JP2012053004A - Three-dimensional-point group synthesis method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a three-dimensional-point group synthesis method capable of automatically synthesizing three-dimensional-point groups obtained by measuring a measuring object in plural directions.SOLUTION: A three-dimensional-point group synthesis method includes: a three-dimensional (3D) measurement step (S2) of measuring plural measuring targets disposed nearby or on a surface of a measuring object for identifying a solid body by using 3D measurement means to obtain them as a 3D-point group; a solid-body point-group discrimination step (S3) of extracting target point groups applicable to measuring targets from the 3D-point group to discriminate which measuring target each of the extracted target point groups is applicable to; a step of repeating the 3D measurement step and solid-body point-group discrimination step in plural directions with positions of the 3D measurement means changed; and a point-group synthesis step (S4) of synthesizing target point groups indicating the same property among target point groups discriminated in one direction and other target point groups discriminated in other directions in the solid-body point-group discrimination step.

Description

  The present invention relates to a method for synthesizing a three-dimensional point group and an object measurement apparatus, and more particularly to a method for synthesizing a three-dimensional point group of a measurement object such as a structure acquired from a plurality of directions using a three-dimensional measurement unit. Is.

Conventionally, when measuring a three-dimensional shape of a measurement object such as a structure, a measurement target is arranged on the measurement object and its vicinity or surface, and the measurement object and the measurement target are imaged using a camera or the like. A measurement method for creating a three-dimensional shape based on the calculated three-dimensional coordinates of each measurement target is known.
As a specific method of creating a 3D shape of a measurement object, the measurement object, target, and reference target are imaged from multiple directions using a 3D shape camera, and the center coordinates of the target and reference target are calculated. In addition, there is known a method for creating a three-dimensional shape of the entire measurement object by superimposing the centers of targets or reference targets that are commonly displayed in a plurality of images (Patent Document 1).
In this way, the measurement target in the image can be recognized more accurately by using the measurement target.

JP 2007-64668 A

On the other hand, a method is also known in which a measurement target and a measurement target are measured using a laser scanner or the like, and shape data of the measurement target and the measurement target is acquired as a three-dimensional point group.
Even in such a case, when creating the three-dimensional shape of the entire measurement object, it is necessary to measure the measurement object from a plurality of directions and synthesize the acquired point cloud.
In this regard, the above-described conventional technique performs composition on image data captured from a plurality of directions, and does not perform composition on a three-dimensional point group obtained by measurement from a plurality of directions. .

When combining each point cloud acquired by measuring a measurement object from multiple directions, place a plurality of measurement targets near or on the surface of the measurement object, and measure and acquire the 3D point cloud From the above, the operator assigns a number to each point cloud corresponding to each measurement target, performs the same processing for each point cloud obtained by measuring from multiple directions, and stacks the point clouds with the same number. It is necessary to perform a process of combining them. However, this method has problems that the worker's work efficiency is poor and that human error cannot be prevented due to manual work by the worker, which can solve these problems. It has been a conventional problem.
The present invention has been made to solve the above-described problems, and an object of the present invention is to automatically synthesize a three-dimensional point group obtained by measuring a measurement object from a plurality of directions. Another object of the present invention is to provide a method for synthesizing a three-dimensional point group.

  In order to achieve the above object, the method for synthesizing a three-dimensional point group according to claim 1 is characterized in that each property data of a plurality of measurement targets for identifying individuals arranged in the vicinity of or on the surface of the measurement object is three-dimensional measurement means. A three-dimensional measurement step to measure and acquire as a three-dimensional point cloud, and extract a target point cloud corresponding to the measurement target from the point cloud, and the target point cloud to which of the measurement targets An individual point group determining step for determining whether it corresponds, and a step of changing the position of the three-dimensional measuring means and repeating the three-dimensional measuring step and the individual point group determining step from a plurality of directions, and then the individual point group Each target point group discriminated in one direction by the discriminating step and each target point group discriminating in the other direction are overlapped with each other. Further comprising a point group synthesizing step of synthesizing Te characterized.

The method for synthesizing a three-dimensional point group according to claim 2 further comprises a shape modeling step for creating a three-dimensional shape of the measurement object from the synthesized point group after the point group synthesizing step. It is characterized by.
In the method for synthesizing the three-dimensional point group according to claim 3, in claim 1 or 2, the plurality of measurement targets are spheres having different radii, and the same property is the same radius. Features.

