Inter-Image Floor Plan Generation Without Depth Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to effectively capture, represent, and utilize building interior information without physical travel, and floor plans are difficult to construct, scale, and maintain for accurate navigation and visualization.
Innovation Solution
An automated system analyzes multiple images of a building using a Pairwise Image Analyzer (PIA) and Bundle Adjustment Pipeline Analyzer (BAPA) to generate a floor plan without depth sensors, determining wall positions and shapes by analyzing pixel columns and using bundle adjustment techniques to refine camera poses and structural elements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional floor plan construction methods are used, then building interior information can be captured, but the process is difficult, time-consuming, and requires physical travel to the building
Solution Approach 1:
The patent replaces manual mechanical floor plan construction with an automated computer vision system. The system uses machine learning models to automatically detect walls, doors, windows, and rooms from building images, eliminating the need for physical travel and manual drafting. This substitution of mechanical processes with automated computational methods directly resolves the contradiction between productivity improvement and time loss.
Solution Approach 2:
The patent creates digital copies of building interiors through image processing and 3D reconstruction. By capturing building images and generating virtual floor plans that replicate the physical space, the system eliminates the need for physical presence while maintaining accurate representation of the building interior, thus resolving the time loss issue while improving productivity.
2Measurement precision
If depth sensors are used to capture building interior information, then measurement precision improves, but device complexity and cost increase
Solution Approach 1:
The patent extracts depth and spatial information from standard 2D building images using computer vision algorithms, eliminating the need for depth sensors. The system processes image data to infer three-dimensional spatial relationships, wall positions, and room layouts, achieving measurement precision without adding device complexity or cost.
Solution Approach 2:
The patent introduces machine learning models as intermediaries between standard images and spatial measurements. These models act as computational mediators that translate 2D image data into accurate 3D building information, achieving precision comparable to depth sensors while avoiding the complexity and cost of actual depth-sensing hardware.
3Reliability
If detailed building interior information is captured, then navigation accuracy improves, but data processing complexity and computational resources increase
Solution Approach 1:
The patent segments the building interior analysis into distinct components: wall detection, door identification, window recognition, and room classification. This segmentation allows the system to process detailed information in manageable stages, improving navigation accuracy through comprehensive analysis while controlling computational complexity through structured processing pipelines.
Solution Approach 2:
The patent performs preliminary processing of building images to extract and organize spatial information before navigation tasks. By pre-processing images to identify and catalog structural elements (walls, doors, windows) and generate floor plan representations in advance, the system reduces real-time computational requirements while maintaining high navigation accuracy.
Data Source
Figure 1
Figure 2A
Figure 2B
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 simultaneously or otherwise concurrently analyzing groups of three or more images having at least pairwise visual overlap between pairs of those images to determine information that includes global inter-image pose and structural element locations, 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.