Key points are not available for this paper at this time.
How to distinguish the camouflage from its natural background is a challenging problem in target detection. As sensing technology advances, more and more information can be extracted from the scenes of interest. This includes spatial information captured by cameras, spectral information retrieved from spectrometers, and polarimetric information obtained by polarimeters. Spatial, spectral, and polarimetric information reveal the different characteristics of objects and background. While the spectral information tend to tell us about the distribution of material components in a scene, polarimetric information tells us about surface feature, shape, shading, and roughness. Polarization tends to provide information that is largely uncorrelated with spectral and intensity images, thus has the potential to enhance many fields of optical metrology. However, both spectral and polarimetric detection systems may suffer from substantial false alarms and missed detection because of their respective background clutter. Since polarimetric and multispectral imaging can provide complementary discriminative information, to distinguish the camouflage target from its natural background, in this paper the visible and near infrared polarimetric information is jointly utilized using imagery fusion technology. A polarimetric imagery fusion algorithm was first proposed based on polarized modified soil adjusted vegetation index to distinguish objects under vegetable environment. Then, the spectral and polarimetric information was fused by using false-color mapping and fuzzy c-means clustering algorithm for more robust object separation. Experimental results have shown that better identification performance was achieved.
Zhou et al. (Thu,) studied this question.