WO2022249617A1 - Three-dimensional model generation support system, program, and recording medium - Google Patents

Abstract

The present invention uses a three-dimensional model generation support system which includes: a partial asset CAD model generation device for generating a partial asset CAD model which has attribute data by using the data from a perspective view which depicts part of the asset; and a three-dimensional pattern matching device for generating an attribute-equipped asset CAD model by identifying the position of the partial asset CAD model using pattern matching between the partial asset CAD model and a three-dimensional image of the asset, and positioning the partial asset CAD model in three-dimensional space. As a result, it is possible to reduce user work and efficiently generate an attribute-equipped three-dimensional model.

Description

3D model creation support system, program and recording medium

The present invention relates to a three-dimensional model creation support system, program, and recording medium.

In recent years, power plants and chemical plants use 3D-CAD (Three-Dimensional Computer-Aided Design) to manage assets.

On the other hand, some of these plants were built a long time ago, do not have 3D-CAD models, and use paper documents for asset management.

Therefore, in recent years, services that acquire 3D point cloud data by laser measurement and create 3D models of plants using that data have become commonplace. Note that, hereinafter, three dimensions may be referred to as "3D" and two dimensions may be referred to as "2D".

With current laser measurement equipment, it is possible to acquire highly accurate 3D point cloud data. A device has also been developed that can acquire color information using a camera at the same time as acquiring coordinate data. Therefore, colored 3D point cloud data can be displayed. Therefore, the human eye can easily identify the pipes, equipment, and the like.

However, those point cloud data are not 3D models with attributes. Therefore, in order to use the tag information of equipment and piping for asset management, it is necessary to convert the point cloud data into a 3D model with attributes.

　In general, the conversion work to such a 3D model is performed by an operator manually converting it into a 3D model while displaying the point cloud data on the screen. Therefore, the conversion requires a great deal of effort.

On the other hand, in recent years, there is a technology that recognizes the shape of piping from point cloud data using software and converts it into a 3D model without attributes. Also, a technique has been developed to automatically assign attributes to 3D models that do not have attributes.

For example, in Patent Document 1, end point connection information of logical connection data, which is a logical connection relationship between plant equipment, is compared with end point connection information of geometric shape data, which is figure shape information in a three-dimensional space. A layout design support apparatus is disclosed that automatically generates a logical connection/geometric shape correspondence table composed of correspondence relationship data in which logical connection data and geometric shape data having matching end point connection information are associated with each other.

Japanese Patent No. 4940267

　Conventional technology that converts point cloud data into a 3D model only targets relatively simple shapes such as piping and steel materials.

Also, in Patent Document 1, it is difficult to convert all devices such as valves into 3D models. As for piping, it is difficult to convert all piping into a 3D model accurately, and it is necessary to correct the created piping model and additionally create a piping model that could not be automatically created.

The purpose of the present invention is to reduce the user's work and efficiently create a 3D model with attributes.

A three-dimensional model creation support system of the present invention includes a partial asset CAD model creation device for creating a partial asset CAD model having attribute data using data of a perspective view showing a part of an asset, and a partial asset CAD model and an asset. a three-dimensional pattern matching device that identifies the position of the partial asset CAD model by pattern matching with the three-dimensional image and arranges the partial asset CAD model in a three-dimensional space to create an asset CAD model with attributes. .

According to the present invention, it is possible to reduce the user's work and efficiently create a 3D model with attributes.

1 is a configuration diagram showing an example of a three-dimensional model creation support system of Example 1; FIG. FIG. 2 is an isometric view showing an example of piping in a plant; It is a flow chart showing a method of creating a piping model and a valve model. It is a figure which shows the example of the line segment of the extracted piping. 1 is a perspective view showing a partial asset CAD model; FIG. It is a table|surface which shows an example of the attribute database regarding piping. FIG. 4 is a diagram showing an example of three-dimensional image data of the entire asset; FIG. 3 is a flow diagram illustrating a method of 3D pattern matching; FIG. 10 is a perspective view showing an example of a partial pipe model in which virtual outer diameters are set; FIG. 4 is a perspective view showing an example of an attributed asset CAD model; FIG. 11 is a configuration diagram showing an example of a three-dimensional model creation support system of Example 2; FIG. 3 is a diagram showing an example of creating a 3D image from multiple 2D images; FIG. 4 is a perspective view showing a 3D image recognized as a bulb; FIG. 10 is a perspective view showing the result of placing the valve and instrument; FIG. 10 is an isometric view showing an example in which information on a connection destination of piping is added; It is a figure which shows the point cloud corresponding to the cross-sectional shape of a piping model.

