Construction robot intelligence system and interpretation and generation method

The construction robot system addresses BIM's environmental parameter mismatch by using point cloud data and machine learning to generate accurate construction schedules, ensuring precise robot operation and reducing errors.

JP2024165988AInactive Publication Date: 2024-11-28FENG CHIA UNIVERSITY
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
JP2023088638
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-05-17
Filing Date
2023-05-30
Publication Date
2024-11-28
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

BIM technology fails to accurately reflect actual construction environment parameters, leading to incorrect construction steps by robots and increased labor costs due to technician assistance.

Method used

A construction robot system that acquires point cloud data, performs machine learning to generate architectural models, and uses a judgment module to determine construction schedules, integrating a positioning system for real-time monitoring and correction.

Benefits of technology

Ensures accurate alignment of construction robots with actual environments, reduces errors, and enhances efficiency by allowing flexible application and real-time adjustments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024165988000001_ABST
    Figure 2024165988000001_ABST
Patent Text Reader

Abstract

To provide a construction robot intelligence system and a generation method for avoiding problems where the construction environment and data during conventional construction do not match each other, affecting construction efficiency.SOLUTION: A construction robot intelligence system of the invention comprises a learning module that performs machine learning on the basis of point group data and building information data in a machine learning stage, and generates and stores a building model, and a determination module that evaluates the similarity of the point group data and each of the building models in a machine interpretation stage, and determines whether to use the building model with the highest similarity or to generate another building model, and the determination module generates a construction schedule corresponding to the building model.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a construction robot, and more particularly to a construction robot intelligence system and a method for interpreting and generating the same. [Background technology]

[0002] With the rapid development of intelligent science and technology, the use of artificial intelligence to replace human labor has become an unstoppable development trend of global industrial automation and smart manufacturing. In the current construction industry, due to the extreme shortage of professional talent and the associated construction safety issues, construction automation is also progressing. Building Information Modeling (BIM) is a common architectural model built by BIM in the current construction industry, and is used to simulate the actual information of a building.

[0003] However, in the field of BIM technology, the architectural environment parameters included in BIM cannot be accurately combined with the construction environment parameters of the actual construction environment (such as the actual size of the space, the actual position of the doors and windows, etc.). If you want to import the current BIM further into a robot to perform construction work in the construction environment, the architectural environment parameters often cannot effectively reflect the actual construction environment parameters, so real-time monitoring by engineers is still required when letting the robot perform the work, and the robot is also prone to performing incorrect construction processes, requiring assistance from engineers, which in turn leads to higher labor costs. Summary of the Invention [Problem to be solved by the invention]

[0004] In view of the above, the objective of the present invention is to provide a construction robot intelligence system that can effectively overcome the problem that the aforementioned BIM cannot effectively reflect the construction environment of an actual building by combining construction robots. [Means for solving the problem]

[0005] The present invention provides a construction robot intelligence system, comprising: a construction robot for acquiring at least one point cloud data; and a control center communicatively connected to the construction robot, wherein the control center includes a learning module that performs machine learning based on at least one of the point cloud data and at least one building information data during a machine learning stage, generates at least one building model, and stores a model database; and a judgment module that evaluates the similarity between the point cloud data and each of the building models during an equipment judgment stage, determines whether to use the building model with the highest similarity or generate another building model, and generates a construction schedule corresponding to the building model.

[0006] The present invention includes a machine learning stage and a machine reading stage, The machine learning stage includes: Inputting at least one building information data and obtaining at least one point cloud data, the point cloud data including construction environment parameters and image data; applying image segmentation techniques to the image data, combining the construction environment parameters and the building information data to perform machine learning to generate an architectural model; storing said architectural model in a model database; Including, The machine reading step includes: Acquiring the point cloud data corresponding to a construction environment, the point cloud data including the construction environment parameters and the image data corresponding to the construction environment; applying the image segmentation technique to the image data, combining the construction environment parameters, evaluating the similarity between each of the architectural models in the model database and the point cloud data, and determining whether to use the architectural model with the highest similarity or to update the architectural model; The present invention further provides a method for interpreting and generating a construction robot intelligence system, including:

[0007] In the above, the judgment module provides construction instructions to the construction robot in response to the construction schedule, and the construction robot includes an execution unit, receives the construction instructions provided by the judgment module, and executes a construction program in a construction environment in response to the construction instructions.

[0008] In the above, the execution unit feeds back the construction progress status to the control center, and the judgment module updates the construction schedule of the architectural model with the construction progress status and stores it in the model database.

