Phase Difference Pixel Interpolation Accuracy in Imaging Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing imaging elements face challenges in maintaining high interpolation accuracy for phase difference detection pixels, especially when densely arranged, leading to decreased autofocus performance due to differences in pixel characteristics and arrangements.
Innovation Solution
An imaging apparatus with a two-dimensional arrangement of phase difference detection pixels and normal pixels, where the phase difference detection pixels have adjacent opening portions for pupil separation, uses pixel value addition and interpolation units to generate accurate pixel values, and adjusts these values using level correction and signal gradient calculations to enhance interpolation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If phase difference detection pixels are densely arranged in the imaging element, then autofocus speed and detection capability are improved, but interpolation accuracy decreases due to differences in pixel characteristics between normal pixels and phase difference pixels
Solution Approach 1:
The patent changes the pixel arrangement pattern by creating a 2×2 block structure where phase difference pixels and normal pixels are grouped together. This structural parameter change allows the system to maintain high-density phase difference pixel arrangement for fast autofocus while preserving sufficient normal pixels for accurate interpolation, thereby resolving the contradiction between autofocus speed and interpolation accuracy
Solution Approach 2:
The imaging element is segmented into multiple 2×2 pixel blocks, each containing both phase difference pixels and normal pixels. This segmentation allows independent optimization of autofocus functionality (using phase difference pixels) and image quality (using normal pixels for interpolation), solving the contradiction by separating the two functions into distinct pixel groups within each block
2Measurement precision
If phase difference pixels are arranged at constant intervals on the image surface, then phase difference detection capability is improved, but device complexity increases due to special arrangement requirements and correction needs
Solution Approach 1:
The 2×2 pixel block design makes each block a universal unit that simultaneously provides phase difference detection capability (through phase difference pixels) and normal imaging capability (through normal pixels). This multi-functionality reduces overall device complexity by eliminating the need for separate correction circuits and complex arrangement patterns, as each block is self-sufficient
Solution Approach 2:
The patent merges phase difference pixels and normal pixels into the same 2×2 block structure, combining two previously separate pixel types into a unified arrangement. This merging simplifies the overall pixel layout by eliminating the need for separate correction mechanisms and complex constant-interval arrangements, reducing device complexity while maintaining detection capability
Data Source
AI summary
An imaging apparatus and an image processing method capable of increasing correction accuracy of a phase difference detection pixel even in a case where the phase difference detection pixel is densely arranged in an imaging element in order to secure AF performance are provided. An imaging element includes normal pixels of RGB and first and second phase difference pixels of which opening portions are adjacently arranged to face each other in a horizontal direction and in which a G filter is arranged. A pixel value addition unit (64) generates an addition pixel corresponding to a virtual G pixel at a pixel position between the first and second phase difference pixels by adding pixel values of the pair of the first and second phase difference pixels. In a case where the first or second phase difference pixel is set as an in-focus pixel that is an interpolation target, an average value interpolation unit (62) uses the normal pixels surrounding a pixel position of the in-focus pixel and the addition pixel in a case of performing an interpolation operation on a pixel value at the pixel position of the in-focus pixel.