A fast stitching method for irregularly broken single-sided images

Abstract

The invention relates to a fast splicing method for irregularly broken single-sided images. First, the fragments of an irregularly broken single-sided image are scanned so that a digital image is obtained, and the counterclockwise pixel grey value sequence of the edge of each image fragment is acquired; each neighboring image fragment is acquired by computing the correlation coefficients between the pixel grey value sequences of the edges of the image fragments, and the irregularly broken image fragments can be spliced fast and automatically by rotating, translating fragments, etc. The method has the advantages that the wealth of edge pixel information of images is utilized fully, and the edge shape matching of irregularly broken images is converted into the matching of edge pixel grey values; thus, the complexity level of splicing computation is lowered largely, and the computation procedures are simplified. The method is the extension of the existing splicing method of regularly broken files and can be realized and promoted easily.

Description

technical field [0001] The invention relates to a fast splicing method for irregularly broken single-sided images. Background technique [0002] The splicing of broken documents has important applications in the fields of forensic evidence recovery, historical document restoration, and military intelligence acquisition. Traditionally, stitching and restoration work needs to be done manually, with high accuracy but low efficiency. When the number of fragments is huge, it is difficult for manual splicing to complete the task in a short time. With the development of computer technology, people try to develop automatic splicing technology of shredded paper to improve the efficiency of splicing restoration. [0003] At present, the research on irregular fragment splicing and matching at home and abroad mainly focuses on the contour curvature matching of image fragments. Many scholars have proposed related algorithms and have achieved certain results. Ying Shan et al. proposed...

AI Technical Summary

Problems solved by technology

[0004] These algorithms have matched the contours to a certain extent, but most of them use the curvature features of the points on the contour curve to match. First, the calculation is large and the steps are cumbersome. Sharp corners are corrupted by smoothing or cannot be represented

And at present, the research on splicing fragments at home and abroad is mainly focused on the research on text shreds, while there are few studies on image splicing and there is a lack of fast and effective algorithms.

Images