JP2018115878A - Measuring apparatus, system and product manufacturing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique advantageous for specifying an abnormal optical unit in a measuring apparatus having a plurality of optical units.SOLUTION: The measuring apparatus measures a position of an object by using a plurality of optical units for measuring a distance on the basis of a triangulation method. The measuring apparatus includes a control unit for specifying an abnormal optical unit among the plurality of optical units on the basis of an evaluation result information group including a plurality of evaluation results obtained by evaluating a plurality of measurement results relating to the position of the object measured by using the plurality of optical units.SELECTED DRAWING: Figure 4

Description

  The present invention relates to a measuring device, a system, and an article manufacturing method.

  Conventionally, there are shape measuring apparatuses provided with two optical units such as cameras and projectors, but in recent years there are shape measuring apparatuses provided with three or more optical units. For example, in the case of a pattern projection system having a projector, a known pattern is projected onto an object and imaged, and the three-dimensional shape of the object is measured from the imaged result. Here, a conventional technique will be described for a pattern projection type shape measuring apparatus.

  Patent Document 1 describes a three-dimensional measuring apparatus having one projector and a plurality of imaging means. In the three-dimensional measuring apparatus of Patent Document 1, a test object projected with pattern light by a projector is imaged by a plurality of imaging units, a three-dimensional point group is obtained based on the captured images, and these are integrated. . At the time of this integration, a value having a large luminance amplitude is selected as having high reliability.

  Patent Document 2 describes a three-dimensional image processing apparatus that includes two light projecting units and a common image capturing unit, and includes two measurement systems. In the three-dimensional image processing apparatus described in Patent Document 2, distance images are obtained by the two measurement systems, and the degree of the state of the three-dimensional image processing apparatus is determined by comparing these distance images. Then, when the degree of state change exceeds a predetermined threshold, calibration of the 3D image processing apparatus is prompted.

JP 2015-21862 A Japanese Patent Laid-Open No. 2015-45587

  In the three-dimensional measuring apparatus described in Patent Document 1, when an imaging unit that outputs a value with a large luminance amplitude is a deteriorated imaging unit, a final three-dimensional measurement is performed based on a value obtained by the imaging unit. A point cloud is determined. Therefore, when the imaging unit that outputs a value with a large luminance amplitude is a degraded imaging unit, correct measurement cannot be performed.

  In the three-dimensional image processing apparatus described in Patent Document 2, this is recognized when the degree of state change exceeds a predetermined threshold, but there is a problem with either the two light projecting means or the common imaging means. It cannot be specified whether it has occurred.

  For example, an object of the present invention is to provide a technique that is advantageous for specifying an abnormal optical unit in a measurement apparatus having a plurality of optical units.

  One aspect of the present invention relates to a measuring device that measures the position of an object using a plurality of optical units for measuring a distance based on a triangulation method, and the measuring device uses the plurality of optical units. A control unit that identifies an abnormal optical unit among the plurality of optical units based on an evaluation result information group including a plurality of evaluation results obtained by evaluating a plurality of measurement results related to the position of the object measured Is provided.

  According to the present invention, for example, a technique advantageous for specifying an abnormal optical unit in a measurement apparatus having a plurality of optical units is provided.

The figure which shows the structure of the measuring device of 1st and 2nd embodiment of this invention. The figure which shows the Example of the measuring device of 1st and 2nd embodiment of this invention. The figure which shows operation | movement of the measuring device of 1st and 2nd embodiment of this invention. The figure which illustrates the several position measurement result (a) in 1st Embodiment of this invention, and the evaluation result information group (b) which is the result of evaluating them. The figure which illustrates reproducibility of a position measurement result. The figure which illustrates reproducibility of a position measurement result. The figure which illustrates the position measurement result (a) in Example 1-1, and the evaluation result information group (b) which is the result of having evaluated them. FIG. 3 is a diagram illustrating an arrangement of a plurality of optical units in Example 1-1. Arrangement (a, b, c) of a plurality of optical units in Example 1-1, position measurement results (d, e, f) by them, and evaluation result information group (g, h) as a result of evaluating them , I). The figure which shows the position measurement result (a, b, c) in Example 1-2, and the evaluation result information group (d, e, f). The figure which shows the measuring device of Example 1-3. The figure which shows the position measurement result (a, b) in Example 1-3, and the evaluation result information group (c, d, e). The figure which illustrates the evaluation result information group in Examples 1-1 to 1-3. The figure which shows the position measurement result (a, b) and the evaluation result information group (c, d) in 2nd Embodiment of this invention. The figure showing the difference amount Z from the reference | standard and reproducibility in 3rd Embodiment of this invention. The figure which shows the structural example of a control part. The figure which shows the system of 4th Embodiment of this invention.

Hereinafter, the present invention will be described through exemplary embodiments thereof with reference to the accompanying drawings.
[First Embodiment]
FIG. 1 shows the configuration of a measuring apparatus 50 according to one embodiment of the present invention. The measuring device 50 includes a plurality (N) of optical units 1-j (j: 1 to N) (N is 4 or more) for measuring a distance based on the triangulation method, and a control unit 9. The control unit 9 is, for example, PLD (abbreviation of Programmable Logic Device) such as FPGA (abbreviation of Field Programmable Gate Array), or ASIC (abbreviation of Application Specific Integrated Circuit). It can be constituted by a computer or a combination of all or part of them.

  The optical unit 1-j includes a lens 6-j and a unit 7-j. When the optical unit 1-j is an imaging unit, the unit 7-j can include an image sensor. When the optical unit 1-j is a projection unit, the unit 7-j may include a light source. The relative positional relationship between the optical units 1-1 to 1-N is calibrated in advance using a calibrator or the like and is known. In addition to controlling the plurality of optical units 1-j, the control unit 9 processes signals acquired from the plurality of optical units 1-j. In FIG. 1, the plurality of optical units 1-j are arranged on the XZ plane, but all or some of the plurality of optical units 1-j may have different positions in the Y direction.

  In the following description, in order to provide a specific example, the measurement device 50 is configured as an active stereo measurement device, one optical unit 1-1 is a projection unit, and the remaining optical units 1-2 to 1- Let N be an imaging unit. A state where the optical unit deviates from the normal state is referred to as an abnormal state. The abnormal state may be a state where at least one of the position and the posture of the optical unit deviates from the normal state. Alternatively, the abnormal state may be a state in which a defect such as fogging has occurred in the lens of the optical unit. An optical unit in an abnormal state is also expressed as an optical unit having an abnormality.

