An Automated Method for the Creation of Oriented Bounding Boxes in Remote Sensing Ship Detection Datasets

Giorgos Savathrakis, Antonis Argyros; Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops, 2024, pp. 830-839

In a variety of maritime applications, the task of accurately detecting ships from remote sensing images is of significant importance. Various object detection algorithms localize objects by identifying either their Horizontal Bounding Boxes (HBBs) or their Oriented Bounding Boxes (OBBs). OBBs provide a far more accurate/tighter localization of object regions as well as their orientation. Several ship detection datasets provide annotations that include both HBBs and OBBs. However, many of them do not include OBB annotations. In this work, we propose a method which takes the ships' HBB annotations as input, and automatically calculates the corresponding OBBs. The proposed method consists of three main parts, (a) object segmentation that is built upon the Segment-Anything Model (SAM) to calculate object masks based on the information provided by the HBBs, (b) morphological filtering which eliminates possible artifacts stemming from the segmentation process, and (c) contour detection applied to the post-processed masks that are used to compute the optimal OBBs of the target objects. By automating the process of OBB annotation, the proposed method permits the exploitation of existing HBB-annotated datasets to train ship detectors of improved performance. We support this finding by reporting the results of several experiments that involve standard datasets, as well as state of the art object detectors.