JPH06207820A - Detecting method for and correcting method for light spot in three-dimensional measurement - Google Patents

Abstract

PURPOSE:To prevent significant increase in processing time by detecting a pixel for receiving light from a light spot for measurement while luminance data from pixels are stored into a frame memory. CONSTITUTION:A spot decision circuit 11 is so arranged to store data alone, for example, of nine pixels 4 (a group of pixels) one-dimensionally in sequence among luminance data to be fed in serial from the pixels 4 of a solid sensor 3 while the luminance data are sent to a frame memory 10. A judging section 13 of the circuit 11 calculates a curved surface based on the luminance data from the nine pixels 4 stored 12 and only when it is judged 15-17 that the inclination, height and center are within a set value, X and Y coordinates of the pixels 4 positioned at the center of the group of the pixels are stored into an FIFO memory 21. This enables the calculating of an output curve only based on the luminance data of the pixels 4 stored in the memory 21 and those 4 in the perimeter thereof thereby enabling significant reduction in computing time required for the detection of the light spot for measurement.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to three-dimensional measurement.
In order to improve the measurement accuracy, even if the number of pixels of the solid-state sensor of the CCD camera is increased, the processing speed does not decrease, and the detection method of the light spot, The present invention relates to a correction method for correcting defects such as insufficient sensitivity.

[0002]

2. Description of the Related Art As one of the measuring methods for three-dimensional measurement of an object to be measured such as an automobile body, the object to be measured is imaged using two or more CCD cameras, and based on the image data at this time, There is a measuring method for calculating the dimensions of the object to be measured by the triangulation method.

In this measuring method, as shown in FIGS. 7 and 8, the solid-state sensor (3) serving as a light-receiving surface is passed through the lens (2) of each CCD camera (1) to a predetermined position on the object (H) to be measured. When a light beam from the measuring light spot (P) formed on the
Among the many pixels (4) that make up (3), the position of the output center from the pixel group that outputs light rays is detected from the center of the solid-state sensor (3), and the light spot for measurement is measured from this position data. Calculate the angle of each CCD camera (P) from each CCD camera (1)
8 (L) is the optical axis of the CCD camera (1)).

Then, the position of the measuring light spot (P) is calculated based on the respective angle data obtained from two or more CCD cameras (1).

In this way, the three-dimensional size of the object to be measured (H) is obtained by sequentially calculating the data from the many measuring light spots (P) on the object to be measured (H). I have to.

It should be noted that a large number of measuring light spots (P) are provided on the object (H) to be measured.
As a means for forming, for example, a laser scanner (5)
And the laser spot may be sequentially irradiated onto the object to be measured (H).

[0007]

The above CCD camera (1)
When detecting the angle of the measuring light spot (P) by
If we try to make the dimension measurement accuracy of the measurement surface very high, such as 0.1 mm or less, the solid-state sensor (3) of each CCD camera (1) has an extremely high pixel count of 1000 × 1000. It is necessary to use a large number of items.

Further, at the time of angle detection, as shown in FIG.
Calculate the output curve corresponding to the distribution of the luminance signal level from each pixel (4), from the position of the maximum value of this output curve,
By detecting which position in the pixel (4) that outputs the maximum value is the center of high output, and detecting the angle of this position, a dimensional measurement accuracy of 0.1 mm or less is guaranteed. It will be.

As described above, when it is attempted to guarantee the dimensional accuracy of 0.1 mm or less by the three-dimensional measurement using the two CCD cameras (1), 1 million pixels (that constitute the solid-state sensor (3) ( After storing each luminance data from 4) once in the frame memory, by obtaining each luminance of one million pixels (4) from the data stored in this frame memory, It is necessary to detect only a few hundred pixel groups that output laser spots from the inside, and to detect the measurement light spot (P) from this pixel group. There was a problem I said.

That is, if the detection accuracy of the three-dimensional measurement is to be doubled, the number of pixels (4) constituting the solid-state sensor (3) will be four times. There is a problem that the processing time required for detecting the light spot increases significantly.