  In the following description, it is assumed that only one of the plurality of optical units 1-j can be in an abnormal state. This is because it is extremely rare for a plurality of optical units to be in an abnormal state at the same time. If at least two optical units of the plurality of units 1-j are in an abnormal state, the situation can be similar to the case where the projection unit is in an abnormal state. Therefore, when it is determined that the projection unit is in an abnormal state, the control unit 9 outputs a message indicating that “the projection unit or at least two imaging units may be in an abnormal state”. Also good.

  FIG. 3 illustrates the operation of the measurement device 50 when one of the plurality of optical units 1-j is in an abnormal state. The operation shown in FIG. 3 is controlled by the control unit 9. First, in step S11, the control unit 9 first causes the optical unit 1-1, which is a projection unit, to project pattern light onto the test object (object) 8, and the optical units 1-2 to 1-N, which are imaging units. Let the subject 8 be imaged. Thereby, the control part 9 acquires the image (image data) which imaged the pattern light projected on the test object 8 from each of the optical units 1-2 to 1-N. Next, in step S11, the control unit 9 obtains a distance image based on the images obtained from the optical units 1-2 to 1-N. Here, the distance image may be an image in which information indicating distances from the reference position to the positions of a plurality of locations of the test object 8 is matched as a pixel value. Since the distance indicates a relative position with respect to the reference position, the distance is equivalent to information indicating the position of each part of the test object 8. Therefore, the distance image indicates the positions of a plurality of locations on the test object 8, that is, the shape of the test object 8. That is, obtaining the distance image corresponds to measuring the shape of the test object 8.

  Here, a method for acquiring a distance image will be described as an example. For example, pattern light such as gray code pattern light is projected onto the test object 8 from the optical unit 1-1 as the light projecting unit, and is applied to the test object 8 by the optical units 1-2 to 1-N as the imaging unit. The projected pattern light is imaged. As a result, (N-1) images are obtained. Next, based on the image obtained by imaging, the address of the pattern setting element (for example, DMD, liquid crystal) of the optical unit 1-1 and the unit (image sensor) of the optical unit 1-j (j = 2 to N) The address 7-j is associated with the spatial coding method. Since the relative positions of the optical unit 1-1 as the projection unit and the optical unit 1-j (j = 2 to N) as the imaging unit are known, the position of each part of the test object 8 based on triangulation A distance image indicating (distance from the reference position) can be obtained.

  The control unit 9 recognizes the test object 8 based on the shape information of the test object 8 (for example, design information such as CAD information) and the distance image, and the position (X, Y, Z) of the test object 8. Can be acquired. The control unit 9 can obtain (N−1) measurement results as measurement results of the position of the test object 8 by using one projection unit and (N−1) imaging units. . When a plurality of test objects 8 exist, the control unit 9 can obtain information indicating the positions of the plurality of test objects 8 based on each distance image.

  As described above, by using the plurality of optical units 1-1 to 1-N, a plurality of measurement results can be obtained as measurement results indicating the position (X, Y, Z) of one test object 8. . If the states of the plurality of optical units 1-1 to 1-N do not change from the calibration state of the measuring device 50, the plurality of measurement results should be equal within the range of measurement reproducibility. However, if any of the plurality of optical units 1-1 to 1-N has changed from the calibration state of the measuring device 50, the plurality of measurement results do not match each other.

  In step S12, the control unit 9 performs a plurality of evaluations on a plurality of position measurement results related to the positions of the test object 8 measured using the plurality of optical units 1-1 to 1-N, and obtains the evaluation. An evaluation result information group indicating the plurality of evaluation results is generated. For example, the control unit 9 can perform the evaluation based on the similarity between the plurality of position measurement results. As a more specific example, the control unit 9 can perform an evaluation according to whether or not the difference between a plurality of measurement results is within a threshold, and the evaluation result information group is used for such an evaluation. It may be a set of evaluation result information determined accordingly. Or the control part 9 can evaluate according to whether the difference of the two measurement results which respectively comprise the some subset obtained from the set which consists of a some measurement result is settled in the threshold value. The evaluation result information group may be a set of evaluation result information determined according to such evaluation. Here, the fact that the difference is within the threshold value indicates that the two measurement results to be compared are similar, and the control unit 9 can determine that the result is acceptable. On the other hand, the fact that the difference does not fall within the threshold indicates that the two measurement results to be compared are not similar, and the control unit 9 can determine that the difference is not acceptable.

  Hereinafter, evaluation of a plurality of measurement results in step S12 (generation of evaluation result information group) will be described in detail. In the following description, the position (X, Y, Z) of the test object 8 will be described as the position Z for simplification of description. Here, the position Z of the test object 8 is expressed as a relative position with respect to a certain optical unit 1-j. FIG. 5A illustrates the distribution of results obtained when the position Z of the test object 8 is measured a plurality of times using a certain optical unit 1-j. In FIG. 5A, the horizontal axis indicates a position obtained by measurement, and the vertical axis indicates establishment or frequency. When noise at the time of measurement is a random component, if measurement is performed a plurality of times, the distribution of measurement results becomes a Gaussian distribution like the curve 20. A range 21 (eg, ± 3σ) of the curve 20 can be an index indicating the reproducibility of the measurement result. The control unit 9 measures the position of the test object 8 using the optical units 1-1 to 1-N, and evaluates the reliability of the position measurement result based on the distribution of the position measurement result obtained thereby. Can do. When the reliability does not reach the predetermined standard, the lighting conditions and the like can be changed.

  Next, when the measurement result of the test object 8 arranged at an arbitrary position is the position measurement result Za in FIG. 5B, the position measurement result Za can correspond to the left end or the right end in the curve 20. There is sex. Therefore, it is necessary to assume a range that is twice the reproducibility range 21 when the measurement is performed a plurality of times as a range where the test object 8 from which the position measurement result Za is obtained may actually exist. is there. The same applies to the position measurement result Zb. In order that the values of the position measurement results Za and Zb do not coincide with each other, Za and Zb must be separated from the range 21 by two times (for example, ± 6σ) or more. Therefore, the threshold value 22 for determining that the two position measurement results are different from each other may be twice the range 21.