A preferred embodiment of the present disclosure will be described.

It is desirable that the 3D model creation support system further include an attribute database having attribute data and asset part control numbers. Then, the partial asset CAD model creation device can acquire the management number displayed in the perspective view, acquire the attribute data associated with the part management number from the attribute database, and add it to the partial asset CAD model. desirable.

It is desirable that the partial asset CAD model creation device acquire dimensions from the dimension lines displayed in the perspective view and create a partial asset CAD model according to the dimensions.

The three-dimensional pattern matching device sets a virtual outer diameter of the partial asset CAD model, identifies a position where the number of point clouds contained inside the partial asset CAD model having the virtual outer diameter is maximum, and identifies It is desirable to place the partial asset CAD model at the location.

The three-dimensional pattern matching device sets virtual cross-sectional dimensions of a predetermined portion of the partial asset CAD model, divides the partial asset CAD model having the cross-sectional dimensions into areas, and divides the partial asset CAD model having the cross-sectional dimensions into areas. It is desirable to identify the location that maximizes the number of point clouds contained within , and place the partial asset CAD model at the identified location.

The three-dimensional pattern matching device sets virtual cross-sectional dimensions of a predetermined portion of the partial asset CAD model, and minimizes the error in the distance between the contour of the cross-section of the partial asset CAD model having the cross-sectional dimensions and the point cloud. It is desirable to identify locations and place partial asset CAD models at the identified locations.

A three-dimensional model creation support system creates a partial asset CAD model having attribute data using data of a perspective view showing a part of an asset including piping and at least one of valves and equipment. Asset CAD model creation device, valve/equipment recognition device that acquires position information of valves and equipment from asset 2D image or 3D image data, and pattern matching between partial asset CAD model and asset 3D image a three-dimensional pattern matching device for creating an attributed asset CAD model by locating the partial asset CAD model and placing the partial asset CAD model in three-dimensional space. Then, the three-dimensional pattern matching device limits the range of pattern matching from the valve and device position information obtained by the valve/device recognition device and the valve and device relative position information obtained from the partial asset CAD model.

Examples will be described below with reference to the drawings.

FIG. 1 is a configuration diagram showing an example of the three-dimensional model creation support system of the first embodiment.

As shown in this figure, the three-dimensional model creation support system includes a partial asset CAD model creation device 2 and a 3D asset pattern matching device 6 (three-dimensional pattern matching device). The asset isometric drawing data 1 is used for predetermined processing in the partial asset CAD model creation device 2 . The partial asset CAD model 3 is created by the partial asset CAD model creation device 2 and used for predetermined processing by the 3D asset pattern matching device 6 . The 3D asset pattern matching device 6 creates an attributed asset CAD model 7 using the partial asset CAD model 3 and 3D image data 5 (three-dimensional image data). The attributed asset CAD model 7 is linked with the attribute database 4 .

In the following, image data may be simply referred to as "image". The term "recognition" refers to identifying pipes, valves, equipment, etc. using numerical data such as two-dimensional or three-dimensional points, lines, and figures.

All of the above devices that constitute the three-dimensional model creation support system may be built in one computer device, or each device may be built in a separate computer device. Each of the above processes is performed by transferring a program recorded in a recording medium such as one or more memories built in the computer device to one or more central processing units built in the computer device. (CPU: Central Processing Unit) performs arithmetic processing. Also, the program may be recorded on a recording medium such as a memory stick that is detachable from the computer device.
The program is for causing the computer device to execute a predetermined process. The recording medium is a computer-readable medium recording the program.

The asset isometric drawing data 1 is input and recorded in one or more recording media such as memories built into the computer device. The partial asset CAD model 3 created by predetermined processing from the asset isometric drawing data 1, the attributed asset CAD model 7 finally created, etc. are stored in one or more memories built in the computer device. recorded on the recording medium. Note that the input to the recording medium may be performed via communication means such as the Internet.

The asset isometric drawing data 1 is the data of the isometric drawing of the asset, and is data capable of displaying a three-dimensional representation of the asset when viewed obliquely from above. Below, the isometric drawing is abbreviated as "Isometric drawing". An isometric view may be an isometric view. Also, the isometric view may be collectively referred to as a "perspective view".