[0011]

An output curve corresponding to the distribution of the luminance signal level from each pixel of the solid-state sensor at this time is obtained by imaging a measuring light spot formed on an object to be measured with a plurality of CCD cameras. Is calculated, the position at which the maximum value of this output curve is output is detected, and the angle at this position is calculated to obtain the three-dimensional dimension of the DUT. While storing the luminance data from the pixel sensor in the frame memory, the luminance data of a predetermined number of pixel groups of the solid-state sensor are sequentially taken into the spot judgment circuit, and the spot judgment circuit distributes the luminance signal level from the predetermined number of pixels. When a curve or curved surface corresponding to is calculated and the curve or curved surface is determined to be from the light spot for measurement, the solid state of the pixel located at the center of the curve or curved surface Continue to store the position on the capacitors,
When the storage of the brightness data from each pixel of the solid-state sensor is completed in the frame memory, the determination of which pixel of the solid-state sensor receives the light from the measurement light spot is completed.

Further, at the time of detecting the light spot, if each pixel of the solid-state sensor has a defect such as insufficient sensitivity, a predetermined value of the solid-state sensor is stored while the brightness data from each pixel of the solid-state sensor is stored in the frame memory. The brightness data of a number of pixel groups are sequentially taken into the spot determination circuit, and this spot determination circuit sequentially calculates a curve or curved surface corresponding to the distribution of the brightness signal level from a predetermined number of pixels,
By setting the position of the defective pixel such as lack of sensitivity among the pixels of the solid-state sensor in advance, the data from the normal pixel is output as it is from the solid-state sensor and the defective pixel is detected. In that case, the influence of the defect of the solid-state sensor is reduced by outputting the data obtained from the curve or curved surface calculated by the spot determination circuit.

[0013]

As described above, which one of the pixels receives the light from the measuring light spot while the luminance data from each pixel of the solid state sensor of the CCD camera is being stored in the frame memory. By detecting, the position of the light spot for measurement is calculated based on the brightness data of each pixel stored in the frame memory, and the calculation operation can be performed based on only the brightness data of the pixel required for this calculation. It was done like this.

[0014]

1 is a block diagram for explaining a light spot detecting method according to the present invention.

In the figure, (3) is a solid-state sensor of a CCD camera consisting of 1 million pixels (4), and (10) is
A frame memory that sequentially stores the brightness data sent serially from each pixel (4) of the solid-state sensor (3), (1
1), while storing the luminance data from each pixel (4) in the frame memory (10), among each pixel (4),
The spot determination circuit detects only the pixel (4) receiving the light from the measurement light spot (P) formed on the object to be measured by the laser spot (5) and stores the position.

This spot determination circuit (11) is a solid-state sensor.
Among the brightness data serially sent from each pixel (4) of (3), only the brightness data of a predetermined number of pixels (4), in this embodiment, nine pixels (4) are sequentially stored. Based on the data of the storage unit (12) and the storage unit (12), it is determined whether or not the pixel group consisting of the nine pixels (4) is receiving the light from the measuring light spot (P). It comprises a discriminating section (13) for discriminating.

The storage unit (12) is a set of three memories (a) to (c), (d) to (f), and (g) to (i) which are arranged in a row in the lateral direction and are stacked in three stages. And between the memory (c) and the memory (d), and the memory
Delay circuits (14a) and (14b) for making a delay corresponding to the number of pixels in the horizontal direction are inserted between (f) and the memory (g), respectively, and serially from the solid-state sensor (3). The brightness data of each pixel (4) sent is as shown by the arrow in FIG.
When sequentially sent from the left to the right of each stage, the storage section (12) always stores the brightness data of the pixel group consisting of nine adjacent pixels (4).

The brightness data is sequentially updated each time the brightness data is sent from one pixel (4) of the solid-state sensor (3), and the solid-state sensor (of the pixel group) stored in the storage unit (12) is stored. 3) The positional relationship with respect to the top moves sequentially, and the brightness data of all pixels (4) on the solid-state sensor (3) is stored in the storage unit (12).
Of the solid-state sensor (3) is sent to the frame memory (10) simultaneously with the above work.

