Indoor 3D Mapping Using Depth Priors Instead of LiDAR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing indoor mapping systems face challenges such as high costs, power demands, bulkiness, and complexity due to reliance on LiDAR sensors, and struggle with occlusions and incomplete data capture using depth cameras, while integration of architectural blueprints requires sophisticated algorithms and computational resources beyond real-time capabilities.
Innovation Solution
A system utilizing RGB cameras and self-supervised learning, combined with geometric and semantic cues, generates a depth prior to refine a depth prediction model, enabling real-time construction of a continuous three-dimensional map by leveraging lightweight sensors and pseudo supervision for adaptability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If LiDAR sensors are used for depth measurement, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent uses depth camera data to create a simplified copy of the depth information that LiDAR would provide, processing this copied data through the monocular depth estimation model to achieve accurate depth measurements without requiring actual LiDAR hardware
Solution Approach 2:
The patent replaces expensive, complex LiDAR sensors with cheaper depth cameras and computational models, using software-based depth estimation as a disposable alternative to expensive hardware solutions
2Measurement precision
If multiple sensors are combined to improve coverage and accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent merges depth camera data with monocular depth estimation model predictions into a unified depth map, combining multiple information sources computationally rather than physically combining multiple sensors
Solution Approach 2:
The system uses a single depth camera that serves multiple functions: capturing visual imagery, providing depth data, and serving as input for the monocular depth estimation model, eliminating the need for separate specialized sensors
3Manufacturing precision
If architectural blueprints are integrated into the mapping system, then manufacturing precision is improved, but device complexity increases
Solution Approach 1:
The patent performs preliminary alignment between the captured environment and architectural blueprints before final mapping, pre-processing the blueprint data to match the captured spatial coordinates and scale, thereby simplifying subsequent integration steps
Data Source
AI summary
A method of and system for generating a three-dimensional map of an environment can include obtaining a first visual data set, generating a depth prior based on the first visual data set, refining a depth prediction model based on the depth prior, generating a layout based on a refined depth prediction model, and constructing a continuous three-dimensional map of the environment based on the layout and aggregated depth measurements. The visual data set can include visual imagery data and depth data. The depth prior can include geometric cues and semantic cues