Image Synthesis Using Pupil Division for Dynamic Range Extension
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional dynamic range extension methods using pupil division images often result in unnatural images due to hand shake and noise issues, and require large cumulative time or readout gain differences, leading to regions with visible blur or noise.
Innovation Solution
An image processing apparatus that acquires and synthesizes image data from sub-pixels sharing a single micro lens, using a first acquirer for underexposure image data and a second acquirer for proper exposure image data, with an image synthesizer generating a synthesized image based on object brightness, thereby avoiding cumulative time or readout gain differences and reducing noise degradation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a larger cumulative time difference is used to extend dynamic range, then the dynamic range extension is improved, but hand shake blur increases leading to unnatural synthesized images
Solution Approach 1:
The patent changes the exposure parameter by capturing images at different exposures using pupil division, then synthesizes them based on object brightness to extend dynamic range while maintaining image naturalness through selective synthesis rather than simple addition
2Manufacturing precision
If a larger readout gain difference is used to extend dynamic range, then the dynamic range extension is improved, but noise increases leading to unnatural synthesized images
Solution Approach 1:
The patent uses different readout gains for different pupil division images, then synthesizes them based on object brightness to extend dynamic range while suppressing noise by selecting the cleaner image for each region
3Manufacturing precision
If shading correction with large correction gain is applied to images with large shading contrast-unevenness, then the shading correction is improved, but noise degradation increases
Solution Approach 1:
The patent performs shading correction on individual pupil division images before synthesis, then combines them based on object brightness, which prevents noise degradation by avoiding large correction gains on the final synthesized image
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 prevents the generation of unnatural images with visible degradation and reduces noise, achieving a natural dynamic range extension without shading correction noise, resulting in an image processing apparatus that effectively extends the dynamic range using pupil division images.
Implementation Method 1
one of a first pixel and a second pixel that share a single micro lens and receive light passing through different pupil area of an optical system
Data Source
AI summary
An image processing apparatus includes a memory that stores instructions, and one or more processors that execute the instructions to cause the image processing apparatus to function as, among other things, an image synthesis unit configured to synthesize first image data acquired by a first acquisition unit and second image data acquired by a second acquisition unit based on an object brightness, so that the first image data is selected in an area in which the object brightness is a first value that is greater than a threshold value, and the second image data is selected in an area in which the object brightness is a second value that is less than the threshold value, and to generate synthesized image data of the first image data and the second image data.