  According to a method for synthesizing a three-dimensional point group according to claim 4, in the first or second aspect, each of the plurality of measurement targets has a luminance in a different luminance band that can identify the measurement target in the individual point group determination step. The same property is a luminance of the same luminance band.

The method for synthesizing a three-dimensional point group according to claim 5 is characterized in that, in claim 4, a plurality of the luminance bands are selected from a range of 10 to 220.
The method for synthesizing a three-dimensional point group according to claim 6 is characterized in that, in the three-dimensional measurement step, the measurement target is measured using a monochrome three-dimensional measurement means.

  According to the method of synthesizing a three-dimensional point group according to the present invention, a three-dimensional point group is obtained by using a three-dimensional measuring unit to obtain a plurality of measurement targets for identifying individuals arranged near or on the surface of the measurement object. A measurement step, a target point cloud corresponding to the measurement target from the point cloud, an individual point cloud determination step for determining which measurement target the extracted target point cloud corresponds to, and a three-dimensional measurement means A target point showing the same property in each target point group in one direction and each target point group determined in another direction by repeating the three-dimensional measurement step and the solid point group determination step from a plurality of directions by changing the position. A point group synthesizing step for superimposing and synthesizing the groups.

  Therefore, by determining the target point cloud corresponding to the measurement target for identifying the individual from the point cloud acquired from one direction, among the target point cloud included in the point cloud measured and acquired from a plurality of directions, Since target point groups having the same properties are searched for and automatically synthesized, it is possible to prevent human error from occurring and improve work efficiency.

It is a schematic block diagram containing the three-dimensional measuring device used for embodiment of this invention. It is a flowchart which shows the synthetic | combination method of the three-dimensional point group of this invention. (A) is a top view of FIG. 1 in the first embodiment of the present invention, (B) is a top view when the position of the three-dimensional measuring means of FIG. 1 is changed, and (C) is (A), (B). It is a top view of the model created by synthesizing the measurement results. In 1st Embodiment, it is a schematic block diagram for investigating the distance of a three-dimensional measuring means and the measurement target, and the precision of the calculated sphere. (A) is a graph showing the measurement result of the measurement target having a radius of 72.5 mm and the distance, and (B) is a graph showing the measurement result of the measurement target having a radius of 19.0 mm and the distance. is there. (A) is a top view of FIG. 1 in the second embodiment of the present invention, (B) is a top view when the position of the three-dimensional measuring means of FIG. 1 is changed, and (C) is (A), (B). It is a top view of the model created by synthesizing the measurement results. In 2nd Embodiment, it is a schematic block diagram for investigating the luminance information of a color sample using a three-dimensional measuring means. (A) is a color sample that shows four levels of luminance bands that can be used for a measurement target-a luminance sample, and (B) is a color sample that shows five levels of luminance bands that can be used for a measurement target- It is a graph showing the relationship of luminance.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a schematic configuration diagram including an object recognition device 1 used in an embodiment of the present invention.
As shown in FIG. 1, the object recognition device 1 is a device that expresses the shape of a measurement object (hereinafter referred to as a workpiece) 2 in a virtual coordinate space. The target object recognition apparatus 1 arranges measurement targets 4 at a plurality of predetermined positions in the vicinity of the workpiece 2 and measures the measurement target 4 supported by the workpiece 2 and the support 6 to obtain shape data (property data). Calculated by the three-dimensional measuring means 8, the calculating means 10 for calculating and synthesizing the workpiece 2 and the measurement target 4 in the virtual coordinate space from the shape data acquired by the three-dimensional measuring means 8, and the calculating means 10. A memory device 12 for storing the synthesized result and the like and an external output means 14 for outputting the synthesized result and the like are provided. The measurement target 4 may be disposed on the surface of the workpiece 2 or a surrounding structure.

The workpiece 2 is, for example, a ship block, but is not limited to this as long as the workpiece 2 maintains a three-dimensional shape.
The measurement target 4 is used when the measured three-dimensional shape of the workpiece 2 is formed in a three-dimensional virtual coordinate space. As shown in FIG. 1, the shape of the measurement target 4 is a sphere.

  In the present embodiment, a laser scanner is used as the three-dimensional measuring unit 8, and the laser scanner functions as a laser distance meter. From the position of the three-dimensional measuring unit 8 by emitting laser light and receiving the reflected laser. By measuring the distance to each measurement point on the workpiece 2 and the measurement target 4 and measuring the direction in which the laser beam is emitted at the same time to obtain a three-dimensional point group, the workpiece 2 and the measurement target 4 are obtained. Get shape data.

