A Priority Algorithm and Image Restoration Method Based on Structural Difference and Edge Structural Coefficient

Abstract

The invention discloses a priority algorithm and an image repair method based on structural differences and edge structure coefficients to be repaired. The algorithm steps include: reading the damaged image and marking the area to be repaired; extracting all known pixel information in the target area ; Construct a structural difference matrix based on the extracted known pixel information; extract a window area of ​​a certain size centered on any point p to be repaired on the boundary of the area to be repaired in the target area, and calculate the edge structure coefficient T of the point to be repaired p ( p), standard deviation parameter S p (p) and confidence parameter C(p); according to the formula P(p)=T(p)·C(p)·S p (p) Calculate the priority value P(p) of the point to be repaired; traverse all points to be repaired on the boundary of the area to be repaired, repair the window area centered on the point to be repaired with the largest priority value, and update the boundary of the area to be repaired ; Repeat the above steps until the area to be repaired is completely repaired. The present invention obtains a stable priority value based on the difference size of each target area, and greatly reduces the priority calculation time.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a priority algorithm and an image repair method based on structural differences and edge structure coefficients to be repaired. Background technique [0002] Image restoration is an ancient art that originated during the European Renaissance, and usually refers to restoring lost or damaged pixels or areas in an image. People are very sensitive to discontinuous boundaries, and the lost area generally includes structure, texture and smoothness, and the structural coherence is closer to the real visual effect, so the researchers separated these three layers, and the restoration structure layer is a more important part of it. The selection of the repair order has a great influence on the repair results, and generally the structural blocks should be repaired first. Criminisi uses the isoluminescence line direction of the local boundary and the unit normal direction to define...

AI Technical Summary

Problems solved by technology

[0005] Although the above algorithms can realize the image repair function to a certain extent, there are some insurmountable problems in the algorithms, such as the filling sequence is not stable enough, the matching criteria are not reasonable enough, and it takes a long time, etc.

Images