  When the difference between the positions Za and Zb measured by the two optical units does not fall within the threshold value 22, the control unit 9 can determine that the two optical units are rejected. On the other hand, as shown in FIG. 5C, when the difference between the positions Za and Zb measured by the two optical units is within the threshold value 22, the two may match. The two optical units can be determined as acceptable. In this specification, “difference” is treated as an absolute value.

  The curve 20 or the threshold value 22 may be determined in advance by actually measuring the test object a plurality of times, or may be determined in advance based on a shot noise coefficient of the sensor or the like. In actual measurement, the brightness of the image, the contrast of the pattern light, and the like may change for each measurement, for each test object 8 to be measured, and for each optical unit 1-j. Therefore, the threshold value may be adjusted or reset for each measurement or for each optical unit. The threshold value can be set by the control unit 9. The threshold value can be determined based on a plurality of differences. Further, the threshold value can be determined based on an average value of a plurality of differences. Here, the threshold value may be the average value itself.

  FIG. 6 illustrates position measurement results Z2 and Z3 by the optical units 1-2 and 1-3. For example, consider a case where the brightness of the image acquired by the optical unit 1-2 is brighter than when the curve 20 shown in FIG. 5 is acquired. In this case, the SN ratio of the optical unit 1-2 is improved, and the range 21-2 of the curve 20-2 representing the measurement reproducibility is narrower than the range 21 of the curve 20 of FIG. 5 as shown in FIG. It is predicted. On the contrary, when the image of the position measurement result Z3 of the optical unit 1-3 is dark, it is predicted that the range 21-3 of the curve 20-3 is widened. In this case, it can be set as the value which added 21-2 and 21-3 as a threshold value for determining both similarity. Thus, since reproducibility changes according to the measurement result for each measurement or each optical unit, the control unit 9 may reset the threshold value according to the measurement result for each measurement or each optical unit.

  In the above example, the threshold value is determined based on reproducibility. However, even if the same test object 8 is simultaneously measured due to the calibration error of the measuring device 50, a slight difference occurs in the absolute values of a plurality of measurement results. Yes. When determining the similarity between two measurement results, a difference in calibration error may be set in addition to a threshold value. Or you may determine a threshold value according to the required precision requested | required of the measuring apparatus 50. FIG.

  Next, a method for evaluating the position measurement result will be described. FIG. 4A illustrates a plurality of position measurement results Z2 to ZN obtained by the control unit 9 processing images captured by the optical units 1-2 to 1-N as the imaging units. Yes. FIG. 4B illustrates a result (evaluation result information group (evaluation result information pattern)) obtained by the control unit 9 evaluating the plurality of position measurement results Z2 to ZN in FIG. In FIG. 4B, the difference (Z2-Z3, Z2-Z4, etc.) between two measurement results respectively constituting a plurality of subsets obtained from a set consisting of a plurality of measurement results Z2-ZN is within the threshold. Cases (if they are similar) are marked with a circle. Further, in FIG. 4B, the difference (Z2-Z3, Z2-Z4, etc.) between two measurement results respectively constituting a plurality of subsets obtained from a set consisting of a plurality of measurement results Z2 to ZN falls within the threshold value. If not (not similar), it is indicated by a cross. For example, Z2 and Z3-Zn are not similar, and Z3-ZN are similar to each other. Therefore, the control unit 9 determines that Z2 is unacceptable and Z3 to Z4 are acceptable. In one example, ◯ (passed) is recorded as evaluation result information = “1”, and X (failed) is recorded as evaluation result information = “0”, and an evaluation result information group can be constituted by these.

  Hereinafter, if it continues description based on FIG. 3, in process S13, the control part 9 will determine whether process S12 was complete | finished about all the test objects 8, and process S12 will be performed about all the test objects 8. If not completed, the process returns to step S12. And process S12 is performed about the to-be-tested object 8 which has not performed process S12 yet. On the other hand, when step S12 is completed for all the test objects 8, the process proceeds to step S14.

  In step S14, the control unit 9 determines whether or not remeasurement is necessary based on the evaluation result information group obtained in the immediately preceding step S13. Specifically, when the control unit 9 determines that the optical unit in the abnormal state cannot be specified in step S15 in the evaluation result information group obtained in the immediately preceding step S13, remeasurement is necessary. Determine and return to step S11. In this case, in step S11, for example, after adjusting the light amount of the optical unit 1-1 as the projection unit, the optical unit 1-1 is caused to project the pattern light onto the test object 8 as the test object, and the imaging unit. The test object 8 is imaged by the optical units 1-2 to 1-N. More specifically, when the brightness of the image acquired by the optical unit 1-3 is dark, it is predicted that the width of the curve 20-3 representing the measurement reproducibility in FIG. In such a case, in step S11, the control unit 9 may not execute the position calculation based on the image acquired by the optical unit 1-3. In this case, part or all of the evaluation result information group illustrated in FIG. 4 may be missing. In such a case, the control unit 9 can determine that remeasurement is necessary.

  Next, in step S15, the control unit 9 identifies an optical unit having an abnormality among the plurality of optical units 1-1 to 1-N based on the evaluation result information group obtained in step S12. Here, since the projection unit affects all the position measurement results Z2 to ZN, if the projection unit is in an abnormal state, an evaluation result indicating that only one of Z2 to ZN is in an abnormal state. An information group (evaluation result information pattern) does not occur. Therefore, as shown in FIG. 4B, the evaluation result information group indicating that only Z2 among Z2 to Z4 is unacceptable is an optical unit 1-1 and an imaging unit as a projection unit related to Z2. This indicates that the optical unit 1-2 is in an abnormal state.

Next, in step 16, the control part 9 performs a post-process based on the result of process S15. For example, when the optical unit 1-2 is identified as an optical unit in an abnormal state, the control unit 9 can be configured to display on the display unit (not shown) that the optical unit 1-2 is in an abnormal state. After that, the position of the test object 8 may be measured by the measurement device 50 on the assumption that the image captured by the optical unit 1-2 is not used. In such a case, information indicating the correct position of the test object 8 can be obtained from the image captured by the optical units 1-3 to 1-N. The position of the test object 8 may be determined based on the average of the position measurement results Z3 to N, or may be determined based on the position measurement result with the highest reliability. In addition, the calibration parameters of the optical unit 1-2 set in advance are updated so that the position measurement result Z2 matches the position measurement results Z3 to N, so that the optical unit 1-2 is in a normal state. The process shown in FIG. 3 may be performed in parallel with the process in which the measuring device 50 measures the position of the test object 8 in order to process the test object 8. It may be performed in 50 test modes.