In the case of this embodiment, the number of pixels of the solid-state sensor (3) is 1000 × 1000 and the matrix calculated by the spot determination circuit (11) is 3 × 3.
Each delay circuit (14a) (14b) is composed of a delay circuit for 997 pixels.

From the luminance data stored in the storage section (12), it is judged whether or not the pixel group receives the light from the measuring light spot (P). If so, the position is determined. The discriminating unit (13) to be stored, based on the luminance data from the nine pixels (4) stored in the storage unit (12), FIGS. 2 (A) to 2 (D) and 3 (A) to (A). As shown in D), the memory is divided into three parts, and eight types of distribution data are calculated.
Calculate the curved surface (U) corresponding to the distribution of the luminance data as shown in, the inclination of the curved surface (U), the inclination determination unit (15) to determine whether it is within the preset value And a height discriminating unit (16) for discriminating whether or not the height of the curved surface (U) is within a preset value, and the vertex of the curved surface (U) has 9 pixels. Of the four pixels (4), the central discriminating unit (17) for discriminating whether or not the pixel is located on the central pixel (4) and the nine pixels (4) in which the luminance data is stored in the storage unit, An X counter (18) and a Y counter (19) for detecting the X coordinate and the Y coordinate of the pixel (4) located at the center on the solid-state sensor (3), an inclination determination unit (15), and a height determination unit ( 16), the X coordinate of the pixel (4) located at the center of the nine pixels (4) by the storage permission signal from the AND circuit (20) only when all the determination results from the center determination unit (17) are OK. And Y coordinate It is made from a FIFO memory (21) go.

The set values (A), (B), (C) and (D) preset in the inclination discriminating section (15) and the height discriminating section (16) are obtained from each pixel (4) of the pixel group. Of the curved surface obtained corresponding to the brightness level of the brightness data, the inclination and height of the curved surface obtained when capturing the measurement light spot (P) by the laser spot irradiated on the DUT are set as the reference. The light spot formed by reflection of external light on the object to be measured and the measurement light spot (P) formed by the laser spot can be clearly separated.

In the above structure, in order to detect which part on the solid-state sensor (3) is detecting the measuring light spot by the light spot detection method according to the present invention, the solid-state sensor (3) is used. The brightness data of each pixel (4) that is serially output is sent to the storage unit (12), and at the same time, sent to the frame memory (10) via the storage unit (12).

Then, the storage section (12) sequentially stores the brightness data of nine adjacent pixels (4) among the brightness data of the pixel (4).

When all of the nine memories (14) of the storage section (12) have finished storing the first brightness data, the first curved surface (U) is calculated based on this brightness data. And then
Whether the inclination of the curved surface (U) is within the set values (A) and (B) is determined by the inclination angle determination unit (15), and the height of the curved surface (U) is set to the set value (C) ( D) Check whether it is inside the height discriminating unit (1
Then, the center discriminating unit (17) discriminates whether the center of the curved surface (U) is located on the center of the nine memories.

When all of the above three determination results are OK, that is, the center of the pixel group currently stored in the storage section (12) receives the light from the measurement light point (P). When it is determined that the X coordinate and the Y coordinate of the pixel (4) located in the center of the pixel group at this time are set to the X counter (1
8) and the Y counter (19) are read, and the X and Y coordinates at this time are stored in the FIFO memory (21).

Thereafter, the brightness data from each pixel (4) is sent from the solid-state sensor (3) to the storage unit (12), and the storage unit
The above operation is repeated every time the brightness data stored in each memory (14) (12) is updated.

As described above, the brightness data from each pixel (4) of the solid-state sensor (3) is sent to the frame memory (10) via the storage unit (12), and the brightness data of each pixel (4) is sent. In the spot determination circuit (11), which pixel (4) among the pixels (4) is parallel to the above work while being stored.
Pixel (4) that detects whether light is being received from the measuring light spot (P) and is receiving light from the measuring light spot (P)
Memorize the X and Y coordinates of.

