Target identification method based on hyperspectral image of unmanned aerial vehicle

The invention belongs to the technical field of image processing, and discloses a target identification method based on a hyperspectral image of an unmanned aerial vehicle, which comprises the following steps: S1, collecting and preprocessing a spectrum of a camouflage material; s2, collecting and preprocessing airborne hyperspectral data; s3, selecting a pure background dictionary Ab based on the spatial spectrum features; and S4, reconstructing the spectrum of the to-be-detected pixel through the background dictionary, and judging whether the to-be-detected pixel belongs to a target or not by using a reconstruction error. According to the invention, through the homogeneity of the image and the Euclidean distance between the image and the laboratory spectrum, the constraint condition of background dictionary selection is established, the spatial spectrum characteristics of the airborne hyperspectral image are fully utilized, and the pure background pixels are selected for background dictionary sparse represe