Image fusion method

Abstract

The invention discloses an image fusion method comprising the steps that the effective area of an image under fusion is determined in the splicing part of the image under fusion; the search criterion E(x,y)=aEcolor(x,y)2+bEgeometry(x,y) of an optimal stitching line is established, and the search direction in which the sum of the intensity values of pixel points in the current search line and the pixels in the following five lines is minimum acts as the extension direction of the optimal stitching line; and the fusion result of the image under fusion is calculated by using a multi-resolution fusion algorithm based on the Laplacian Pyramid according to the position of the optimal stitching line. According to the image fusion method, only the pixel points in the effective area act as the extension direction of the optimal stitching line and the influence of color and structural difference is both considered so that the weight of feature points is increased, the priority extension mode under the condition of multiple possible extension directions is also clear and thus the image fusion effect is enhanced.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to an image fusion method. Background technique [0002] Image stitching and fusion technology is an important branch of image processing research, and has important applications in remote sensing, aerial photography, panoramic cameras and other fields. Image mosaic and fusion technology includes two key steps: image matching and image fusion. Image matching is the process of registering several images that have the same part in the shooting content, and these matched parts are called splicing parts. Because the image to be matched, that is, the original image is an image with pixel differences due to different acquisition time, shooting angle or / and lighting conditions, the matched image, that is, the image to be fused, is likely to be in some positions in the above overlapping part. Image ghosting occurs. Please refer to Figure 1, this figure includes two original imag...

AI Technical Summary

Problems solved by technology

[0004] Under the guidance of the above core ideas, some phased results have been achieved in the selection of fusion algorithms, the establishment of search criteria, and the expansion of seam lines. The seam problem, the fusion effect needs to be improved urgently

Images