Imaging Apparatus Motion Vector Detection for Blur Correction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image blur correction systems face challenges in accurately detecting motion vectors with low noise and high frequency resolution, leading to instability and reduced performance due to time delays and low accuracy in noise separation.

Innovation Solution

The system improves image blur correction by using multiple motion vector calculations from images of different sizes, combining control elements with varying gain contributions based on frequency resolutions to reduce calculation delay and enhance detection accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If motion vectors are calculated from high-resolution images to improve detection accuracy, then measurement precision is improved, but calculation time increases causing time delay

Engineering Contradiction:
Improvemotion vector detection accuracyVSAvoidcalculation delay
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent segments the image processing task by calculating motion vectors from both high-resolution and low-resolution images simultaneously. The low-resolution motion vectors are computed first and used as initialization for the high-resolution calculation, dividing the overall computation into manageable stages that reduce total calculation time while maintaining accuracy.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary action by calculating motion vectors from low-resolution images before using them to guide the calculation from high-resolution images. This preliminary calculation provides initial estimates that accelerate the subsequent high-precision computation, reducing the overall time delay.

Inventive Principle:
Principle #10Preliminary action

2Productivity

If motion vectors are calculated from low-resolution images to reduce calculation time, then productivity is improved, but measurement precision deteriorates

Engineering Contradiction:
Improvecalculation speedVSAvoidmotion vector detection accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent merges the results from both low-resolution and high-resolution motion vector calculations. The low-resolution results provide fast initial estimates, while the high-resolution results provide precise corrections. By combining both approaches, the system achieves both speed and accuracy that neither method could achieve alone.

Inventive Principle:
Principle #5Merging (Combining)

3Reliability

If feedback control gain is increased to improve image blur correction performance, then correction effectiveness is improved, but system stability deteriorates due to phase delay

Engineering Contradiction:
Improveimage blur correction performanceVSAvoidfeedback control system stability
Core Design Contradiction:
ReliabilityVSStability of the object's composition

Solution Approach 1:

The patent implements a feedback mechanism where motion vectors from both resolution levels are continuously calculated and used to adjust the feedback control gain. The system monitors the phase delay introduced by calculation processing and dynamically adjusts the gain to maintain stability while maximizing correction performance.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3454544B1Imaging apparatus and control method for the same
Publication Date: 2023.08.02 CANON KK
  • EP3454544B1 patent drawingFigure 1
  • EP3454544B1 patent drawingFigure 2A~2B
  • EP3454544B1 patent drawingFigure 3

AI summary

In an imaging apparatus, a first motion vector detector calculates a first motion vector on the basis of images with a low spatial resolution and a second motion vector detector calculates a second motion vector on the basis of images with a high spatial resolution. A first control amount having a great weight for information with a low frequency resolution with respect to frequency components of the first motion vector information and a second control amount having a great weight for information with a high frequency resolution with respect to frequency components of the second motion vector information are calculated. A third control amount is calculated through multiplication by a predetermined weight regardless of the frequency resolution with respect to frequency components of the first or second motion vector information. Drive control of an image blur correction lens is performed according to the first to third control amounts.