Image Matching Using SAD and Gradient Information
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image matching methods are influenced by illumination, occlusion, noise, and have low robustness, leading to inaccurate image matching.
Innovation Solution
A method of image matching that involves obtaining a reference image and an image to be matched, determining a template image block, and then finding a first image block with the smallest sum of absolute differences (SAD) and a second image block with the smallest gradient information difference, to determine a matching image block.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image matching methods are used, then the processing speed is fast, but the accuracy and robustness are low due to influence from illumination, occlusion, and noise
Solution Approach 1:
The patent combines multiple feature types (SAD information and gradient information) into a unified matching criterion. The gradient information module calculates gradient values in multiple directions and combines them with SAD values to form a comprehensive matching score, thereby improving robustness against illumination changes and noise while maintaining matching accuracy
Solution Approach 2:
The patent transforms the matching criterion from using only intensity differences (SAD) to using a combination of intensity differences and gradient information. By calculating gradient values in multiple directions (horizontal, vertical, diagonal) and combining them with SAD values, the method changes the parameters used for matching to achieve better robustness and accuracy
2Measurement precision
If multiple image blocks are compared to find the best match, then the matching accuracy improves, but the computational complexity increases
Solution Approach 1:
The patent segments the comparison process into two distinct modules: an SAD information module that evaluates intensity differences, and a gradient information module that evaluates texture differences. This segmentation allows each module to focus on specific features, improving matching accuracy while making the overall process more manageable and efficient
Solution Approach 2:
The patent calculates gradient information for multiple image blocks (K blocks where K≥1) around the first image block. By performing partial comparisons with a limited number of candidate blocks rather than exhaustively comparing all possible blocks, the method achieves good matching accuracy while controlling computational complexity
3Reliability
If gradient information calculation is performed for multiple image blocks, then the robustness against noise and illumination improves, but the processing time increases
Solution Approach 1:
The patent applies gradient information calculation selectively to K image blocks that are spatially close to the first image block (within a certain distance threshold). This local approach ensures that gradient information is computed only where most relevant, improving robustness against noise and illumination in the critical matching region while minimizing unnecessary computations elsewhere
Solution Approach 2:
The patent performs gradient information calculation for a limited number of candidate image blocks (K blocks) rather than all possible blocks in the search region. This partial action approach achieves sufficient robustness by focusing computational resources on the most promising candidates, thereby reducing overall processing time
Data Source
AI summary
A method of image matching and an image processing device are disclosed. The method comprises: obtaining a reference image and an image to be matched; determining a template image block in the reference image, wherein the template image block is an image block located in a matching window in the reference image; determining a first image block in the image to be matched, wherein the first image block is an image block in the image to be matched having a smallest sum of an absolute difference (SAD) with the template image block; determining a second image block in the image to be matched, wherein the second image block is an image block in the image to be matched having a smallest gradient information difference with the template image block; and determining a matching image block in the image to be matched based on the first image block and the second image block.