JP4391137B2 - Measuring apparatus and measuring method for three-dimensional curved surface shape - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method and apparatus for measuring the shape of a three-dimensional curved surface in a non-contact manner.
[0002]
[Prior art]
As the non-contact three-dimensional shape measurement technique using light, the “light cutting method” shown in FIG. 9 (see, for example, “Image Processing Handbook (Shokendo) pp. 398-399”) was the mainstream. As shown in FIG. 9, the light beam pattern (light cutting line) formed on the object surface when the object to be measured 51 is irradiated with the slit light 53 from the slit light source 52 is different from the irradiation direction. This method focuses on the phenomenon of corresponding to the cross-sectional shape at the slit light irradiation position of the measured object 51 when observed from the direction, and is a method that has been widely used because of its simplicity, non-contact property, and quantitativeness. It is.
[0003]
However, in this method, when the reflectance of the measurement object is not uniform, the brightness of the light section line is not uniform, and therefore it is not always easy to extract the light section line itself from the background. For example, when the slit light is strengthened to detect a light cutting line in a portion having a low reflectance, there is a problem that halation occurs in a portion having a high reflectance and the resolution of the light cutting line is lowered.
[0004]
Therefore, the measurement accuracy is easily affected by the unevenness of reflectance and large unevenness of the measurement object, and it is difficult to measure under background light. In the field where “flexibility of measurement (anytime, anywhere, anything can be measured)” is required, there is a limit to its application.
[0005]
In contrast, in the 1960s, a group of methods known as so-called “spatial coding methods” was proposed. The “spatial coding method” is a method in which, for each point on the measurement target surface, first, a projection angle of light to be projected is obtained, and a shape is obtained from the angle based on the triangulation principle. This method is a group of methods having high “robustness” in measurement because a process of extracting “light cutting lines” unlike the “light cutting method” is not required for shape reproduction.
[0006]
The “spatial coding method” used here is used in a broad sense to refer to all methods corresponding to the above definition. Therefore, in addition to the “binary code pattern projection method” sometimes referred to as “spatial coding method” in a narrow sense, the “image encoding method” described below is also treated as a “space coding method” in a broad sense.
[0007]
As one of the system groups, an “image encoding method” disclosed in Patent Document 1 has been devised. A conceptual diagram of this method is shown in FIG. As shown in FIG. 10, a slit light 3a that spreads in a direction perpendicular to the plane of the paper is projected obliquely from above onto the surface of the measurement object 2 placed on the reference surface 1, and this slit light 3a is, for example, a rotating mirror 4 For example, the television camera 8 captures an image from directly above the object to be measured 2 while moving in the horizontal direction of the drawing.
[0008]
At this time, on the monitor television 8a connected to the television camera 8, it is observed that the linear reflection pattern of the slit light on the object surface moves in the horizontal direction of the screen. As described above, the shape of the curve drawn by the reflection pattern of the slit light 3a reflects the unevenness information on the object surface, and in the conventional light cutting method, the line shape of the reflection pattern is extracted from time to time, and this is extracted. By reconstructing, the three-dimensional shape of the object to be measured was measured.
[0009]
In this method, a three-dimensional shape is reconstructed based on a video signal output from the television camera 8 that shows a state in which the linear reflection pattern of the slit light 3a moves on the object surface. First, for each pixel in the screen, an image is synthesized with the slit light projection angle at the moment when the slit light passes through the position of the object surface corresponding to that pixel as the value of that pixel.
[0010]
FIG. 11 shows the configuration of an “image encoder” that is an image processing apparatus that is the core of this system. In FIG. 11, the image composition circuit 13 processes the video signal input from the TV camera 8 and calculates the brightness of the moment when the brightness becomes the brightest for each pixel. And an image composition calculation unit 19 that performs image composition calculation using the slit light projection angle θ at the moment when the maximum luminance is obtained as the value of the pixel. A generation circuit 21 and an output control circuit 22 are provided.
[0011]
The maximum luminance calculation unit 18 performs A / D conversion on the video signal based on the timing signal output from the synchronization circuit 20 around the maximum luminance image memory 23 which is a buffer memory for maximum luminance image calculation, and digitizes the video signal. The maximum luminance image memory 25 for controlling the reading and writing of the data of the address of the maximum luminance image memory designated by the conversion circuit 24 and the memory address generation circuit 21, and the image inputted from the TV camera and the image of the maximum luminance memory The comparison circuit 26 and the switch circuit 27 are configured to compare corresponding pixel values and selectively output the larger value.
[0012]
On the other hand, the composite image calculation unit 19 is configured around a composite image memory 28 that stores a composite image calculation result, and based on an output signal of the comparison circuit 26 in the maximum luminance image calculation unit 18, from the TV camera. A composite image memory control circuit 29 having a function of writing the slit light projection angle θ into the composite image memory 28 when the input signal level is larger than the corresponding pixel value of the address of the maximum brightness image memory 23 I have.