Hereinafter, Examples 1-1, 1-2, and 1-3 of the first embodiment will be described.
[Example 1-1]
In Example 1-1, as shown in FIG. 2, a measuring apparatus 50 having one optical unit 1-1 configured as a projection unit and four optical units 1-2 to 1-5 configured as imaging units. Will be described. In Example 1-1, it is assumed that images captured by the four optical units 1-2 to 1-5 are images captured with an appropriate amount of light.

  First, in step S11, the control unit 9 obtains four images by causing the four optical units 1-2 to 1-5 configured as imaging units to image the test object 8, and based on the respective images. To obtain four position measurement results Z2 to Z5. Next, in step S12, the control unit 9 determines whether or not the difference between the position measurement results Z2 to Z5 of the test object 8 is within the threshold (determines the similarity between the position measurement results Z2 to Z5). Thus, the position measurement results Z2 to Z5 are evaluated. In Example 1-1, a threshold value 22 (range of ± 6σ) is set as a threshold value for evaluation.

  FIG. 7A illustrates the position measurement results Z2 to Z5 and the threshold value T for evaluation in Example 1-1. FIG. 7B shows an evaluation result information group of the position measurement results Z2 to Z5 in FIG. Differences between two measurement results (ie, differences between Z2 and Z3, differences between Z2 and Z4, and differences between Z2 and Z5) that respectively constitute a plurality of subsets obtained from a set consisting of a plurality of position measurement results Z2 to Z5 The difference, the difference between Z3 and Z4, the difference between Z3 and Z5, the difference between Z4 and Z5) is within the threshold value T. The difference between the two measurement results constituting each of the plurality of subsets obtained from the set of the plurality of position measurement results Z2 to Z5 is within the threshold T. The plurality of position measurement results Z2 to Z5 are similar to each other. Means that In addition, the difference between two measurement results constituting a plurality of subsets obtained from the set of position measurement results Z2 to Z5 (that is, the difference between Z2 and Z3, the difference between Z2 and Z4, the difference between Z2 and Z5, (Difference, difference between Z3 and Z4, difference between Z3 and Z5, difference between Z4 and Z5) is not within the threshold value T. The difference between the two measurement results constituting each of the plurality of subsets obtained from the set of the plurality of position measurement results Z2 to Z5 does not fall within the threshold T. The plurality of position measurement results Z2 to Z5 are similar to each other. Means not.

  FIG. 7B visually shows the evaluation result information group. The position measurement result Z2 is rejected, and the position measurement results Z3, Z4, and Z5 are acceptable. The evaluation result information group (evaluation result information pattern) is stored in the storage area of the control unit 9 as a two-dimensional array of 4 rows and 4 columns. For example, ◯ is indicated by “1”, and × is indicated by “0”. sell.

  In Example 1-1, since there is one test object 8, step S13 passes. In step S14, it is determined that re-measurement is unnecessary. In step S14, the control unit 9 determines that the optical unit 1-2 that can affect the position measurement result Z2 is in an abnormal state based on the evaluation result information group generated in step S12.

  Here, the position measurement result Z2 can also be affected when the optical unit 1-1, which is the projection unit, is in an abnormal state, but when the optical unit 1-1 is in an abnormal state, the evaluation result information group is This is different from that shown in FIG. This point will be described below.

  Here, FIG. 8 shows a state in which the arrangement of the optical units 1-1 to 1-5 of the measuring device 50 shown in FIG. 2 is viewed from above (a state in which the Z axis is viewed from the negative side in the positive direction). Suppose we have such an arrangement. In FIG. 8, the symbol x indicates the optical unit 1-x. For example, the code | symbol 1 has shown the optical unit 1-1. As a state in which the position of the optical unit 1-1 has changed, (State A) a state of approaching an imaging unit, (State B) a state of approaching an intermediate position of two imaging units, (State C) (A) And three states between (B) are conceivable.

  First, in (State A), as shown in FIG. 9A, the optical unit 1-1 approaches the direction of the optical unit 1-4, and the base line length between the optical unit 1-1 and the optical unit 1-4 is short. On the contrary, the base line length of the optical unit 1-1 and the optical unit 1-2 becomes longer. Therefore, the position measurement result is as shown in FIG. 9D, and the evaluation result information group is as shown in FIG. Next, in (state B), as shown in FIG. 9B, the optical unit 1-1 approaches the intermediate position between the optical unit 1-3 and the optical unit 1-4, and the optical unit 1-1 and the optical unit 1 are moved. The baseline length with -3 / 1-4 is shortened. On the other hand, the base line length of the optical unit 1-1 and the optical unit 1-2 / 1-5 becomes long. Therefore, the position measurement result is as shown in FIG. 9E, and the evaluation result information group is as shown in FIG. 9H. Finally, in (State C), the base line lengths between the optical unit 1-1 and the other optical units 1-2 to 1-5 are different from each other because they are intermediate between (State A) and (State B). The position measurement result is as shown in FIG. 9 (f), and the evaluation result information group is as shown in FIG. 9 (i). Δ in FIG. 9 (i) becomes “◯” or “X” depending on whether the optical unit 1-1 is closer to the change direction of the position (state A) or (state B). Or, if the reproducibility is poor, the circles and the crosses change for each measurement and cannot be determined immediately, and there is a possibility that multiple measurements may be required.

  9 (g), (h), and (i), the combination of Z2 and Z4 is x, the combination of Z2 and Z5 is ◯, and the combination of Z3 and Z4 is ◯. 9D, 9E, and 9F, the difference value between the position measurement results Z2 and Z4 is always large, and the difference value between the position measurement results Z3 and Z5 is always small. Therefore, when the position of the projection unit changes, the position measurement results Z2 and Z4 by the imaging units arranged at the diagonal positions are x. Further, the position measurement results Z2 and Z5 by the image pickup units arranged at the adjacent positions become ◯, and the position measurement results Z3 and Z4 by the image pickup units arranged at the adjacent positions become ◯. Therefore, when measurement is performed a plurality of times, the result of the position measurement by the image pickup unit arranged at the diagonal position is x, and the result of the position measurement result by the image pickup unit arranged at the adjacent position is ◯ is accumulated. The Therefore, when the position of the projection unit changes, such an abnormal state of the projection unit can be identified based on the accumulation of results having the above-described tendency. Therefore, when the evaluation result information group as shown in FIG. 7B is obtained, it can be specified that the optical unit 1-2 is in an abnormal state.

