Feature Point Matching Using Geometric Line Segments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image matching methods using multi-sensors require complex calculations and increased memory capacity due to the need for similarity measurements between image patches from different wavelength bands, such as visible and thermal images.
Innovation Solution
An image matching method that utilizes coordinate information of feature points to match images without relying on patch images, by selecting and connecting feature points to generate straight lines and searching for matching lines based on length and angle criteria between the reference and target images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image patches are used to measure similarity between candidate feature points, then matching accuracy is improved, but calculation complexity and memory requirements increase
Solution Approach 1:
The patent extracts only the essential geometric information (coordinates of feature points) from the image data, eliminating the need to process entire image patches. By taking out only the necessary coordinate information and using it to form straight lines between feature points, the method achieves matching without the computational burden of patch-based similarity calculations.
Solution Approach 2:
The patent replaces the mechanical/computational process of comparing image patches (pixel-by-pixel similarity measurement) with a geometric approach using straight lines connecting feature points. This substitution uses simple geometric relationships (length and angle of lines) instead of complex image processing operations, dramatically reducing calculation complexity.
2Measurement precision
If image patches are used for similarity measurement, then matching accuracy is improved, but memory capacity and I/O bandwidth increase
Solution Approach 1:
The patent extracts only the coordinate information of feature points from the images, storing and processing merely these coordinate values rather than entire image patches. This extraction principle dramatically reduces the quantity of data that needs to be stored in memory and transferred through I/O channels, while still providing sufficient information for accurate matching through geometric relationships.
3Measurement precision
If complex similarity calculations are performed on image patches, then matching precision is improved, but processing time increases
Solution Approach 1:
The patent substitutes complex similarity calculation operations with simple geometric computations involving straight lines between feature points. Instead of performing computationally intensive pixel-wise comparisons, the method calculates only the length and angle of lines connecting feature points, which are much faster computations that maintain matching precision while significantly reducing processing time.
Solution Approach 2:
The patent changes the parameters used for matching from pixel intensity values and similarity metrics to geometric parameters (length and angle of straight lines between feature points). This parameter transformation simplifies the computation while preserving the essential matching information, thereby reducing processing time without sacrificing precision.
Data Source
AI summary
An image matching method includes: extracting a plurality of feature points from a reference image; selecting a first feature point from the feature points, and selecting a first reference search area comprising the first feature point; setting a first matching candidate search area corresponding to the first reference search area from a target image, and extracting a plurality of feature points from the first matching candidate search area; selecting a second feature point closest to the first feature point in the first reference search area, and selecting a first straight line connecting the first and second feature points; generating a plurality of segments from the feature points extracted from the first matching candidate search area; and determining a first matching straight line matching a length and an angle of the first straight line, from the segments generated from the feature points extracted from the first matching candidate search area.