Sparse Color Image Sensor With Multi-Frame Color Reconstruction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Electronic image sensors with color filter arrays (CFAs) face inefficiencies in light capture, leading to reduced resolution, increased noise, and decreased image quality, particularly in low-light conditions, due to wavelength-specific filters blocking significant light and causing chromatic aberrations.

Innovation Solution

A sparse color filter array (CFA) with a majority of panchromatic pixels and a minority of wavelength-filtered pixels is used to capture multiple frames, which are then computationally aligned and merged to generate a single color image, undersampling chromatic data and compensating for spatial offsets.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If a Color Filter Array (CFA) with wavelength-specific filters is used to capture chromatic information, then color image creation is enabled, but optical efficiency is significantly reduced and light reception is blocked

Engineering Contradiction:
Improvechromatic informationVSAvoidlight reception
Core Design Contradiction:
Loss of informationVSLoss of energy

Solution Approach 1:

The CFA is segmented into multiple subunits, each containing a mix of wavelength-filtered pixels and panchromatic pixels. This segmentation allows different regions to capture different types of information (chromatic and luminance) simultaneously, resolving the contradiction between capturing color information and maintaining light efficiency.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different pixels within the CFA have different local qualities - some pixels have wavelength-specific filters for chromatic information while others are panchromatic for luminance information. This local differentiation allows the system to capture both color and light efficiency in different locations, resolving the overall contradiction.

Inventive Principle:
Principle #3Local quality

2Loss of information

If wavelength-specific filters are disposed on individual pixels in a repeating pattern, then chromatic information is captured, but spatial resolution is reduced and noise increases

Engineering Contradiction:
Improvechromatic informationVSAvoidspatial resolution
Core Design Contradiction:
Loss of informationVSManufacturing precision

Solution Approach 1:

The CFA structure is divided into subunits with specific patterns of filtered and panchromatic pixels. This segmentation creates a more efficient sampling pattern that improves spatial resolution while maintaining chromatic information capture, compared to traditional repeating patterns.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces temporal dimension by capturing multiple image frames over time. This allows the system to overcome spatial sampling limitations by combining information from multiple temporal instances, thereby improving effective spatial resolution while maintaining chromatic accuracy.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Manufacturing precision

If multiple image frames are captured and processed computationally, then image resolution and quality are enhanced, but processing complexity increases

Engineering Contradiction:
Improveimage resolutionVSAvoidprocessing complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The CFA is pre-configured with specific subunit patterns that optimize the capture of both chromatic and luminance information. This preliminary structural arrangement reduces the complexity of subsequent processing by ensuring that complementary information is captured in an organized manner from the outset.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

Multiple identical or similar subunit patterns are replicated across the CFA structure. This copying approach allows the system to capture redundant information in different locations, which simplifies processing by providing multiple sources for the same type of data, making the computational merging more straightforward.

Inventive Principle:
Principle #26Copying

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 enhances image resolution and detail while maintaining high light sensitivity, reducing noise, and improving image quality by computationally combining chromatically undersampled and shifted information from multiple frames.

Implementation Method 1

each subunit including a minority of one or more wavelength-filtered adjacent pixels and a majority of remaining panchromatic pixels

Methodology Applied
Scientific EffectWavelength filtering: Filter (optical)

Implementation Method 2

the majority of pixels of each subunit are panchromatic pixels... Whereas the panchromatic pixels provide luminance information for a captured image frame

Methodology Applied
Scientific EffectPanchromatic detection:

Implementation Method 3

electronic image sensors in mobile and other imaging systems typically include a Color Filter Array (CFA)... To capture chromatic information

Methodology Applied
Scientific EffectPhotoelectric effect: Photoelectric Effect

Data Source

PatentEP4281927B1Sparse color image sensor system
Publication Date: 2025.07.02 GOOGLE LLC
  • EP4281927B1 patent drawingFigure 1
  • EP4281927B1 patent drawingFigure 2
  • EP4281927B1 patent drawingFigure 3

AI summary

Systems and techniques are described for generating a color image by computationally combining chromatically undersampled and shifted information present in multiple component image frames captured via a sparse color filter array that includes a minority of wavelength-filtered picture elements and a remaining majority of panchromatic picture elements. A burst capture is initiated of multiple image frames via a color filter array comprising a plurality of subunits, each subunit including a minority of one or more wavelength-filtered adjacent pixels and a majority of remaining panchromatic pixels. Each of the multiple image frames is processed to generate a resulting color image.