Deep Learning Based Ships Detections from ALSAT-2 Satellite Images

Object detection is defined as the task of classification and localization of objects in images that got a lot of attention in remote sensing community. In the past few years, it has gained prominence in recent years due to its widespread applications. Deep learning (DL) techniques were nearly the best performing methods among all others for this task. In this paper, a multimodal DL framework for ships detection from ALSAT-2A multispectral images is developed. Unet and Mask-RCNN models are trained separately and then a majority voting (MV) decision fusion is performed to enhance the results. Experimental results prove the effectiveness of the method developed under numerous evaluation metrics.