Automated Floor Plan Generation from Paired Building Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods face challenges in accurately and efficiently generating and utilizing building information from visual data of images, particularly in creating detailed floor plans without physical presence, and maintaining and updating these plans effectively for navigation and other applications.
Innovation Solution
The use of a trained neural network to analyze pairs of images, determine inter-image pose and co-visibility, and generate floor plans by identifying structural elements and layouts, without requiring depth sensors, using RGB pixel data for visual data analysis and combining information from multiple image pairs to create comprehensive building information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used to capture and analyze building interior information, then detailed building information can be obtained, but the process is time-consuming and requires physical presence in the building
Solution Approach 1:
The patent uses image data as a copy or representation of the physical building interior. By analyzing photographs and visual data of the building interior, the system can extract building information without needing physical presence or direct measurement in the building space, thus reducing time while maintaining accuracy
Solution Approach 2:
The patent replaces physical mechanical inspection methods with automated image analysis systems. Instead of requiring physical travel and manual measurement, the system uses computational algorithms to analyze visual data and automatically extract building information, significantly reducing time consumption
2Measurement precision
If floor plans are manually constructed and maintained, then detailed building layout information can be obtained, but the process is complex and difficult to update
Solution Approach 1:
The system automatically generates and updates floor plans by analyzing new image data without requiring manual intervention. The automated analysis process extracts building layout information from photographs and updates the floor plan representation, making the system self-maintaining and eliminating complex manual update procedures
Solution Approach 2:
The system continuously analyzes new image data and uses this feedback to automatically update and refine the floor plan representation. By comparing new visual information with existing floor plan data, the system can detect changes and update the representation accordingly, maintaining accuracy without manual complexity
3Measurement precision
If multiple image pairs are analyzed to generate comprehensive building information, then detailed floor plan data can be obtained, but computational resources and processing time increase
Solution Approach 1:
The patent divides the building into multiple rooms or areas and processes image data for each segment separately. By analyzing image pairs corresponding to different spatial segments and then integrating the results, the system can generate comprehensive building information while managing computational resources through divide-and-conquer processing
Solution Approach 2:
The system analyzes multiple image pairs beyond what might be strictly minimal, using redundant visual data to improve the robustness and completeness of building information extraction. This partial excessive action ensures comprehensive coverage and allows for cross-validation of extracted information, improving reliability
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 multiple types of building information (e.g., to include a floor plan for the building), such as by using a trained neural network to jointly generate the multiple types of building information by combining visual data from pairs of the images, and for subsequently using the generated building information in one or more further automated manners, with the building information generation further performed in some cases without having or using information from any distance-measuring devices about distances from an image's acquisition location to walls or other objects in the surrounding room. The automated operations may include generating some types of building information with respect to each image pixel column, and other types of building information by combining data from both images of a pair.