Given a target prior information, our goal is to propose a method forautomatically separating known targets of interests from the background inhyperspectral imagery. More precisely, we regard the given hyperspectral image(HSI) as being made up of the sum of low-rank background HSI and a sparsetarget HSI that contains the known targets based on a pre-learned targetdictionary constructed from some online Spectral libraries. Based on theproposed method, two strategies are briefly outlined and evaluatedindependently to realize the target detection on both synthetic and realexperiments.
展开▼