[0013]
This circuit starts from the state in which the maximum luminance image memory 13 and the composite image memory 28 are cleared to zero at the start timing of the operation, and uses the A / D conversion circuit 24 for the video signal input from the TV camera. Compare the value of the video signal with the value of the pixel of the maximum luminance image memory 13 corresponding to the position of the pixel while digitizing, and the value of the pixel of the maximum luminance image memory 13 only when the value of the video signal is larger Is updated with the value of the video signal, and at the same time, the slit light projection angle θ at that time is written into the corresponding pixel of the composite image.
[0014]
As described above, the above-described calculation is performed while instructed by the calculation control signal from the outside. As a result, the predetermined image described above is generated in the composite image memory 28 at the end of the calculation. The composite image calculated in this way is transferred to the next arithmetic circuit via the output control circuit 22. This technology has realized shape measurement that combines measurement accuracy and reliability (robustness).
[0015]
For the purpose of shortening the measurement time, an “image encoding method” using “multi-slit scanning” disclosed in Patent Document 2 has been devised. In this method, as shown in FIG. 12, a plurality of slit light beams 3a are applied to the surface of the object 2 to be measured placed on the reference surface 1 so as to spread obliquely from above and perpendicular to the paper surface and intersect on the same straight line. For example, the slit light is imaged by the television camera 8 from, for example, directly above the object to be measured 2 while moving the slit light in the horizontal direction of the paper using the rotating mirror 4.
[0016]
At this time, on the monitor television 8a connected to the television camera 8, it is observed that the reflection patterns of the plurality of slit lights move on the surface of the measurement target. Then, based on the video signal output from the TV camera 8, for each pixel in the screen, the slit light scanning angle at that moment when the slit light passes through the position of the object surface corresponding to the pixel. Is synthesized with an image whose value is the pixel.
[0017]
As described above, this conventional technique makes it possible to limit the scanning angle by compressing a plurality of slits, and to compress the measurement time.
[0018]
On the other hand, a “binary code pattern projection method” disclosed in Patent Document 3 is known as a shape measurement method classified as the same “space coding method”. As shown in FIG. 14, this method projects a binary stripe pattern coded in association with the projection angle on the surface to be measured, processes the image taken by the television camera, and decodes the code. In this method, the projection angle at each location on the target surface is obtained and the shape is obtained by the triangulation principle.
[0019]
[Patent Document 1]
Japanese Patent Publication No. 6-25655
[Patent Document 2]
Japanese Patent Laid-Open No. 7-260443
[Patent Document 3]
JP-A 64-54208
[0020]
[Problems to be solved by the invention]
Among the prior arts, the “light cutting method” is simple, but as mentioned earlier, the measurement accuracy is easily affected by uneven reflectance and large unevenness in the measurement target, and even under background light. There is a problem that there is a difficulty in so-called “robustness” such as difficulty in measurement.
[0021]
On the other hand, the “image encoding method” described in Patent Document 1 is a method having the features of “high accuracy”, “absolute shape measurement”, and “robustness”, but generally has a problem that the measurement time is long. Therefore, a higher-speed shape measurement method has been demanded for FA applications that require responsiveness and human body shape measurement applications that cannot withstand long periods of rest.
[0022]
As shown in FIG. 13, the technique “Multi-slit image encoding method” described in Patent Document 2 compresses the measurement time. Creates uncertainty. For this reason, the features of “high accuracy” and “robustness” are followed, but when viewed from the viewpoint of “absolute shape measurement”, constrains are generated. In other words, measurement results that do not cause any practical problems can be obtained by connecting the shapes between the slit scanning areas with a relatively flat shape, but in principle, a shape with a step cannot be distinguished from adjacent slits. There was a problem that reproduction might become impossible.
[0023]
The technique “binary code pattern projection method” described in Patent Document 3 belongs to the broadly defined “space coding method”, has essentially “robustness”, and is configured in consideration of high-speed measurement. It is a method and “high speed” is also essentially satisfied. Furthermore, since the illumination projection angle of each point on the surface to be measured can be coded as “absolute value”, there is an advantage of “absolute value measurement” which is not found in “multi-slit image encoding method”. However, when the stripe width becomes narrower, there is a problem that “measurement accuracy” cannot be expected due to the resolution limit that stripes with a pitch less than the camera resolution cannot be separated from each other when shooting with a television camera.
[0024]
As described above, there is no method in the prior art that satisfies all the conditions such as `` high speed '', `` high accuracy '', `` absolute value measurement '', `` robustness '' that the shape measurement method should originally have, The current situation is that there are advantages and disadvantages.
[0025]
Furthermore, in the adjustment of the optical system of the shape measuring apparatus based on such “triangulation principle”, “simplification of calibration work” has been left as a common problem.
[0026]
In the shape measurement method based on the triangulation principle, the relative positional relationship between the light projecting means and the target reference surface (eg, the angle between the reference surface and the light projecting means optical axis, the distance between the reference surface center and the light projecting means) ) Is an important parameter that determines the accuracy of the shape reproduction calculation. Furthermore, the deviation of the light projection means from the ideal optical system (eg, slit distortion / parallelism / twist in the case of slit projection, pattern distortion / twist due to lens aberration during pattern projection) causes measurement errors. It becomes.
