Image Sensor Motion Correction via Segmented Pixel Readout

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current digital cameras face challenges in low-light conditions due to camera motion during exposure, leading to poor image quality and noise amplification, with existing solutions being costly or inefficient, such as optical image stabilization, mechanical methods, and blind deconvolution algorithms.

Innovation Solution

The method involves defining multiple sets of image element acquisition devices on an image sensor, with signals from these devices being read out multiple times to process and correct for camera motion through deblurring, spatial alignment, and sharpening, using techniques like block-matching and integral projections to estimate and correct for motion.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Illumination intensity

If exposure time is increased to boost photons in low-light conditions, then image brightness is improved, but image sharpness deteriorates due to camera motion

Engineering Contradiction:
Improveimage brightnessVSAvoidimage sharpness
Core Design Contradiction:
Illumination intensityVSManufacturing precision

Solution Approach 1:

The patent divides the image sensor into multiple sets of image pixels (e.g., first set, second set, third set) that are read out at different times during the exposure period. This segmentation allows the system to capture motion information from multiple temporal snapshots while maintaining the total exposure duration, thereby preserving both brightness and sharpness.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary readouts of selected image pixel sets before the final image is complete. These preliminary readouts capture motion information that is used to estimate and correct camera motion during the exposure, allowing motion compensation to be applied before the final image is finalized.

Inventive Principle:
Principle #10Preliminary action

2Illumination intensity

If digital gain factor is applied to boost light intensity, then image brightness is improved, but noise levels increase

Engineering Contradiction:
Improveimage brightnessVSAvoidnoise levels
Core Design Contradiction:
Illumination intensityVSObject-generated harmful factors

Solution Approach 1:

By segmenting the sensor into multiple pixel sets read at different times, the system captures sufficient signal information during the total exposure period without needing to amplify noise through digital gain. The motion estimation from segmented reads enables motion compensation that reduces the need for aggressive gain application.

Inventive Principle:
Principle #1Segmentation

3Measurement precision

If optical image stabilization with gyroscopic measurement and lateral actuators is used, then camera motion correction is improved, but device cost and complexity increase

Engineering Contradiction:
Improvecamera motion correctionVSAvoiddevice cost
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent replaces mechanical gyroscopic measurement and lateral actuator systems with an electronic solution. By using multiple image pixel sets read at different times and applying block-matching algorithms to estimate motion, the system achieves camera motion correction without mechanical components, significantly reducing cost and complexity.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent creates multiple temporal copies of image data by reading different pixel sets at different times during the exposure. These copies are then processed using block-matching to estimate motion, providing a software-based alternative to hardware stabilization systems.

Inventive Principle:
Principle #26Copying

4Measurement precision

If mechanical methods with gyroscopic measurements are used to track camera motion, then motion measurement precision is improved, but device cost and complexity increase

Engineering Contradiction:
Improvemotion measurementVSAvoiddevice cost
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent substitutes mechanical gyroscopic measurement systems with an electronic motion estimation approach. By reading multiple pixel sets at different times and applying block-matching algorithms to the captured data, the system achieves accurate motion measurement without mechanical gyroscopes, reducing both cost and complexity.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

5Measurement precision

If strips of sensor pixels are dedicated to motion estimation, then camera motion tracking is improved, but spatial resolution of output image decreases

Engineering Contradiction:
Improvecamera motion trackingVSAvoidspatial resolution
Core Design Contradiction:
Measurement precisionVSManufacturing precision

Solution Approach 1:

The patent segments the sensor into multiple pixel sets that are read at different times during the exposure. This temporal segmentation allows motion estimation using the same spatial pixels that contribute to the final image, avoiding the need to dedicate separate spatial regions for motion tracking and thereby preserving spatial resolution.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent moves motion estimation from the spatial dimension to the temporal dimension by reading different pixel sets at different times. This allows motion information to be extracted from temporal variations in the same spatial pixels, preventing loss of spatial resolution while maintaining motion tracking capability.

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

6Device complexity

If blind deconvolution algorithms are used to reduce camera motion blur, then image processing is simplified, but correction accuracy decreases due to lack of motion knowledge

Engineering Contradiction:
Improveprocessing simplicityVSAvoidcorrection accuracy
Core Design Contradiction:
Device complexityVSManufacturing precision

Solution Approach 1:

The patent performs preliminary readouts of multiple pixel sets during the exposure to capture motion information. This preliminary action provides actual motion data that is used to guide the deconvolution process, improving correction accuracy compared to blind deconvolution while maintaining reasonable processing complexity.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent uses block-matching algorithms to estimate motion from the multiple pixel set readouts, providing feedback about actual camera motion during the exposure. This feedback information is then used to guide the image processing, improving correction accuracy over blind deconvolution methods that lack such feedback.

Inventive Principle:
Principle #23Feedback

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 reduces camera motion blur and noise in low-light conditions, improving image quality without the need for additional costly hardware or assumptions about motion during exposure.

Implementation Method 1

an image sensor is defined. Multiple sets of image pixels are defined on the image sensor. Signals from a first set of image pixels are read out multiple times. Signals from a second set of image pixels are read out multiple times.

Methodology Applied
Scientific EffectPhotoelectric Effect: Photoelectric Effect

Data Source

PatentEP2191639B1Correcting imaging device motion during an exposure
Publication Date: 2018.11.07 OMNIVISION TECHNOLOGIES INC
  • EP2191639B1 patent drawingFigure 1
  • EP2191639B1 patent drawingFigure 2
  • EP2191639B1 patent drawingFigure 3

AI summary

A system and method for determining and correcting for imaging device motion during an exposure is provided. According to various embodiments of the present invention, multiple sets of image pixels are defined on an image sensor, where each set of pixels is at least partially contained in the output image area of the image sensor. Signals from each set of image pixels are read out once or more during an exposure, motion estimates are computed using signal readouts from one or more sets of image pixels, and signal readouts from one or more sets of the image pixels are processed to form the final output image.