3D Shape Measurement Using Stripe Patterns for Faster Pixel Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing three-dimensional shape measuring methods using active stereo techniques face computational complexity and long computing times due to the use of sparse estimation for Light Transport Matrix (LTM) calculations, particularly when dealing with complex light reflections, and lack parallel calculation capabilities.
Innovation Solution
A method involving projecting row-direction and column-direction stripe patterns on a target object, followed by sparse optimization to identify corresponding pixel rows and columns, and applying an epipolar constraint to extract appropriate projection pixels, thereby reducing computational complexity and enabling faster 3D shape measurement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sparse estimation is used to calculate Light Transport Matrix for handling complex light reflections, then measurement precision is improved, but computing time increases significantly
Solution Approach 1:
The patent segments the LTM calculation process into two independent stages: first calculating row-direction LTMs using row stripe patterns, then calculating column-direction LTMs using column stripe patterns. This segmentation allows each stage to be processed independently and enables parallel computation, significantly reducing the overall computing time while maintaining measurement precision for complex light reflections.
Solution Approach 2:
The patent employs periodic stripe patterns (row stripes and column stripes) that are projected sequentially onto the target object. By using these periodic patterns in alternating sequences, the system can capture multiple reflected light components over time and reconstruct the complete LTM through periodic sampling, thereby improving precision without requiring excessively long continuous computation.
2Reliability
If random patterns are used for LTM calculation, then comprehensive light reflection information is captured, but computational complexity increases
Solution Approach 1:
The patent changes the pattern parameters from random values to structured stripe patterns (row-direction and column-direction stripes). This parameter change transforms the LTM calculation into a more efficient form where the structured patterns enable direct mapping between projection pixels and captured pixels, reducing computational complexity while still capturing comprehensive light reflection information through the systematic coverage of the target surface.
3Productivity
If multiscale LT Matrix estimation is used to reduce computational complexity, then calculation efficiency is improved, but parallel calculation capability is lost
Solution Approach 1:
The patent segments the LTM calculation into independent row-direction and column-direction components that can be computed in parallel. Each segmentation operates on separate data sets (row stripe captures and column stripe captures) and produces independent intermediate results that are subsequently combined, thereby maintaining parallel calculation capability while achieving the efficiency gains of reduced computational complexity.
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 significantly shortens computing time for 3D shape measurement by utilizing stripe patterns and sparse optimization, allowing for efficient identification of corresponding pixels and reducing measurement errors from indirect reflections.
Implementation Method 1
a projector configured to project projection images
Implementation Method 2
captures an image of the target object on which the projection image is projected
Implementation Method 3
obtains respective 3D points by using the principle of the triangulation
Data Source
AI summary
A three-dimensional shape measuring method includes: a first step to: project first and second projection images on a target object, wherein the first projection images include row-direction stripe patterns different from each other, and the second projection images include column-direction stripe patterns different from each other, and obtain first and second captured images of the target object, wherein the first captured images are captured while the first projection images are projected, and the second captured images are captured while the second projection images are projected; a second step to identify a corresponding projection pixel of the first and second projection images, wherein the corresponding projection pixel corresponds to each of captured pixels of the first and second captured images; and a third step to identify a 3D shape of the target object based on a result of the second step.