[0027]
The “calibration” operation of the optical system of the shape measuring device was not able to completely eliminate the “disturbance factors” of these shape measurements as much as possible, while specifying the positional parameters of the light projecting means and the target reference surface. This is an operation for determining correction parameters for “disturbance factors” (eg, lens aberration). In general, this work is performed by a “temporary” measuring apparatus using a plurality of reference planes having different positions and orientations (the shape of each reference plane and the relative positions and orientations between the reference planes are measured in advance by another means). Based on the measurement results (results measured with parameters that have not been completely adjusted) ", the position and orientation of the light projecting means, as well as the position and correction parameters, are driven by" cut and try " ， Because it takes a lot of time and effort, its “simplification” and “simplification” have been demanded.
[0028]
The present invention has been made in view of the above-mentioned problems of the “optical cutting method”, “(multi-slit) image encoding method”, “binary code pattern projection method”, and other conventional techniques. The purpose is to solve the problems of the technology. At the same time, various features such as “high precision”, “robustness”, “high speed” or “absolute shape measurement” which are the advantages of the conventional technology can be realized by combining them according to the situation. A measuring device and a measuring method for a three-dimensional curved surface shape are provided.
[0029]
The present invention also provides a measuring device and measuring method for a three-dimensional curved surface shape that can realize “simple calibration”.
[0030]
[Means for Solving the Problems]
First, as a method for solving a problem as a “method” of the prior art, a solution method by “composite” and “fusion” of these methods can be considered. By combining each method well, it is possible to extract only the “advantages” of each method and hide the defects. For example, by combining the “multi-slit image encoding method” and the “binary pattern projection method”, it is possible to realize “high speed”, “high accuracy”, “robust”, and “absolute value” measurement by combining the advantages of each. .
[0031]
On the other hand, as a solution for “simplification of calibration work”, which is another problem of the prior art, each position of the projection pattern (pattern on the projection means) and each projection pattern (pattern on the target reference plane) If "1: 1 perfect correspondence" with the position is attached, "automatic calibration" becomes possible.
[0032]
These “system combination” and “automatic calibration” are solved by the following invention. The invention is a pattern projecting means capable of projecting an arbitrary pattern on the surface of the object to be measured, and projected on the surface of the object to be measured. One or more types of patterns including at least a pattern projected multiple times while shifting one or more slits in a direction orthogonal to the slits An image pickup means for picking up a pattern from a direction different from the projection direction, and a shape calculation means for calculating the three-dimensional curved surface shape of the object to be measured by performing image processing on the picked-up pattern projection image. 3D curved surface shape measuring device.
[0033]
Here, the arbitrary pattern is a pattern including a two-dimensional pattern as well as a pattern that changes only one-dimensionally, such as a simple linear or belt-like pattern. Note that even if the pattern is a line or a band, the pattern in which the amount of light is changed along the line or band is a two-dimensional pattern, and the pattern of the prior art (which changes only in one dimension) is Differently, it can be said to be any pattern of the present invention.
[0034]
Thus, the brightness of the surface of the object to be measured can be made uniform by changing the light quantity two-dimensionally. As a result, the measurement problem due to the uneven brightness of the projection pattern in the prior art can be solved, and “robustness” can be ensured.
[0036]
In the present invention, by projecting various patterns, it is possible to select an appropriate one depending on the situation such as the surface condition of the object to be measured, the complexity of the shape, the limitation of the measurement time, etc. A combination of methods can be used. As a result, the problems of the individual methods can be compensated for each other. The pattern to be projected may be one type of pattern such as a slit pattern, but a plurality of types of patterns can be freely combined and projected.
[0037]
For example, as two types of patterns, a pattern for measuring the overall rough shape of the surface of the object to be measured and a pattern for measuring the precise shape of each part are combined into the overall rough shape and the partial precision shape. By doing so, it is possible to obtain the entire precise shape of the surface of the object to be measured. In this case, a code pattern or the like is projected as a pattern for measuring the overall rough shape, and a pattern (multi-slit pattern or the like) repeated at a predetermined spatial period is projected as a pattern for measuring a partial precision shape. To do.
[0038]
As the image processing, for example, a “binary code pattern projection method” can be used for the overall schematic shape, and a “multi-slit encoding method” can be used for the partial precision shape. In addition, when the reflectance of the surface of the object to be measured is uniform, the amount of information of an image to be captured can be increased using a pattern in which the intensity of the projected light is several steps. By making the brightness of the projection pattern multi-valued, the number of projections can be reduced and the measurement time can be shortened.
[0040]
The present invention projects a slit pattern composed of one slit light (single slit light) or a plurality of slit lights (multi-slit light) from the projector onto the surface of the object to be measured, and these projection patterns are different from the projection direction. Images are taken with a camera from the direction to obtain an image corresponding to each projection pattern. In the present invention, the number of slits or the slit width can be controlled by a computer by using the pattern projection means as a projector. This can be easily performed by using a liquid crystal projector or the like as the projector.
