Range Measurement Using Non-Projected Region Edge Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for measuring the position and orientation of objects, especially those with a spread in the image depth direction, suffer from precision drops due to the quality of range images generated from pattern light, leading to inaccurate measurements.
Innovation Solution
A range measurement apparatus and method that senses an image projected with pattern light, specifies the projected region, searches for geometric features from non-projected regions, and derives the object's position and orientation using these features and a three-dimensional model, enhancing precision by avoiding pattern light noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pattern light is projected to generate range image, then range measurement can be obtained, but measurement precision deteriorates due to pattern light noise especially for objects with spread in depth direction
Solution Approach 1:
The patent divides the image processing into two separate processing paths: one for the pattern light projected region (for range measurement) and one for the non-projected region (for edge detection). This segmentation allows each path to be optimized independently, with the non-projected region free from pattern light noise interference, thereby resolving the contradiction between obtaining range data and maintaining measurement precision.
Solution Approach 2:
The patent extracts and removes the harmful pattern light noise by excluding the pattern light projected region from edge detection processing. By taking out the noisy region and processing only the clean non-projected region for edge-based feature extraction, the measurement precision is maintained while still utilizing range data from the projected region.
2Measurement precision
If edge extraction is performed on pattern light projected region, then edge detection precision is improved, but range image quality deteriorates
Solution Approach 1:
The patent segments the image processing by directing the pattern light projected region to range image generation while directing the non-projected region to edge extraction. This segmentation prevents the degradation of range image quality that would occur if edge extraction processing interfered with the pattern light region data.
Solution Approach 2:
The patent applies different processing qualities to different regions: the pattern light projected region undergoes range measurement processing to maintain its quality, while the non-projected region undergoes edge extraction processing. This local differentiation of processing quality ensures that each region is handled appropriately without mutual interference.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach achieves higher precision in position and orientation measurements, particularly for objects with a spread in the image depth direction, by effectively isolating and minimizing the impact of pattern light noise on the measurement process.
Implementation Method 1
an image sensing unit for sensing an image of a target object projected with pattern light
Data Source
AI summary
A region projected with pattern light in an image sensed by an image sensing unit which senses an image of a target object projected with pattern light is specified. Geometric features of the target object are searched for from a region other than the region projected with the pattern light in the sensed image based on the specified region. A position and orientation of the target object are derived based on the searched geometric features of the target object and geometric features of a model of the target object.