KR-20260063035-A - A Method for Automatic Labeling of Synthetic Data for Artificial Intelligence Learning of Building Elements based on Cloud BIM Model

Abstract

A commercial BIM repository that stores multiple building information; A database that stores shape image information and metadata information; A token authentication generator that generates an authentication token required to access the platform; and A cloud-based BIM data extraction and analysis platform for AI learning that monitors and analyzes automatically labeled data in real time, including a screen interface where the user selects model cases and enters chat commands; and an automatic for AI learning

Inventors

• 김명환

Assignees

• 주식회사 상상진화

Dates

Publication Date

20260507

Application Date

20241030

Claims (5)

1. As a synthetic data automatic labeling method for AI learning of building elements based on cloud BIM models, Step of selecting a learning model case of a cloud platform interface (S100); Step of entering training data generation command into the chat command input window (S200); A training data generation step (S300) in which shape images and metadata are generated as a training dataset; and Including the generated training data storage step (S400), A method characterized by creating shape images generated from various angles using cloud linkage technology for 3D shape information of a BIM model, and automatically labeling metadata on the shape images in the cloud.
2. In claim 1, A method characterized by further including a security authentication processing step for data linkage.
3. A cloud-based artificial intelligence BIM data processing method that recognizes and classifies building characteristics and components by applying automatically labeled training data to a machine learning algorithm according to the method of claim 1 or 2.
4. Cloud-based artificial intelligence BIM data extraction and analysis platform for monitoring and analyzing automatically labeled training data in real time using the method of claim 1 or 2
5. A commercial BIM repository that stores multiple building information; A database that stores shape image information and metadata information; A token authentication generator that generates an authentication token required to access the platform; and A cloud-based BIM data extraction and analysis platform for AI learning that monitors and analyzes automatically labeled data in real time, including a screen interface where users select model cases and enter chat commands.

Description

A Method for Automatic Labeling of Synthetic Data for Artificial Intelligence Learning of Building Elements based on Cloud BIM Model The present invention relates to a method for automatically labeling synthetic data for artificial intelligence learning for automatic recognition of building elements using shape information and attribute information of a BIM (Building Information Modeling) model stored in the cloud. From the perspective of AI model training, numerous training datasets are required to automatically recognize images of building components. Utilizing a sufficient amount of training data enables the AI model to recognize building component images more accurately and minimize errors. The more training data is used, the more continuously the performance of the AI model can be improved, which enhances usability and reliability in real-world applications. However, creating training datasets manually is costly and time-consuming. Manually classifying and labeling images is a very cumbersome and time-consuming task; consequently, building large-scale training datasets requires a significant workforce and a high level of professional skill and experience to maintain quality. According to Korean Published Patent 10-2023-0061164 (Device and method for constructing an artificial intelligence learning dataset based on a 3D authoring tool for object recognition), The present invention relates to an apparatus and method for constructing an AI training dataset based on a 3D authoring tool for object recognition, wherein image data is modeled using a 3D authoring tool, the model is rendered as if it were a real photograph, and a dataset for AI object recognition is constructed through data cleaning and processing. A method for constructing an artificial intelligence training dataset based on a 3D authoring tool for object recognition is implemented by receiving an image for object recognition, modeling the received image using a 3D authoring tool, rendering the modeled model, constructing a dataset through data synthesis, cleaning, and processing, and training the constructed dataset with a deep learning algorithm, which is an object recognition algorithm, to recognize objects. Figure 1 is a flowchart for generating a synthetic dataset based on a BIM model according to the present invention. FIG. 2 is an example of a cloud platform interface according to the present invention. Figure 3 is an example of a synthetic dataset generation command input according to the present invention. Figure 4 is an example of the result of generating a synthetic dataset according to the present invention. FIG. 5 is an example of a cloud-based artificial intelligence BIM data extraction and analysis platform according to the present invention. Figure 6 is an example of a synthetic dataset according to the present invention. The present invention is capable of various modifications and may have various embodiments, and specific embodiments are illustrated in the drawings and described in detail. However, this is not intended to limit the invention to specific embodiments and should be understood to include all modifications, equivalents, and substitutions that fall within the spirit and scope of the invention. Similar reference numerals have been used for similar components in the description of each figure. Terms including ordinal numbers, such as second, first, etc., may be used to describe various components, but said components are not limited by said terms. The terms used in this specification are used solely for the purpose of distinguishing one component from another. For example, without departing from the scope of the invention, a second component may be named a first component, and similarly, a first component may be named a second component. The term "and/or" includes a combination of a plurality of related described items or any one of a plurality of related described items. When it is stated that one component is "connected" or "connected" to another component, it should be understood that while it may be directly connected or connected to that other component, there may also be other components in between. On the other hand, when it is stated that one component is "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between. The terms used in this application are used merely to describe specific embodiments and are not intended to limit the invention. The singular expression includes the plural expression unless the context clearly indicates otherwise. In this application, terms such as "comprising" or "having" are intended to specify the presence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof. Unless otherwi