Image Blur Estimation Using Motion Vector and Gyroscope Offset Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image stabilization techniques for image capturing apparatuses face challenges in accurately estimating and removing offset components from gyroscope sensor signals, especially when capturing moving objects, due to mixed motion components and temperature variations, leading to increased estimation errors.
Innovation Solution
An image blur information acquisition system that includes a processor or circuit configured to acquire shake detection results from both motion vectors and angular velocity sensors, with an estimation unit that updates the offset amount based on the difference between these results to suppress erroneous offset estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion vector is used for offset estimation, then offset removal is enabled, but motion components of apparatus and object get mixed preventing correct estimation
Solution Approach 1:
The patent segments the shake detection into two independent sources: motion vector-based detection and gyroscope sensor-based detection. By separating these detection methods and comparing their differences, the system can identify and remove offset components without being confused by object motion, thus resolving the contradiction between enabling offset removal and maintaining reliability when capturing moving objects
Solution Approach 2:
The patent introduces an intermediary mechanism by using the difference between motion vector-based shake detection and gyroscope-based shake detection as an indicator of offset. This intermediary approach allows the system to estimate and remove offset components while filtering out object motion interference, thereby maintaining both offset removal capability and estimation reliability
2Measurement precision
If predetermined range limit is set for offset estimation, then estimation error increase is prevented, but actual offset variation due to temperature cannot be handled
Solution Approach 1:
The patent applies dynamics by making the offset estimation process adaptive rather than static. Instead of using a fixed predetermined range, the system dynamically estimates offset based on the real-time difference between motion vector-based and gyroscope-based shake detection. This dynamic approach allows the system to maintain precision while adapting to temperature variations and other environmental changes
3Measurement precision
If gyroscope sensor output is integrated to obtain angle, then rotational blur detection is achieved, but integration error accumulates causing drift
Solution Approach 1:
The patent implements feedback by continuously monitoring the difference between motion vector-based shake detection and gyroscope-based shake detection. This feedback mechanism allows the system to identify and correct offset components in real-time, preventing integration error accumulation and drift while maintaining accurate rotational blur detection over extended periods
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 erroneous offset estimation during image stabilization, ensuring accurate image capture even when capturing moving objects by dynamically adjusting the estimation based on the difference between shake detection results, thereby improving image stabilization performance.
Implementation Method 1
a plurality of image signals output from an image sensor configured to photoelectrically convert an object image
Data Source
AI summary
An image blur information acquisition apparatus includes a first acquisition unit configured to acquire a first shake detection result from a first shake detection unit configured to detect a shake amount of an image capturing apparatus based on a motion vector, a second acquisition unit configured to acquire a second shake detection result from a second shake detection sensor configured to detect a rotational angular velocity, and an estimation unit configured to estimate an amount of offset included in the second shake detection result, based on the first shake detection result and the second shake detection result, wherein the estimation unit is configured to set degree of change of estimation value by updating the offset amount based on a difference between the second shake detection result and the first shake detection result.