  Here, the case where the baseline length is changed by changing the position of the optical unit 1-1 has been described. Based on the same principle, an abnormal optical unit can be identified based on the evaluation result information group even when the convergence angle changes due to a change in the attitude of the optical unit 1-1.

  Hereinafter, a configuration example of the control unit 9 and an operation example of the control unit 9 for specifying an optical unit in an abnormal state will be described. Here, a template for specifying an optical unit in an abnormal state is prepared, and a method for specifying an optical unit in an abnormal state by pattern matching between a table result information group (evaluation result information pattern) and a template is described. To do. FIG. 16 shows a configuration example of the control unit 9.

  The control unit 9 includes, for example, a position calculation unit 91, an evaluation unit 92, a template generation unit 93, a template storage unit 90, a unit specifying unit 94, and a post-processing unit 95. The template generation unit 93 generates a plurality of templates TMP to be referred to by the unit specifying unit 94 in order to specify which of the plurality of optical units 1-1 to 1-5 is in an abnormal state, and stores the templates. Stored in the unit 90. The template TMP, for example, uses the same information as the evaluation result information group (evaluation result information pattern) as exemplified in FIGS. 7B, 9G, 9H and 9I. It is included in a format comparable to the result information group. The template TMP may be generated by an external device of the measurement device 50, provided to the control unit 9 of the measurement device 50, and stored in the template storage unit 90.

  The position calculation unit 91 is a component that executes step S11 of FIG. 3, calculates the position of the test object 8 based on each of the plurality of images provided from the optical units 1-2 to 105, Output measurement results. The evaluation unit 92 is a component that executes step S12 of FIG. 3, evaluates a plurality of measurement results provided from the position calculation unit 91, and generates an evaluation result information group indicating the results. The unit specifying unit 94 is a component that executes step S15 of FIG. 3, and compares the evaluation result information group generated by the evaluation unit 92 with a plurality of templates TMP stored in the template storage unit 90 (pattern matching). To identify the optical unit in an abnormal state. The post-processing unit 95 is a component that executes step S16 of FIG.

  Here, in order to increase the accuracy of identifying an abnormal optical unit by the unit specifying unit 94, the position calculation by the position calculation unit 91 and the evaluation by the evaluation unit 92 (generation of the evaluation result information group) are repeated a plurality of times. May be.

  Next, post-processing in step S15 will be described. In step S15, for example, the control unit 9 notifies the operator of an abnormal optical unit through an output unit such as a display unit (not shown). Here, as illustrated in FIG. 7, the process that can be executed in step S15 when the optical unit 1-2 is in an abnormal state will be described as an example. When the optical unit 1-2 is in an abnormal state, the position measurement result Z2 of the test object 8 obtained using the image of the optical unit 1-2 is not correct, so the optical unit 1-2 is not used. On the other hand, since the position measurement result of the test object 8 obtained using the images of the remaining optical units 1-3 to 1-5 is correct, the test obtained using the images of the optical units 1-3 to 1-5. The average value of the position measurement results of the object 8 can be output as the final position measurement result of the test object 8. Here, instead of the average value, for example, a position measurement result with the highest reliability among a plurality of position measurement results obtained by the optical units 1-3 to 1-5 may be output.

  In addition, the amount of difference between the position measurement result by the optical unit 1-2 and the position measurement result by the other optical units 1-3 to 1-5 has been found. Therefore, the parameters of the optical unit 1-2 that have been calibrated in advance are corrected (recalibrated), and the optical unit 1-2 is recalibrated to match the outputs of the other optical units 1-3 to 1-5. It is also possible to do. However, an extra calculation time for correction may occur and affect the tact, so it may be possible to select whether or not to enable the function for performing such correction.

  As described above, according to Example 1-1, when the optical unit 1-2 as the imaging unit is in an abnormal state, the control unit 9 notifies the occurrence of the abnormal state and then remains. Measurement of the position of the test object 8 can be continued using the optical units 1-3 to 1-5. Further, when the parameters of the optical unit 1-2 can be corrected, the measurement of the position of the test object 8 can be continued using the optical units 1-2 to 1-5. Of course, the control unit 9 may end the measurement and notify the operator to adjust the optical unit 1-2.

  When the optical unit 1-1 as the projection unit is in an abnormal state, all of the plurality of position measurement results for the test object 8 are not correct. Therefore, in this case, the control unit 9 notifies the operator that the optical unit 1-1 as the projection unit is in an abnormal state, and ends the measurement. However, as described with reference to FIG. 9, when the position of the projection unit changes, the two position measurement results change to the positive and negative sides, so the average of all the position measurement results may be output. . Further, when the optical unit 1-1 can be recalibrated so that all the position measurement results match, the measurement can be continued after the recalibration.

  The range 21 (index indicating reproducibility) of the curve 20 can be arbitrarily determined, and in addition to the above 3σ, ranges such as 2σ and 6σ can be determined.

[Example 1-2]
In Example 1-2, as shown in FIGS. 10A and 10D, the difference between the position measurement result and the threshold value is small, so that an optical unit that is in an abnormal state is identified in one measurement. The case where it is not possible will be described. Here, FIG. 10A shows the position measurement result, and FIG. 10D shows the evaluation result information group generated in step S12 of FIG. 3 based on the position measurement result of FIG. . In the example of FIG. 10D, the absolute value of the difference between Z2 and Z5 does not fall within the threshold value, and is determined to be rejected (x), and otherwise, it is determined to be passed (O).

  When the position measurement result as shown in FIG. 10A is obtained in step S11 of FIG. 3 and the evaluation result information group as shown in FIG. 10D is obtained in step S12, only such evaluation result information group is obtained. In step S15, the optical unit in the abnormal state cannot be specified. Therefore, in step S14, it is determined that remeasurement is necessary, and the process returns to step S11. Instead of providing step S14, the procedure may be changed to return to step S11 when the optical unit in an abnormal state cannot be specified in step S15.

