Abstract:

The paper proposes a new method of generating imitation digital camouflage based on the background images. First, convert the color space of the background pattern. Dominant colors are extracted using the K-means clustering analysis. The minimum size of digital unit is determined by the detection distance. Finally, the digital camouflage spots are generated. The effect of camouflage is assessed by Canny edge detection. Experimental result shows that the generated digital camouflage can effectively change the original contour of the target, and can be perfectly blended with the background to achieve a good camouflage effect.

Key words:  background images, imitation digital camouflage, cluster analysis, spots, effect evaluation