JP7573385B2 - Design Support Equipment - Google Patents

Description

The present invention relates to a design support device.

A design support system for steel structures is known that includes a process for inputting basic data about the steel structure and a process for generating structural component data for the steel structure based on the basic data and predetermined component placement rules that have been set up in advance (for example, Patent Document 1).

There is also known a structural analysis information generating device that inputs design CAD information and generates information indicating the center line of a building based on information indicating the shape, dimensions, and positions of multiple columns or multiple walls contained in the design CAD information (for example, Patent Document 2).

A building CAD system is also known that determines the number of each product component used in all planar modules included in a building, the number of structural components obtained from each product component by cutting the product component, and information indicating the size and placement position of each structural component (for example, Patent Document 3).

JP 2004-245040 A JP 2008-158793 A JP 2017-10413 A

In the technology described in Patent Document 1, after basic data is input on the input screen, structural member data for a steel structure is generated and the generated results are displayed. However, Patent Document 1 does not describe generating an image showing the positions where structural members should be added from an image showing the volume.

In the technology described in Patent Document 2, data for structural analysis is created in a continuous process from CAD data of the architectural design, including the volume of the building. However, Patent Document 2 does not describe generating an image showing the positions where structural members should be added from an image showing the volume.

Furthermore, the above-mentioned Patent Document 3 does not describe generating an image showing the position where a structural member should be added from an image showing a volume.

Taking the above facts into consideration, the present invention aims to easily generate appropriate structural member arrangements and assist in the structural design of buildings.

The design support device according to the present invention includes an image generation unit that receives an image representing the volume of a building to be designed as an input, and generates an additional image using a generation model for generating an additional image to which position information to which structural members should be attached is added, and a structural member generation unit that generates the arrangement of the structural members relative to the volume of the building to be designed based on the generated additional image.

According to the design support device of the present invention, an image generation unit receives an image representing the volume of a building to be designed as an input, and generates the additional image using a generation model for generating an additional image to which position information to which a structural member should be added is added. Then, a structural member generation unit generates the arrangement of the structural members relative to the volume of the building to be designed based on the generated additional image.

In this way, an image showing the volume of the building to be designed is input, an additional image is generated using a generative model for generating an additional image with added position information for where the structural members should be attached, and the placement of the structural members relative to the volume of the building to be designed is generated based on the additional image. This makes it possible to easily generate an appropriate placement of the structural members and to assist in the structural design of the building.

In the design support device according to the present invention, the generative model is trained in advance using pairs of training data, each of which is an image showing the volume of a building obtained from actual building information, and an image in which the positions of structural members have been added to the image showing the volume. This makes it possible to train the generative model from data obtained from actual building information.

In the design support device according to the present invention, the generative model is an autoencoder including an encoder and a decoder, and is trained in advance using GAN (Generative Adversarial Networks). This makes it possible to train a generative model that generates additional images with high accuracy.

In the design support device according to the present invention, the input image is color-coded according to the attributes of the room, and the additional image is the image color-coded according to the attributes of the room to which pixel values corresponding to the degree to which structural members should be added are added. This makes it possible to generate additional images according to the attributes of the room.

In the design support device according to the present invention, the generative model receives an image representing the volume and design information as input and generates the additional image, and the image generation unit receives an image representing the volume of the building to be designed and design information of the building to be designed as input and can generate the additional image using the generative model. This makes it possible to generate the additional image taking into account the design information of the building.

The design support device according to the present invention can further include a display unit that displays the generated arrangement of the structural members superimposed on the volume. This makes it possible to more effectively support the structural design of buildings.

As described above, the design support device of the present invention takes an image representing the volume of the building to be designed as input, generates an additional image using a generative model for generating an additional image with added position information for where the structural members should be attached, and generates the arrangement of the structural members relative to the volume of the building to be designed based on the additional image, thereby providing the effect of easily generating an appropriate arrangement of structural members and supporting the structural design of the building.

1 is a block diagram showing a learning device and a design support device according to an embodiment of the present invention; 1 is a functional block diagram showing a learning device according to an embodiment of the present invention; FIG. 1 is a diagram showing an example of an image representing a volume of a building. FIG. 13A is a diagram showing an example of an additional image to which the positions of beams have been added, and FIG. 13B is a diagram showing an example of an additional image to which the positions of pillars have been added. FIG. 1 is a diagram illustrating an example of a trained model. FIG. 1 illustrates an example of a GAN. 1 is a functional block diagram showing a design support apparatus according to an embodiment of the present invention; FIG. 1 is a diagram illustrating an example of a building volume. FIG. 13 is a diagram showing an example of a screen that displays the arrangement of columns superimposed on the volume of a building. 4 is a flowchart showing the contents of a learning process routine of the learning device according to the embodiment of the present invention. 4 is a flowchart showing the contents of a design support processing routine of the design support device according to the embodiment of the present invention. FIG. 13 is a diagram showing an example of an image representing a building volume, color-coded according to room attributes.

