND circuit 301, and the VRAM input from the terminal DA.
The output data from (A) 213 is sequentially input to the shift register 303 through the switch 304. The shift register 3030 corresponds to the number of data of (x _P ± ω) (2ω
The shift registers of +1) stages are arranged in parallel by the same number as the number of bits forming the data, and the data contents are shifted in parallel. On the other hand, when no signal is input from the AND circuit 232 to the terminal A, the switch 304 supplies the output data of the shift register 303 to the input of the shift register 303. Therefore, the switch 304 and the shift register 303
Form a ring-shaped shift register, that is, a ring register.

When a signal is input from the comparator (y _P ) 226 to the terminal B,
The counter (1) 307 is in the reset state, the RAM 305 is in the writing state, and the counter (2) 309 is in the counting state by receiving the clock signal through the AND circuit 308. That is, since the count of the counter (1) 307 is 0, R
The AM 305 receives the content of the counter (2) 309 as it is as an address signal through the adder (1) 310, and writes the data from the terminal DB to the address position according to the output of the counter (2) 309. The comparator (END) 311 is
The number n of data for one scan which is a target for detecting the corresponding points is compared with the value from the adder (1) 310, and when they match, the counter (2) 309 is reset.

When a signal is input to the terminal B from the comparator (y _P ) 226, the AND circuit 30 is operated by the inverter 306.
2 and the AND circuit 321 are in a prohibited state, and they do not generate a signal. Here, the delay element 322 is a counter
(2) 309, adder (1) 310 and comparator (END) 31
The AND circuit 321 is prevented from generating a signal due to the delay time of 1.

As described above, while the address data on the y-axis from the video signal control unit 252 is y _P , a signal is generated at the terminal B, and the address data on the x-axis is (x _P ± ω) during the generation period. A signal is generated at terminal A while in range. Therefore, the RAM 305 has VRAM (B) 214
Image data for one scan of y _P rows of the image data (B) stored in is input to the shift register 3
In the image data (A) 213, the image data in the range “x _P ± ω” of the y _P row is sequentially input to the VRAM (A) 213.

The end of signal generation at terminal B means the start of correlation processing. That is, when the signal is no longer applied to the terminal B from the comparator (y _P ) 226, the counter (1) 307 starts counting the clock signal, and the ring register constituted by the switch 304 and the shift register 303 has the AN signal.
When the prohibition of the D circuit 302 is released, it operates according to the clock signal.

Completion of the signal generation at the terminal B also puts the RAM 305 in the reading state and delays the AND circuit 321 for a while to enable transmission of the inhibition release state, that is, the output state of the comparator (END) 311 to the latch 319.

The counter (1) 307 is a counter having a count of (2ω + 1) and performing one cycle, and operates in synchronization with the ring register. The count value of the counter (1) 307 means i in formula (9), and the correlator 234 calculates G ₁ in formula (9) every time the counter (1) 307 makes one cycle.

The count value of the counter (2) 309 means (x _P ′ −ω),
The counter (1) 307 is incremented by 1 every one cycle.
The adder (1) 310 has a counter (1) 307 and a counter (2) 3
The count value of 09 is added, and the added value is input to the RAM 305 as address data. Therefore, the RAM 305 stores B ₀ , B ₁ , and B ₁ in the first cycle of the counter (1) 307.
, Output data of B _2ω , output data of B ₁ , B ₂ , ..., B _{2ω + 1 in} the next cycle, and output data of B ₂ , B ₃ , ..., B _{2ω + in the} next cycle. Output the output data of ₂ . That is, the output of the RAM 305 is a data string deviated by one data for each cycle of the counter (1) 307.

The shift register 303 repeatedly outputs the data sequence of A ₀ , A ₁ , ..., A _2ω for each cycle of the counter (2) 309. This operation of the counter (1) 307 is performed by the adder (1)
The output value of 310 becomes n, and the process is continued until the comparator (END) 311 generates an output signal and the correlation process ends.

The subtractor 312 subtracts the output data of the RAM 305 from the output data of the shift register 303 and outputs the subtracted data. When the output data of the subtractor 312 is negative, the absolute value calculator 313 converts it to positive and outputs it.

The cumulative adder (1) 314 is composed of an adder and a latch,
While the counter (1) 307 performs one cycle, the output data of the absolute value calculator 313 is sequentially added into the latch. The combiner 316 will be described in detail later, but here, for convenience of explanation, the output of the cumulative adder (1) 314 is input to the minimum value detector 317.

