Image Stabilization Control Apparatus Using Predicted Shake Signal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image stabilization techniques for digital cameras face challenges in accurately detecting camera shake during still image capturing, as low-frequency noise from shake sensors overlaps with actual camera shake, and motion vector detection is not feasible during exposure, leading to insufficient image stabilization control.
Innovation Solution
An image stabilization control apparatus that combines a shake detection signal with a predicted shake signal using a motion vector detection signal, extracting high and low-frequency components to produce a combined shake signal for effective image stabilization, and adjusts the cut-off frequency based on exposure time to minimize noise and improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a shake sensor is used to detect camera shake, then the detection can be performed during still image capturing, but low frequency noise (drift) overlaps with the sensor output, deteriorating detection accuracy
Solution Approach 1:
The shake detection signal is segmented into high-frequency components (actual camera shake) and low-frequency components (drift noise) using frequency-based separation. The shake extracting unit extracts high-frequency components from the shake detection signal, effectively isolating the useful signal from the noisy low-frequency drift component.
Solution Approach 2:
A predicted shake signal is introduced as an intermediary to bridge the gap between the shake detection signal and the motion vector detection signal. This predicted signal, generated by the predicting unit using motion vectors from video capturing, serves as a reference to reconstruct the complete shake information during still image capturing.
2Measurement precision
If motion vector detection is used to accurately detect camera shake, then detection accuracy is improved, but motion vector cannot be detected in image sensor exposure time for still image capturing, disabling image stabilization control
Solution Approach 1:
Motion vectors are detected during video capturing (still image capturing preparation period) before the actual still image exposure. The predicting unit uses these pre-detected motion vectors to generate a predicted shake signal that is valid during the exposure period, enabling image stabilization control even when motion vectors cannot be detected during exposure.
Solution Approach 2:
The system uses feedback from motion vector detection during video capturing to continuously update and refine the predicted shake signal. This feedback mechanism allows the system to maintain accurate shake prediction during still image capturing by leveraging information from the preceding video capturing phase.
3Measurement precision
If high frequency components are extracted from shake detection signal, then drift noise is reduced, but low frequency camera shake information is lost
Solution Approach 1:
The combining unit merges the high-frequency first shake signal (extracted from shake detection) with the low-frequency predicted shake signal (generated from motion vectors). This combination reconstructs the complete shake signal spectrum, preserving both high-frequency actual shake and low-frequency shake information while eliminating drift noise.
Data Source
AI summary
The image stabilization control apparatus includes a shake extracting unit extracting a first shake signal whose frequency is higher than a predetermined frequency from a shake detection signal, a predicting unit producing, by using a motion vector detection signal, a predicted shake signal indicating a predicted value of shaking of an optical apparatus, a combining unit combining the first shake signal with the predicted shake signal to produce a second shake signal, a controlling unit performing an image stabilization control using the second shake signal, and a motion vector extracting unit extracting a specific motion vector signal whose frequency is lower than the predetermined frequency from the motion vector detection signal. The controlling unit performs the image stabilization control using the second shake signal in still image capturing, and the predicting unit produces the predicted shake signal by using the specific motion vector signal.