The computing means 10 and the memory device 12 are provided in a computer (not shown). The calculation means 10 is constituted by a central processing unit of the computer, and the memory device 12 is constituted by including RAM, ROM and the like.
The external output unit 14 outputs a result of synthesis by serial communication, LAN network communication, screen output, external memory writing, file output, or the like.

Hereinafter, a method for synthesizing a three-dimensional point group of the object recognition apparatus 1 according to the first embodiment of the present invention configured as described above will be described.
FIG. 2 shows a flowchart of a method for synthesizing a three-dimensional point group. Each process from step S2 to S6 described below is performed by executing each program stored in the memory device 12 by the calculation means 10.

  In step S <b> 1, a plurality of measurement targets 4 are installed around the work 2. Specifically, as shown in the top view of FIG. 1 in FIG. 3A, measurement targets 4a, 4b, 4c, and 4d having different radii (properties) are arranged. The measurement targets 4a to 4d are spheres of a size that can be identified by the three-dimensional measurement means 8. The sizes of the measurement targets 4a to 4d used here are determined depending on the type and wavelength of the three-dimensional measurement means 8 to be used, or the measurement environment.

  In step S2, the shape data of the workpiece 2 and the measurement targets 4a to 4d are acquired as a three-dimensional point group using the three-dimensional measurement means 8 (three-dimensional measurement step). For example, as shown in FIG. 3A, the measurement is performed by arranging the three-dimensional measuring means 8 on the measurement targets 4c and 4d side. Here, the viewpoint of the three-dimensional measuring apparatus 8 is on the workpiece 2 side. Note that at the time of measurement, at least three of the measurement targets 4 a to 4 d need to be measured by the three-dimensional measurement means 8.

In step S3, the point group constituting the sphere is determined as the target point group from the measurement point group acquired in step S2, and the center coordinates and radius of the target point group are calculated to determine the measurement targets 4a to 4d. Which one corresponds is determined (individual point group determination step).
When step S3 is completed and the data of the workpiece 2 from a plurality of directions necessary for point group synthesis described later has not been acquired, the process returns to step S2. At this time, as shown in FIG. 3B, for example, after the three-dimensional measuring device 8 is moved to the measurement targets 4b and 4d, step S2 is executed.
On the other hand, if the data necessary for the point cloud synthesis is acquired, the process proceeds to step S4.

In step S4, the point cloud acquired in step S2 is synthesized (point cloud synthesis step).
For example, as shown in FIG. 3C, a top view of a model created by combining the point clouds obtained by measuring FIGS. 3A and 3B, each direction obtained by repeating step S2 above. The point groups of the workpiece 2 and the measurement targets 4a to 4d are combined so that the target point groups having the same radius overlap according to the radius of each target point group determined in step S3.

In step S5, the three-dimensional coordinates of the point group constituting the work 2 are calculated from the point group synthesized in step S4 to perform shape modeling, and a model of the work 2 is created (shape modeling step).
In step S6, the dimension measurement of the model of the workpiece 2 created by the shape modeling and the shape evaluation of the model are performed.

Here, by the method described above, the measurement target 4 is measured by changing the distance L to the measurement target 4 from 2 m to 32 m using the three-dimensional measuring means 8 as shown in the schematic configuration diagram of FIG. The accuracy of the radius of the sphere calculated from the target point group corresponding to the measurement target 4 was investigated. Note that the wavelength of the three-dimensional measuring means 8 used is around 650 nm.
The investigation results are shown in FIGS. 5 (A) and 5 (B). Here, FIG. 5A is a graph when a sphere with a radius of 72.5 mm is used as the measurement target 4, and FIG. 5B is a case when a sphere with a radius of 19.0 mm is used as the measurement target 4. It is a graph of.

From FIG. 5A, the number of points that can be acquired as the distance L increases decreases in proportion to L ^−1.8, and the number of points that can be acquired at 32 m is about 70 points. However, although the number of points that can be acquired decreases as the distance L increases, the calculated variation in the radius of the measurement target 4 is small, and the radius can be calculated with high accuracy.
The same can be said for FIG. 5B as described above, but since the number of points acquired at 32 m was about 5, the sphere could not be determined from the point group and the radius could not be calculated. It was.