The minimum value detector 317 is composed of a comparator and a latch.
The latch has the largest value in the reset state (for example, 8bit
In the case of “FF”), the comparator compares the input data with the contents of the latch, and if the input data is small, the input data is input to the latch and an output signal indicating the data change in the latch is output. The minimum value detector 317 outputs it to the address latch 318. The output signal of the counter (1) 307 is supplied to the cumulative adder (1) 314 and the minimum value detector 317, and the minimum value detector is generated each time the counter (1) 307 makes one cycle. 317 operates. Then, when the generation of the output signal of the counter (1) 307 is completed, the cumulative adder (1) 314 is reset and the output signal of the minimum value detector 317 is stopped.

In the address latch 318, the value obtained by adding ω to the count value of the counter (2) 309 by the adder (2) 323, that is, the count value of the counter (2) 309 is (x _P ′ −ω).
The value x _P ′ −ω + ω = x _P ′ is supplied, and this x _P ′ is stored in the latch 319 as the output signal of the minimum value detector 317 is generated. Therefore, the value of x _P ′ at which a smaller G ₁ is obtained is stored in address latch 318. The output signal of the comparator (END) 311 is supplied to the latch 319 and the minimum value detector 317 through the AND circuit 321. This causes the contents of address latch 318 to
19 and is output as output data to the outside of the correlator 234. On the other hand, the minimum value detector 317 enters the reset state and prepares for the next correlation processing.

By the way, the wider the window width (2ω) in the correlation process, the less the mismatching of the corresponding points with respect to the measurement point, that is, the less mismatching, but the smoother G ₁ value with respect to the change of x _P ′, and the lower the detection sensitivity. On the other hand, when the window width (2ω) is narrow, the detection sensitivity is high, but the amount of data to be subjected to the correlation calculation is small, resulting in a large number of mismatches. In order to eliminate this inconvenience, the correlator 234 of the present embodiment is configured to use two or more window widths, synthesize correlation states obtained from each of them, and detect a final corresponding point. The combiner 316 and the cumulative adder (2) 31 described in
5, the comparator 320 acts for this.

In the comparator 320, the output value of the counter (1) 307 is (ω ±
Generate an output signal in the range of ω '). Counter (1) 30
When 7 is ω, the output of the shift register 303 is x _P , so (ω ± ω ′) means (x _P ± ω ′) (where ω> ω ′). Cumulative adder (2) 315 is comparator 320
Only makes cumulative additions when is producing an output signal. The combiner 316 includes a cumulative adder (1) 314 and a cumulative adder (2) 31.
The values of 5 are combined and output to the minimum value detector 317.

Next, the synthesizer 316 will be described. As shown in FIG. 6, the combiner of the first embodiment inputs the comparator 401 to which the output of the cumulative adder (1) 314 is input, and the output of the comparator 401 and the cumulative adder (2) 315. And a gate 402. Then, the comparator 401 compares the output from the cumulative adder (1) 314 shown in FIG. 6 with a preset threshold level, and the cumulative adder (1) 314
If the output from is smaller, the gate 402 is released and the output of the accumulator (2) 315 is released. On the other hand, if the output is larger than the above threshold level, the gate 4
02 is prohibited and the largest value (in case of 8bit, "F
F ″) is output to the minimum value detector 317.

Therefore, in the combiner of the first embodiment, when the correlation state with a large window width is not good, the correlation state with a small window width is ignored and the correlation state is not established, while the correlation state with a large window width is When it is good, the corresponding points are detected by the correlation state with a small window width that can be expected for more accurate correlation processing.

The synthesizer of the second embodiment is an adder 40 as shown in FIG.
The output of the cumulative adder (1) 314 and the output of the cumulative adder (2) 315 are added by the adder 403, and the output is output to the minimum value detector 317.

The above-described method of detecting the final corresponding point by synthesizing the correlation states in each window width using two or more window widths can also be applied to the correlator using the correlation coefficient method. As the synthesizing method in this case, as in the synthesizer of the first embodiment, a correlation state with a large window width is determined at a threshold level, and when this is good, a correlation state with a small window width is used, and a large window width is used. A suitable method is to multiply the correlation states with a small window width.

