Distilled Heterogeneous Feature Alignment Network for SAR Image Semantic Segmentation.

Synthetic aperture radar (SAR) image semantic segmentation has attracted increasing attention in the remote-sensing community recently, due to SAR’s all-time and all-weather imaging capability. However, SAR images are generally more difficult to be segmented than their electro-optical (EO) counterparts, since speckle noises and layovers are inevitably involved in SAR images. On the other hand, EO images could only be obtained under cloud-free conditions, which limits their applications. To this end, this letter investigates how to introduce EO features to assist the training of an SAR-segmentation model so that the model could segment SAR images without their EO counterparts in the application and proposes a distilled heterogeneous feature alignment network (DHFA-Net), where an SAR-segmentation student model learns and aligns the features from a pretrained EO-segmentation teacher model. In the proposed DHFA-Net, both the student and teacher models employ an identical architecture but different parameter configurations, and a heterogeneous feature distillation module (HFDM) is explored for transferring latent EO features from the teacher model to the student model through heterogeneous feature distillation and then supervising the training of the SAR-segmentation model. Moreover, a heterogeneous feature alignment module (HFAM) is designed to aggregate multiscale features for segmentation by the feature alignment approach in each of the student and teacher models. By enabling the multiscale heterogeneous feature aggregation, the SAR segmentation performance could be boosted. Experimental results on two public datasets demonstrate the superiority of the proposed DHFA-Net.