Pseudo-Random Filter for High Dynamic Range Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image reconstruction techniques using spatially regular filters result in loss of spatial resolution and high-frequency components due to grid-like sampling, leading to suboptimal image quality, especially in high dynamic range (HDR) imaging.
Innovation Solution
The use of a spatially pseudo-randomly ordered filter with multiple intensity attenuation levels allows for the acquisition of pseudo-randomly filtered images, which can be reconstructed into high-quality HDR images using mathematical techniques like dictionary learning and ℓ1/TV minimization, leveraging sparsity and incoherency properties to achieve higher resolution and dynamic range.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a spatially regular filter is used for image acquisition, then the filter structure is simple and easy to manufacture, but the reconstructed image loses spatial resolution and high spatial frequency components
Solution Approach 1:
The patent applies asymmetry by replacing the symmetric, regular grid pattern of conventional filters with an asymmetric pseudo-random pattern. This pseudo-random arrangement of filter elements breaks the periodicity that causes aliasing and resolution loss, while maintaining manufacturability through computational design of the filter pattern.
Solution Approach 2:
The patent changes the spatial distribution parameter of the filter from a regular periodic arrangement to a pseudo-random arrangement. This parameter change transforms the sampling pattern in a way that preserves high spatial frequency information while maintaining the filter's physical manufacturability.
2Reliability
If multiple exposures of different exposure times are taken, then high dynamic range is achieved, but image degradation occurs especially for moving objects
Solution Approach 1:
The patent applies preliminary action by encoding multiple intensity levels (including different exposure information) into a single filter pattern that is applied before the exposure. This allows the system to capture HDR information in a single exposure rather than requiring multiple sequential exposures, thereby avoiding motion-related image degradation.
Solution Approach 2:
The patent replaces the mechanical approach of taking multiple sequential exposures with a computational approach using a pseudo-random filter pattern. Instead of physically moving the camera or changing exposure settings between shots, the system uses a designed filter pattern combined with computational reconstruction to achieve HDR imaging in a single static exposure.
3Reliability
If multiple sensors and beam splitters are used, then high dynamic range imaging is achieved, but device cost increases
Solution Approach 1:
The patent applies universality by designing a single filter pattern that serves multiple functions simultaneously - it encodes multiple intensity levels, enables HDR reconstruction, and works with a standard single sensor array. This eliminates the need for multiple dedicated sensors and beam splitters, reducing system complexity and cost while maintaining HDR capability.
Solution Approach 2:
The patent merges the functions of multiple sensors and beam splitters into a single integrated system consisting of one sensor array and one pseudo-random filter pattern. By combining what would traditionally require separate optical paths and multiple sensors into a single computational framework, the system reduces component count and overall system cost.
4Reliability
If FPA with multiple sized pixels is fabricated, then high dynamic range is achieved, but spatial resolution decreases
Solution Approach 1:
The patent applies local quality by using a uniform sensor array where all pixels have the same size and characteristics, but the pseudo-random filter pattern applied to them creates local variations in the captured information. This allows different regions of the image to be reconstructed with appropriate detail while maintaining overall high spatial resolution, unlike approaches that use physically different sized pixels.
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 enables the reconstruction of high-quality images with higher spatial resolution and dynamic range from a single exposure using a single sensor array, reducing costs and avoiding image degradation, while maintaining high speed and accuracy.
Implementation Method 1
The use of a spatially pseudo-randomly ordered filter with multiple intensity attenuation levels allows for the acquisition of pseudo-randomly filtered images
Data Source
Figure 1
Figure 2~4
Figure 5
AI summary
An image processing system comprises a filter 40 associated with a sensor array 20, which is operable to capture an image. The filter is provided to separate a plurality of distinct qualities of light from the scene to be captured. The filter has filter portions associated with the plurality of distinct qualities of light which are spatially pseudo-randomly ordered relative to each other. The image processing system also comprises an image reconstruction algorithm specifically designed to operate with the filter.