Image Processing Method for Rotational Pose Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image detection algorithms face challenges in rotational invariance when detecting target images with varying poses, leading to low reliability and complexity in feature extraction and matching.
Innovation Solution
An image processing method that divides detection and template images into subregions, calculates the principal rotation direction by averaging differences in radial and gradient directions of pixel points on the outer contour lines, and adjusts the detection image's pose to match the template image, simplifying the identification process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If template matching is performed directly on images with different poses, then the detection algorithm can be simpler, but the reliability of image identification decreases
Solution Approach 1:
The patent applies preliminary action by calculating the rotation angle between the detection image and template image before performing template matching. This pre-alignment step ensures that both images have consistent orientations, which improves the reliability of feature extraction and matching while maintaining algorithmic simplicity.
2Reliability
If the detection algorithm adjusts for rotation angle, then the reliability of image identification improves, but the algorithm becomes more complicated
Solution Approach 1:
The patent segments the image processing task into distinct steps: first calculating the rotation angle between detection image and template image, then adjusting the detection image based on this angle, and finally performing template matching. This segmentation makes the algorithm more structured and manageable while improving reliability through proper alignment.
Solution Approach 2:
The patent replaces complex mechanical trial-and-error alignment methods with a mathematical approach using gradient directions and angle calculations. By computing the rotation angle analytically from image gradients, the system achieves reliable alignment without exhaustive search or complex iterative adjustments.
3Loss of time
If feature extraction is performed on images with misaligned poses, then processing time is reduced, but the accuracy of matching decreases
Solution Approach 1:
The patent performs preliminary rotation angle calculation and image adjustment before feature extraction and template matching. This ensures that features are extracted from properly aligned images, significantly improving matching accuracy while the efficient angle calculation method keeps the time overhead minimal.
Data Source
AI summary
The present disclosure discloses an image processing method and device. The image processing method includes: dividing a detection image into a plurality of first subregions, dividing a template image into a plurality of second subregions, calculating a principal rotation direction of each first subregion with respect to the corresponding second subregion; and calculating a principal rotation direction of the detection image according to the principal rotation directions of the plurality of first subregions.