  Therefore, the sphere used for the measurement target 4 is preferably large. For example, it is preferable to use a sphere having a conventionally used radius larger than 77 mm. In addition, since the accuracy varies depending on the wavelength of the three-dimensional measuring unit 8, the method of adding the difference in radius between the measurement targets 4a to 4d is determined by the wavelength of the three-dimensional measuring unit 8 to be used.

  As described above, according to the present embodiment, the workpiece 2 and the measurement targets 4a to 4d using the three-dimensional measurement means 8 with respect to the measurement targets 4a to 4d having different radii installed around the workpiece 2 are used. Shape data is acquired as a three-dimensional point group, and the point group constituting the sphere corresponding to the measurement targets 4a to 4d is determined as the target point group from the point group, and the center coordinates and the radius of each sphere are determined. calculate. Then, after changing the position of the three-dimensional measuring means 8 and repeating the measurement and discrimination of the workpiece 2 and the measurement targets 4a to 4d from a plurality of directions, the target point group having the same radius in each point group acquired from each direction Are combined so that they overlap.

Thereby, by using a sphere having a radius that can be individually identified as the measurement target 4, the calculation means 10 can measure the measurement targets 4 a to 4 d from the three-dimensional point group obtained by using the three-dimensional measurement means 8. The target point cloud corresponding to each can be automatically discriminated and the point cloud acquired from each direction is synthesized based on the target point cloud of the same radius, so that human error can be prevented from being mixed. , Work efficiency can be improved.
Furthermore, since the shape modeling of the workpiece 2 is performed from the automatically synthesized point group, the reliability of the created model of the workpiece 2 can be further improved.

Next, a second embodiment of the present invention will be described.
The method of synthesizing the three-dimensional point group of this embodiment is different from the first embodiment in that the radii of the measurement targets 4a to 4d are substantially the same and the spheres have different hues. The configuration is the same. Therefore, the description of the common part is omitted, and the difference is described.

FIGS. 6A and 6B are top views of FIG. 1 in the second embodiment of the present invention. This will be described below together with a flowchart showing the method for synthesizing the three-dimensional point group in FIG. 2 described above.
In step S <b> 1, measurement targets 4 a ′ to 4 d ′ having different hues and thus different luminances (properties) are arranged around the work 2. The measurement targets 4a ′ to 4d ′ used here are determined by the type and wavelength of the three-dimensional measurement means 8 to be used or the reflection of the workpiece 2 to be measured.

In step S2, the shape data of the workpiece 2 and the measurement targets 4a ′ to 4d ′ are acquired as a three-dimensional point group by using the three-dimensional measuring unit 8.
In step S3, the luminance information of the measurement targets 4a ′ to 4d ′ is acquired from the point cloud acquired using the three-dimensional measurement unit 8, and it is determined which of the measurement targets 4a ′ to 4d ′ corresponds. (Individual point group discrimination step). Here, since the three-dimensional measuring unit 8 is a laser scanner, luminance information is acquired by processing the acquired point group by the calculating unit 10.

When step S3 is completed and data of the workpiece 2 from a plurality of directions necessary for the point group synthesis has not been acquired, the process returns to step S2. At this time, as shown in FIG. 6B, for example, step S2 is executed after the three-dimensional measuring device 8 is moved to the measurement targets 4b and 4d.
On the other hand, if the data necessary for the point cloud synthesis is acquired, the process proceeds to step S4.
In step S4, each point cloud obtained by measuring the workpiece 2 and the measurement targets 4a ′ to 4d ′ from a plurality of directions is obtained as a target point cloud having the same luminance band as shown in FIG. 6C. The overlapping is performed (point group combining step).

  Here, as shown in the schematic configuration diagram of FIG. 7, the color sample 4 e was measured by the three-dimensional measuring unit 8 and the luminance band that can be used as the measurement target 4 was investigated by the method described above. Here, the difference can be easily recognized by using the value of color information (R / G / B). However, since the data capacity is large and the point cloud compositing process requires a lot of time, the monochrome laser scanner uses it. The brightness information was investigated by measuring. Note that the wavelength of the three-dimensional measuring unit 8 used is near 650 nm, and the distance between the three-dimensional measuring unit 8 and the color sample 4e is constant.

