Image Sensor Movement for Quad Bayer Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Quad Bayer coding in image sensors reduces effective resolution, compromising camera sensitivity and dynamic range, while existing methods to address this issue, such as interpolation, introduce color artifacts and do not adequately improve image quality.
Innovation Solution
Physically moving the image sensor to different positions to capture images and generating a composite image from these positions, allowing pixels to detect different colors effectively, thereby increasing resolution without the drawbacks of Quad Bayer coding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If Quad Bayer coding is used to improve camera sensitivity and dynamic range, then sensitivity and dynamic range are improved, but effective resolution is reduced
Solution Approach 1:
The patent applies the Dynamics principle by making the image sensor physically movable between different positions. The sensor transitions from a static arrangement where pixels are fixed to specific locations, to a dynamic system where the sensor can be repositioned to capture images from multiple positions. This allows the same physical pixel to effectively sample different spatial locations at different times, thereby achieving higher effective resolution without sacrificing the sensitivity benefits of Quad Bayer coding.
Solution Approach 2:
The patent applies the Another dimension principle by adding the temporal dimension to the spatial sampling process. Instead of relying solely on the spatial arrangement of pixels in a single static image, the system captures images at multiple time points with the sensor at different positions. This temporal dimension allows the system to effectively increase the number of sampled spatial locations beyond what a single static sensor arrangement can provide, resolving the resolution limitation of Quad Bayer coding.
2Measurement precision
If physical movement of the sensor is used to increase resolution, then resolution is improved, but device complexity increases
Solution Approach 1:
The patent applies the Segmentation principle by dividing the image capture process into multiple discrete steps. Instead of attempting to achieve high resolution through a single complex static sensor design, the system segments the capture into multiple simpler shots taken at different positions. Each individual shot uses the same relatively simple Quad Bayer sensor, but the sequence of segmented shots collectively achieves the high resolution outcome when combined through de-mosaicing algorithms.
3Loss of information
If interpolation is used to reproduce color in under-sampled pixels, then color information is recovered, but color artifacts are introduced
Solution Approach 1:
The patent applies the Preliminary action principle by capturing additional color information through physical sensor movement before the final image processing occurs. Instead of relying on post-capture interpolation to estimate missing color data, the system preliminarily acquires the actual color information by positioning the sensor to capture images where different pixels sample different locations. This preliminary data collection provides authentic color measurements that eliminate the need for artifact-prone interpolation.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for color imaging. In one aspect, a method includes obtaining, by an image sensor that includes pixels for detecting a first color and pixels for detecting a second color, a first image of a scene while the image sensor is in a first position, moving the image sensor to a second position, wherein, in the second position, a particular pixel for detecting the first color is located where a particular pixel for detecting the second color was previously located when the image sensor was in the first position, obtaining, by the image sensor, a second image of the scene while the image sensor is in the second position, generating a composite image based on the first image and the second image, and providing the composite image for output.