"Superpixel-Driven Graph Adaptive Transformer for Hyperspectral Target Detection"

"The San Diego Airport dataset is the first experimental dataset used to validate the proposed method. It was acquired by the airborne visible\/infrared imaging spectrometer (AVIRIS) over the San Diego Airport area in California, USA. The original hyperspectral data cover a wavelength range of 370\\text{-}2510 nm, consisting of 224 spectral bands, with a spatial resolution of approximately 3.5 meters. During preprocessing, invalid bands were removed, resulting in 189 bands retained for experimental analysis. In the experiments, multiple aircraft within the scenes were selected as detection targets, while the remaining regions were treated as background. Notably, the airport area contains a large number of metallic structures, such as rooftops and runways. The spectral similarity between these background pixels and aircraft targets often leads to significant spectral confusion, causing false detections and a higher false alarm rate. "