TSP Image Restoration Model Based on NCC Correlation
Weining Feng, Lingxuan Weng · 2022
When restoring a large number of image fragments, the use of automatic splicing technology can improve the efficiency of image restoration. This paper constructs a TSP image restoration model based on NCC correlation, and proposes different restoration schemes for different image restoration scenarios. The modeling is carried out according to the process of image preprocessing, feature matching, problem conversion and image stitching. Firstly, each small photo is converted from RGB color space to HSV color space, and the V monochromatic channel image is separated, then the gray value matrix is constructed. Secondly, the edge feature vector of the picture is extracted and the NCC edge correlation matrix between the pictures is calculated. Finally, according to the correlation of the upper and lower azimuth between the pictures, the small photos are classified in columns, and then the column stitching and row stitching are completed in turn. In addition, according to different types of small photos, such as partial repetition, different sizes, more fragments and so on, this paper obtains the restored images through K-means clustering algorithm and image enhancement processing. The image restoration system constructed in this paper has better processing effect and higher efficiency.