Vector spaces, invariance, and camouflage

摘要

We present a method based on an orthonormal vector space basis representation to detect camouflaged targets in natural environments. Because the method is intensity invariant we detect camouflage targets independently of the illumination conditions. The detection technique docs not require knowing the exact camouflage pattern, but only the class of patterns (foliage, netting, woods...). We used nonlinear filtering based on the calculation of several correlations. Moreover, the nonlinearity of the filtering process allows a high discrimination against false targets. Several experiments confirm the target detectability where strong camouflage might delude even human viewers.