In the embodiment of the coordinate measuring device described above,
It is configured to input measurement points with a tablet or a key, but instead of this, set some measurement points in advance and configure to detect corresponding points in sequence for these measurement points. Good. Furthermore, the input of the measurement point is such that the illumination unit 108 projects a grid pattern onto the target object 3, the grid pattern position is extracted from the image data (A), and the extracted grid pattern position is taken as the x coordinate of the measurement point. , Y-coordinates can also be determined by using predetermined ones.

The correlator 234 described above is configured by a circuit as shown in FIG. 5 and performs the correlation processing, and performs the correlation processing in real time. When high-speed correlation processing is not required, the correlation processing can be performed by an arithmetic unit such as a microcomputer. Further, the correlation coefficient method has many kinds of arithmetic functions as is clear from the equation (10), and the processing by the program is advantageous. Therefore, the correlation processing of the correlation coefficient method by the arithmetic unit will be described with reference to the flowchart of FIG. As described above, the correlation coefficient method uses the equation (10).
The following is an expansion of equation (10).

Is. Therefore, the correlation coefficient G is calculated as the elements CA, CB, C.
It is obtained by finding C, WX and WY and substituting them into the equation (11).

Now, referring to FIG. 9, in step, the setting unit 2
The large and small window widths ω and ω ′ are read from 16, and in step, from the setting unit 216, the coordinates x _{P of the} measurement point,
Read y _P. In step, from the image data (Ax _P −ω) of the y _P row set in the setting unit 216 to (A x _P
+ Ω), and B ₀ to B _n to VRAM (A),
(B) Read from 213 and 214. In the step, a value 0 when there is no correlation is put in G _m as an initial value.

In step, add ω is 1/2 of the window width as the initial setting of the x _P '. In the step, 0 is entered as the initial value of i. In step C
A ₂ , CA ₃ , CB ₂ , CB ₃ , CC ₂ , CC ₃ , WX ₂ , W
Insert 0 into each of X ₃ , WY ₂ and WY ₃ . In steps ,, and, CA ₁ , CB ₁ , and CC ₁ are (Ax _P -ω + i · Bx _P -ω + i), (Ax _P -ω + i) ² , (Bx _P -ω + i), respectively.
Calculate and write ² .

In step, it is determined whether i <ω−ω ′. If YES, the process proceeds to step, and if NO, the process proceeds to step. In steps, i> (ω +
ω ') is determined, and if YES, the process proceeds to step, and if NO, the process proceeds to step. In Step, CA ₁ is added to the CA _3, CB ₁ is added to the CB _3, CC ₁ is added to CC _3, WX ₁ is added to the WX _3, WY ₁ to WY ₃ is added, CA ₃ and C respectively
B ₃ , CC ₃ , WX ₃ , and WY ₃ are rewritten. In Step, CA ₁ is added to the CA _2, CB ₁ to CB ₂
There are added, CC ₁ is added to the CC _2, WX ₁ is added to the WX _2, WY ₁ is added to WY _2, respectively CA _2,
CB ₂ , CC ₂ , WX ₂ and WY ₂ are rewritten.

In the step, 1 is added to the value of i, which advances the processed data by one step. I =
It is determined whether or not it is 2ω, that is, whether or not all the data in the small window has been processed. If YES, the process proceeds to the step, and if NO, the process returns to the step. In Step, CA ₂ is added to the CA _3, CB ₂ to CB ₃
There are added, CC ₂ is added to CC _3, WX ₂ is added to the WX _3, WY ₂ is added to WY _3, the elements corresponding to each of a large window.

_{CA 2, CB 2, CC 2} , WX 2, WY 2 shows the elements corresponding to a small window, wherein the value at step
The correlation coefficient g ₂ in a small window is obtained by substituting in (11) and calculating. In the step, the elements CA ₃ , CB ₃ , CC ₃ , WX ₃ , WY obtained in the step
The correlation coefficient g ₁ in a large window is obtained by substituting the value of ₃ into the equation (11) and performing calculation.

In step, small window correlation coefficient g
G ₀ , which is a composite correlation coefficient between ₁ and a large window correlation coefficient g ₂ , is obtained from g ₁ · g ₂ . In step, it is determined whether G ₀ > G _m . If yes, then go to step; if no, then go to step. In the step, the contents of G _m are replaced with G _0, and in the step, x _P ′ is input into X _P ′.

In step, it is determined whether or not x _P ′ = n−ω, that is, whether or not the search is completed. If YES, the process proceeds to step, and if NO, the process proceeds to step. In the step, the value of x _P ′ is incremented by 1, and then the process proceeds to the step. In the step, X _P ′ that has obtained the maximum correlation coefficient G ₀ , that is, the corresponding point is output, and the correlation processing ends.

