Building Floor Plan Generation Using Diffusion and Bundle Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in effectively capturing, representing, and utilizing building interior information from visual data to generate accurate floor plans without physical presence, especially in constructing and maintaining precise polygonal room shapes and scaling room interiors.
Innovation Solution
A combination of a diffusion transformer machine learning model and a bundle adjustment optimizer model is used to analyze visual data from multiple images, determining global inter-image pose and wall locations to generate a precise floor plan, even with limited image overlap, by using geometric priors and denoising diffusion architecture to inpaint missing elements and constrain model outputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional floor plan construction methods are used, then accuracy of wall locations and room shapes can be maintained, but the process becomes time-consuming and requires significant manual effort
Solution Approach 1:
The system performs preliminary actions by automatically detecting walls, doors, and windows from images, and pre-processing image data to extract structural information before floor plan construction. This preliminary detection and extraction of geometric features from images significantly reduces the time required for manual floor plan construction while maintaining accuracy through algorithmic analysis of architectural elements.
2Loss of information
If manual floor plan construction methods are used, then detailed information about room interiors can be captured, but the process becomes complex and difficult to maintain
Solution Approach 1:
The system replaces manual mechanical construction processes with automated computer-based image analysis and processing algorithms. By substituting human operators with automated systems that detect walls, doors, and windows from images, the complexity of the construction process is reduced while maintaining comprehensive capture of building information through systematic image processing and automated feature extraction.
3Ease of operation
If existing image-based floor plan generation is used, then visualization of building interiors can be provided, but the accuracy of polygonal room shapes and scaling is insufficient
Solution Approach 1:
The system changes key parameters by implementing automated detection of wall locations, door positions, and window placements from images, and by applying algorithms that calculate accurate polygonal room shapes and scaling factors. These parameter changes in the processing methodology—moving from manual measurement to automated image-based detection—significantly improve the accuracy of room shapes and dimensions while maintaining ease of visualization through digital floor plan generation.
Data Source
AI summary
Techniques are described for automated operations to analyze visual data from images acquired in multiple rooms of a building to generate building information that may include a floor plan for the building, such as by analyzing visual overlap between those images to determine information that includes global inter-image pose and locations of walls and optionally other structural elements, and by using the generated building information in further automated manners. In some situations, the described techniques include using a combination of a trained diffusion transformer machine learning model and a bundle adjustment optimizer to determine global inter-image pose and wall location data and to use that data to generate a resulting floor plan for the building, such as to operate in parallel or with the bundle adjustment optimizer as a layer within the diffusion model that provides guidance for its automated determinations.