[0041]
Further, in the present invention, by using a projector for the projection of the slit pattern, the brightness of the slit pattern on the surface of the object to be measured can be made uniform by adjusting the light quantity of the slit according to the projection location. This is achieved, for example, by projecting light from the entire surface of the projector, observing the object to be measured with a monitor or the like, and adjusting the amount of light at each part of the projector so that the brightness of the surface of the object to be measured is uniform. .
[0042]
In this case, the luminance of the surface of the object to be measured does not need to be completely uniform, and may be sufficiently higher than a predetermined threshold, for example. In general, since the brightness of the slit pattern decreases at portions that are inclined with respect to the imaging means (for example, the peripheral portion of the object to be measured), the amount of light projected on these portions may be set high. This can be performed by operating a mouse or the like by applying an interface screen for gradation adjustment in image processing. As described above, in the present invention, it is easy to extract the light cutting line itself from the background by projecting the slit pattern using the projector.
[0043]
From these projection images, the three-dimensional position (height from the reference surface, etc.) of the surface of the object to be measured is calculated based on the geometric relationship between the projecting means, the imaging means, and the surface of the object to be measured. Among these, the straight line connecting the projection means and the surface of the object to be measured is included in the slit surface (plane consisting of slit light). Since the intersection of the slit surface and the incident light beam is the position of the surface of the object to be measured, the measurement of the three-dimensional position of the surface to be measured results in the problem of obtaining the intersection of the straight line and the plane in principle.
[0044]
A light ray incident on the image pickup means (incident light beam) is specified by an address of an element of the image pickup means, and a slit surface is specified by a relative position in the slit pattern of the projection means. Therefore, the slit surface and the incident light beam can be expressed by linear and plane equations using these geometrical arrangements as parameters. The intersection of these lines and planes can be determined by solving both equations as simultaneous equations.
[0046]
The present invention projects a plurality of times while shifting the slit pattern on the surface of the object to be measured. The slit pattern may be shifted sequentially or randomly. In the case of a single slit, the entire slit projection area, in the case of multi-slit light, the area between adjacent slits, or more (overlapping near the boundary). It is sufficient if it covers the area.
[0047]
Further, in the above invention, the shape calculating means obtains the position of the pixel corresponding to the slit pattern from the slit pattern image, and from the geometrical relationship between the relative position of the slit in the slit pattern, the projector, and the imaging means. In addition, it is possible to provide a three-dimensional curved surface shape measuring apparatus that calculates the position of the surface of the object to be measured corresponding to each slit by a calculation process and obtains the shape of the surface of the object to be measured.
[0048]
Although the present invention is the same as the conventional “light cutting method” as the principle of shape calculation, as described above, the light quantity of the slit can be adjusted according to the projection location in the present invention, so that the slit pattern on the surface of the object to be measured can be adjusted. Can be made uniform. As a result, even when it is difficult to detect the light cutting line in the prior art, the light cutting line can be detected.
[0049]
As an invention using a shape calculation means different from this invention, the shape calculation from the slit pattern image is performed by selecting a slit pattern in which each pixel has the maximum luminance for each pixel, and the relative position of the slit in the slit pattern. From the geometric relationship between the slit relative position memory, the projector, the slit position corresponding to each pixel, and the imaging means, the surface of the object to be measured corresponding to each pixel is stored. It is also possible to provide a three-dimensional curved surface shape measuring apparatus characterized in that the position is calculated by a calculation process and the shape of the surface of the object to be measured is obtained.
[0050]
In this invention, the position at which the slit is projected is not obtained from the position of the pixel, but the slit pattern at the time when the brightness becomes highest during the projection of the slit pattern is selected for each pixel, and the slit within the slit pattern is selected. Is the slit relative position corresponding to the pixel. The slit relative position may be the slit projection position on the reference plane, but a slit number indicating the number of the slit in the slit pattern may be used.
[0051]
The synchronization with the rotation angle of the mirror is not required as in the prior art, and it is not necessary to perform the calculation of the projection angle of the slit pattern at the same time as the photographing, so that the photographing time can be shortened.
[0052]
In addition, in the above-described invention, a three-dimensional curved surface shape measuring apparatus characterized by having a plurality of pattern projecting means, a three-dimensional curved surface shape measuring apparatus characterized by having a plurality of imaging means, or projection / imaging A three-dimensional curved surface shape measuring apparatus characterized by using invisible light as the means can also be used.
[0053]
The invention of the method for measuring a three-dimensional curved surface shape using this apparatus is as follows. It uses a projector connected to a computer to project a slit pattern consisting of one or more slit lights on the surface of the object to be measured, shifting it in a direction orthogonal to the slit, and projecting it multiple times.