The partial asset CAD model creation device 2 creates a 3D model from the assets included in the asset isometric drawing data 1. Normally, only a part of the asset is recorded like a piping model, so the partial asset is converted to CAD. A detailed CAD method will be described later. In addition, since the asset isometric drawing data 1 usually includes the control number of the target part, it is possible to assign the control number to the CAD model. In the example of piping, piping numbers are given as management numbers.

The partial asset CAD model 3 is a model created by the partial asset CAD model creation device 2. In the piping example, it is a 3D-CAD model of a series of piping having a certain piping number. Further, the partial asset CAD model 3 is given a pipe number as attribute data.

The attribute data of each part of the asset is recorded in the attribute database 4 together with the management number of the part. In the case of piping, there is a piping number as a management number, and attribute information (attribute data) such as the outer diameter, wall thickness, and material of the piping is recorded in association with the piping number. Therefore, a 3D-CAD model having pipe numbers can acquire all related attributes by linking with the attribute database 4 .

The 3D image data 5 is three-dimensional image data of the entire asset. For example, it can be created from multiple images taken from a drone.

The 3D asset pattern matching device 6 matches the partial asset CAD model 3 and the 3D image data 5, and arranges the partial asset CAD model 3 in the three-dimensional space where the 3D image data 5 is drawn. Since the positional information of the partial asset CAD model 3 is usually unknown, it is placed at a location where the shape matches the 3D image.

The attributed asset CAD model 7 is created by arranging all the partial asset CAD models 3. Since the partial asset CAD model 3 is given the management number of the partial asset as an attribute, the management number is also given to all of the attributed asset CAD models 7 which are aggregates thereof. Therefore, by linking with the attribute database 4, an attributed asset model including other attribute data is obtained.

The details of this embodiment will be described below with an example of a plant.

Fig. 2 is an isometric view showing an example of plant piping.

This figure shows what is created from the asset isometric drawing data 1 in FIG.
Since isometric drawings of plants are often used for construction of piping, they are generally drawn by extracting a series of pipes.

For the sake of explanation, directions are defined by the x-axis, y-axis, and z-axis shown in the figure. The isometric drawing is the most basic drawing method. That is, the z-axis in the vertical direction is directed upward in the figure, the x-axis and the y-axis are each inclined 30 degrees with respect to the horizontal direction in the figure, and the angle between them is 120 degrees. .

Most of the tubing is arranged along the x-, y-, or z-axis. The bent portion of the pipe is precisely a bent pipe, but in this embodiment, the bent pipe portion is schematically drawn as a right-angled pipe. If there is a valve in the middle of the pipe, it is also displayed in the isometric drawing. In addition, the dimensions necessary for construction, such as the length of the straight pipe, the length to the branch pipe, and the position from the end to the valve, are indicated by dimension lines. Both ends of the dimension line are arrows, and the lead lines of the dimension line are line segments.
However, the thickness of the line segment is different from the piping so that it can be distinguished from the piping. In addition, a pipe number is displayed on the pipe.

In Figure 2, "PAS-01-300" displayed on the rightmost pipe is the pipe number. In recent years, isometric drawings are often created with CAD software, but here, those created with paper were targeted.

The partial asset CAD model creation device 2 creates a partial asset CAD model 3, which is a three-dimensional CAD model, from the isometric drawing.

In this example, since the objects of the model are piping and valves, a piping model was created according to the following steps, and then a valve model was created.

　Fig. 3 is a flow diagram showing a method of creating a piping model and a valve model.

As shown in this figure, the isometric drawing shown in FIG. 2 is converted into data by image recognition, and only the line segments with the same thickness as the piping are extracted (step S21).

FIG. 4 is an example of the result of step S21, showing only the line segments of the extracted piping.

Next, one end point is selected (symbol A in FIG. 4), its coordinates are set to (0, 0, 0), and the slope and length of the extracted line segment on the drawing are used to determine the other points ( are indicated by symbols B, C, D, E and F.) are determined (step S22).

For example, if the horizontal right direction of the line segment AB is 0 degrees, the point B is oriented 210 degrees from the point A and has a length of 58 mm. Since the direction of 210 degrees is the x-axis negative direction in three-dimensional coordinates, the coordinates of B are (-58, 0, 0).

