Image Registration Method Using Pixel Count-Based Preprocessing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image registration methods, particularly for satellite images, face challenges in efficiently processing large-sized images due to the high number of feature points, leading to increased processing time and resource wastage when preprocessing is unnecessary.
Innovation Solution
A method that determines whether preprocessing is required based on the number of pixels in the images, generating divided images through reduction and clustering techniques, and registering them using neural network-based feature point extraction models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If preprocessing is performed on all images through feature point extraction and matching, then image registration accuracy is improved, but processing time and computational resources increase significantly for large-sized satellite images
Solution Approach 1:
The patent divides the image registration process into two segments: a preprocessing stage for extracting and matching feature points, and a main registration stage that uses only the matched feature points. This segmentation allows the system to handle large images by processing them in manageable portions rather than attempting to process all pixels simultaneously.
Solution Approach 2:
The patent extracts only the essential information needed for registration (feature points and their matches) from the large satellite images, discarding unnecessary data. By taking out only the critical feature points rather than processing the entire image, the system reduces computational burden while maintaining registration accuracy.
2Quantity of substance
If preprocessing is performed on all images, then feature point extraction completeness is improved, but computational resources are wasted on images that do not require preprocessing
Solution Approach 1:
The patent implements a dynamic decision-making process where the system determines whether to perform preprocessing based on image characteristics and registration requirements. This dynamic approach allows the system to adapt its processing behavior, performing preprocessing only when necessary and using alternative methods when possible, thereby optimizing computational resource usage.
Solution Approach 2:
The patent changes the processing parameters and methods based on image size and complexity. For large satellite images, the system uses different processing parameters compared to smaller images, adjusting the level of preprocessing applied to match the specific requirements and computational capabilities, thus avoiding unnecessary resource expenditure.
3Measurement precision
If the SIFT algorithm is used to extract feature points from satellite images, then feature point extraction capability is improved, but the high number of objects in satellite images makes feature point extraction difficult
Solution Approach 1:
The patent segments the image into smaller regions or uses multiple scales when extracting feature points from satellite images. This segmentation approach helps the SIFT algorithm handle the high number of objects by processing them in manageable portions, reducing the difficulty of detecting and measuring feature points across the entire large image.
Solution Approach 2:
The patent applies partial preprocessing and feature point extraction only to the extent necessary for successful registration, rather than attempting to extract all possible feature points. This partial action approach reduces the difficulty of detection by focusing computational effort on the most critical areas while avoiding unnecessary processing of less relevant regions.
Data Source
AI summary
Disclosed is a method for image registration performed by a computing device including at least one processor according to some exemplary embodiments of the present disclosure. The method for image registration may include: determining whether to perform preprocessing on a first image and a second image, based on at least one of the number of first pixels of the first image or the number of second pixels of the second image; when performing the preprocessing, generating a first divided image and a second divided image from each of the first image and the second image through a preprocessing process; and registering the first image and the second image, based on the first divided image and the second divided image.