Digital Camouflage Pattern Generation Approach Based on Fragmentation Edges

This paper demonstrates a camouflage pattern generation approach based on fragmentation edges. The proposed camouflage pattern generation method features the K-means algorithm improved by the sparrow search algorithm (SSA-Kmeans) to extract the background dominant colors, and background pattern segmentation using the simple linear iterative clustering with random matrix algorithm according to the fragmentation weight. Compared with conventional methods, namely the watershed segmentation algorithm and the recursive overlapping of pattern templates algorithm, the proposed adaptive camouflage pattern generation approach shows obvious advantages in texture camouflage effect, and the difference is reduced by more than 64% compared with the other two algorithms.