  As usable luminance bands, graphs showing the relationship between the color sample and the luminance are shown in FIGS. A color chip from PANTONE was used as a color sample. Each color code is shown in FIGS. 8A and 8B, and the average value represents the average value of luminance in each color code. As shown in FIG. 8A, the luminance bands of 10 to 210 are divided into four stages, that is, the respective luminance bands divided into 10 to 60, 61 to 110, 111 to 160, and 161 to 210, or FIG. As shown, it was possible to discriminate by dividing the luminance band of 20 to 220 into five levels, that is, each luminance band divided into 20 to 60, 61 to 100, 101 to 140, 141 to 180, and 181 to 220. . However, if the luminance band in the above-described range is further increased, there is a possibility that the determination cannot be made.

Here, when the wavelength of the three-dimensional measuring unit 8 changes, it is assumed that the same hue results in different luminances. Therefore, the method of adding the luminance difference of the measurement target 4 is determined by the wavelength of the three-dimensional measuring unit 8 to be used.
Furthermore, since the luminance varies depending on the reflectance of the measurement target 4, it is necessary to use a material that does not reflect sunset or the like for the measurement target 4.

  As described above, in this embodiment, the measurement targets 4a ′ to 4d ′ are obtained from the three-dimensional point group measured and acquired by the three-dimensional measurement unit 8 using the measurement targets 4a ′ to 4d ′ having different luminances. Point groups obtained by measuring the workpiece 2 and the measurement targets 4a ′ to 4d ′ from a plurality of directions based on the luminance information acquired from the measurement targets 4a ′ to 4d ′ for the target point group corresponding to Are combined so that target point groups having the same luminance band overlap.

Thereby, since the target point group is discriminated based on the luminance information of the measurement targets 4a ′ to 4d ′ and the three-dimensional point group is synthesized, the same effect as in the first embodiment can be obtained.
In addition, since the luminance used for the measurement targets 4a ′ to 4d ′ is selected from the luminance bands in the range of 10 to 220, each luminance band can be reliably determined, so that the target point group is automatically set. Thus, the working efficiency can be improved.

  Furthermore, by processing the point cloud of the measurement targets 4a ′ to 4d ′ measured by the monochrome laser scanner and acquiring the luminance information, the data capacity is smaller than when acquiring data as color information, and the point cloud is reduced. Since the process of synthesizing becomes faster, the working efficiency can be improved.

In addition, this invention is not limited to embodiment mentioned above.
For example, in the first embodiment, the measurement target 4 having a different radius is used, and in the second embodiment, the target point group is determined using the measurement target 4 having a different luminance. You may comprise combining different brightness | luminance. Thereby, the discrimination accuracy of the target point group can be further improved.
Moreover, although each said embodiment demonstrated the workpiece | work 2 as a ship block, the workpiece | work 2 is not restricted to this, It is applicable also to a bridge and a building.

1 Object Recognition Device 2 Workpiece (Measurement Object)
4, 4a, 4b, 4c, 4d, 4a ', 4b', 4c ', 4d' Measuring target 8 Three-dimensional measuring means 10 Arithmetic means 12 Memory device

Claims (6)

A three-dimensional measurement step of measuring each property data of a plurality of measurement targets to be identified that are arranged near or on the surface of the measurement object using a three-dimensional measurement means and obtaining them as a three-dimensional point group;
An individual point group determination step of extracting a target point group corresponding to the measurement target from the point group and determining to which measurement target the target point group corresponds;
After changing the position of the three-dimensional measurement means and repeating the three-dimensional measurement step and the individual point group determination step from a plurality of directions, each target point group determined in one direction by the individual point group determination step and A point group synthesis step for superposing and synthesizing target point groups showing the same properties among the target point groups determined in the other direction,
A method for synthesizing a three-dimensional point group.   The method for synthesizing a three-dimensional point group according to claim 1, further comprising a shape modeling step for creating a three-dimensional shape of the measurement object from the synthesized point group after the point group synthesizing step. .   3. The method for synthesizing a three-dimensional point group according to claim 1, wherein the plurality of measurement targets are spheres having different radii, and the same property is the same radius.   The plurality of measurement targets are spheres each having a luminance in a different luminance band capable of determining the measurement target in the individual point group determination step, and the same property is a luminance in the same luminance band. The method for synthesizing a three-dimensional point group according to claim 1 or 2.   The method for synthesizing a three-dimensional point group according to claim 4, wherein a plurality of the luminance bands are selected from a range of 10 to 220.   6. The method for synthesizing a three-dimensional point group according to claim 4, wherein, in the three-dimensional measurement step, the measurement target is measured using a monochrome three-dimensional measurement means.