A panoramic super-resolution image stitching method, device, equipment and storage medium

By planning the optimal stitching path in panoramic super-resolved image stitching, and combining Gaussian mixed model and registration algorithm for noise characteristics, the problems of stitching misalignment and cumulative errors are solved, and high-precision panoramic super-resolved image stitching is achieved.

CN116862774BActive Publication Date: 2025-03-25HAINAN UNIV
2 Cites 0 Cited by

Patent Information

Application Number
CN202310889839.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-07-19
Publication Date
2025-03-25
Estimated Expiration
2043-07-19

AI Technical Summary

Technical Problem

When splicing panoramic super-resolved images, it is difficult to calculate the position of the intercellular gap, resulting in misalignment and cumulative errors, affecting the splicing effect.

Method used

By planning the stitching path, the optimal stitching path is generated, and the algorithm combined with Gaussian hybrid model and noise characteristics is used to downsample and register the stitching area of ​​the registration sub-image, and finally deduplication and panoramic fusion are performed to obtain the panoramic super-resolved image completed.

Benefits of technology

The splicing accuracy is improved, and the super-resolution positioning image of large field of view can be obtained, which solves the problem that the Gaussian hybrid model cannot calculate in areas without structural points, and realizes the analysis and research of microstructures of cells with panoramic and nano-level levels.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The present invention relates to the field of image stitching, and discloses a panoramic super-resolution image stitching method, device, equipment and storage medium, the method comprising: planning a stitching path to generate an optimal stitching path; downsampling the stitching area of ​​the sub-image to be registered; according to the optimal stitching path, registering the down-sampled sub-images using an algorithm combining a Gaussian mixture model with noise point features; deduplicating and panoramically fusing the registered images to obtain a stitched panoramic super-resolution image. In this way, the cumulative error is reduced, the calculation efficiency is improved, and the noise point features in the super-resolution positioning imaging process are fully utilized in the registration process to assist in the stitching of super-resolution positioning images, solve the problem that the Gaussian mixture model cannot be calculated in areas without structural points, improve the stitching accuracy, and can obtain a large-field panoramic super-resolution positioning image, so as to realize the analysis and research of panoramic and nano-level cell fine structures.
Need to check novelty before this filing date? Find Prior Art