Inter-Image Building Analysis for Floor Plans 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, particularly in generating accurate floor plans and navigating buildings, as they often require depth sensors and are computationally intensive.
Innovation Solution
An automated system using a Pairwise Image Analyzer (PIA) and Graph Neural Network-Based Analyzer (GNNBA) processes overlapping visual data from multiple images to generate building information, including a floor plan, without depth sensors, by analyzing pairwise and group image overlaps to determine structural elements and global pose, using a multi-layer graph neural network.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If depth sensors are used to capture building interior information, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent extracts and removes the depth sensor component from the system, achieving building interior mapping using only visual data from standard cameras. This eliminates the need for complex depth sensing hardware while maintaining measurement capabilities through image analysis alone.
Solution Approach 2:
The patent replaces mechanical/optical depth sensing systems with a computational vision system that uses image processing and neural networks to infer spatial information. This substitutes physical measurement devices with algorithm-based measurement approaches.
2Manufacturing precision
If traditional floor plan construction methods are used, then manufacturing precision is improved, but productivity and time consumption worsen
Solution Approach 1:
The patent performs preliminary actions by capturing multiple images of the building interior before processing. These images are pre-captured with sufficient overlap and coverage, enabling subsequent automated processing to generate floor plans rapidly without manual measurement during the construction phase.
Solution Approach 2:
The system performs self-service through automated image analysis and floor plan generation. The neural network automatically processes images, identifies structural elements, and constructs floor plans without requiring manual intervention, thereby dramatically increasing productivity while maintaining precision.
3Loss of information
If comprehensive building information is captured, then information completeness is improved, but computational resources and processing time increase
Solution Approach 1:
The patent segments the building interior into multiple overlapping image regions that are processed independently. By dividing the comprehensive capture task into manageable image segments, the system reduces computational complexity while maintaining overall information completeness through the aggregation of segmented results.
Solution Approach 2:
The patent uses partial action by capturing images with sufficient overlap beyond what a single comprehensive view would provide. This excessive sampling in the form of overlapping images enables more efficient processing through parallelization and reduces the need for complex global processing of entire building interiors.
4Ease of operation
If manual floor plan construction is performed, then ease of operation is improved, but productivity worsens
Solution Approach 1:
The system performs self-service by automatically generating floor plans from captured images without requiring manual construction. The neural network autonomously processes visual data, identifies walls and structural elements, and produces floor plans, thereby maintaining operational simplicity while dramatically increasing productivity.
Solution Approach 2:
The patent replaces manual mechanical floor plan drawing with automated computational processes. This substitution eliminates manual labor while maintaining ease of operation through automated systems that require minimal user intervention, thereby resolving the contradiction between simplicity and efficiency.
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 one or more types of building information (e.g., global inter-image pose data, a floor plan for the building, etc.), 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.