AR Information Processing Device Structural Feature Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In augmented reality, the generation of environment maps with noise from real-world measurements leads to decreased matching accuracy and increased calculation requirements when recognizing the correspondence between real and reference spaces.
Innovation Solution
An information processing device that acquires structural feature points from captured images and generates coordinate transformation information by matching structure data, reducing noise impact and calculation complexity through the use of structural feature points and their classification information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If environment map is generated by measuring the real world with a camera, then the real world space can be captured, but noise from non-recognition target objects is included which decreases matching accuracy and increases calculation amount
Solution Approach 1:
The invention extracts only the necessary structural feature points from the environment map that are relevant to the recognition target, separating them from noise caused by non-recognition target objects. This extraction process removes harmful information while preserving useful spatial structure data, thereby improving matching accuracy without requiring excessive calculation resources.
Solution Approach 2:
The environment map processing is segmented into distinct steps: first identifying structural feature points, then extracting only those relevant to the recognition target. This segmentation allows the system to handle large amounts of environmental data efficiently by processing only the necessary portions, reducing overall calculation complexity while maintaining measurement precision.
2Loss of information
If environment map includes all objects in the real world, then complete spatial information is obtained, but noise from non-recognition target objects decreases matching accuracy
Solution Approach 1:
The invention extracts only the structural feature points related to the recognition target from the complete environment map, removing noise from non-recognition target objects. This selective extraction maintains the essential spatial information needed for accurate matching while eliminating detrimental noise that would otherwise degrade matching precision.
3Adaptability or versatility
If feature data from all objects is used for matching, then comprehensive spatial recognition is achieved, but calculation amount increases significantly
Solution Approach 1:
The invention extracts only the structural feature points relevant to the specific recognition target from the comprehensive environment map, rather than processing feature data from all objects. This extraction maintains the system's ability to perform spatial recognition accurately while significantly reducing the calculation amount by focusing only on necessary data.
Solution Approach 2:
The matching process is segmented to first identify structural feature points and then selectively process only those relevant to the recognition target. This segmentation enables comprehensive spatial recognition capability while managing calculation resources efficiently by avoiding processing of irrelevant object data.
Data Source
AI summary
The information processing device 4 includes an acquisition unit 41A and a structure matching unit 43A. The acquisition unit 41A is configured to acquire, from a captured image captured by a photographing unit 15A of a display device configured to display a virtual object superimposed on a view, multiple combinations of classification information of a structural feature point that is a feature point, in structure, of a target structure and position information of the structural feature point. The structure matching unit 43A is configured to generate coordinate transformation information by matching structure data “Ds” with the multiple combinations, the structure data Ds including information associated with a position and a class of each of structural feature points of the target structure, the coordinate transformation information relating to a coordinate transformation between a first coordinate system that is a coordinate system referred to by the display device and a second coordinate system that is a coordinate system adopted in the structure data Ds.