Array Camera Parallax Convergence Sub-Pixel Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional array cameras suffer from suboptimal image resolution, which limits their commercial applicability, especially in capturing images involving motion.
Innovation Solution
An array camera system paired with an image processor that leverages parallax convergence differences between cameras to create a sub-pixel sampling pattern, increasing resolution by at least three times through predefined offset points within a Hilbert sampling pattern.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional array cameras are used to capture images from multiple perspectives, then motion capture capability is improved, but image resolution deteriorates
Solution Approach 1:
The image processing is segmented into multiple passes: first parallax convergence to create intermediate images at different offset points, then super-resolution processing to combine these intermediates into a final high-resolution image. This multi-stage segmentation allows each processing stage to focus on specific aspects of resolution enhancement.
Solution Approach 2:
The patent processes images in the sub-pixel domain by creating intermediate images at different offset points (e.g., 1/4 pixel offsets vertically, horizontally, and diagonally). This dimensional approach to sampling below the pixel grid enables resolution enhancement beyond the original sensor limitations.
2Manufacturing precision
If multiple cameras are used to increase resolution, then image resolution can be improved, but device complexity increases
Solution Approach 1:
Instead of using many physical cameras, the system creates virtual copies of camera views through computational processing. Intermediate images are synthesized at different offset points by processing data from existing cameras, effectively creating additional virtual camera positions without adding physical hardware.
Solution Approach 2:
The patent changes the sampling parameters by processing images at sub-pixel offset points (e.g., 1/4 pixel offsets in vertical, horizontal, and diagonal directions). This parameter transformation in the image processing domain achieves resolution enhancement without requiring proportional increases in physical camera count.
3Manufacturing precision
If parallax convergence is applied at multiple offset points to create sub-pixel sampling, then resolution enhancement is achieved, but processing complexity increases
Solution Approach 1:
The complex super-resolution processing is divided into manageable segments: first performing parallax convergence at each offset point to create intermediate images, then performing the computationally intensive super-resolution combining step. This segmentation makes the overall complex process more tractable and efficient.
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
The solution enhances image resolution significantly, achieving at least three times the original resolution by effectively combining image data from multiple cameras, improving the camera's ability to capture detailed images, especially in motion scenarios.
Implementation Method 1
make use of parallax convergence differences between different cameras or groups of cameras within an array camera
Data Source
AI summary
Aspects of the disclosed technology relate to an imaging system and method in which an array camera is employed along with an image processor to make use of parallax convergence differences between different cameras or groups of cameras within an array camera to provide improved super resolution performance.