[0009] In the above, the multiple construction robots are connected to the control center in communication, and the judgment module generates corresponding architectural models based on the multiple point cloud data, and simultaneously gives the construction commands to the multiple construction robots individually, and moves each construction robot between each of the architectural models based on each construction schedule updated according to the construction progress status of each of the architectural models.

[0010] In the above, a positioning system including a total station, at least one environmental reflecting member, and at least one dynamic reflecting member is provided, the total station is communicatively connected to the control center, and when the total station is installed in the construction environment, the absolute position of the total station is recorded in the architectural model, the at least one environmental reflecting member is installed at an arbitrary location in the construction environment, the total station uses an optical path to obtain an absolute position point of the environmental reflecting member corresponding to the architectural model, the at least one dynamic reflecting member is installed at an arbitrary location of the construction robot, and the total station uses the principle of optical signals and the absolute position point to obtain a relative position of the dynamic reflecting member corresponding to the architectural model. Effect of the Invention

[0011] Based on the above description of the implementation manner and embodiments, the construction robot intelligence system and generation method provided by the present invention can achieve the following advantages. 1. Through the machine learning stage, the architectural model generated by the construction robot intelligent system is equipped with the actual construction environment parameters of the construction environment and the architectural data of the corresponding data at the same time, thereby avoiding the problem that the construction environment and data during traditional construction are not consistent with each other, which affects construction efficiency. 2. The machine reading stage can be performed through the judgment module, and a corresponding architectural model can be formed using only the point cloud data. In this way, even if the construction robot intelligent system is used in a new construction environment, a new architectural model can be directly generated, without having to go back to the machine learning stage to combine the data, allowing the construction robot intelligent system to be flexibly applied and saving preparation time in the early stages. 3. When multiple construction robots work together, the judgment module can use each updated construction schedule to coordinate each construction robot between the building model and the multiple construction schedules. 4. Through the construction technology combined with the positioning system and the architectural model, the construction robot intelligence system can achieve the effect of monitoring the relative position points of each of the construction robots in real time and making instant corrections, thereby reducing the error rate. [Brief description of the drawings]

[0012] [Figure 1] FIG. 1 is a system diagram of a preferred embodiment provided by the present invention. [Diagram 2] 1 is a flow chart of a preferred embodiment of the machine learning stage provided by the present invention. [Diagram 3] 4 is a flow chart of a preferred embodiment of the machine reading step provided by the present invention; [Figure 4] 1 is an explanatory diagram of a preferred embodiment of a positioning system provided by the present invention; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] 1 is a block diagram of a preferred embodiment of a construction robot intelligence system 100 provided by the present invention. The construction robot intelligence system 100 includes a control center 10 and a construction robot 20. The control center can be executed by a computer device (not shown), which refers to electronic software and hardware having input, output, calculation and control functions, and preferably has a visual interface. Since this is a common technique, it will not be described in detail here. The control center 10 is communicatively connected to the construction robot 20, and can be combined with the construction robot 20 in a form in which the construction robot 20 is equipped with a computer device.

[0014] 2, the control center 10 includes a learning module 11, a judgment module 12, and a model database 13. Here, the learning module 11 is used by the construction robot intelligence system 100 to execute a machine learning stage S1, in which a building model A is generated and stored in the model database 13. Here, the method and steps for the learning module 11 to execute the machine learning stage S1 will be described in detail in the following paragraphs.

[0015] The model database 13 can be loaded into a building information modeling (BIM) database by linking or importing information, and will be described below using a BIM database. The BIM database stores multiple BIM data and is used when the learning module 11 executes the machine learning stage S1. At the same time, the model database 13 stores multiple building models A generated from the learning module 11.

[0016] 3, the judgment module 12 is used by the construction robot intelligence system 100 to execute the machine reading stage S2, and when the construction robot intelligence system 100 is in the machine reading stage S2, reads one of the architectural models A in the model database 13, automatically generates another new architectural model A', and / or provides a construction command O to the construction robot 20. Here, the method and steps of the judgment module 12 to execute the machine reading stage S2 will be described in detail in the following paragraphs.

[0017] The construction robot 20 includes a point cloud sensor 21 and an execution unit 22. The point cloud sensor 21 extracts construction environment parameters of the construction environment, forms point cloud data C, and transmits it to the control center 10. Here, the point cloud sensor 21 can acquire the construction environment parameters using a scanning method such as lidar scanning, infrared photography, depth camera, general camera, or binocular photography, but is not limited thereto.

