Dynamic Correction Gain for Image Shake Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image pick-up apparatuses face challenges in accurately detecting and correcting for image shake and attitude changes, particularly when bias errors in inertial sensors are significant or when sudden camera movements occur, leading to delayed and incorrect position and attitude estimation.
Innovation Solution
An image pick-up apparatus equipped with processors that acquire shake information and subject movement data, using band-limited information and image signal processing to estimate and correct position and attitude, incorporating vibration sensors and feature point tracking for enhanced accuracy through a feature coordinate map and position and attitude estimation unit.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If correction using image information is performed to correct inertial sensor bias error, then position and attitude estimation accuracy is improved, but if the bias error varies greatly or sudden camera work occurs, correction may not be correctly performed or takes long time to converge
Solution Approach 1:
The patent dynamically adjusts the correction gain based on the magnitude of position and attitude deviations. When deviations are large (indicating sudden camera work or environmental changes), the correction gain is reduced to prevent over-correction and instability. When deviations are small (indicating stable conditions), the correction gain is increased to improve convergence speed and accuracy. This dynamic adjustment resolves the contradiction by adapting the correction behavior to current operating conditions.
Solution Approach 2:
The patent changes the correction parameter (correction gain) based on the state of the system. By monitoring position and attitude deviations and adjusting the correction gain accordingly, the system optimizes correction performance under different conditions. This parameter change strategy allows the system to maintain both high accuracy during stable periods and reliability during transient disturbances.
2Loss of time
If correction gain is increased to speed up bias error correction, then convergence time is reduced, but correction accuracy may deteriorate when large deviations occur
Solution Approach 1:
The correction gain is made dynamic rather than fixed. The system automatically increases correction gain when deviations are small (enabling fast convergence) and decreases it when deviations are large (maintaining accuracy and stability). This dynamic behavior resolves the contradiction between convergence speed and correction accuracy under different operating conditions.
Solution Approach 2:
The correction gain parameter is adjusted based on the magnitude of position and attitude deviations. This parameter change strategy enables the system to achieve fast convergence during normal operation while maintaining accuracy during transient disturbances or large deviations.
3Speed
If inertial sensor data is used for position and attitude estimation, then real-time estimation is achieved, but bias error variation due to environment changes reduces estimation accuracy
Solution Approach 1:
The patent implements feedback control by continuously monitoring position and attitude deviations and using this information to adjust the correction gain. The feedback mechanism allows the system to maintain high estimation accuracy under varying environmental conditions while preserving real-time estimation capability. The correction process is continuously adapted based on actual system performance.
Solution Approach 2:
The system performs self-adjustment by automatically modifying the correction gain based on observed deviations without external intervention. This self-service capability enables the system to maintain optimal performance across different environmental conditions while preserving real-time estimation speed.
Data Source
AI summary
A vibration sensor of an image pick-up apparatus detects shake of an image pick-up apparatus and acquires shake information. An imaging unit outputs an image signal of a subject to an imaging signal processing unit. A motion vector detection unit calculates a motion vector according to an image signal after imaging. A feature point tracking unit performs feature point tracking by calculating a coordinate value of a subject on a photographing screen using the motion vector. A feature coordinate map and position and attitude estimation unit estimates a position or attitude of the image pick-up apparatus on the basis of information obtained by a band limit filter performing band limitation on shake information from a vibration sensor and an output of the feature point tracking unit. An estimation unit evaluates an estimation error, and changes a band limited by the band limit filter on the basis of the calculated evaluation value or changes a correction magnification at the time of correcting the estimation value according to a correction value.