Imaging Device Motion Compensation via Pixel Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image stabilization techniques for stationary imaging devices are computationally costly and time-consuming, and can reduce image quality, hindering 3D structure determination due to reliance on re-sampling and the assumption of camera motion.
Innovation Solution
A system for imaging device motion compensation that determines and updates orientation parameters by capturing a reference frame, correlating subsequent frames, and applying pixel-shift and rotation data to maintain accurate external orientation, thereby compensating for camera motion without re-sampling and costly hardware.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If re-sampling image data is used for stabilization, then image stabilization is achieved, but image quality is reduced and computational cost increases
Solution Approach 1:
The patent replaces the mechanical re-sampling process with a computational approach using image correlation and pixel-shift calculation. Instead of physically re-sampling the image data, the system correlates feature points between frames to determine motion parameters and applies geometric transformations to compensate for camera movement, thereby avoiding the quality degradation associated with re-sampling while achieving stabilization.
Solution Approach 2:
The patent creates a virtual copy of the image data through correlation matching. By identifying corresponding feature points between consecutive frames and calculating their relative positions, the system generates a transformed version of the image that compensates for motion without physically re-sampling the original data, thus preserving image quality while achieving stabilization.
2Stability of the object's composition
If re-sampling is used for image stabilization, then stabilization is achieved, but computational cost and processing time increase
Solution Approach 1:
The patent applies partial action by focusing correlation computation only on key feature points rather than the entire image data. By identifying and correlating distinctive features (such as corners or edges) between frames, the system determines motion parameters more efficiently without processing every pixel, thereby reducing computational time and cost while achieving effective stabilization.
Solution Approach 2:
The patent substitutes the computationally intensive re-sampling process with a correlation-based approach that calculates pixel-shift and rotation parameters. This method uses geometric transformations on the original image data rather than generating new samples, significantly reducing processing time and computational resources required for stabilization.
3Stability of the object's composition
If existing stabilization methods are used, then camera motion is compensated, but the methods assume camera motion which does not apply to stationary imaging devices
Solution Approach 1:
The patent inverts the conventional approach by not assuming camera motion occurs and then trying to compensate for it. Instead, the system correlates image data from consecutive frames to detect and quantify any motion that does occur, then applies compensation geometrically. This inverted approach allows the method to handle both stationary and moving camera scenarios uniformly, making it adaptable to stationary imaging devices while still providing motion compensation when needed.
Data Source
AI summary
Generally discussed herein are systems, apparatuses, and methods for. In one or more embodiments, a method can include recording imaging device external orientation parameters of a rigidly mounted imaging device, capturing a reference frame using the imaging device in an orientation corresponding to the recorded parameters, capturing a later frame after the reference frame, correlating the later frame to the reference frame to determine pixel orientation in the reference frame relative to pixel orientation in the later frame, determining an imaging device movement required to return the imaging device back to the orientation associated with the recorded imaging device external orientation parameters based on the determined pixel orientation in the reference frame relative to the pixel orientation in the later frame, and updating the imaging device external orientation parameters to account for the determined imaging device movement in accord with the imaging device external orientation parameters.