CMOS Dual-Photodiode Pixel Segmentation for High Dynamic Range Imaging

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Digital imaging systems, particularly CMOS sensors in mobile devices, face challenges in capturing outdoor scenes due to insufficient dynamic range, leading to loss of detail in images with high contrast environments, and existing methods to extend dynamic range often result in motion artifacts, resolution loss, or increased memory requirements.

Innovation Solution

Incorporating dual-diode pixels with different integration times in CMOS image sensors, where a larger photodiode with longer integration time captures details in dark areas and a smaller photodiode with shorter integration time captures details in bright areas, allowing for increased dynamic range without significant memory requirements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If dual-frame approaches with different exposure times are used to extend dynamic range, then detail in both bright and dark regions is improved, but motion artifacts increase and full frame memory is required

Engineering Contradiction:
Improvedetail captureVSAvoidmotion artifacts
Core Design Contradiction:
Measurement precisionVSObject-generated harmful factors

Solution Approach 1:

The pixel array is segmented into first and second groups of pixels, where each group captures images with different exposure times simultaneously. This spatial segmentation eliminates the need for temporal sequencing, thereby removing motion artifacts while maintaining the ability to capture detail in both bright and dark regions.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from temporal dimension (sequential frame capture) to spatial dimension (simultaneous pixel group capture). By assigning different exposure times to different spatial groups of pixels rather than capturing frames sequentially, the system extends dynamic range without introducing motion artifacts.

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

2Measurement precision

If dual-frame approaches with different exposure times are used to extend dynamic range, then detail in both bright and dark regions is improved, but memory requirements increase

Engineering Contradiction:
Improvedetail captureVSAvoidmemory requirements
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The pixel array is segmented into first and second groups of pixels, where each group captures images with different exposure times simultaneously. This spatial segmentation eliminates the need for temporal sequencing, thereby removing motion artifacts while maintaining the ability to capture detail in both bright and dark regions.

Inventive Principle:
Principle #1Segmentation

3Object-generated harmful factors

If different integration times are implemented within the same readout frame with row-based segmentation, then motion artifacts are minimized and memory requirements are reduced, but vertical resolution is lost

Engineering Contradiction:
Improvemotion artifactsVSAvoidvertical resolution
Core Design Contradiction:
Object-generated harmful factorsVSManufacturing precision

Solution Approach 1:

The patent segments the pixel array into first and second groups that are interleaved throughout the array, rather than dividing the array into separate row bands. This fine-grained segmentation allows each pixel group to contribute to the complete image, preserving vertical resolution while still enabling different integration times for motion artifact reduction.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different regions of the pixel array (first and second groups) are assigned different integration times based on local lighting conditions. This local quality approach allows each pixel group to be optimized for its specific lighting environment while maintaining overall image resolution through proper interpolation during image generation.

Inventive Principle:
Principle #3Local quality

4Manufacturing precision

If tightly interleaved long and short integration times are implemented pixel-by-pixel, then vertical resolution is improved, but some resolution loss remains in both vertical and horizontal directions

Engineering Contradiction:
Improvevertical resolutionVSAvoidoverall resolution
Core Design Contradiction:
Manufacturing precisionVSMeasurement precision

Solution Approach 1:

The patent segments the pixel array into first and second groups that are interleaved throughout the array, rather than dividing the array into separate row bands. This fine-grained segmentation allows each pixel group to contribute to the complete image, preserving vertical resolution while still enabling different integration times for motion artifact reduction.

Inventive Principle:
Principle #1Segmentation

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 effectively extends the dynamic range of image sensors, minimizing motion artifacts and resolution loss while maintaining high signal-to-noise ratio, enabling accurate capture of both bright and dark areas within the same image frame.

Implementation Method 1

Each pixel may include a large photodiode and a small photodiode

Methodology Applied
Scientific EffectPhotoelectric effect: Photoelectric Effect

Data Source

PatentEP3036894B1System and method for capturing images with multiple image sensing elements
Publication Date: 2023.01.18 QUALCOMM INC
  • EP3036894B1 patent drawingFigure 1
  • EP3036894B1 patent drawingFigure 2
  • EP3036894B1 patent drawingFigure 3

AI summary

Methods, devices, and computer program products for CMOS visible image sensors incorporating multiple image sensing elements with different integration times to extend dynamic range are disclosed herein. In one aspect, a method of taking an image using a CMOS visible image sensor that includes at least one first light sensing element with a first well capacity and at least one second light sensing element with a second well capacity, where the second well capacity is greater than the first well capacity is disclosed. The method includes determining a first integration time for each of the at least one first light sensing elements. The method further includes determining a second integration time for each of the at least one second light sensing elements, where the second integration time is different than the first integration time.