  In Example 1-2, as described above, when the optical unit in the abnormal state cannot be specified only by the evaluation result information group already prepared in step S14, that is, when the number of data is insufficient, the processing is repeated. The process returns to step S11 for measurement. As another example of returning to step S11 for remeasurement, a case where the accuracy of the evaluation result information group is insufficient can be cited. For example, a determination threshold value for determining accuracy is determined in advance, and the measurement accuracy predicted from the position measurement result in FIG. 7A is lower than the determination threshold value. The determination threshold value for determining the accuracy can be determined based on, for example, measurement reproducibility predicted from image brightness and pattern light contrast.

  When returning to step S11 for re-measurement, the test object 8 that has already been measured may be re-measured, or another test object 8 may be measured. Here, description will be made assuming that the test object 8 that has already been measured is further measured multiple times.

  When the position of the test object 8 is measured a plurality of times, the position measurement result can vary within the range of reproducibility. For example, position measurement results as shown in FIGS. 10B and 10C are obtained in addition to the position measurement results shown in FIG. When the position measurement results of FIGS. 10B and 10C are evaluated in step S12, an evaluation result information group as shown in FIGS. 10E and 10F is obtained. In the evaluation result information group of FIGS. 10C, 10E, and 10F, only the position measurement result Z2 is not similar to the other position measurement results Z3, Z4, and Z5, and is determined as rejected. . In other words, only a subset composed of two elements with the position measurement result Z2 as one element is determined as rejected.

  When multiple evaluation result information groups are obtained by measurement and evaluation multiple times, a plurality of evaluation result information groups can be combined, and an optical unit in an abnormal state can be identified based on the combined evaluation result information group . As a method of synthesizing a plurality of evaluation result information groups, for example, when the pass is “1” and the reject is “0”, the corresponding columns in FIGS. 10 (d), (e), and (f) There is a method of obtaining an evaluation result information group synthesized by ANDing values. FIG. 10G shows an example of the evaluation result information group generated by such a method. When the evaluation result information group that can identify the optical unit in the abnormal state is obtained as shown in FIG. 10G, the control unit 9 ends the remeasurement. However, there is a possibility that the evaluation result information group that can identify the optical unit in the abnormal state may not be obtained even if the remeasurement is repeated, so the remeasurement is stopped when the number of remeasurements reaches the preset number. May be.

[Example 1-3]
There may be a case where an image that can obtain a distance image from the optical units 1-2 to 1-5 to the control unit 9 is not provided due to the posture of the test object or the influence of ambient light. In that case, an evaluation result information group cannot be obtained, and data shortage may occur. In such a case, it is not possible to obtain distance images from all optical units with a single test object. Therefore, in Example 1-3, by measuring a plurality of test objects, or by measuring a single test object a plurality of times while changing measurement conditions such as the amount of light, all the optical units are measured. A position measurement result is acquired and an evaluation result information group is acquired based on the position measurement result. Hereinafter, an example in which a plurality of test objects exist as shown in FIG. 11 will be described.

  Hereinafter, description will be made based on the flowchart of FIG. First, step S11 will be described. When the test object is glossy, the amount of diffused light on the surface of the test object decreases, so the specularly reflected light having the same reflection angle as the incident light on the surface has a large amount of light, but the reflected light at other angles Reduces the amount of light. Therefore, when a glossy test object is imaged, an appropriate amount of light may not be acquired by some imaging units. If the amount of light is small, the SN deteriorates, and the reproducibility may not satisfy the required accuracy. For example, when an appropriate amount of light cannot be obtained in the vicinity of the test object in the image captured by the optical unit 1-3, there is a possibility that the position of the test object cannot be measured.

  When there are a plurality of specimens, it is necessary to determine which specimen is the same after recognizing the plurality of specimens for each optical unit. As shown in FIG. 11, when there are four optical units 1-2 to 1-5 as imaging units, by recognizing the test objects 8-1 and 8-2 in each of the four distance images, Up to eight position measurement results can be obtained. Then, the eight position measurement results are classified into a position measurement result for the test object 8-1 and a position measurement result for the test object 8-2.

  For example, the position of the test object 8-1 cannot be measured based on the image captured by the optical unit 1-3, and the position of the test object 8-2 based on the image captured by the optical unit 1-5. Suppose that cannot be measured. In this case, the optical units 1-2 and 1-4 recognize two objects, and only one object can be recognized from the images of the optical units 1-3 and 1-5. The number of test objects can be determined by comparing the test objects respectively recognized by the four optical units 1-2 to 105. In Example 1-3, there are two specimens, and the two specimens are recognized based on the images of the optical units 1-2 and 1-4. Executed over and over.

  In the above example, the case where the position of the test object cannot be recognized when the amount of light is small is described. In addition to this, when the difference between the CAD information of the test object and the distance image is large, the position of the test object may not be recognized even when the contrast of the projected pattern light is not in the proper range.

  Assume that the position measurement result shown in FIG. 12A is obtained in step S11, and the evaluation result information group shown in FIG. 12C is obtained in step S12. In FIG. 12C, the evaluation is performed based on the mutual difference between the position measurement results Z2, Z4, and Z5 and the threshold value T. In FIG. 12C, a blank (“”) indicates that the evaluation was impossible due to the absence of the position measurement result Z3. In step S13, it is determined that measurement of all the specimens has not been performed, and in step S12 executed for the second time, based on the position measurement result shown in FIG. 12B for the second specimen, The evaluation result information group shown in FIG.

  In the next step S14, the control unit 9 generates a combined evaluation result information group based on the evaluation result information group shown in FIGS. 12C and 12D obtained so far. Here, the pass (◯) and blank (“”) are set to “1”, and the reject is set to “0”, and the values of the corresponding columns in FIGS. 12 (c) and 12 (d) are ANDed. The evaluation result information group can be obtained. FIG. 12E shows an example of the evaluation result information group generated by such a method. Hereinafter, in step S15, the control unit 9 identifies an optical unit in an abnormal state based on the synthesized evaluation result information group, and in step S16, the control unit 9 performs post-processing.

  In Examples 1-2 and 1-3, an evaluation result information group is generated in step S12, and two evaluation result information groups are synthesized in step S14. Instead, in step S12, the difference amount between the two position measurement results in the subset is recorded, and in step S14, the difference amount recorded in the two steps S12 is averaged to obtain the averaged difference amount. An evaluation may be performed based on this.

