Line-by-Line Camera Shake Correction in CMOS Video Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
CMOS image sensors in portable video cameras face challenges in correcting image distortions caused by camera shake due to differences in exposure time across lines, leading to incomplete elimination of shake effects and increased memory requirements for storing image data.
Innovation Solution
An image processing apparatus that calculates and interpolates corrections for each line based on camera shake information, using a combination of imaging, vibration detection, conversion, calculation, and interpolation to accurately correct pixel shifts and reduce memory capacity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single value calculated from camera shake information is used to correct image distortion, then the correction process is simple, but the image distortion caused by camera shake cannot be completely eliminated due to differences in exposure time between lines
Solution Approach 1:
The patent divides the image into multiple lines and calculates correction values separately for each line based on their respective exposure times. This segmentation allows the system to account for the time difference between top and bottom lines, thereby eliminating image distortion while maintaining a relatively simple correction process.
Solution Approach 2:
The patent dynamically adjusts the correction value for each line according to its exposure time. Instead of using a static single value for the entire image, the system varies the correction parameter line by line, adapting to the changing exposure conditions and achieving complete elimination of camera shake distortion.
2Manufacturing precision
If correction data for all lines is stored in memory, then complete camera shake correction can be achieved, but the memory capacity requirement increases
Solution Approach 1:
The patent calculates and stores only the essential correction parameters (such as shake amount and exposure time differences) in advance, rather than storing complete correction data for all lines. This preliminary calculation of key parameters reduces memory requirements while still enabling accurate line-by-line correction during image processing.
Solution Approach 2:
The patent introduces intermediary parameters (shake amount, exposure time difference) that serve as mediators between the raw camera shake information and the final correction values. By storing and processing these intermediate representations rather than full correction datasets, the system achieves complete correction with reduced memory capacity.
Data Source
AI summary
The present invention relates to an image processing apparatus and method enabling to correct for camera shake given in image capturing, a recording medium, and a program. Discrete data pieces indicating shifts are obtained. The discrete data pieces are interpolated to obtain data indicating a shift every line. By the interpolation, a correction is calculated every line on the basis of the corresponding shift. On the basis of the corrections, pixel data read positions are determined. Pixel data is read based on the determined read positions. When a pixel to be read overlaps two pixels, pixel data of the pixel to be read is produced (interpolated) from pixel data pieces. In this manner, the effects of camera shake are corrected. The present invention is applicable to video cameras.