Balance Scene Learning Mechanism for Offshore and Inshore Ship Detection in SAR Images

Huge imbalance of different scenes’ sample numbers seriously reduces synthetic aperture radar (SAR) ship detection accuracy. Thus, to solve this problem, letter proposes a balance scene learning mechanism (BSLM) for offshore and inshore in SAR images. BSLM involves three steps: 1) based on unsupervised representation learning, generative adversarial network (GAN) is used extract the features images; 2) using these features, binary cluster (offshore/inshore) conducted by ${K}$ -means; 3) finally, small cluster’s samples (inshore) are augmented via replication, rotation transformation or noise addition another big (offshore), so as eliminate bias obtain balanced ability that can enhance benefits improve This applies four widely open-sourced deep detectors, i.e., faster regions-convolutional neural (Faster R-CNN), Cascade R-CNN, single shot multibox detector (SSD), RetinaNet, verify its effectiveness. Experimental results open data set (SSDD) reveal greatly accuracy, especially more complex scenes.