The solid-state sensor is added to the frame memory (10).
When all the luminance data of each pixel (4) for one frame of (3) is stored, the solid-state sensor () is stored in the FIFO memory (21) of the spot determination circuit (11) when the storage work is completed. Pixel receiving light from the measuring light spot (P) in 3)
The X and Y coordinates of (4) are stored.

Therefore, after the above storage work is completed, an output curve corresponding to the distribution of the brightness signal level of the brightness data of each pixel (4) stored in the frame memory (10) is calculated, and the maximum value of this output curve is calculated. From the position, when detecting which position in the pixel (4) that is outputting the maximum value is the center of high output and performing the task of detecting the angle of this position, as in the past, , It is not necessary to calculate the output curve from the luminance data of all pixels (4), and the pixels whose X and Y coordinates are stored in the FIFO memory (21)
The output curve may be calculated only from the luminance data from (4) and the pixels (4) located around it.

In the above embodiment, an example in which the curved surface (U) corresponding to the luminance distribution is obtained based on the luminance data of the nine memories (a) to (i) of the storage section (12) has been described. , 9 memories
An output curve is obtained from the luminance data of (a) to (i), and nine memories (a) to (a) are obtained from the slope, height and position of the apex of this output curve.
It may be possible to determine whether or not (i) receives light from the laser spot.

Next, in the above embodiment, the solid-state sensor
The sensitivity distribution of each pixel (4) in (3) is not uniform and
If only the column with (4) is extremely insensitive, the luminance data from pixel (4) in that column is not used and the prohibited pixel (4) is used when calculating the curved surface (U). If the pixel is included in the pixel group, an embodiment will be described in which interpolation is performed based on the luminance data of the pixels (4) at both ends of the pixel (4).

In order to implement the correction method according to this embodiment, the discriminator (13) shown in FIG. 1 is provided with a defect position setting unit (22) for pre-storing the position of the defective pixel, and an X position.
Position data from the counter (18) and Y counter (19),
Comparison for comparing with the defective pixel position data stored in the defect position setting section (22) to determine whether or not the pixel corresponding to the data currently stored in the storage section (12) is a defective pixel Add circuit (23) and.

Then, this comparison circuit (23) is
When it is detected that defective pixel data is input to the pixel (4) corresponding to each memory (a) to (i) in 2), the following interpolation work is performed.

For example, among the memories (a) to (i), the memory
If (e) stores the data from the defective pixel, the memory at both ends of the memory (e) corresponding to the defective pixel.
Instead of using the average of the brightness data of (d) and (f) as the brightness data of the memory (e) corresponding to the defective pixel, the average shape of the curved surface (U) obtained at the time of detecting the measurement light spot is stored. The interpolation is performed so that the curved surface (U) obtained by performing the interpolation approaches the average curved surface (U).

That is, when the memory (e) corresponds to a defective pixel, when the average of the luminance data of the memories (d) and (f) at both ends is used as the luminance data of the memory (e), the interpolation is performed. = (D + f) / 2, as shown in FIG.
The distribution data greatly deviates from the actual distribution state shown by the broken line (L).

Therefore, in the present invention, when the above interpolation is performed, the average of the luminance data of the memories (d) and (f) at both ends of the memory (e) corresponding to the defective pixel (4) is averaged in the memory (e). The average shape of the curved surface (U) obtained at the time of detecting the light spot for measurement is stored, and the curved surface (U) obtained by interpolation is stored as shown in FIG. ,solid line
The interpolation is performed so as to approach the average curved surface (U) shown in (M).

Further, when the interpolation work is performed in this way, the determination unit (13) makes a determination based on the interpolated data and at the same time sends the data to the frame memory (10). Thus, the interpolated data is also stored in the frame memory (10).

