JPH02110305A - Three-dimensional measuring device - Google Patents

Abstract

PURPOSE:To measure the whole shape of an object only by one pattern projection and one image pickup, by using a pattern wherein a large number of slits formed of segments whose colors are different and whose lengths are varied at random are constructed and disposed so that no segments of the same color are located at the same position. CONSTITUTION:Segments using three colors of R, G and B are disposed on one slit. The length of each segment is decided at random by using a reference length and uniform random numbers. The segments are disposed so that no segments of the same color are located at the same position. The segments in a pattern which have the possibility of corresponding to the image 12 of some segment in an image are checked up in a parallel direction as indicated by broken lines and are limited to the one that has the same color and is located at the same position. In the present case, the segment denoted by 13 corresponds to said segment in the image.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to a three-dimensional measuring device that measures the shape of an object by projecting a light pattern.

[Prior Art] Conventionally, there has been a three-dimensional measuring device of this type as shown in Fig. i13.

Figure 13 is 190 of Digital Image Processing Engineering, written by Kei Tezuka, Tadahiro Kitahashi, and Hideo Ogawa, published by Nikkan Kogyo Shimbun in 1985.
A perspective explanatory diagram showing a method called slit light projection method performed by a conventional three-dimensional measuring device using light pattern t2 shadow shown on the page, and FIG. 11 is a block diagram showing the operation of the three-dimensional measuring device shown in FIG. 13. It is a diagram. In the figure, (1) is a mask pattern for the projected slit pattern, (2)
) is a mask pattern driving device for changing the slit position, (3) is a pattern projection device for projecting the slit pattern onto an object, (4) is a television camera for reading an image of the object onto which the slit pattern is projected, ( 5) is a storage device for storing image data, and (6) is a slit pattern that is projected onto the object and the slit pattern stored in the storage device (5) is used to calculate the slit pattern from the projected image of the object when the slit butter hits it. This is a calculation device that measures the three-dimensional distance of the area.

FIG. 12 is a plan view showing the slit pattern mask pattern (1) of FIG. 11.

Next, the operation will be explained. A slit pattern projection mask pattern (1) is moved to an appropriate position by a mask pattern driving device (2), and a slit-shaped light pattern is applied to an object by a pattern projecting device (3). Next, this image is read by a television camera (4) and stored in an image data storage device (5), and an image of the portion illuminated by light is detected by a calculation device (6). From the detected image of the slit pattern and the position of the mask pattern (1) used for projection, 3
Measure the distance to the part of the object that is illuminated by angle measurement. The above operation is carried out by the mask pattern driving device (2).
The distance of the entire object is measured by moving the mask pattern (1) and repeating the measurement using .

[Problems to be Solved by the Invention] Since the conventional three-dimensional measuring device using slit pattern projection is configured as described above, in order to measure the shape of the entire object, the mask pattern for slit pattern projection must be sequentially moved. In order to measure the distance, many images are required, and the measurement takes a long time.

This invention was made to solve the above-mentioned problems, and its purpose is to provide a three-dimensional measuring device that can measure the shape of an entire object with just one pattern projection and one image capture. do.

[Means for Solving the Problems] The three-dimensional measuring device according to the present invention uses a large number of slits each consisting of randomly changing length segments of different colors as a mask pattern for pattern projection, and slits of the same color at the same position. This uses a pattern configured and arranged so that there are no segments.

[Function] A large number of slits in the mask pattern of this invention make it possible to measure the entire object with one image capture, and the length varies randomly and the slits are arranged so that no two slits have the same color at the same position. The slit, which is made up of segments, provides an index of which slit on the mask pattern the slit image on the image corresponds to, making it easy to associate the slit image with the mask pattern, and making it possible to easily associate the slit image with the mask pattern. Realizes three-dimensional distance measurement of the entire object.

[Example] Hereinafter, an example of the present invention will be described with reference to the drawings. 1st
The figure is a plan view showing one embodiment of a mask pattern for slit pattern projection used in the present invention. In the diagram,
A case is shown in which three colors, R (red), G (edge), and B (blue) are used. Segments using the three colors R, G, and H are arranged on one slit. The length 2 of each segment is determined by using the reference length 2° and the −-like random number ε, 2=2°+ε
is randomly determined as. Arrange each segment so that there are no segments of the same color in the same position. This will be explained with reference to FIG.