Photograph the light pattern projected on the surface of the object to be measured from a direction different from the projection direction,
For the slit pattern showing the maximum brightness for each pixel from among the images of the multiple slit patterns taken, the relative position of the slit in the slit pattern is the relative position of the slit corresponding to that pixel. Store in memory,
A three-dimensional curved surface shape measuring method, wherein the shape of the object to be measured is calculated by arithmetic processing from the geometrical relationship between the slit relative position corresponding to each pixel, the pattern projection means, and the imaging means.
[0054]
In the present invention, a projector connected to a computer is used to shift and project a slit pattern on the surface of an object to be measured. At this time, these projection patterns are photographed by a camera from a direction different from the projection direction, and images corresponding to the individual projection patterns are obtained. For the projected image, the relative position of the slit in the slit pattern where the pixel shows the maximum luminance is set as the relative slit position corresponding to the pixel.
[0055]
In general, projection of slit light is performed using a laser. In the present invention, a projector connected to a computer is employed. The projector connected to the computer can freely switch the pattern by a computer program, and the switching speed has been higher than the scanning speed of the TV camera in the recent progress of technology.
[0056]
By using a projector connected to a computer, even for slit light scanning, a pattern consisting of multiple slits can be easily realized by sequentially displaying lines for several scanning lines on the screen and shifting the display position. I can do it.
[0057]
Further, if the projection pattern is projected by a digital projector, the pattern itself can be handled as array data (i rows and j columns), so that calculation processing such as the projection angle of the pattern becomes easy. Furthermore, it is easy to calibrate the pattern, etc., and to adjust the pattern itself such as the slit interval and slit width according to the measurement target (resolution), the arrangement data can be obtained without mechanical setting or adjustment of the projection means. You can easily change it.
[0058]
Further, in the method of measuring a three-dimensional curved surface shape using the present invention or the above-described measuring apparatus, a projector connected to a computer is used as the projecting unit, and one or more patterns for alignment are arbitrarily selected from the projecting unit. The projection pattern projected onto the reference plane, the projection pattern projected onto the reference plane is captured by the imaging unit from a direction different from the projection direction, and the projected pattern image is displayed, thereby adjusting the positional relationship between the projection unit and the imaging unit. It can also be set as the measuring method of the three-dimensional curved surface shape characterized by this.
[0059]
Similarly, using a projector connected to a computer as projection means, one or more calibration patterns are projected from the projection means onto an arbitrary reference plane, and the projection pattern projected onto the reference plane is defined as the projection direction. Is taken by the imaging means from different directions, the positional relationship between the projection means and the imaging means is determined from the projection pattern image by arithmetic processing, and the surface shape of the object to be measured is calculated using the data of the positional relationship. It can also be set as the measuring method of the characteristic three-dimensional curved surface shape.
[0060]
In these inventions, by using a projector connected to a computer, it is possible to project an apparatus adjustment pattern or a data calibration pattern onto a reference plane or the like. As a result, the apparatus can be manually or automatically adjusted based on the projected data.
[0061]
Further, the distortion of the pattern of the projection means can be calculated by capturing the pattern for data calibration by the imaging means and calculating the shape of the reference plane itself by the shape calculation means. In addition, it is possible to reversely calculate the geometric arrangement of the apparatus by a process similar to the calculation process of image analysis, or to calculate a correction coefficient for correcting the luminance of the surface of the object to be measured.
[0062]
As described above, according to the present invention, various patterns are projected according to the purpose, and the captured image is subjected to arithmetic processing so that "high accuracy", "robustness", "high speed", "absolute shape measurement", etc. Various technical requests can be realized. Furthermore, it enables “automatic calibration” of the optical system of the shape measuring apparatus.
[0063]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 shows an example of the configuration of the present invention. The projector is connected to a computer and arranged so that a pattern generated by the computer can be projected onto the surface of the measurement object. The camera is arranged to observe the measurement target surface from a direction different from the projection direction of the projector, and the photographed image is digitized and captured in the computer.
[0064]
As for the projector and the camera, the former can be easily realized by liquid crystal or DLP (Digital Light Processing), and the latter can be realized by a device using a CCD element or the like. The DLP projector performs projection by controlling the reflection angle of a minute mirror surface DMD (Digital Micromirror Device), and enables high-speed control of the projection pattern.
[0065]
The number of pixels of these elements can be appropriately selected according to the target measurement accuracy. For example, if the projector is 1024 × 768, the camera is 600 × 480, etc., it is sufficient for the usual purpose at the time of filing. Further, from the viewpoint of a balance between the two precisions, the number of pixels of the projector may be about 1 to 2 times each of the vertical and horizontal directions relative to the number of pixels of the camera.
[0066]
FIG. 2 is a block diagram illustrating a configuration example of the arithmetic processing system of the apparatus of the present invention. The arithmetic processing system is realized by a computer, and includes a pattern projection block 30, an image input block 40, a shape arithmetic block 50, and sequence control means 90. Here, the sequence control means 90 operates these blocks, the projection means (projector) 7 and the image pickup means (camera) 8 in a predetermined sequence, and also synchronizes them. Hereinafter, each block will be described.
