Rolling Shutter Image Stabilization via Acceleration Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
CMOS image sensors using the rolling shutter method face challenges in capturing clear images of moving subjects due to image distortion, with existing digital image stabilizer techniques requiring high computational effort and assuming constant camera motion, leading to accuracy issues.
Innovation Solution
A photographing apparatus and method that continuously captures images using a rolling shutter method, estimating and compensating for uniformly accelerated motion by comparing pixel lines and calculating uniform acceleration functions to correct image distortions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If the rolling shutter method is used to continuously capture images, then the image capture speed and continuous shooting capability are improved, but image distortion occurs when the subject or camera moves
Solution Approach 1:
The patent replaces the mechanical/optical image stabilization system with a digital image processing system. By using digital image stabilization algorithms that analyze multiple sequentially captured images and calculate motion compensation based on pixel line comparisons, the system corrects rolling shutter distortion through computational methods rather than mechanical adjustments, thereby maintaining high capture speed while improving image quality
Solution Approach 2:
The patent applies preliminary action by capturing multiple images in sequence before performing motion compensation. The system预先 captures a series of images with the rolling shutter method, then uses these pre-captured images to calculate motion vectors and apply distortion correction, allowing the high-speed capture capability to be preserved while correcting the distortion in post-processing
2Manufacturing precision
If digital image stabilization is applied to compensate for rolling shutter distortion, then image quality is improved, but large computational effort is required
Solution Approach 1:
The patent applies segmentation by dividing the image into multiple pixel lines and processing them separately in sequence. Instead of analyzing the entire image at once, the system processes each pixel line independently to calculate motion vectors, significantly reducing the computational complexity and power requirements while still achieving effective distortion correction
Solution Approach 2:
The patent uses partial action by focusing computational resources only on the essential task of calculating motion compensation for pixel lines, rather than performing comprehensive image analysis. The system applies just enough computational effort to determine motion vectors and apply correction, avoiding unnecessary calculations while maintaining image quality
3Device complexity
If motion compensation is based on constant velocity assumption, then calculation complexity is reduced, but accuracy deteriorates when the camera undergoes uniformly accelerated motion
Solution Approach 1:
The patent applies dynamics by transitioning from a static constant velocity model to a dynamic uniformly accelerated motion model. The system calculates motion compensation by considering acceleration components in addition to velocity, allowing the motion estimation to adapt to changing motion conditions. This dynamic approach accurately handles scenarios where the camera undergoes acceleration while maintaining reasonable calculation complexity through efficient algorithm design
Data Source
AI summary
A photographing apparatus, a motion estimating apparatus, an image compensating method, a motion estimating method, and a non-transitory computer-readable recording medium are provided. The photographing apparatus includes: an image sensing unit which continuously captures a plurality of images by using a rolling shutter method; and an image processor which compensates for a uniformly accelerated motion of the photographing apparatus by using the plurality of images.