Abstract
The segmentation of ship targets in remote sensing images is of great importance in both military and civil fields. The existing methods often suffer from a low overlap between region proposals, which are constrained to be the horizontal rectangles and the ground truth. This enhances the problems of background noise and results in more false positives and a lower recall. Hence, we propose a method to generate the rotated bounding boxes that are more appropriate for ship segmentation. The proposed method is based on the Mask R-CNN framework, and the key contribution lies in the approach to predefine the rotated bounding boxes and generate the rotationally unconstrained region proposals. Specifically, for a double-stage proposal generation, a multiangle region proposal layer is designed; following this, an adapted alignment approach is proposed to extract the features for each proposal. In addition to the above-mentioned two proposed methods, we adopt a second stage to refine the proposals with rotation regression and mask prediction. We evaluated the proposed method using a remote sensing dataset with extensively labeled ship targets, and the experimental results show that the proposed method performs better than its competitors. The proposed method significantly increases the recall of densely arranged ships while substantially reducing the number of false positives.
Highlights
Ship target detection and segmentation in remote sensing imagery plays an important role in various applications both in military and civil fields, e.g. search and rescue of ships, monitoring of entry and exit ships and national defense [1]–[6]
The contributions of this paper can be summarized as follows: (1) the proposed methods represent a step toward the simultaneous detection and segmentation of ship targets in remote sensing images using rotationally unconstrained region proposal, which have not been exploited in the literature, for this we proposed a new approach of labeling and created a dataset with pixel-level annotations; (2) we modified the Mask R-convolutional neural networks (CNNs) method to generate rotationally unconstrained bounding boxes, which is shown to be effective and beneficial for locating densely arranged ship targets; (3) we evaluated the use of joint learning and rotationally unconstrained proposals in the task of optical remote sensing ship target detection and segmentation, and conclude that, in remote sensing, joint learning with the segmentation task
As the results shown in Table.5, the proposed method yields a better result, especially with respect to number of False Negatives (FN) and Average Precision (AP).As the experimental result shows, Mask R-CNN [11] performs the best of these three state-ofthe-art methods, but there are still a large number of FNs in the results, especially in the side-by-side docked ships
Summary
Ship target detection and segmentation in remote sensing imagery plays an important role in various applications both in military and civil fields, e.g. search and rescue of ships, monitoring of entry and exit ships and national defense [1]–[6]. Zhang et al.: Rotationally Unconstrained Region Proposals for Ship Target Segmentation in Optical Remote Sensing the features of each proposal are extracted from a CNN feature map by RoI-Pooling This is followed by two processing branches, one for bounding box classification and regression, and another for mask prediction, which indicates whether each pixel in the bounding box lies in the object or not. The contributions of this paper can be summarized as follows: (1) the proposed methods represent a step toward the simultaneous detection and segmentation of ship targets in remote sensing images using rotationally unconstrained region proposal, which have not been exploited in the literature, for this we proposed a new approach of labeling and created a dataset with pixel-level annotations; (2) we modified the Mask R-CNN method to generate rotationally unconstrained bounding boxes, which is shown to be effective and beneficial for locating densely arranged ship targets; (3) we evaluated the use of joint learning and rotationally unconstrained proposals in the task of optical remote sensing ship target detection and segmentation, and conclude that, in remote sensing, joint learning with the segmentation task.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