[0038]

As described above, according to the present invention, the CCD
While storing the brightness data from each pixel of the solid state sensor of the camera in the frame memory,
Since it is detected which pixel receives the light from the measurement light spot, when calculating the position of the measurement light spot based on the brightness data of each pixel stored in the frame memory, the conventional As described above, it is not necessary to calculate the output curve from the brightness data of all the pixels, and the calculation operation can be performed based on only the brightness data of the pixels necessary for this calculation.

Therefore, in order to improve the measurement accuracy in the three-dimensional measurement, even if the solid-state sensor of the CCD camera has a very large number of pixels, for example, one million, the measuring light is used. The calculation time required for point detection can be prevented from significantly increasing, and high-accuracy and high-speed three-dimensional measurement can be realized.

Further, when the position of the defective pixel is stored in advance and the data from this pixel is sent to the storage section, the data from this pixel is not used, but the data from the pixels located at both ends thereof is not used. By performing the interpolation operation based on the average distribution data stored in advance, more accurate three-dimensional measurement is possible.

[Brief description of drawings]

FIG. 1 is a block diagram for explaining the present invention.

FIG. 2 is a graph showing a distribution example of luminance data from each pixel.

FIG. 3 is a graph showing a distribution example of brightness data from each pixel.

FIG. 4 is a diagram showing a curved surface created based on the data of FIGS. 2 and 3;

FIG. 5 is a graph for explaining an example of conventional interpolation work.

FIG. 6 is a graph for explaining an interpolation operation according to the present invention.

FIG. 7 is a perspective view showing an example of three-dimensional measurement.

FIG. 8 is a perspective view showing a state in which light from a measurement light spot is incident on a solid-state sensor during three-dimensional measurement.

FIG. 9 is a perspective view showing an example of an output curve.

[Explanation of symbols]

 H object to be measured P measurement light spot 1 CCD camera 3 solid state sensor 4 pixel 10 frame memory 11 spot determination circuit 12 storage unit 13 determination unit 15 inclination determination unit 16 height determination unit 17 center determination unit 18 X counter 19 Y counter 21 FIFO memory 22 Defect position setting section 23 Comparison circuit a memory b memory c memory d memory e memory f memory g memory h memory i memory

Claims (2)

[Claims] 1. A measurement light spot formed on an object to be measured is imaged by a plurality of CCD cameras, and an output curve corresponding to a distribution of a luminance signal level from each pixel of a solid-state sensor at this time is calculated, The luminance data from each pixel of the solid-state sensor is used to obtain the three-dimensional dimension of the DUT by detecting the position where the maximum value of this output curve is output and calculating the angle of this position. While storing in the frame memory, the brightness data of a predetermined number of pixel groups of the solid-state sensor are sequentially taken into the spot determination circuit, and in this spot determination circuit, a curve corresponding to the distribution of the brightness signal level from the predetermined number of pixels. Alternatively, when the curved surface is sequentially calculated and it is determined that the curved surface or curved surface is from the measurement light spot, the position of the pixel located at the center of the curved surface or curved surface on the solid-state sensor is determined. When the luminance data from each pixel of the solid-state sensor is stored in the frame memory, the determination of which pixel of the solid-state sensor receives the light from the measuring light spot is ended. A method for detecting a light spot in three-dimensional measurement characterized by the above. 2. While the brightness data from each pixel of the solid-state sensor is stored in the frame memory, the brightness data of a predetermined number of pixel groups of the solid-state sensor are sequentially taken into a spot determination circuit, and this spot determination circuit determines a predetermined value. By sequentially calculating the curve or curved surface corresponding to the distribution of the luminance signal level from several pixels, and by setting the position of the defective pixel such as lack of sensitivity among the pixels of the solid-state sensor in advance. , The data from the normal pixel is output as it is from the solid-state sensor, and if it is a defective pixel, the data obtained from the curve or curved surface calculated by the spot determination circuit is output to detect the defect of the solid-state sensor. A method of correcting a light spot in three-dimensional measurement, which is characterized by reducing the influence of the light.