  In Example 1-2, 1-3, the example which produces | generates several evaluation result information groups by the loop respectively comprised by process S13, S14 was demonstrated. The loop configured in step S13 generates an evaluation result information group over a plurality of times when a plurality of test objects are simultaneously recognized by one measurement in step S11. The loop configured in step S14 generates an evaluation result information group based on a plurality of measurements irrespective of the number of objects to be simultaneously recognized by one measurement.

  Through Examples 1-1, 1-2, and 1-3, the method for specifying an optical unit in an abnormal state based on the evaluation result information group has been described. In the case of the active stereo method, as shown in FIG. 13A, when all of the plurality of subsets having a specific position measurement result as a constituent element indicate failure (dissimilarity), The optical unit that provided the position measurement result is in an abnormal state. In other words, in FIG. 13A, the measurement result by one imaging unit of the plurality of imaging units is not similar to the measurement result by another imaging unit, and the measurement results by the other imaging units are similar. The case where the evaluation result information group indicates that the In such a case, the control unit 9 can identify the one imaging unit as an abnormal optical unit.

On the other hand, as illustrated in FIG. 13B, when only a subset of two specific position measurement results indicates failure (dissimilarity), the projection unit is in an abnormal state. In other words, FIG. 13B shows the evaluation result information group that the measurement result by one imaging unit among the plurality of imaging units is not similar only to the measurement result by the other imaging unit. The case is illustrated. In such a case, the control unit 9 can specify the projection unit as an abnormal optical unit.
[Second Embodiment]
Hereinafter, a measuring apparatus 50 according to a second embodiment of the present invention will be described with reference to FIG. Matters not mentioned in the first embodiment can follow the first embodiment. In the second embodiment, the measurement device 50 is configured as a passive stereo measurement device in which all of the plurality of optical units 1-1 to 1-N (N is 4 or more) is configured by an imaging unit. In the passive stereo method, after an image is picked up by the optical units 1-1 to 1-N, the control unit 9 specifies a characteristic portion on the test object in each image. The position measurement by the passive stereo method is based on the premise that the same feature portion exists in the image A and the image B by the optical units 1-1 and 1-2, respectively. Under this premise, the position of the feature (reference position) based on the address (position) of the feature in image A, the address (position) of the feature in image B, and the relative position of optical units 1-1 and 1-2. Distance). In the passive stereo system, N (N−1) / 2 position measurement results can be obtained using N optical units.

  When all of the N optical units are in the normal state, N (N−1) / 2 position measurement results are obtained, and the difference between these position measurement results falls within the threshold T. On the other hand, consider a case where one optical unit is in an abnormal state. In this case, (N−1) (N−2) / 2 position measurement results are obtained by another normal optical unit, and the difference between them falls within the threshold value T, and is determined to be acceptable. . On the other hand, (N−1) position measurement results are obtained using the image of the optical unit in the abnormal state, and the difference between them does not fall within the threshold value T, and it can be determined that the result is unacceptable. Therefore, the control part 9 can identify the optical unit in an abnormal state based on the determination result that it is unacceptable.

[Example 2-1]
Example 2-1 of the second embodiment will be described with reference to FIGS. In Example 2-1, a measuring apparatus 50 having five optical units 1-1 to 1-5 each configured as an imaging unit will be described as shown in FIG. When five optical units 1-1 to 1-5 are used, ten position measurement results are obtained in step S11. Steps S12 to S14 are the same as Examples 1-1, 1-2, and 1-3 of the first embodiment.

  In step S15, the control unit 9 identifies an optical unit in an abnormal state. This will be specifically described. First, it is assumed that the position measurement result as shown in FIG. The horizontal axis represents the position measurement result Z of the test object 8, and “Zmn” represents the position measurement result obtained from the images of the optical units 1-m and 1-n. FIG. 14C shows an evaluation result information group obtained by evaluating the position measurement result of FIG. Referring to FIG. 14C, there is a failure only in the position measurement result obtained by the combination with the optical unit 1-5. Therefore, the control unit 9 can specify that the optical unit 1-5 is abnormal information by template matching between the evaluation result information group in FIG. 14C and a template (not shown).

On the other hand, in the evaluation result information group in FIG. 14D generated based on the position measurement result in FIG. 14B, the position measurement result obtained by the combination with the optical unit 1-4 and the optical unit 1-5. There is a failure in the position measurement result obtained by the combination. Therefore, the control unit 9 cannot specify which of the optical units 1-4 and 1-5 has the abnormal state. Therefore, the control unit 9 determines that re-measurement is necessary in step S14, and measures a plurality of test objects, or changes the measurement conditions such as the amount of light for one test object a plurality of times. By measuring, an evaluation result information group capable of specifying an optical unit having an abnormal state is obtained. Thereafter, the control unit 9 identifies an optical unit in an abnormal state based on the new evaluation result information group.
[Third Embodiment]
In the first and second embodiments, similarity is determined by comparing the difference between a plurality of position measurement results with a threshold value. In the third embodiment, the difference between each position measurement result and an index is set as a threshold value. An evaluation result information group is generated by comparison. The index is, for example, an average value of a plurality of position measurement results. The threshold value can be determined based on, for example, measurement reproducibility.

  In the case of the active stereo method, when the imaging unit is in an abnormal state, only a specific position measurement result has a large difference from an average value of a plurality of position measurement results. For example, in the example of FIG. 7, only the position measurement result Z2 has a difference from the average value larger than the other position measurement results. On the other hand, when the projection unit is in an abnormal state, only the two specific sets of position measurement results have a large difference from the average value. For example, in the example of FIG. 9, only the position measurement results Z2 and Z4 have a difference from the average value larger than the other position measurement results.

  As described above, when the imaging unit is in the abnormal state, the specific position measurement result does not fall within the threshold value, and when the projection unit is in the abnormal state, only the specific two sets of position measurement results do not fall within the threshold value. Probability is high. Therefore, an evaluation result information group can be generated using this, and an optical unit in an abnormal state can be specified based on the evaluation result information group.