Next, the line segment BC has an inclination of 150 degrees on the paper surface and a length of 30 mm. Since the line segment BC is in the positive direction of the y-axis in three-dimensional coordinates, the coordinates of the point C are (-58, 30, 0). Similarly, the coordinates of points D, E and F can be obtained. However, the length is the length on paper and differs from the actual length. Since the ratio between the actual length and the length on the paper surface is constant, the actual length can be obtained by converting the coordinates by the constant ratio. In this embodiment, the actual length is 100 times the length on the paper surface, so if the actual coordinates are A (0, 0, 0), it becomes B (-5800, 0, 0). .

Next, the outer diameter of the pipe to be created is obtained (step S23). As for the outer diameter, the corresponding outer diameter is obtained by reading the piping number displayed along the piping in the isometric drawing by OCR technology. In this embodiment, the end of the pipe number represents the outer diameter. In the example of FIG. 2, 300 of the piping number "PAS-01-300" represents the outer diameter. However, since this means that the nominal diameter is 300A, the actual outer diameter is 318.5 mm.

Next, a piping model is created from the coordinates obtained in step S22 and the outer diameter obtained in step S23 (step S24). Using this information, it is possible to create a 3D-CAD model of piping using a general 3D-CAD system API (Application Programming Interface).

Finally, place the valve model in the created piping model (step S25). The position coordinates of the valve can be specified in the same manner as the method for determining the end point of the pipe in step S22. Once the position of the valve is known, the CAD software API can be utilized to locate the valve model.

The valve model differs depending on the type of valve. In this example, the valve model was selected from the model number of the valve displayed in the isometric drawing. Since VG represents a gate valve in the example of FIG. 2, a model of the gate valve is arranged.

As described above, the partial asset CAD model 3 of FIG. 1 can be created by the method shown in FIG.

FIG. 5 is a perspective view showing the partial asset CAD model created in this embodiment.

In the above example, the dimensions of the pipe are the length of the line segment that represents the pipe. Values may indicate correct dimensions. Therefore, instead of reading the length of each line segment, the OCR function may be used to read the dimension values and convert them into coordinate values.

Next, the attribute database 4 in FIG. 1 will be explained.

　In the attribute database 4, the attribute data of each part of the asset is recorded together with the management number of the part.

FIG. 6 shows an example of the attribute database 4 related to piping.

In this figure, attribute information data such as the outer diameter, wall thickness, and material of the pipe are displayed together in a table using the pipe number as an index.

The 3D image data 5 (Fig. 1) is three-dimensional image data of the entire asset. In this embodiment, 3D image data 5 is created from a plurality of images taken by a drone.

FIG. 7 shows an example of the 3D image data 5.

As shown in this figure, in addition to the piping 11, there is point cloud data of the tank 12a, the pump 12b, and the valve 13.

The 3D asset pattern matching device 6 matches the partial asset CAD model 3 and the 3D image data 5, and arranges the partial asset CAD model 3 in the three-dimensional space where the 3D image data 5 is drawn. Since the position information of the partial asset CAD model 3 is usually unknown, it is arranged in a place where the shape matches the 3D image by the following method.

FIG. 8 is a flow diagram showing a method of 3D pattern matching.

A virtual outer diameter that is a constant multiple of the outer diameter is set with respect to the center line of the created partial pipe model (step S61).

Fig. 9 shows an example of a partial piping model with virtual outer diameters set.

In this figure, the virtual outer diameter is set to 1.5 times the outer diameter. The partial pipe model is a pipe 75 having a virtual outer diameter arranged around a straight pipe center line 71 .

Here, the center line refers to the line connecting the center of the cross section of the pipe in the 3D image. A center line is a line segment that is not cut by a valve installed in the middle of a pipe and that connects a device at an end point of the pipe to a device at another end point. A connected line segment is called a “polyline”.

A piping model having a virtual outer diameter is placed on the 3D image of FIG. 7, and the number of point groups included in the virtual outer diameter is counted (step S62). The orientation of the piping model with the outer diameter is arranged according to the orientation on the 3D image with reference to the isometric drawing.

A piping model having a virtual outer diameter is translated, and the number of point groups included in the virtual outer diameter is counted (step S63).

The target space is exhaustively searched, and the piping model is placed at the position with the largest number of point groups (step S64).