After determining the length of each segment using random numbers, the color is determined in order from the reference slit. In the reference slit, the color of each segment is determined so as to match the color adjacent above and below. How to determine the color of each segment in the other slits will be explained with reference to FIG. 3, with reference to an example in which the leftmost slit is used as the reference slit. Assume now that the color of each segment of the i-th slit has already been determined.

The color of the segment shown in (7) of the i+1th slit is determined by checking whether there is a segment in the same position as (7) in the single slit determined so far, as shown by the broken line in Figure 2. If there is a segment whose end points match, choose a different color. If segments of all color types already exist at the same position, select a color that maximizes the slit interval between segments of the same color. If there is no segment whose end points coincide with each other at the same position, the segment that occupies the largest proportion of the portion indicated by the broken line in FIG. 2 in each slilat is selected, and its color is examined to determine the color. This will be explained with reference to FIG.

The length of the segment in FIG. 2 (7) is assumed to be 2. The segment with the maximum intersection with segment (7) in the j-th slit is indicated by (8). Measure the length of segment (8), set the length of the intersection of segment (8) with (7) by 2°, and set the color of segment (8) to C. Color C is either R9G or B in this case.

The color of the segment in (7) is mR=c−nim(j
%c), the color in (7) is determined to be c (c is either R, G, or B). However, at this time, IIIR+ 'M O+
” If any @ is 0, choose that color.

FIG. 4 is a block diagram of the three-dimensional measuring device of the present invention using the mask pattern determined in this way.
9) is the mask pattern shown in FIG. 1, (lO) is a pattern projection device for projecting the pattern onto an object, (11)
) is a color camera for reading the projected image of the projected colored pattern, (12) is a storage device for storing image data, and (13) is the image of the slit pattern projected onto the object and the mask pattern (9). ) and measures the three-dimensional distance to an object by triangulation. It is assumed that the pattern projection device (10) and the color camera (11) are arranged in parallel.

Next, the function and operation will be explained. Before projecting a pattern onto an object, each segment in the pattern is classified. When a large number of slit patterns are projected onto an object, it is necessary to determine which slit in the pattern each slit projected onto the object corresponds to. Segments in the pattern that may correspond to segments in the slit image projected onto the object are detected by a pattern projection device (10
) and the color camera (11) are arranged in parallel, the search is performed in the parallel direction, and the search is limited to those that match. This will be explained with reference to FIG. Segments in the pattern that may correspond to the image (12) of a certain segment in the image are examined in parallel directions, as indicated by broken lines, and are limited to those of the same color and at the same position. In this case (13
) is the corresponding segment. When two or more such segments exist, it is not possible to determine the correspondence by this alone.In order to investigate the nature of such a pattern, the segments in the pattern are classified before projecting the pattern. This will be explained with reference to FIG. The segments in the pattern are classified into the 3111 class. (a) The figure shows a case in which there are no segments with the same color and at the same position even when examined in the parallel direction, and there is no match at either end point. The figure (b) is similar to the figure (a) in that there are no segments with the same color and the same position, but there are segments with the same end point. (C) The figure shows a case where segments of the same color and at the same position exist. Since such pattern classification can be performed before measurement, there is no effect on the time required for measurement. After the patterns are classified, in step (14) in the flowchart shown in FIG. 7, a pattern as shown in FIG. 1 is projected onto the object. By projecting a pattern as shown in FIG. 1, a large number of slit images consisting of segments of three colors, RlG and B, are generated on the object. This is read into the storage device (13) by the color camera (11), the image of the slit pattern on the object is detected in step (15), and then in step (16) the slit image is converted into R (red), R (red), Classify by G (edge) and B (blue) and proceed to step (17
), the segments in the pattern are associated with each segment of the slit image on the object for each color classified.

In step (18), if the correspondence is not uniquely determined from the individual correspondence for each color,
In step (19), a correspondence is established based on peripheral consistency such as the color and positional relationship of the upper and lower and left and right segments. From the correspondence obtained in steps (17) and (1g), the object is Calculate the 3D distance to.

Next, the correspondence in step (17) and the correspondence in step (18) will be explained.