[0018] Furthermore, the construction environment parameters contained in the point cloud data C include accurate dimensions of each object (walls, beams, columns, etc.) in the construction environment.

[0019] Furthermore, the point cloud data C may further include image data, which represents the construction environment in a 2D format, such as a drawing or a photograph, making the point cloud data C more complete and accurate.

[0020] Here, when the learning module 11 executes the machine learning stage S1, the point cloud data C is used to compare and combine with one of the BIM data to form the architectural model A, and the simulation environment parameters provided by the BIM data in the architectural model A can be calibrated by the point cloud data C. At the same time, the architectural model A can be provided with the architectural data provided by the BIM data, such as the construction steps corresponding to each object in the construction environment parameters, the materials or material suppliers used during construction, etc.

[0021] Here, when the point cloud data C is used by the judgment module 12 to perform the machine reading step S2, the judgment module 12 selects one of the architectural models A that is most similar to the point cloud data C in terms of the construction environment parameters provided by the point cloud data C, generates a construction schedule based on the architectural data of the architectural model A, and provides the construction command O to the construction robot 20 corresponding to the construction schedule.

[0022] Furthermore, the attributes of each object in the construction environment parameters can be used to sequence the construction schedule. For example, the attributes of each object in the construction environment parameters include the positions of walls, planned doors and windows, beams installed on the ceiling, and columns connecting the ceiling and the ground, and the judgment module 12 can set the construction schedule in the order of beams, columns, walls, doors, and windows according to the attributes of each object in the architectural data.

[0023] Furthermore, the judgment module 12 compares each building model A with the point cloud data C to evaluate the similarity between the point cloud data C and each building model A, and determines one building model A that is most similar according to the similarity.

[0024] Furthermore, the judgment module 12 judges whether or not to generate a new architectural model A according to the similarity.

[0025] Furthermore, the similarity can be displayed in the form of a number, a percentage, a text or an image, and the similarity evaluation method can be based on the principles of calculations such as repeat percentage, distance / vector similarity, cosine of included angle or correlation coefficient.

[0026] Preferably, the judgment module 12 sets a similarity threshold as necessary, which is set in combination with the above-mentioned method for displaying and / or evaluating the similarity. The judgment module 12 judges whether to generate another new architectural model A' according to the similarity threshold. For example, when the similarity is shown as a percentage and the similarity threshold is set to 90%, the judgment module 12, in addition to evaluating the architectural model A that is most similar to the point cloud data C, further checks whether the similarity between the architectural model A and the point cloud data C is higher than the similarity threshold, and if the similarity exceeds the similarity threshold, the architectural model A can be directly extracted and used, and if not, another new architectural model A is generated.

[0027] When the point cloud data C is used by the judgment module 12 to execute the machine reading step S2 to generate another architectural model A', the judgment module 12 selects the architectural model A with the highest similarity, combines it with the architectural model A using the point cloud data C, generates an updated architectural model A', and stores it in the model database 13.

[0028] Here, whether the judgment module 12 uses the architectural model A or generates an updated architectural model A', the construction robot intelligence system 100 simultaneously allows the user to make appropriate modifications to the architectural model A or the architectural model A', and similarly, the judgment module 12 further stores the modified architectural models A, A' in the model database 13.

[0029] The execution unit 22 receives the construction command O provided by the judgment module 12, executes a construction program in the construction environment according to the construction command O, and feeds back the construction progress status F of the construction program to the control center 10. At this time, the judgment module 12 further updates the construction schedule of the architectural models A, A' with the construction progress status F, and stores it in the model database 13, so that the latest progress status of the construction schedule can be immediately obtained when the subsequent architectural models A, A' are extracted again.

[0030] Furthermore, since the construction robot intelligence system 100 can communicate with a plurality of the construction robots 20 simultaneously via the control center 10, the construction robots 20 are placed in a plurality of different construction environments simultaneously to acquire the point cloud data C of each of the construction environments. In this manner, the construction robot intelligence system 100 can generate the corresponding architectural models A, A' based on the plurality of point cloud data C, and give the construction command O to the plurality of construction robots 20 individually and simultaneously.

[0031] In addition, when multiple construction robots 20 work in cooperation with each other, each construction robot 20 can feed back the construction progress status F to each of the architectural models A, A', and move each of the construction robots 20 via each of the construction schedules updated in the judgment module 12.

