Oriented Bounding Box Heading Angle Determination via Lattice Map Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for determining the orientation of an object using oriented bounding boxes (OBBs) face challenges such as increased computation requirements, skewing of external shapes due to curved objects, and inaccurate extraction of heading angles.
Innovation Solution
A method that generates pixel data by creating a lattice map of unit pixels from a point cloud, calculates candidate box costs based on center coordinates and point inclusion, and selects the candidate box with the smallest cost to determine the heading angle of an OBB, thereby generating an oriented bounding box that accurately represents the object's orientation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If oriented bounding box is generated using points of point cloud, then orientation information can be obtained, but computational complexity increases
Solution Approach 1:
The patent segments the continuous point cloud data into discrete lattice map cells (unit pixels). Each cell aggregates multiple points into a single representative structure, transforming the problem from processing individual points to processing grid cells. This segmentation reduces computational complexity while preserving orientation information through the lattice structure that maintains spatial relationships.
Solution Approach 2:
The patent introduces a lattice map as an intermediary data structure between the raw point cloud and the oriented bounding box generation. The lattice map serves as a mediator that organizes points into a structured grid format, enabling more efficient computation of orientation parameters without directly processing the raw point cloud data.
2Measurement precision
If oriented bounding box is generated using points, then orientation can be determined, but external shape becomes skewed by curved object surfaces
Solution Approach 1:
By dividing the point cloud into lattice map cells, the patent creates discrete geometric units that represent local regions of the object. This segmentation allows the system to determine orientation based on the distribution of lattice cells rather than being influenced by local curved surface variations, thereby preventing shape skewing while maintaining accurate heading angle extraction.
3Reliability
If conventional point-based OBB generation is used, then orientation information is obtained, but heading angle extraction becomes inaccurate
Solution Approach 1:
The lattice map segmentation groups points into cells that represent stable geometric units. This grouping improves the reliability of heading angle extraction by reducing sensitivity to individual point variations, while the regular grid structure of the lattice map provides computational efficiency compared to processing raw point clouds.
Solution Approach 2:
The patent changes the fundamental parameter representation from individual point coordinates to lattice cell indices and aggregated point counts. This parameter transformation enables more stable and reliable orientation determination by working with discretized, binned data that is less sensitive to measurement noise and variations.
Data Source
AI summary
An object orientation determination method includes generating pixel data on each of a plurality of unit pixels included in a region of interest of a point cloud acquired from an object, generating a plurality of candidate boxes using the generated pixel data, and determining, as a heading angle of an oriented bounding box, an inclination of a candidate box having a smallest cost among costs calculated on the plurality of candidate boxes. A cost of each of the plurality of candidate boxes is calculated based on positions of respective sides of a corresponding one of the plurality of candidate boxes and the pixel data.