Transfer Learning With Nonlinear Spectral Synthesis for Hyperspectral Target Detection

Spectral distortion severely limits detection performance in hyperspectral imagery, while feature learning with neural networks could provide sufficient capacity to enhance spectral consistency. This article designs an end-to-end hyperspectral target detection (HTD) network based on transfer learning and nonlinear spectral synthesis (TLNSS). We first utilize bilinear mixture model (BMM) to synthesize nonlinear target and background spectra for training sample augmentation, which could better characterize ground objects in complex environments. Due to the mutual constraints between the quantity and diversity of the synthesized spectra, transfer learning is introduced to further address data insufficiency. Specifically, we propose an asymmetric autoencoder with a particularly designed multilevel loss to maximally distinguish the reconstruction residuals of background and target, where the multiscale feature extraction subnetwork is trained with abundant reference data, and the simple restoration subnetwork is updated with the simulated spectra. To effectively reconstruct the input as expected, the features extracted from different blocks are complementarily integrated through residual attention. Finally, we accumulate reconstruction residuals across all levels for final detection. The experimental results and ablation analysis of single-data detection on three hyperspectral images verify the superiority and effectiveness of the proposed method and further cross-data detection consolidates the satisfactory tolerance of TLNSS to spectral variation.