Fuzzy c-means clustering based digital camouflage pattern design and its evaluation

Abstract

Optical camouflage is one of the significant ways to protect military weapons in modern wars. Effective camouflage design can largely reduce the conspicuity of the fixed targets in a real battlefield. In this paper, an innovative digital camouflage pattern design based on fuzzy c-means clustering and its evaluation is presented to imitate the background both in color and texture. In this new method, histogram quantization and matching between the real background and I military standard colors is first carried out in HSV color space. Then camouflage pattern is abstained by a combination of the fuzzy c-means clustering and morphological operations, such as open operation and close operation. And the camouflage scene is composed with tank using new digital camouflage pattern. Finally, several common edge detectors are utilized to give an objective evaluation to the new digital camouflage pattern which is painting on the target in a composed camouflage scene. Assessment results also show the effectiveness and potency of the new proposed method.