[0032] For example, after the first robot completes a program for applying cement in the first construction environment based on a first construction schedule and feeds back a construction progress status F to the control center 10, the cement material needs to be left to stand for a certain period of time to harden by natural drying, and the judgment module 12 gives the construction command 0 to the first robot based on the first construction schedule and executes the next construction schedule. When the standing time has passed and the second robot has completed the corresponding construction program in the second construction environment based on the second construction schedule, the judgment module 12 simultaneously judges the first construction schedule and the current construction progress status F of the first robot, and judges whether to move the second robot to the first construction environment to continue executing the first construction schedule, for example, the work of applying paint to a wall on cement.

[0033] In this way, when the construction robot intelligence system 100 operates the multiple construction robots 20 in cooperation with each other, each of the construction robots 20 is not limited to executing the construction command O in only one of the architectural models A, A' (i.e., one of the construction schedules), and the judgment module 12 may simultaneously judge the multiple architectural models A, A' and the multiple construction schedules, and adjust each of the construction robots 20 to execute the construction program during each construction schedule. Each of the construction schedules is not limited to one of the construction robots 20, and may have multiple construction robots 20 to simultaneously execute one construction schedule and update the construction progress status F.

[0034] The construction robot intelligence system 100 provided by the present invention realizes a technology that can automatically generate the architectural model A by using machine learning through the cooperation of the control center 10 and the construction robot 20, and further achieves the scheduling and updating of each of the construction programs, achieves the effect of multiple construction robots 20 working together simultaneously, and achieves the characteristics of multiple construction robots 20 moving flexibly, thereby effectively improving the overall efficiency.

[0035] Referring to FIG. 4, the construction robot intelligence system provided by the present invention further includes a positioning system 40 constructed based on the building data of the building models A and A', and performs real-time positioning and dynamic calibration of the construction robot 20.

[0036] In one embodiment, the positioning system 40 includes a total station 41 and at least one dynamic reflecting member 43. The total station 41 is in communication with the control center 10, and when installed in the construction environment, the building models A, A' record the absolute position of the total station 41, where the absolute position of the total station 41 is fixed, and the dynamic reflecting member is a prism or a seal with a reflective effect.

[0037] Here, the method for recording the absolute position of the total station 41 on the building models A, A' is not limited, and the total station 41 may be manually operated to mark the absolute position on the building models A, A', or the total station 41 may determine the absolute position in a manner similar to the point cloud sensor 21 described above.

[0038] The dynamic reflecting member 43 is disposed at any position on the construction robot 20, and when the construction robot 20 performs the construction program in the construction environment, the total station 41 can receive the feedback of the dynamic reflecting member 43 on the construction robot 20 by utilizing the principle of optical signals, thus the total station 41 can calculate the absolute position and the relative position point of the dynamic reflecting member 43 corresponding to the building models A, A'. Due to the combination of the positioning system 40 and the building models A, A', and the excellent characteristics of high-speed optical transmission, the total station 41 can detect and calculate the relative position point of the dynamic reflecting member 43 in real time, and feedback it to the control center 10 and record it in the building models A, A'.

[0039] At the same time, the judgment module 12 uses the relative position point to verify whether the position of the construction robot 20 performing the construction program accurately corresponds to the construction schedule, and if the relative position of the construction robot 20 does not match the construction schedule, it immediately provides the construction command O to the construction robot 20 to perform calibration, thereby improving the construction quality of the construction robot 20 and effectively reducing the error rate.

[0040] Preferably, the dynamic reflecting member 43 is installed on the mechanical arm 23 that executes the construction program of the construction robot 20, and it can be confirmed that the construction robot 20 corresponds to the construction position in the construction environment at the precise relative position point. For example, the dynamic reflecting member 43 is installed at the end of the free end of the mechanical arm 23 of the construction robot 20, and thus, when the construction robot 20 paints, it can be confirmed that the paint is applied to the actual construction position in the construction environment using the relative position point.

[0041] In another embodiment, the positioning system 40 further includes at least one environmental reflecting member 42, and the total station 41 can move freely within the construction environment. The environmental reflecting member 42 is installed at an arbitrary location in the construction environment, and the total station 41 transmits an optical signal such as a microwave or an infrared ray using an optical path, receives feedback reflected by the environmental reflecting member 42, and calculates the absolute position of the environmental reflecting member 42 corresponding to the building models A, A'. Here, the environmental reflecting member 42 may be a prism or a seal having a reflective effect, like the dynamic reflecting member 42.

