Image splicing method for random texture scene and application thereof

Abstract

The invention discloses an image splicing method for a random texture scene, and the method comprises the steps: obtaining a current frame image, carrying out the position marking, so as to determine an image group which needs to be spliced with the current frame image, and enabling the to-be-spliced frame images in the image group to be combined with the current frame image, and carrying out the combination marking of the to-be-spliced frame images and the current frame image as calculation pairs; selecting a plurality of target pixel points according to the overlapped part of the current frame image and the to-be-spliced frame image in each group of calculation pairs to independently calculate offset values; calculating a similarity value according to the overlapped part, and selecting a group of target pixel points with the highest similarity value to calculate an offset value as an accurate offset value of the corresponding calculation pair; and according to the accurate deviation value, carrying out fusion processing on the overlapping region of the calculation pair so as to splice the current frame image into a complete image. The problems that an existing image splicing method of a random texture scene is low in splicing precision, large in calculated amount and poor in splicing real-time performance due to the fact that operation such as coordinate transformation needs to be carried out can be solved.

Description

technical field [0001] The present application relates to the technical field of wafer defect detection, and more particularly, to an image stitching method for random texture scenes, an image stitching device for random texture scenes, an electronic device, and a computer-readable storage medium. Background technique [0002] In recent years, with the rapid development of semiconductor integrated circuits, the performance of integrated circuit devices is also rapidly improving, and the manufacturing process thereof has also become more and more complicated. In the manufacturing process of semiconductor integrated circuits, it is particularly important to timely discover the problems existing in the product manufacturing process. In the process of semiconductor wafer inspection, the field of view of the camera is limited, and it is a feasible way to construct a multi-camera array. Therefore, the rapid and accurate stitching of array industrial images has become a key step in...

AI Technical Summary

Problems solved by technology

[0005] Aiming at at least one defect or improvement requirement of the prior art, the present invention provides an image mosaic method, device, electronic equipment and computer-readable storage medium for random texture scenes, aiming to solve the existing image mosaic method of random texture scenes The stitching accuracy is low, and operations such as coordinate transformation are required, resulting in a large amount of calculation and poor real-time stitching.

Images