The introduction of sparse representation provides a new way for the detection of hyperspectral images. In the detection process, the structure of the dictionary is obtained directly from the hyperspectral image and there are uncertainties. It is very likely that the dictionary may not contain any spectral characteristic information of target pixels, which affects the accuracy in target detection. In order to solve the following problem, this paper proposes a novel method of hyperspectral imagery target detection based on sparse representation, using an unsupervised method to complete the construction of dictionary to ensure that the dictionary contains some spectral information of target pixels. Experiments have been carried out on one hyperspectral image, which reveals that the method we proposed shows an outstanding detection performance.
展开▼
