A multi-image automatic registration and fusion stitching method

Abstract

A multi-image automatic registration and fusion stitching method. Firstly, the out-of-order images are enhanced by the multi-scale Retinex method; then the phase correlation algorithm improved by logarithmic polar coordinate transformation is used to match the images, and the scaling, rotation and translation parameters are calculated. Estimate the overlapping area between images, and formulate the sorting rules according to the peak energy of the impulse function; then extract the feature points of the overlapping area to be registered through the SUFR algorithm, improve the random sampling consensus algorithm to accurately purify the matching point pairs, optimize the model parameters, and establish the transformation between images Matrix, sequentially splicing sequence images; finally, use NSCT transformation algorithm and formulate fusion strategy to further fuse the spliced ​​images, and output the fused spliced ​​image; The impact of high noise and uneven illumination on image registration and mosaic can help save the cost of imaging equipment and make accurate diagnosis for medical staff. It has a good application prospect in the field of auxiliary medical imaging.

Description

technical field [0001] The invention relates to the technical fields of machine vision and image processing, in particular to a multi-image automatic registration and fusion splicing method; Background technique [0002] At present, most of the images taken by medical imaging equipment are partial pictures, the field of view is limited, and the panoramic image equipment is expensive, which is difficult to promote and use; the use of image stitching technology to generate panoramic pictures has low cost and good effect, and provides an overall target image, which is convenient for medical treatment It can be used in X-ray imaging, ultrasonography, etc.; in recent years, stitching technology has made great progress at home and abroad, and commonly used registration methods include transform domain matching, grayscale matching, and feature matching. , template matching, etc., by calculating the transformation parameters between images, transform the coordinate system of multipl...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the shortcomings of the existing registration and fusion stitching methods due to the particularity of medical imaging, the present invention provides a multi-image automatic registration and fusion stitching method, which greatly improves the manual sorting of multiple images. Difficulty, fuzzy image details, low resolution, high noise, uneven illumination and other problems have an impact on image registration and mosaic, which helps save the cost of imaging equipment and accurate diagnosis for medical staff. The field of medical imaging has a good application prospect

Images