Fragment reconstruction method based on local texture pattern

Abstract

The invention proposes a fragment reconstruction method based on local texture patterns, which belongs to the technical field of digital image processing. The method aims to extract a local binary pattern with normal direction based on the local texture features of fragments, so as to automatically stitch Broken and messy document (or image) fragments caused by factors, so that the original information can be restored to the greatest extent. First, image preprocessing is performed on the fragments, then the ND‑LBP unit is calculated, and the weight histogram of ND‑LBP is obtained, and finally the fragment reconstruction is completed by judging the matching relationship according to the ND‑LBP weight histogram. This method has the characteristics of small calculation, strong anti-noise ability and high discrimination.

Description

technical field [0001] The invention relates to a fragment reconstruction method based on a local texture mode, and belongs to the technical field of digital image processing. Background technique [0002] The problem of splicing fragments is a challenge encountered in many research fields, for example: damaged ancient calligraphy and painting classics need to be repaired by archaeologists, messy documents and photo fragments and other physical evidence need to be combined and restored by public security organs, and precious works of art are damaged. Afterwards, artists are required to repair and imitate. When the number of fragments is small, it can be spliced ​​by hand, but in the face of a large number of fragments, especially when irrelevant fragments may be mixed, manual splicing is an extremely heavy physical and mental work. At the beginning of the birth of computer vision technology, experts in related fields hoped to use the large-scale information processing capab...

AI Technical Summary

Problems solved by technology

[0002] The problem of splicing fragments is a challenge encountered in many research fields, for example: damaged ancient calligraphy and painting classics need to be repaired by archaeologists, messy documents and photo fragments and other physical evidence need to be combined and restored by public security organs, and precious works of art are damaged. Afterwards, artists need to repair the imitation

The challenge of matching based on edge information is that when facing fragments with similar edge shapes or missing edge parts, the matching accuracy is very low

In the actual situation, there are indeed such problems: First, when people shred a piece of paper, they often fold and tear it repeatedly, and many pieces with similar shapes will be torn out; this is especially true for documents shredded by a shredder. almost identical

This method takes into account the global nature of fragments, but the problem is: first of all, the content of many images and documents cannot be obtained in advance

[0009] This kind of LBP does not have rotation invariance. When the image is rotated, the LBP value also changes.

Images