Assessment method for camouflage performance based on visual perception

Abstract

Camouflage stealth is of great significance in concealing military targets and countering military reconnaissance, but the assessment methods for camouflage performance remain some problems, such as incomplete evaluation indexes and inconsistency with the visual perception. This study proposes a comprehensive assessment model, named comprehensive similarity, CSsub for short. The model bases on subjective visual evaluation and contains three metrics of image color appearance, structure, and texture. The metrics are analysed by the image color appearance similarity (ICAS), structure similarity (ISS), and texture similarity (ITS), respectively, to measure the camouflage effectiveness between background and camouflage target images. The three weights, in CSsub, are fitted through the results of subjective experiments. In the experiments, five ships are camouflaged in six sea states, a total of 30 camouflage scenes. Finally, seven evaluation methods, including the proposed model, are analyzed, and compared to predict camouflage effectiveness. The results show that CSsub model has the best prediction performance. And the ICAS index, which is applied in camouflage assessment for the first time, accounts for a significant weight in the model, indicating that color has a great impact on image perception. So, this method based on visual perception has more superiority than others and can be used as an accurate and reliable assessment model in predicting camouflage performance.