FIG. 8 is a flowchart showing the correspondence processing in step (17). After classifying the slit images by color in step (16), the process shown in FIG. Search in the direction, check whether there is a pattern in which both endpoints correspond, and if there is one such pattern,
The pattern class I checked in advance is the 6th one.
When the situation is as shown in figure (a), step (23)
, the response shall be deemed finalized. If not, that is, the pattern class is as shown in Fig. 6(b),
In the case of (C), the correspondence is determined to be undetermined in step (24), the obtained correspondence candidate is stored in step (25), and the consistency of the surroundings is checked in step (18). If there is no correspondence candidate corresponding to both endpoints, step (2)
In C), check to see if there is one that corresponds to one side. If one such pattern exists, check the class of that pattern in step (27) (a)
or (b), the response shall be deemed finalized.

In other cases, the correspondence is left undetermined, and the consistency of the surroundings is checked in step (18). Next, step (1
The process 8) will be explained. Figure 9 shows the step (18
) is a flowchart showing the processing. As for the segment whose correspondence has not yet been determined by the processing in step (17), it is checked in step (28) whether or not its left and right neighbors have been dealt with in step (17), and if they have been dealt with, in step (29) Check its consistency, and if it is consistent, step (30)
), the response is considered finalized. This is shown in Figure 10(d).
Let me explain with an example. In this figure, 2 is the segment whose correspondence is currently being searched, and the correspondences have already been determined on both sides of the segment, which is the t-1th slit and the i÷1th slit. At this time, it can be seen that 2 is the i-th slit. If both adjacent segments are not matched, check the correspondence between the upper and lower segments in (31), and if - both are matched and have the same slit number, the one currently being checked is also considered to belong to the same slit. . In addition, even if both are not already supported, check whether the matching color change is the only one among the matching candidates in (32), and if it is the only one, the matching is confirmed in (33). shall be.

This will be explained using an example in FIG. 1O(e). 2 is the segment currently being searched, its color is R (red), x
l, f' are the upper and lower segments, and their colors are B(
blue), G (green), and the correspondence candidates for 2 are il,
Suppose that there are three such as j' and nl, and the appearance of the color is the first one.
Suppose that the situation is as shown in Figure O (e). At this time, j' is selected as the corresponding one based on color consistency. In this way, by projecting a large number of slits consisting of color-divided segments whose lengths vary randomly, it is possible to reduce correspondence errors due to the randomness of segment lengths. By encoding using colors, even if the end points match, the correspondence can be made from the colors on both sides and the colors above and below. This-
By projecting such a pattern, the entire object can be measured with one pattern projection and one image capture.

In addition, in the above example, the colors used are RlG, 83
Although the case of colors is shown, more colors may be used.

In addition, a pattern projection device (lO) and a color camera (11
) are arranged in parallel, but they do not necessarily have to be parallel.

[Effects of the Invention] As described above, according to the present invention, a mask pattern for pattern projection is made up of segments of random length and arranged so that no two slits have the same color at the same position so that each slit can be distinguished. Since it is composed of many slits, 1
It is possible to measure the three-dimensional shape of an object by projecting a pattern and photographing images twice, which has the effect of requiring fewer images and a shorter measurement time.

[Brief explanation of the drawing]

FIG. 1 is a plan view showing an example of a mask pattern for projecting a slit pattern used in the present invention, FIGS. 2 and 3 are explanatory diagrams showing a method of constructing the pattern shown in FIG. 1, and FIG. is a block diagram showing a three-dimensional measuring device according to an embodiment of the present invention, and FIGS. 5, 6, and 10 are explanations of patterns for explaining the operation of the three-dimensional measuring device according to an embodiment of the present invention. 7, 8, and 9 are flowcharts showing the operation of a three-dimensional measuring device according to an embodiment of the present invention, FIG. 11 is a block diagram showing the operation of a conventional three-dimensional measuring device, and FIG. 1 is a plan view showing a conventional mask pattern for projecting a slit pattern, and FIG. 13 is a perspective explanatory view showing the operation of a conventional three-dimensional measuring device. In the figure, (7) (a) (12) (13) indicate patterns. In addition, in the figures, the same reference numerals indicate the same or equivalent parts.

Claims (1)

[Claims] In an active measurement device that measures the shape of a three-dimensional object by projecting a light pattern onto the object, it is possible to measure the shape of the object by capturing a single image by projecting multiple slit patterns encoded using colors. A three-dimensional measurement device characterized by performing three-dimensional measurement.