The following describes in detail an embodiment of the present invention with reference to the drawings.

<Configuration of the learning device according to the present invention>
As shown in FIG. 1, a learning device 100 according to an embodiment of the present invention includes a CPU 12, a graphics card 13, a GPU 14, a RAM 16, a HDD 18, a communication interface 21, and a bus 23 for interconnecting these components.

The CPU 12 and GPU 14 execute various programs. The RAM 16 is used as a work area when the CPU 12 executes the various programs. The HDD 18, which serves as a recording medium, stores various programs and data, including a program for executing the learning processing routine described below.

The learning device 100 in this embodiment is shown in FIG. 2 as a functional block diagram in accordance with a program for executing a learning processing routine. The learning device 100 includes an input unit 10, a calculation unit 20, and an output unit 50.

The input unit 10 receives as input learning data including pairs of images obtained from the building's performance information, each of which is a pair of an image showing the volume of the building and an image in which the positions of structural members are added to the volume image.

Specifically, from the building's performance information, an image showing the building volume as shown in Figure 3 and an image with the positions of structural members added as shown in Figure 4 are created. Figure 3 shows an example of an image showing the building volume, where each pixel value is 0 or 1 indicating whether it corresponds to the volume or not. Figure 4(A) shows an example of an image showing the building volume with the positions of beams added, and Figure 4(B) shows an example of an image showing the building volume with the positions of columns added. Then, learning data including pairs of an image showing the building volume and an image with the positions of structural members added to the volume image are accepted.

In this embodiment, this learning data is accepted for each type of structural member (pillar, beam, wall, brace, etc.).

The calculation unit 20 includes a learning unit 22.

The learning unit 22 obtains a learned model for each type of structural member based on multiple pieces of learning data received by the input unit 10.

Specifically, the trained model uses an image representing the volume of a building as input data, and uses an additional image in which a pixel value according to the degree to which a structural member should be added is added to a pixel corresponding to a position where the structural member should be added as output data. For example, as shown in FIG. 5, an autoencoder including an encoder and a decoder can be used as an example of the trained model, and a GAN can be used as an example of the learning algorithm (FIG. 6). FIG. 5 shows an example in which each pixel value x ₁ , x ₂ , ..., x _m of an image representing the volume of a building is used as input data, and the encoder converts it into a latent variable y(x), and the decoder outputs each pixel value of the additional image. FIG. 6 shows an example in which an additional image generated by inputting an image z made of noise into a trained model, and a true additional image obtained from actual information on the building are input into a discrimination model that identifies whether or not the data is true, and the discrimination model is trained so that the discrimination model can be correctly identified, and the generation model is trained so that the generated additional image is identified as true data by the discrimination model.

In addition, the generative model, which is a trained model, is trained so that when an image representing the volume of a building is input as training data, an additional image of the training data is output.

In this embodiment, a trained model is generated for each type of structural member (column, beam, wall, brace, etc.).

<Configuration of the design support device according to the present invention>
As shown in FIG. 1 above, the design support device 200 according to an embodiment of the present invention, like the learning device 100, comprises a CPU 12, a graphics card 13, a GPU 14, a RAM 16, a HDD 18, a communication interface 21, and a bus 23 for connecting these to each other.

The CPU 12 and GPU 14 execute various programs. The RAM 16 is used as a work area when the CPU 12 executes the various programs. The HDD 18, which serves as a recording medium, stores various programs and data, including a program for executing the design support processing routine described below.

When the design support device 200 in this embodiment is represented by functional blocks along with a program for executing a design support processing routine, it becomes as shown in FIG. 7. The design support device 200 has an input unit 110, a calculation unit 120, and an output unit 150.

The input unit 110 receives as input an image representing the volume of the building to be designed. For example, it receives as input an image in which the three-dimensional space of the volume of the building to be designed is projected onto a two-dimensional plane, as shown in FIG. 8.

The calculation unit 120 includes an image generation unit 122, a model storage unit 124, a structural member generation unit 126, and a display unit 128.

The image generation unit 122 uses a trained model to generate additional images for each type of structural component from the input image representing the volume of the building to be designed.

The model storage unit 124 stores trained models trained by the learning device 100 for each type of structural member.

The structural member generation unit 126 generates the layout of the structural members for each type of structural member based on the generated additional image. Specifically, for each type of structural member, the structural member layout is generated so that the structural members are provided at positions from the additional image where the degree to which the structural member should be added is equal to or greater than a threshold value.

The display unit 128 superimposes the generated arrangement of structural members for each type of structural member on the volume and displays it via the output unit 150 (Figure 9). Figure 9 shows an example in which the generated arrangement of columns is superimposed on the volume of the building to be designed.

<Operation of the learning device>
Next, the operation of the learning device 100 according to the embodiment of the present invention will be described.