In the above embodiment, the minimum value detection unit 3 of the correlator 234 is used.
Reference numeral 17 stores data indicating the correlation state of the corresponding points, but this correlation state can also be displayed by displaying a mark indicating the measurement point, for example, by changing the color or the like. It is also possible to connect an adder to the output of the correlator 234 so that the corresponding point coordinates can be corrected by the adder when it is determined that the correlation processing is not performed with the desired accuracy.

[Brief description of drawings]

FIG. 1 is a block diagram of a data reading system of a coordinate measuring apparatus according to an embodiment of the present invention, FIG. 2 is a block diagram of the same processing output system, FIG. 3 is a waveform diagram of the data reading system, and FIG. 4 is coordinate calculation. 5 is a block diagram of a correlator, FIG. 6 is a block diagram of a combiner of the first embodiment, FIG. 7 is a block diagram of a combiner of the second embodiment, and FIG. FIG. 9 is a waveform diagram of the combiner of the first and second embodiments, and FIG. 9 is a chart of an arithmetic unit for performing correlation processing. DESCRIPTION OF SYMBOLS 1 ... Main control part, 3 ... Target object 20 ... Control part, 30 ... Storage part 103, 104 ... Linear CCD 105, 106 ... Objective lens 123, 124 ... RAM 128 ... Timing pulse generation part 222 ... Setting part, 234 ... Correlation 254 ... Monitor TV 303 ... Shift register 316 ... Synthesizer, 317 ... Minimum value detector

Claims (8)

[Claims] 1. A storage unit for storing first image data and second image data for forming a stereo image, a setting unit for setting a measurement point for the first image data, and a unit other than the stereoscopic view. Correspondence point detection unit that finds a correspondence point in the second image data with respect to the measurement point, and measurement point mark data is added to the first image data based on the data related to the measurement point from the setting unit, and the correspondence from the corresponding point detection unit A marker section for adding corresponding point mark data to the second image data based on point data and outputting the first point, and a first measuring point mark for receiving the output of the marker section.
An image forming unit that forms a stereo image based on the image and the second image to which the corresponding point mark is added, three-dimensional data is obtained by the measurement point data and the corresponding point data obtained by the corresponding point detection unit. A coordinate measuring apparatus comprising: a calculating unit for obtaining. 2. The coordinate measuring device according to claim 1, wherein the corresponding point detecting unit includes a first data string included in a first window centered on a measuring point provided in the first image data, A coordinate measuring apparatus characterized in that the corresponding point is obtained by performing a correlation process on the second image data based on a second data string included in the second window having a width narrower than that. 3. The coordinate measuring device according to claim 1 or 2, wherein the corresponding point detecting unit is configured to perform processing by an absolute value method. . 4. The coordinate measuring apparatus according to claim 1 or 2, wherein the corresponding point detecting unit is configured to perform processing by a correlation coefficient method. apparatus. 5. The coordinate measuring apparatus according to claim 1, wherein the image forming unit includes a first image based on the first image data input from the marker unit on the same screen, and the second image. An image forming apparatus that alternately forms a second image based on image data, and a scope that is arranged in front of the left and right eyes of a viewer and observes either the first or second image with the left eye and the other with the right eye. And a coordinate measuring device. 6. The coordinate measuring device according to claim 1, wherein when the corresponding point obtained by the corresponding point detecting section is judged to be unsuitable by the stereo image formed by the image forming section, the corresponding point is dealt with. A coordinate measuring device comprising a correction means for correcting a point. 7. A first step of storing first image data and second image data forming a stereo image, a second step of setting a measurement point in the first image data, and the measurement by means other than stereoscopic vision. A third step of obtaining a corresponding point in the second image data corresponding to the point, a measuring point mark is given to the first image data based on the measuring point, and a corresponding point mark is given to the second image data based on the corresponding point And a fifth step of forming an image based on the first image data having the measurement point mark and the second image data having the corresponding point mark, the measurement point data and the corresponding point And a sixth step of calculating three-dimensional data of the measurement point based on the data and the coordinate measuring method. 8. A coordinate measuring method according to claim 7, wherein a method of obtaining a corresponding point in the second image data corresponding to the measuring point by means other than the stereoscopic vision in the third step is a correlation process. A coordinate measuring method characterized in that.