[0042] In this way, when the construction environment is relatively large, the total station 41 has the ability to move, achieves automatic detection without blind spots, and uses the absolute position of the environmental reflecting member 42 to determine the relative position of the dynamic reflecting member 43, so as to achieve the functions of monitoring and calibrating the construction robot 20.

[0043] Furthermore, since the relative position points are simultaneously recorded on the building models A, A', the positioning system 40 can determine the moving path of the total station 41 in the construction environment through the control of the judgment module 12. For example, the judgment module 12 gives a detection command to the total station 41, so that the total station 41 moves toward the relative position points marked on the building models A, A', detects the dynamic reflective member 43 of the current construction robot 20, obtains the relative position points of the current dynamic reflective member 43, and finally updates the building models A, A'.

[0044] The steps of the machine learning stage S1 of the construction robot intelligence system 100 further provided by the present invention include: S1-1: Load the building information data (Building Information Modeling, BIM), the BIM data will be described later, and obtain the point cloud data C of the construction environment, where the point cloud data C includes the construction environment parameters and image data corresponding to the construction environment. S1-2: Use image segmentation technology on the image data, combine the construction environment parameters and the BIM data, and perform machine learning to form the architectural model A. S1-3: The architectural model A is stored in the model database 13.

[0045] The steps of the machine reading step S2 of the construction robot intelligence system 100 further provided by the present invention include: S2-1: The point cloud data C of the construction environment is acquired, and the point cloud data C includes the construction environment parameters and image data corresponding to the construction environment. S2-2: Use image segmentation technology on the image data, combine it with the construction environment parameters, evaluate the similarity between the architectural model A and the point cloud data C in the model database 13, and select the architectural model A that has the highest similarity to the point cloud data C. S2-3: Evaluate whether the similarity exceeds a predetermined similarity threshold and decide whether to generate another new architectural model A'; if the similarity exceeds the similarity threshold, enter step S2-4; if the similarity does not exceed the similarity threshold, enter step S2-5. S2-4: If the similarity exceeds the similarity threshold, extract the architectural model A directly, and create the construction schedule based on the architectural model A. S2-5: If the similarity does not exceed the similarity threshold, select the architectural model A with the highest similarity, combine it with the architectural model A using the point cloud data C, generate an updated architectural model A', store it in the model database 13, and create the construction schedule according to the updated architectural model A'.

[0046] Based on the above description of the implementation manner and embodiments, the construction robot intelligence system and generation method provided by the present invention can produce the following effects. 1. Through the machine learning stage S1, the architectural model A generated by the construction robot intelligence system 100 is simultaneously equipped with the actual construction environment parameters of the construction environment and architectural data of the BIM data, thereby avoiding the problem that the construction environment and BIM data during traditional construction do not match each other, which affects construction efficiency. 2. The machine reading step S2 can be performed through the judgment module 12, and the corresponding architectural models A, A' can be formed using only the point cloud data C. In this way, even if the construction robot intelligence system 100 is used in a new construction environment, a new architectural model A' can be directly generated, and there is no need to return to the machine learning step S1 to combine BIM data, allowing the construction robot intelligence system 100 to be flexibly adapted and saving preparation time in the initial stage. 3. When multiple construction robots 20 work in coordination, the judgment module 12 can use each updated construction schedule to coordinate each construction robot 20 between the building models A, A' and the multiple construction schedules, thereby improving overall efficiency. 4. Through the construction technology combined with the positioning system 40 and the building models A, A', the construction robot intelligence system 100 can achieve the effect of monitoring the relative position points of each of the construction robots 20 in real time and correcting them instantly, thereby reducing the error rate. [Explanation of symbols]

[0047] 100 Construction Robot Intelligence System 10 Control Center 11 Learning Modules 12 Judgment Module 13 Model Database 20 Construction Robot 21 Point Cloud Sensor 22 Execution Unit 23 Mechanical Arm 40 Positioning System 41 Total Station 42 Environmental Reflective Materials 43 Dynamic Reflective Members S1 Machine Learning Stage S1‐1 Step S1‐2 Step S1‐3 Step S2 Machine-readable stage S2‐1 Step S2‐2 Step S2‐3 Step S2‐4 Step S2‐5 Step A Architectural Model A' Architectural Model C Point cloud data F. Construction progress ○ Construction instructions

Claims

1. A construction robot for acquiring at least one point cloud data; A control center that is communicatively connected to the construction robot; Equipped with The control center a model database for storing at least one architectural model generated by the control center through machine learning based on at least one of the point cloud data and at least one of the architectural information data; a judgment module that evaluates a similarity between one of the point cloud data and each of the architectural models, determines whether to use the architectural model with the highest similarity or to generate another architectural model, and generates a construction schedule corresponding to the architectural model; Construction robot intelligence system including.