If there are multiple positions where the number of point clouds is the largest, reduce the virtual outer diameter of the pipe by a predetermined value ΔD, compare the number of point clouds at the target multiple locations, and find the number of point clouds that is the largest. Positions with a high frequency are identified (step S65).

　Step S65 is repeated until the virtual outer diameter becomes equal to the outer diameter (step S66). If it cannot be identified, place it in one of multiple candidates.

If the partial pipe model is in the correct position, all the point group data will be included in the virtual outer diameter model, so the number of point groups will be the maximum.

Here, the point cloud was counted in a cylindrical region with a virtual outer diameter, but it may be a rectangular parallelepiped with a square cross section and a side length of a constant multiple of the outer diameter of the pipe. Furthermore, instead of counting the number of point groups in the area, a method may be used in which the cylinder or rectangular parallelepiped is divided into N in the longitudinal direction, and whether or not each divided area has a point group is determined. In other words, a virtual cross-sectional dimension of a predetermined part of a partial asset CAD model such as a partial piping model (for example, if the cross-sectional shape is a square, the length of one side of the square is a constant multiple of the outer diameter of the pipe) Assuming a rectangular parallelepiped, the length of one piece thereof is set as a representative dimension of a virtual cross section (that is, a virtual cross-sectional dimension)), the partial asset CAD model having the cross-sectional dimension is divided into areas, A position where the number of points included in the partial asset CAD model having the same cross-sectional dimension is maximized is specified, and the partial asset CAD model is placed at the specified position.

By repeating the above steps for all partial asset models, an attributed asset CAD model 7 can be created.

FIG. 10 is a perspective view showing an example of the attributed asset CAD model 7 created in this embodiment.

As described above, in this example, the attributed asset CAD model 7 was created by utilizing the isometric drawing. This is the main model of the plant, excluding structures such as pipe supports that support the pipes, and can create a 3D model more efficiently than conventional manual model creation.

In the second embodiment, the target assets are plant facilities composed of pipes, valves, equipment, and the like.

FIG. 11 is a configuration diagram showing an example of the three-dimensional model creation support system of this embodiment.

In this figure, the three-dimensional model creation support system includes a partial plant CAD model creation device 22, a valve/equipment recognition device 26, and a 3D pattern matching device 27.

The piping isometric drawing data 21 (piping perspective drawing data) is used for predetermined processing in the partial plant CAD model creation device 22 . The partial plant CAD model 23 is created by the partial plant CAD model creation device 22 and used for predetermined processing by the 3D pattern matching device 27 . The 2D/3D image data 25 (two-dimensional/three-dimensional image data) is used for predetermined processing by the valve/equipment recognition device 26 . The 3D pattern matching device 27 uses the partial plant CAD model 23 and data processed by the valve/equipment recognition device 26 using the 2D/3D image data 25 to create a plant CAD model 28 with attributes. The attributed plant CAD model 28 is linked with the attribute database 24 .

Also in Example 1, since plant equipment is taken as an example, Example 2 has many parts in common with Example 1. 2D/3D image data 25 , valve/equipment recognition device 26 and 3D pattern matching device 27 are different. These will be further described below.

The 2D/3D image data 25 includes 2D image data obtained by photographing the same part from multiple angles and 3D image data created based thereon. These are saved in a recording medium such as a database. Here, along with the data of each point of the 3D image, the correspondence relationship of the data of the 2D image used when creating them is also stored. The 2D image data includes an image file name and a two-dimensional coordinate format on the image file.

FIG. 12 shows an example of creating a 3D image from multiple 2D images.

In this figure, a 3D image is created from N 2D images.

The valve/equipment recognition device 26 recognizes valves and equipment from 2D or 3D images of the plant and acquires position information of the recognized valves and equipment. In this example, a 2D image was used to recognize the valve.

Specific procedures are shown below.

Recognize valves and equipment from the used 2D image (step S251). Deep learning, which is widely used, is used as a method for recognizing objects in images. By learning images of valves and equipment in advance, images of valves and equipment can be identified from all 2D images.

Using the correspondence relationship between the 3D image and the 2D image stored in the 2D/3D image data 25, a 3D image corresponding to the valve and device images identified on the 2D image is identified (step S252).