[0067]
In the pattern projection block 30, a slit pattern (sequence) generation circuit 31a generates a slit pattern having a predetermined slit interval and slit width, and further generates a plurality of slit patterns with the slit positions shifted.
[0068]
These slit patterns are stored in a memory or the like in a predetermined order, and are output to the projection means 7. The series of arithmetic processing described above is executed under the control of the pattern projection control circuit 39.
[0069]
Various patterns projected from the projection means (projector) 7 onto the measurement target are picked up by the image pickup means (camera) 8, and the obtained image signal is input to the image input block 40 as needed.
[0070]
In the image input block 40, the image signal is subjected to data conversion as necessary and stored in an image memo. First, the camera input circuit 41 converts the image input (signal) from the camera into a data format (image data) that can be processed in the image input block 40. Next, the image data is transferred to the image memory 43.
[0071]
The sequence image memory 43 for slit patterns stores these image data together with projection pattern projection order data (identification ID) 48. The series of arithmetic processing described above is executed under the control of the image input control circuit 49.
[0072]
In the shape calculation block 50, the shape of the surface of the object to be measured is obtained by calculation processing using a series of image data stored in each of the image memories 43. The slit projection angle image synthesis circuit 51 reads image data (slit image data) from the slit pattern image memory 43 and obtains a slit pattern identification ID when the slit is projected on each pixel. From the slit pattern identification ID, the projection position of the slit is calculated.
[0073]
Using the obtained projection position of the slit, the precision partial shape calculation circuit 52 calculates the geometric relationship (intersection of the slit surface and the incident light beam) by calculation processing to obtain the shape of the surface of the object to be measured. As a result, a precise partial shape (corresponding to the resolution of the projection means) can be obtained for each band-like region (corresponding to the minimum width stripe pattern) having the width between adjacent slits.
[0074]
In the precision overall shape calculation circuit 53, the precise partial shapes obtained by the precision partial shape calculation circuit 52 are sequentially connected to synthesize a precise overall shape. The obtained shape is stored in the shape image memory 54, and output such as monitor display is performed as necessary. The series of arithmetic processing described above is executed under the control of the shape arithmetic control circuit 59.
[0075]
In the example of the arithmetic processing system described above, each process of the computer is expressed by various circuits and memories. However, these individual processes may be configured by either hardware or software. When configured by software, the calculation processing speed is lower than that by hardware, but as long as the number of pixels is not particularly high, the processing is practically possible in a sufficiently short time.
[0076]
FIG. 3 shows a pattern projection / image input procedure. For the measurement, multi-slit pattern projection is performed. As described above, the multi-slit pattern projection is realized by displaying and projecting the multi-slit pattern on the projector while shifting the multi-slit pattern in a direction orthogonal to the slit (in this figure, the vertical direction).
[0077]
The series of multi-slit pattern images shown in FIG. 3 are obtained by converting the slit relative position (for example, 0 to 15), which is the maximum luminance for each pixel, into brightness and displaying it.
[0078]
Here, if the relative positions of the slits are connected in the direction orthogonal to the slits by the “scaling” method, the entire slit position information can be obtained. When it is uncertain which slit position information is due to which slit position information is due to “discontinuity” (large step) in the target shape, increase the slit interval to increase the number of slits. Indeterminate between.
[0079]
Further, for an object in which “shape uncertainty” still remains even by the above-described method, the “absolute shape” is always obtained by using, for example, the “binary code pattern projection” method together with the same optical system. It can also comprise.
[0080]
In addition, when the reflectance of the surface of the object to be measured is uniform, adjacent slits can be identified by increasing the light quantity of the slit pattern instead of widening the slit interval. As a result, an increase in the number of slit pattern projections, that is, an increase in measurement time can be avoided.
[0081]
If the projection angle 3 is obtained, the shape image (the value of each pixel corresponds to the pixel is measured based on the geometrical arrangement of the projector, the measurement object, and the relative position of the camera. An image representing the amount of unevenness on the object can be easily obtained. FIG. 4 shows an example of a procedure for calculating and reproducing a shape from a series of binary code pattern images and multi-slit pattern images obtained in this way.
[0082]
In this example, the pattern projection means is configured by one set. However, in order to eliminate the “dead angle” of measurement, pattern projection is performed from a plurality of directions, and the resulting shape is synthesized to form a shape with less “dead angle”. It is easily possible to obtain. FIG. 5 is a diagram showing an apparatus configuration example in which a plurality of projectors (1) and (2) 7 are provided as pattern projection means.
[0083]
In this example, the pattern projection block ((1), (2)) 30 generates slit patterns in the same manner as described above, and two projectors (1), (2) (projection means (1), (2)). 7 project patterns in a predetermined sequence. The image signal obtained by the camera 8 is converted into image data by the image input block 40.
