Image splicing method and device, terminal and computer readable storage medium

Abstract

The embodiment of the invention discloses an image splicing method and device, a terminal and a computer readable storage medium. According to the embodiment of the invention, the method comprises thesteps: respectively obtaining overlapping regions in a plurality of to-be-spliced images; determining a global homography matrix between the overlapping regions according to the feature point pairs between the overlapping regions; meshing processing is carried out on the overlapping region to obtain a plurality of overlapping sub-regions; based on the global homography matrix, determining a localhomography matrix between the overlapped sub-regions according to the feature line pairs between the overlapped sub-regions; and finally, splicing the images to be spliced according to the local homography matrix to obtain a spliced image. According to the method, the global homography matrix between the overlapping regions is calculated according to the feature point pairs between the overlapping regions, and then the local homography matrix between the overlapping sub-regions is calculated according to the feature line pairs on the basis of the global homography matrix, so that fine splicing of the to-be-spliced images is realized, and the splicing effect of the images can be improved.

Description

technical field [0001] The present application relates to the technical field of image processing, and in particular to an image splicing method, device, terminal, and computer-readable storage medium. Background technique [0002] In the field of new retail (such as unmanned vending smart retail), when taking pictures of commodity shelves, it is difficult to take a complete image of the commodity shelf at one time due to the space between the shelves and the limitation of space, which will affect the possession of the commodity shelf. Therefore, it is necessary to stitch the images of commodity shelves. [0003] However, because there are many repeated products in the images of the commodity shelves and the similarity is large, it is easy to cause image stitching misalignment by using the traditional image stitching method, and the stitching effect is poor. Contents of the invention [0004] Embodiments of the present application provide an image stitching method, device...

AI Technical Summary

Problems solved by technology

[0002] In the field of new retail (such as unmanned vending smart retail), when taking pictures of commodity shelves, it is difficult to take a complete picture of commodity shelves at one time due to the distance between shelves and space constraints, which will affect the possession of commodity shelves. rate and display rate, etc., therefore, it is necessary to splicing images of commodity shelves

[0003] However, because there are many repeated products in the images of the commodity shelves and the similarity is large, it is easy to have image stitching misalignment by using the traditional image stitching method, and the stitching effect is poor.

Images