A 2D image corresponding to a 3D image of a certain valve spans multiple image files. And the points of the 3D image and the points of the 2D image are not in one-to-one correspondence. Therefore, among the valve images recognized in certain 2D image data, the pixels corresponding to the 3D image are limited. However, by recognizing the valve and equipment in all the 2D images used to create the 3D image and identifying the corresponding 3D image, the 3D image of the valve in the 3D image can be identified. As a method of finding the corresponding 3D image, a method of determining whether or not the corresponding 2D image is included in the area specified in step S251 for all the points of the 3D image may be used.

Positional information of the valve and equipment is obtained from the coordinate data of the valve and equipment specified on the 3D image (step S253).

FIG. 13 is a perspective view showing a 3D image recognized as a valve.

In this figure, the part including the valve is indicated by a dashed rectangular parallelepiped 110 . Devices may also be indicated by rectangular parallelepipeds.

Since the 3D image includes multiple valves and devices, it is necessary to separate the 3D image identified in step S252 into individual valves and devices. In the present example, a point cloud whose intervals are continuous within a threshold value is separated as one valve or device.

In addition, rectangular parallelepipeds created by this process whose volume was smaller than a predetermined threshold were excluded from the position information of valves and equipment. By this process, the data erroneously recognized as valves or devices in step S252 can be deleted.

By the above method, the three-dimensional positional information of valves and equipment can be obtained.

In this embodiment, valves and equipment are recognized using 2D images, but they may be recognized using only 3D images. That is, the process of step S252 may be performed using a plurality of 2D images with viewpoints at various positions of the 3D images. Also, if there is sufficient 3D data of valves and equipment as teaching data, the valves and equipment may be identified by three-dimensional shape recognition.
Next, the 3D pattern matching device 27 will be explained.

The 3D pattern matching device 27 uses the valve and device position information obtained by the valve/device recognition device 26 to determine which position on the 3D image the partial plant CAD model 23 corresponds to. By using valve and instrument position information, the search range can be greatly reduced.

The specific processing procedure is shown below. As the piping isometric drawing data 21, FIG. 2 is used, and as the partial plant CAD model 23, FIG.

First, the valve and device position information obtained by the valve/device recognition device 26 is arranged on the 3D image (step S261).

FIG. 14 is a perspective view showing the result of arranging the valves and equipment.

In this figure, the position information of the tank 121, the valves 122, 123, 124, 125 and the pumps 126, 127 are displayed.

Next, the relative positions of valves and equipment that can be acquired from the partial plant CAD model 23 are obtained (step S262). In the example of Fig. 2, when VG-01-50 is used as the reference point (0, 0, 0) for the placement positions (points on the piping center line) of the two valves VG-01-50 and VG-02-50, , VG-02-50 becomes (0,2400,0).

Here, the valve and equipment position information obtained in step S261 and the valve and equipment relative position information obtained in step S262 are used to narrow down the search range for pattern matching of the partial plant model (step S263). In other words, it limits the scope of pattern matching.

In this example, as shown in FIG. 14, the areas where valves may be arranged are the areas of valves 122, 123, 124, and 125. Also, from step S262, the relative positions of the two valves must be 2400 mm apart in the y-axis direction. Therefore, the coordinates (X1, Y1, Z1) of V1 are changed in the entire area of the bulbs 122, 123, 124, 125, and (X1, Y1+2400, Z) is included in the area of the bulbs 122, 123, 124, 125. is the scope of the search.

Limiting the search range obtained in step S263, setting the virtual outer diameter shown in FIG. The position with the maximum number is identified (step S264).

In addition, in the following cases, it is possible to further narrow down the search range.

FIG. 15 shows an example of the isometric view of FIG. 2 to which the information on the connection destination of the piping is added.

As shown in FIG. 15, if there is device information at the pipe connection destination, the relative position information of the valve and the device can also be acquired in step S262, so it is possible to further narrow down the search range in step S263.

In addition, in Examples 1 and 2, as a method of pattern matching, a method using the virtual outer diameter shown in FIG. 9 is shown, but a method different from this is also possible.

Fig. 16 shows a point cloud corresponding to the cross-sectional shape of the piping model.

As shown in this figure, the distance between the contour 140 of the cross section of the piping model and the point cloud 141 may be matched by a statistical method so that the error is minimized. A method of least squares or the like can be used as a statistical method.

In addition, when the pipe is wrapped with heat insulating material, the shape of the outer surface of the pipe including the heat insulating material may be matched as the contour 140 .