[0084]
In the shape calculation block ((1), (2)) 50, the partial shapes are calculated by calculation processing in the same manner as described above, the partial shapes are synthesized, and the precision based on the respective image data (1), (2) is obtained. The overall shape ((1), (2)) is calculated. In the shape synthesizing block 60, with respect to the obtained precision overall shapes (1) and (2), the portions which are both blind spots are supplemented with the other shapes to obtain a shape with less “dead spots”.
[0085]
In these examples, an example in which the imaging means is configured by one set has been shown, but in order to eliminate the “dead angle” of measurement, imaging is performed from a plurality of directions, and the shapes obtained are combined to obtain a shape with less “dead angle” That is easily possible. FIG. 6 is a diagram showing an apparatus configuration example in which a plurality of cameras ((1), (2)) 8 are provided as imaging means.
[0086]
In this example, a slit pattern is generated in the pattern projection block 30 as described above, and each pattern is projected by the projector (projection means) 7. Image signals obtained by the camera (1), (2) 8 are converted into image data (1), (2) by an image input block (1), (2) 40.
[0087]
In the shape calculation block ((1), (2)) 50, the partial shapes are calculated by calculation processing in the same manner as described above, the partial shapes are synthesized, and the precision based on the respective image data (1), (2) is obtained. The overall shape ((1), (2)) is calculated. In the shape synthesizing block 60, with respect to the obtained precision overall shapes (1) and (2), the portions which are both blind spots are supplemented with the other shapes to obtain a shape with less “dead spots”.
[0088]
In the present invention, since the projector is used as the projection means, the alignment becomes easy. For example, an alignment pattern is projected from the projection means and displayed on a monitor or the like by the imaging means. While watching the alignment pattern displayed on the monitor or the like, the adjustment position adjustment screw of the projection means and the imaging means is turned to adjust the individual position and the expected angle.
[0089]
It is also possible to calibrate the arithmetic processing means by pattern projection. In this case, a calibration pattern is projected from the projection means, and the image is picked up by the image pickup means to capture the calibration pattern. For example, by determining the position x = s (k) of the slit k on the reference plane by this method, it is possible to calibrate the device geometry.
[0090]
Consider the case where the pattern display element (i, j) of the projection means is projected onto the reference plane as shown in FIG. Here, for simplicity, the reference plane is assumed to be perpendicular to the center line of the imaging means 8.
[0091]
FIG. 8 shows a projection pattern on the reference plane. When there is no projection distortion, the projection pattern is composed of straight lines as shown in FIGS. As shown in FIGS. 4A and 4B, the slit pattern is scanned in the vertical direction (a) and the horizontal direction (b), and image processing is performed by the arithmetic processing means. In this way, the element (i, j) and the projection position (c) on the reference plane can be made to correspond one-to-one.
[0092]
Also, the symmetry and parallelism of the pattern with respect to the reference plane can be easily set by adjusting the angle φ around the axis of the projection means. Note that in the conventional projection means capable of projecting only the slit pattern, it is necessary to rotate the projection means by 90 degrees in order to perform vertical scanning (a) and horizontal scanning (b). It is possible to project a slit pattern in two vertical and horizontal directions while fixing.
[0093]
8D to 8F show the projection patterns of the display element (i, j) when the lens has aberration. In this case, in the conventional projection means that can project only the slit pattern, it is necessary to perform the vertical scanning (d) and the horizontal scanning (e) by rotating the projection means by 90 degrees. A slit pattern in two vertical and horizontal directions can be projected with the means fixed. In this way, the projection position on the reference plane of the display element (i, j) is associated, and it is possible to easily determine what aberration is present from the projection pattern. Furthermore, if the deviation of the display element (i, j) due to this aberration is measured by the imaging means and stored as correction data δ (i, j), the accuracy of the shape measurement can be further improved.
[0094]
In addition, this structure shows one Example of this invention until it gets tired, The variation of the following structures can be considered easily and can be implemented.
[0095]
(1) In this example, “multi slit” is projected as “arbitrary pattern”. However, the pattern may be “single slit”, “binary code pattern”, or any other pattern or combination thereof. Good.
[0096]
(2) In this example, “multi-slit” is projected as “calibration pattern”, but the pattern is “single slit”, “binary code pattern”, any pattern, or a combination of them. Also good.
[0097]
(3) In this example, the direction of the projection pattern is “horizontal direction”, but this direction may be “vertical direction”, diagonal direction, or any direction.
[0098]
(4) In this example, a computer for pattern projection control, a computer for image input control, and a computer for shape calculation are implemented by a single computer. It can be easily realized by a computer.
[0099]
(5) As in the “security field”, measure the shape by projecting and imaging invisible light such as infrared light so that the subject is not aware of the measurement. It is also possible to configure easily. This can also be applied to the case where a natural expression of the measurement subject is obtained without being conscious of measurement, such as in the beauty field.
[0100]
【The invention's effect】
The present invention makes it possible to solve the problems of the prior art by projecting various patterns using projection means capable of projecting an arbitrary pattern. As a result, “high precision”, “high speed”, “robust” “absolute shape” measurement can be realized with a “compact” optical system configuration by combining and using appropriate methods according to technical demands. Is possible. Furthermore, a simple “automatic calibration” of the optical system is possible.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration of an invention.