2. 2. The construction robot intelligence system according to claim 1, wherein the judgment module includes a predetermined similarity threshold, and the judgment module compares the similarity with the similarity threshold and determines whether to use the architectural model with the highest similarity or to update the architectural model.

3. 3. The construction robot intelligence system according to claim 2, wherein when the judgment module determines to update the architectural model, the judgment module selects the architectural model having the highest similarity, combines the point cloud data with the architectural model to generate an updated architectural model, and stores the updated architectural model in the model database.

4. The judgment module provides a construction command to the construction robot in response to the construction schedule; The construction robot includes an execution unit that receives the construction command provided by the judgment module and executes a construction program in a construction environment in response to the construction command; The construction robot intelligence system according to any one of claims 1 to 3, wherein the execution unit feeds back the construction progress status to the control center, and the judgment module updates the construction schedule of the architectural model with the construction progress status and stores it in the model database.

5. 5. The construction robot intelligence system according to claim 4, wherein the plurality of construction robots are connected in communication with the control center, and the judgment module generates the corresponding architectural models based on the plurality of point cloud data, and simultaneously gives the construction instructions to the plurality of construction robots individually, and moves each of the construction robots between the architectural models based on each of the construction schedules updated according to the construction progress status of each of the architectural models.

6. a positioning system including a total station and at least one dynamic reflective member; The total station is connected to the control center for communication, and when the total station is fixedly installed in the construction environment, the absolute position of the total station is recorded in the architectural model; 6. The construction robot intelligence system according to claim 5, wherein the at least one dynamic reflective member is installed at any location on the construction robot, and the total station acquires the relative position of the dynamic reflective member corresponding to the building model by utilizing the principle of optical signals and the absolute position.

7. A positioning system including a total station, at least one environmental reflector, and at least one dynamic reflector, The total station is connected to the control center for communication, and when the total station is movably installed in the construction environment, the absolute position of the total station is recorded in the architectural model; The at least one environmental reflection member is installed at an arbitrary location in the construction environment, and the total station acquires an absolute position point of the environmental reflection member corresponding to the architectural model by using an optical path; 6. The construction robot intelligence system according to claim 5, wherein the at least one dynamic reflective member is installed at any location on the construction robot, and the total station obtains the relative position of the dynamic reflective member corresponding to the building model by utilizing the principle of optical signals and the absolute position point.

8. 8. The construction robot intelligence system according to claim 7, wherein the judgment module provides a detection command to the total station to move the total station toward the relative position point marked on the architectural model, and the total station detects the current detection command of the current dynamic reflective member and updates the architectural model.

9. Includes a machine learning stage and a machine reading stage, The machine learning stage includes: Inputting at least one building information data and obtaining at least one point cloud data, the point cloud data including construction environment parameters and image data; applying image segmentation technology to the image data, combining the construction environment parameters and the building information data to perform machine learning to generate an architectural model; storing said architectural model in a model database; Including, The machine reading step includes: Acquiring the point cloud data corresponding to a construction environment, the point cloud data including the construction environment parameters and the image data corresponding to the construction environment; applying the image segmentation technique to the image data, combining the construction environment parameters, evaluating the similarity between each of the architectural models in the model database and the point cloud data, and determining whether to use the architectural model with the highest similarity or to update the architectural model; The present invention relates to a method for interpreting and generating a construction robot intelligence system, the method including:

10. In the machine interpretation step, evaluating whether the similarity exceeds a predetermined similarity threshold; If the similarity exceeds the similarity threshold, directly extract the architectural model and form a construction schedule; 10. The method of claim 9, further comprising: if the similarity does not exceed the similarity threshold, selecting the architectural model with the highest similarity, combining the architectural model with the point cloud data to generate an updated architectural model, and storing the updated architectural model in the model database.

Citation Information

Patent Citations

  • Construction management system, construction management support system, and program

    JP2020140257A

  • Construction work supporting system and construction work supporting method

    JP2021156016A

  • In-building structure recognition system and in-building structure recognition method

    JP2023066126A