As shown in the above embodiment, by utilizing the positional information of valves and devices, the position of the partial plant model can be efficiently specified, and the attributed plant CAD model 28 can be created with a small number of man-hours. can.

1: asset isometric drawing data, 2: partial asset CAD model creation device, 3: partial asset CAD model, 4: attribute database, 5: 3D image data, 6: 3D asset pattern matching device, 7: asset CAD model with attributes, 11: Piping, 12a: Tank, 12b: Pump, 13: Valve, 21: Piping isometric drawing data, 22: Partial plant CAD model creation device, 23: Partial plant CAD model, 24: Attribute database, 25: 2D/3D image Data, 26: valve/equipment recognition device, 27: 3D pattern matching device, 28: plant CAD model with attributes, 71: straight pipe centerline, 75: piping.

Claims (9)

1. a partial asset CAD model creation device for creating a partial asset CAD model having attribute data using perspective view data showing a part of the asset;
   The position of the partial asset CAD model is identified by pattern matching between the partial asset CAD model and the 3D image of the asset, and the attributed asset CAD model is generated by arranging the partial asset CAD model in a 3D space. A three-dimensional model creation support system, comprising a three-dimensional pattern matching device for creating.
2. further comprising an attribute database having the attribute data and a management number of the part of the asset;
   The partial asset CAD model creation device acquires the management number displayed in the perspective view, acquires the attribute data associated with the management number of the part from the attribute database, and generates the partial asset CAD model. 3. The three-dimensional model creation support system according to claim 1, which is added to.
3. The three-dimensional model creation support system according to claim 1, wherein said partial asset CAD model creation device acquires dimensions from dimension lines displayed in said perspective view and creates said partial asset CAD model according to said dimensions.
4. The three-dimensional pattern matching device sets a virtual outer diameter of the partial asset CAD model, and identifies a position where the number of point groups included inside the partial asset CAD model having the virtual outer diameter is maximum. 2. The three-dimensional model creation support system according to claim 1, wherein said partial asset CAD model is placed at said identified position.
5. The three-dimensional pattern matching device sets virtual cross-sectional dimensions of a predetermined portion of the partial asset CAD model, divides the partial asset CAD model having the cross-sectional dimensions into areas, and divides the divided cross-sectional dimensions into areas. 2. The three-dimensional model creation support according to claim 1, wherein a position where the number of point groups contained inside said partial asset CAD model is maximum is specified, and said partial asset CAD model is placed at said specified position. system.
6. The three-dimensional pattern matching device sets a virtual cross-sectional dimension of a predetermined portion of the partial asset CAD model, and calculates an error in the distance between the contour of the cross section of the partial asset CAD model having the cross-sectional dimension and the point cloud. 2. The three-dimensional model creation support system according to claim 1, wherein a position where is the smallest is specified, and the partial asset CAD model is placed at the specified position.
7. a partial asset CAD model creation device for creating a partial asset CAD model having attribute data using data of a perspective view showing a portion of an asset including piping and at least one of valves and equipment;
   a valve/equipment recognition device that acquires position information of the valve and the equipment from 2D image or 3D image data of the asset;
   The position of the partial asset CAD model is identified by pattern matching between the partial asset CAD model and the 3D image of the asset, and the attributed asset CAD model is generated by arranging the partial asset CAD model in a 3D space. a three-dimensional pattern matching device that creates
   The three-dimensional pattern matching device uses the position information of the valve and the equipment acquired by the valve/equipment recognition device and the relative position information of the valve and the equipment acquired from the partial asset CAD model. A 3D model creation support system that limits the range of matching.
8. to the computer,
   A procedure for creating a partial asset CAD model having attribute data using perspective view data showing a part of the asset;
   The position of the partial asset CAD model is identified by pattern matching between the partial asset CAD model and the 3D image of the asset, and the attributed asset CAD model is generated by arranging the partial asset CAD model in a 3D space. A procedure to create and a program to execute it.
9. to the computer,
   A procedure for creating a partial asset CAD model having attribute data using perspective view data showing a part of the asset;
   The position of the partial asset CAD model is identified by pattern matching between the partial asset CAD model and the 3D image of the asset, and the attributed asset CAD model is generated by arranging the partial asset CAD model in a 3D space. A computer-readable recording medium that records a procedure to create and a program for executing it.

Images