FIG. 2 is a block diagram showing a configuration example of a control system.
FIG. 3 is a diagram showing a shape calculation procedure.
FIG. 4 is a diagram showing an example of shape calculation.
FIG. 5 is a diagram showing a configuration example using a plurality of projection means.
FIG. 6 is a diagram illustrating a configuration example using a plurality of imaging units.
FIG. 7 is a diagram showing a geometric relationship between a pattern display element (i, j) of a projection unit and a projection pattern.
FIG. 8 is a diagram showing a projection position of a pattern display element (i, j) on a reference plane.
FIG. 9 is a diagram showing a configuration of a light cutting method.
FIG. 10 is a diagram showing a configuration of an image encoding method.
FIG. 11 is a diagram illustrating a configuration of an image encoder.
FIG. 12 is a diagram showing a configuration of a multi-slit image encoding method.
FIG. 13 is a diagram showing a problem of the multi-slit image encoding method.
FIG. 14 is a diagram showing a configuration and problems of a binary pattern projection method.
[Explanation of symbols]
1 Reference plane
2 Object to be measured
3a slit light
4 Rotating mirror
7 Projection means
8 TV camera
8a Monitor TV
13 Image composition circuit
18 Maximum brightness image calculator
19 Image composition calculator
20 Synchronous circuit
21 Memory address generator
22 Output control circuit
23 Maximum brightness image memory
25 Maximum brightness image memory control circuit
26 Comparison circuit
27 Switch circuit
28 Composite image memory
30 pattern projection block
31 Pattern (sequence) generation circuit
32 Projector output switching circuit
39 Pattern projection control circuit
40 Image input block
41 Camera input circuit
42 Image memory switching circuit
43 Image memory
49 Image input control circuit
50 Shape calculation block
51 Angle image composition circuit
52 Precision partial shape calculation circuit
53 Precision overall shape calculation circuit
54 Shape image memory
59 Shape calculation control circuit
60 shape composite block
90 Sequence control means
θ slit light projection angle (conventional technology: angle with reference surface, the present invention:
The angle formed with the normal)
角度 Angle between the center line of the projection means and the reference plane
Z Height from reference plane

Claims (8)

  Pattern projection means capable of projecting an arbitrary pattern on the surface of the object to be measured, and a pattern projected multiple times while shifting one or more slits projected on the surface of the object to be measured in a direction perpendicular to the slits An imaging unit that captures at least one type of pattern including at least one direction from a direction different from the projection direction, and a shape calculation unit that calculates the three-dimensional curved surface shape of the object to be measured by performing image processing on the captured pattern projection image And a three-dimensional curved surface shape measuring apparatus.   2. The three-dimensional curved surface shape measuring apparatus according to claim 1, wherein the pattern projection unit is a projector connected to a computer.   The shape calculating means obtains the position of the pixel corresponding to the slit pattern from the slit pattern image, and corresponds to each slit from the relative position of the slit in the slit pattern, the geometric relationship between the projector and the imaging means. 3. The apparatus for measuring a three-dimensional curved surface shape according to claim 1, wherein the position of the surface of the object to be measured is calculated by a calculation process to obtain the shape of the surface of the object to be measured.   For each pixel of the slit pattern image, the shape calculating means has a slit relative position memory for storing the relative position in the slit pattern of the slit corresponding to the pixel from the slit pattern when each pixel shows the maximum luminance. And calculating the position of the surface of the object to be measured corresponding to each pixel from the geometric relationship between the slit relative position corresponding to each pixel, the projector, and the imaging means, and the shape of the surface of the object to be measured 4. The apparatus for measuring a three-dimensional curved surface shape according to claim 1, wherein:   5. The three-dimensional curved surface shape measuring apparatus according to claim 1, further comprising a plurality of pattern projection means.   6. The three-dimensional curved surface shape measuring apparatus according to claim 1, comprising a plurality of imaging means.   7. The three-dimensional curved surface shape measuring apparatus according to claim 1, wherein invisible light is used as the projection / imaging means. Using a projector connected to a computer, a slit pattern consisting of one or more slit lights is projected on the surface of the object to be measured, shifted in a direction perpendicular to the slit, and projected multiple times.
Photograph the light pattern projected on the surface of the object to be measured from a direction different from the projection direction,
A slit pattern showing the maximum brightness is selected for each pixel from the captured images of the plurality of slit patterns, and the relative position of the slit corresponding to the pixel in the slit pattern is stored in the slit relative position memory. And
From the relative position of the slit corresponding to the pixel, the geometrical relationship of the pattern projection means, and the imaging means, the position of the surface of the object to be measured corresponding to each pixel is calculated by calculation processing, and the shape of the surface of the object to be measured is calculated. A method for measuring a three-dimensional curved surface shape, comprising measuring.

Images