Room Segment Refinement Using Polygon Mesh Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating aesthetically pleasing room segments from noisy and incomplete 2D LiDAR scan maps are inadequate, as they often result in noisy, incomplete, and unsuitable outputs with irregular boundaries, which are not suitable for user-facing applications.
Innovation Solution
A method involving a processor-based pipeline to refine room segments by determining room polygons, generating a polygon mesh, and aligning edges to produce clean, smooth, and properly aligned boundaries, suitable for user interfaces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If a machine learning-based segmentation model is used to segment the 2D LiDAR scan map, then the segmentation can be performed automatically, but the output remains noisy, incomplete, and not suitable for user-facing applications
Solution Approach 1:
The refinement process segments the polygon mesh into individual room polygons, then processes each polygon's boundary independently through vertex simplification and alignment operations, finally reassembling them into refined room segments
Solution Approach 2:
The patent introduces an intermediary processing pipeline between the machine learning segmentation model and the final user-facing output. This pipeline includes polygon extraction, mesh generation, vertex simplification, and alignment operations that transform the raw model output into clean, aesthetically pleasing room segments
2Device complexity
If polyline simplification methods are applied to individual polylines, then the boundary complexity is reduced, but the topology of the polygon mesh is broken and edges of adjacent polygons are disrupted
Solution Approach 1:
The patent merges adjacent polygons into a unified polygon mesh structure before applying simplification operations. This allows vertex simplification to be performed across polygon boundaries while maintaining topological integrity, as shared vertices are preserved and adjacent edges are coordinated through the mesh structure
Solution Approach 2:
The patent transitions from operating on individual 2D polylines to operating on a connected polygon mesh structure. By elevating the operational dimension from single-boundary to multi-boundary mesh, the system can simplify vertices while maintaining adjacency relationships through the mesh's topological constraints
3Productivity
If the direct output of the segmentation model is used, then the processing time is minimized, but the output contains small or ill-shaped segments that are not suitable for user display
Solution Approach 1:
The patent performs preliminary filtering and evaluation of room polygons before final output generation. Small segments are identified and merged with adjacent polygons, and ill-shaped segments are evaluated and refined through vertex alignment operations, ensuring only aesthetically suitable segments reach the user interface
Data Source
AI summary
The disclosure presents an efficient method for improving the aesthetic quality of noisy room segments corresponding to a scanned environment, such as those used by mobile robots to navigate an environment to perform a task. Room polygons are extracted from the noisy room segments and visual scores of the extracted room polygons are assessed. Based on the visual score, low scoring room polygons are removed or merged. A polygon mesh is formed from the room segments, which is then simplified and aligned using efficient methods. Finally, room segments and boundaries between rooms are recovered to generate refined room segments from the polygon mesh. Compared to the noisy room segments, the refined room segments are better suited for visual presentation to users, such as in mobile applications for operating such mobile robots.