In the second embodiment, a plurality of evaluation result information groups may be combined after data accumulation, but the difference amounts may be averaged after the difference amounts are accumulated. First, measurement is performed a plurality of times, or measurement is performed on a plurality of test objects, and a difference amount between the obtained position measurement result and the index is accumulated. And the value which averaged the accumulated some difference amount for every position measurement result increases the precision of difference amount by the averaging effect. For example, it is assumed that the variation of the multiple measurement of the difference amount Z in FIG. At this time, if the results of the number of measurements N are averaged, the range of the probability of existence of the difference amount Z is narrowed by (N) ^−1/2 of the range 31 and the accuracy of the difference amount that can be acquired is increased, and the evaluation result information with high accuracy You can get a group.

  In the above, an average value of a plurality of position measurement results is used as an index. However, in step S11 of FIG. 3, when the reliability of the output of a certain optical unit is low and the position measurement result cannot be obtained based on this output, the average value changes greatly. In that case, data may be accumulated by the following method. First, measurement conditions such as the amount of light are determined based on a specific reference imaging unit, and the reliability of the measurement results obtained by this reference imaging unit is always ensured. And the method of accumulating the difference with other measurement results can be considered using this measurement result as the first index. However, since there is a possibility that the reference imaging unit is shifted, before the evaluation, the average value of the accumulated total difference amounts is set as the second index, and the difference between the second index and the total difference amount is set as the second difference amount. It is desirable to determine the similarity from the second difference amount.

[Supplement to the first to third embodiments]
In the present invention, four or more optical units can be used as an optical unit such as an imaging unit and a projection unit. When there are three optical units, the number of output results of the measuring device is two in the active stereo method. Therefore, when the two values are different, it is impossible to determine which is correct. In the case of the passive stereo method, the number of output results of the measurement apparatus is three, but two of the three output results are calculated based on the output of one optical unit. Therefore, if there is even one abnormal optical unit, the values of the three output results may all be different, and it is impossible to determine which one is correct. Therefore, four or more optical units are required.

[Fourth Embodiment]
The above-described measuring device 50 can be used while being supported by a certain support member. In the present embodiment, as an example, a system used by being installed in a robot arm 300 (gripping device) as shown in FIG. 17 and an article manufacturing method using the system will be described. The measuring apparatus 100 projects and images pattern light on the test object 210 placed on the support base 350, and acquires an image. Then, the control unit of the measurement apparatus 100 or the control unit 310 that has acquired the image data from the control unit of the measurement apparatus 100 obtains the position and orientation of the test object 210 and controls the obtained position and orientation information. Acquired by the unit 310. Based on the position and orientation information, control unit 310 sends a drive command to robot arm 300 to control robot arm 300. The robot arm 300 holds the test object 210 with a robot hand or the like (gripping part) at the tip, and moves it such as translation or rotation. Further, by assembling the test object 210 to other parts by the robot arm 300, an article composed of a plurality of parts, such as an electronic circuit board or a machine, can be manufactured. Further, an article can be manufactured by processing the moved test object 210. The control unit 310 includes an arithmetic device such as a CPU and a storage device such as a memory. Note that a control unit for controlling the robot may be provided outside the control unit 310. In addition, measurement data measured by the measurement apparatus 100 and an obtained image may be displayed on the display unit 320 such as a display.

  As mentioned above, although preferable embodiment of this invention was described, this invention is not limited to these embodiment, A various deformation | transformation and change are possible within the range of the summary.

50: Measuring device, 1-1 to 1-5: Optical unit, 8: Test object, 9: Control unit

Claims (15)

A measuring device that measures the position of an object using a plurality of optical units for measuring distance based on triangulation,
Based on an evaluation result information group including a plurality of evaluation results obtained by evaluating a plurality of measurement results related to the position of the object measured using the plurality of optical units, the abnormal among the plurality of optical units. A control unit that identifies the optical unit;
A measuring device characterized by that. The evaluation result information group is a set of evaluation result information determined according to the difference between the plurality of measurement results.
The measuring apparatus according to claim 1. The evaluation result information group is a set of evaluation result information determined according to a difference between two measurement results that respectively constitute a plurality of subsets obtained from the set of the plurality of measurement results.
The measuring apparatus according to claim 2. The evaluation result information group is a set of evaluation result information determined according to the difference between each of the plurality of measurement results and the index.
The measuring apparatus according to claim 1. The evaluation result information group is a set of evaluation result information determined by comparing the difference with a threshold value.
The measuring apparatus according to any one of claims 2 to 4, wherein the measuring apparatus is characterized in that The threshold is determined based on a plurality of measurement results measured using the plurality of optical units.
The measuring apparatus according to claim 5. The plurality of optical units includes a projection unit that projects pattern light onto the object, and a plurality of imaging units.
When the evaluation result information group indicates that the measurement result by one imaging unit of the plurality of imaging units and only the measurement result by another one imaging unit are not similar, Identifying the projection unit as the abnormal optical unit;
The measuring apparatus according to any one of claims 1 to 6, wherein The plurality of optical units includes a projection unit that projects pattern light onto the object, and a plurality of imaging units.
The control unit is configured such that the measurement result by one imaging unit of the plurality of imaging units is not similar to the measurement result by another imaging unit, and the measurement results by the other imaging units are similar to each other. When the evaluation result information group indicates, the one imaging unit is specified as the abnormal optical unit.
The measuring apparatus according to any one of claims 1 to 6, wherein The measurement by the plurality of optical units is performed a plurality of times,
The control unit identifies the abnormal optical unit based on the evaluation result information group obtained by the measurement over the plurality of times.
The measuring apparatus according to claim 1, wherein the measuring apparatus is one of the following. Measurements by the plurality of optical units are performed on a plurality of objects,
The control unit identifies the abnormal optical unit based on the evaluation result information group obtained by measuring the plurality of objects.
The measuring apparatus according to claim 1, wherein the measuring apparatus is one of the following. The control unit determines the threshold based on a plurality of the differences.
The measuring apparatus according to claim 5. The control unit determines the threshold based on an average value of a plurality of the differences.
The measuring apparatus according to claim 5. The control unit determines the index based on the plurality of measurement results.
The measuring apparatus according to claim 4. The measurement apparatus according to any one of claims 1 to 12, which measures the position of a test object;
A robot that holds and moves the test object based on the measurement result of the measurement device;
The system characterized by having. A step of measuring the position of the object using the measuring device according to claim 1;
A step of manufacturing the article by processing the object based on a measurement result by the measurement device;
A method for producing an article, comprising:

Images