Distorted Image Registration via Integral Projection Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image registration techniques face challenges in accurately registering distorted, rotated, translated, and differently scaled images, particularly in correcting distortions and estimating scale factors, which affects the precision of image alignment and quality of reconstructed pixels.
Innovation Solution
The method involves computing horizontal and vertical integral projection vectors for distorted images, applying an approximation of the inverse distortion function to generate distorted integral projection vectors, and estimating scale factors by resampling images with various factors to minimize absolute differences between corresponding vectors, allowing for precise registration without initial distortion correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional image registration techniques are used on distorted images, then the processing is simpler, but the registration precision deteriorates due to inaccurate scale factor estimation and distortion effects
Solution Approach 1:
The patent applies distortion correction to images before performing registration operations. By pre-correcting the distortion in each image using distortion correction parameters, the subsequent registration process operates on corrected images, achieving accurate scale factor estimation and precise alignment without being hindered by distortion effects.
Solution Approach 2:
The patent separates the registration process into distinct stages: first correcting distortion independently for each image, then performing registration on the corrected images. This segmentation allows each step to be optimized independently, maintaining precision while managing complexity through modular processing.
2Measurement precision
If distortion correction is applied before registration, then registration precision improves, but processing time and complexity increase
Solution Approach 1:
The patent performs distortion correction as a preliminary step before registration, using pre-determined distortion correction parameters. This approach improves scale factor estimation accuracy by ensuring images are properly corrected before comparison, while the use of pre-computed parameters helps minimize additional processing time.
Solution Approach 2:
The patent utilizes distortion correction parameters that characterize the distortion of each image. By applying these parameters to correct the images beforehand, the system achieves more accurate registration results. The parameters are applied efficiently to balance correction quality with processing speed.
3Manufacturing precision
If images are corrected for distortion before registration, then the quality of reconstructed pixels improves, but the device complexity increases
Solution Approach 1:
The patent applies distortion correction to images before registration and reconstruction operations. By pre-correcting the distortion using distortion correction parameters, the quality of reconstructed pixels is significantly improved, as the registration is performed on geometrically accurate images rather than distorted ones.
Solution Approach 2:
The patent segments the image processing pipeline into distinct modules: distortion correction, registration, and reconstruction. This segmentation allows each module to perform its function optimally, with distortion correction preparing the images for subsequent processing steps, thereby improving overall system efficiency and output quality.
Data Source
AI summary
An image registration method involves computing horizontal and vertical integral projection vectors for first and second distorted or partially distorted images or distortion-corrected images, or both. The images are registered by applying a translation, rotation and/or scale factor estimation between the first and second images on the horizontal and vertical integral projection vectors.