The input unit 10 accepts as input a number of learning data sets, each set including a pair of an image showing the volume of a building, obtained from the building's performance information, and an image in which the position of the structural member is added to the volume image, for each type of structural member. The learning device 100 then executes the learning process routine shown in FIG. 10.

First, in step S100, the learning unit 22 learns a generative model, which is a trained model, for each type of structural member based on the multiple pieces of learning data received by the input unit 10, so that when an image of the learning data representing the volume of a building is input, an additional image of the learning data is output.

In step S102, the learning unit 22 inputs an additional image generated by inputting an image z consisting of noise into a generative model, which is a trained model, and an additional image of the learning data into a discriminative model for each type of structural member, and trains the discriminative model so that it can make correct discrimination, and also trains the generative model, which is a trained model, so that it can deceive the discriminative model.

In step S104, it is determined whether a predetermined iteration end condition is satisfied. If the iteration end condition is not satisfied, the process returns to step S100. On the other hand, if the iteration end condition is satisfied, the learning process routine is terminated. Here, the iteration end condition is, for example, that the number of iterations has reached an upper limit.

<Operation of the design support device>
Next, the operation of the design support device 200 according to the embodiment of the present invention will be described.

First, data on the trained models for each type of structural member trained by the learning device 100 is stored in the model storage unit 124.

Then, when the input unit 110 receives as input an image representing the volume of the building to be designed, the design support device 200 executes the design support processing routine shown in FIG. 11.

First, in step S110, the image generation unit 122 acquires an image representing the volume of the building to be designed.

In step S112, the image generation unit 122 generates additional images for each type of structural component from the input images representing the volume of the building to be designed, using the trained model.

In step S114, the structural member generation unit 126 generates the arrangement of the structural members for each type of structural member based on the generated additional image.

In step S116, the display unit 128 superimposes the generated arrangement of structural members for each type of structural member on the volume and displays it via the output unit 150, and then ends the design support processing routine.

As described above, the design support device according to the embodiment of the present invention takes an image representing the volume of the building to be designed as input, generates an additional image using a generative model for generating an additional image with added position information to which structural members should be attached, and generates the arrangement of structural members for the volume of the building to be designed based on the additional image, thereby easily generating an appropriate arrangement of structural members and supporting the structural design of the building.

The present invention is not limited to the above-described embodiment, and various modifications and applications are possible without departing from the spirit of the invention.

For example, in the above embodiment, an image in which only the volume of a building is projected onto a two-dimensional plane is used as input, but this is not limited to the above. As shown in FIG. 12, an image representing the volume of a building and color-coded for each room attribute may be input. In this case, an additional image in which pixel values according to the degree to which the value should be added are added to the image color-coded for each room attribute may be output. For example, an image color-coded for each design condition, such as the purpose of the room, the load capacity, and whether it is a column-free space, may be input.

The generative model may also receive an image representing the volume of a building and design information as input, and generate an additional image. In this case, design information of the building to be designed may be further received as input, and the image representing the volume of the building and the design information may be input to the generative model to generate an additional image.

In addition, although an example has been described in which the learning device and the design support device are configured as separate devices, this is not limited thereto, and the learning device and the design support device may be configured as a single device.

The program of the present invention may also be provided by storing it on a storage medium.

10, 110 Input unit 20, 120 Calculation unit 22 Learning unit 50, 150 Output unit 100 Learning device 122 Image generation unit 124 Model storage unit 126 Structural member generation unit 128 Display unit 200 Design support device

Claims (4)

a learning unit that uses a pair of an image representing the volume of a building obtained from actual building information and an image in which the positions of structural members are added to the image representing the volume as learning data, and learns a generation model for generating an added image to which position information of a structural member is added by inputting an image representing the volume of the building, using a plurality of learning data;
an image generating unit that receives an image representing the volume of a building to be designed as an input, and generates an additional image by using a generating model for generating an additional image to which position information of a structural member is added;
a structural member generation unit that generates an arrangement of the structural members with respect to the volume of the building to be designed based on the generated additional image;
Including,
the generative model is an autoencoder including an encoder and a decoder;
the learning unit learns the generative model based on the plurality of learning data such that, when an image of the learning data representing a volume of a building is input, an additional image of the learning data is output;
An additional image generated by inputting an image consisting of noise into the generative model and the additional image of the learning data are input into a discriminative model, and the discriminative model is trained to be able to correctly discriminate, and the generative model is trained to be able to fool the discriminative model;
Repeat until a predetermined iteration termination condition is met.
Design support device. The input image is color-coded according to the attributes of the room,
2. The design support system according to claim 1 , wherein said additional image is obtained by adding pixel values according to the degree to which a structural member should be added to said image color-coded for each attribute of the room. the generative model receives an image representing the volume and design information as input, and generates the additional image;
3. The design support device according to claim 1, wherein the image generation unit receives an image representing a volume of the building to be designed and design information of the building to be designed as input, and generates the additional image using the generative model. 4. The design support device according to claim 1, further comprising a display unit that displays the generated arrangement of the structural members by superimposing it on the volume.