AI-Assisted BIM Model Conversion for Ambiguous Code Checking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The manual process of reviewing building and land development designs for compliance with codes and regulations is time-consuming, error-prone, and costly, and existing automated systems lack a generalized framework to handle subjective and ambiguous regulations, making them difficult to modify and maintain.
Innovation Solution
A system utilizing AI and Rule-Based algorithms transforms civil and architectural drawings into BIM or CAD models, employing Machine Learning platforms like Fast R-CNN and Mask R-CNN to extract elements, create project databases, and apply generative AI tools to ensure compliance with building codes, using neural networks and fuzzy logic to handle subjective rules.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If manual review process is used for code compliance checking, then flexibility in handling subjective and ambiguous regulations is maintained, but time consumption and cost increase significantly
Solution Approach 1:
The patent introduces an AI-based intermediary system that bridges the gap between automated compliance checking and human expert judgment. The AI system processes building designs against code regulations and provides preliminary compliance assessments, while human reviewers focus on interpreting subjective and ambiguous regulations. This intermediary approach maintains flexibility in handling complex regulatory interpretations while significantly reducing the time and cost associated with manual review processes.
2Loss of time
If existing automated compliance systems are used, then time consumption is reduced, but they lack a generalized framework to handle subjective and ambiguous regulations
Solution Approach 1:
The patent implements a dynamic compliance checking framework where the AI system can adapt its interpretation and evaluation methods based on the complexity and subjectivity of the regulations being applied. The system dynamically adjusts its analysis depth, interpretation strategies, and compliance determination criteria to appropriately handle both objective and subjective regulatory requirements, thereby achieving both time efficiency and adaptability.
Solution Approach 2:
The system changes its operational parameters and processing approaches based on the nature of the regulations being evaluated. When encountering subjective or ambiguous regulations, the AI system modifies its analysis parameters to incorporate multiple interpretation scenarios, expert judgment weights, and flexible compliance criteria, enabling it to handle diverse regulatory complexities effectively.
3Reliability
If manual compliance checking is performed, then comprehensive evaluation of all code requirements is possible, but skill and know-how requirements involve much effort and cost
Solution Approach 1:
The AI-based compliance checking system performs self-service by automatically evaluating building designs against code requirements without requiring extensive manual intervention. The system independently processes design documents, extracts relevant information, applies compliance rules, and generates comprehensive evaluation reports, thereby maintaining comprehensive evaluation capability while significantly reducing the skill and effort requirements compared to manual processes.
4Extent of automation
If existing automated systems are used, then some compliance checking is achieved, but they are difficult to modify and maintain
Solution Approach 1:
The patent segments the compliance checking system into modular components, including separate modules for different code regulations, AI processing functions, and compliance evaluation logic. This segmentation enables independent modification and maintenance of specific modules without affecting the entire system, thereby improving ease of manufacture and maintenance while maintaining a high level of automation.
Data Source
AI summary
The present disclosure presents systems, methods, and non-transitory computer-readable media for transforming civil, architectural, or engineering drawings into Building Information Modeling (BIM) or Computer-Aided Design (CAD) models. One such method comprises training an artificial intelligence algorithm on a collection of 2D civil and architectural blueprint designs to enable the artificial intelligence algorithm to efficiently extract elements from the blueprint designs and map them into 3D data model objects; creating a 3D data model using the trained artificial intelligence algorithm, from an input collection of 2D blueprint designs; creating a project database by adding metadata from the input collection of 2D blueprint designs; generating a database of code rules and regulations; performing, using the project database, and the database of code rules and regulations, a compliance check to determine if the 3D data model complies with the database of code rules and regulations; and presenting results of the compliance check.