3D Point Cloud Fitting via Reference Plane Projection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques require a reference object model for fitting three-dimensional point cloud data, limiting flexibility as shapes must be similar or identical for accurate alignment, which is not feasible in all scenarios.
Innovation Solution
An information processing method that acquires reference three-dimensional model data and corresponding data from a sensor, sets a reference search range based on specific coordinate information, and retrieves corresponding coordinate information within that range, enabling fitting without direct use of a reference object model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If three-dimensional point cloud data is used directly for fitting with a reference object model, then the fitting process requires shape similarity between models, but this limits the adaptability to objects with different shapes
Solution Approach 1:
The patent introduces a reference plane as an intermediary element between the actual object model and the reference object model. Instead of directly comparing three-dimensional point cloud data which requires shape similarity, the system extracts two-dimensional coordinate information from both models based on a reference plane, enabling comparison and fitting without requiring the objects to have similar three-dimensional shapes.
Solution Approach 2:
The patent transforms the fitting problem from three-dimensional space to two-dimensional space by extracting coordinate information based on a reference plane. This dimensionality reduction allows comparison of objects with different three-dimensional shapes while maintaining fitting accuracy through the two-dimensional projection.
2Loss of information
If three-dimensional point cloud data is used for fitting, then complete shape information is preserved, but the complexity of finding corresponding points increases significantly
Solution Approach 1:
The patent extracts only the necessary coordinate information from the three-dimensional point cloud data based on a reference plane. Instead of processing all three-dimensional point data to find corresponding points, the system extracts two-dimensional coordinate information, significantly reducing the complexity of the fitting process while retaining sufficient shape information for accurate comparison.
3Manufacturing precision
If shape similarity is required for fitting, then accurate alignment can be achieved, but the method cannot handle objects with different shapes
Solution Approach 1:
The reference plane serves as a mediator that enables accurate alignment without requiring shape similarity. By projecting both the actual object and reference object onto the same reference plane, the system achieves accurate alignment through two-dimensional coordinate comparison, while accommodating objects with different three-dimensional shapes.
Data Source
AI summary
An objective of the present invention is to substantially enable fitting without using a reference object model being three-dimensional point cloud data directly for fitting and enable comparison between an actual object model and the reference object model. There is provided an information processing method that is a principal aspect of the present invention, the information processing method including: a step of acquiring reference three-dimensional model data from a reference object with a sensor; a step of acquiring, by a reference-point-coordinate acquisition unit, any number of pieces of reference-point-coordinate information from the reference three-dimensional model data, the any number being three or more; a step of acquiring corresponding three-dimensional model data from a measurement object with the sensor; a step of setting, by a reference-search-range setting unit, a reference search range based on the reference-point-coordinate information; and a step of acquiring, by a corresponding-point-coordinate-information acquisition unit, corresponding-point-coordinate information within the reference search range, in the corresponding three-dimensional model data.