Rapid panoramic stitching method for microscopic blood cell images

Abstract

The invention discloses a rapid panoramic stitching method for microscopic blood cell images. The method includes fixing a microscopic camera, enabling the mobile platform to continuously collect blood sample slides according to a bow-shaped route, obtaining microscopic blood cell image data, and performing filtering and denoising processing on the to-be-spliced image by using a Gaussian operator;extracting image feature points by using an SURF feature method; performing feature matching by using an FLANN method, performing comparison with a set value, removing duplicate images, and obtainingcoarse matching results; solving a coordinate transformation homography matrix of the to-be-spliced image and the spliced image by adopting a random sampling consistency method; using a random sampling consistency method again to obtain accurate matching point pairs; calculating the position relationship between the image A and the image B according to the accurate matching point pairs, completing the unification of coordinates, carrying out the fusion transition of the overlapping region, completing the stitching, traversing all blood sample slides, sequentially carrying out the stitching calculation, and outputting a panoramic blood cell image, thereby improving the stitching quality and speed.

Description

technical field [0001] The invention relates to the technical field of image splicing, in particular to a method for rapid panorama splicing of microscopic blood cell images. Background technique [0002] In the analysis of blood cell morphology, due to the limitation of the field angle of the microscope, the collected images cannot reflect all the information of the microscopic objects. Microscopic objects have many targets on the same focal plane, and even under low-magnification objective lenses, it still takes several images to fully reflect the object information, let alone under high-magnification objective lenses. In order to obtain a large-field microscopic image, multiple high-magnification images need to be stitched together to form a full-scene large-field image, which is convenient for verifying and confirming the shape and type of blood cells. At the present stage, image stitching needs to detect the overlapping areas between images first, and seamlessly merge ...

AI Technical Summary

Problems solved by technology

At the present stage, image stitching needs to detect the overlapping areas between images first, and seamlessly merge the overlapping areas according to the mutual positional relationship, so as to obtain images of panoramas or large scenes. However, in the entire image stitching operation, hundreds of thousands of Image stitching is very time-consuming, feature extraction requires a lot of calculations, and requires a high overlap ratio of overlapping areas

After stitching multiple images continuously, due to the existence of cumulative errors, the scale of the panorama will change, making the scale of the subsequent images too large or too small compared to the first image, and the stitching quality and speed are low

Images