A Deep Neural Network Based on an Attention Mechanism for SAR Ship Detection in Multiscale and Complex Scenarios

Chen Chen,Chuan He,Changhua Hu,Licheng Jiao,Hong Pei

doi:10.1109/access.2019.2930939

Abstract

Synthetic aperture radar (SAR) ship detection based on deep learning has been widely applied in recent years. However, two main obstacles are hindering SAR ship detection. First, the identification of ships in a port is seriously disrupted by the presence of onshore buildings. It is difficult for the existing detection algorithms to effectively distinguish the targets from such a complex background. Additionally, it appears more complicated to accurately locate densely arranged ships. Second, the ships in SAR images exist at a variety of scales due to multiresolution imaging modes and the variety of ship shapes; these pose a much greater challenge to ship detection. To solve the above problems, this paper proposes an object detection network combined with an attention mechanism to accurately locate targets in complex scenarios. To address the diverse scales of ship targets, we construct a loss function that incorporates the generalized intersection over union (GIoU) loss to reduce the scale sensitivity of the network. For the final processing of the results, soft nonmaximum suppression (Soft-NMS) is also introduced into the model to reduce the number of missed detections for ship targets in the presence of severe overlap. The experimental results reveal that the proposed model exhibits excellent performance on the extended SAR ship detection dataset (SSDD) while achieving real-time detection.

Highlights

With the advancement of the ocean industry, ships are playing an increasingly essential role in marine development and transportation
We propose a single-stage object detection algorithm combined with an attention mechanism to solve the current problems arising in the context of ship detection in Synthetic aperture radar (SAR) images
This paper proposes a method of ship detection in SAR images based on an attention mechanism

Summary

INTRODUCTION

With the advancement of the ocean industry, ships are playing an increasingly essential role in marine development and transportation. Most SAR ship detection algorithms based on deep learning adopt a two-stage detection framework based on Faster R-CNN, which emphasizes detection accuracy while ignoring the detection speed This results in failure to detect the targets in real time. We propose a single-stage object detection algorithm combined with an attention mechanism to solve the current problems arising in the context of ship detection in SAR images. We integrate an attention mechanism into the network to obtain salient feature maps at different depths and fuse corresponding multiscale features, thereby improving the accuracy of the network in detecting and locating densely arranged ship targets against complex backgrounds. 4) The proposed model is based on a single-stage object detection algorithm and can achieve a good detection effect while maintaining a fast detection speed It can support real-time ship target detection.

METHODS

LOSS FUNCTION

INTRODUCTION TO THE EXPERIMENTAL PLATFORM AND DATASET

EXPERIMENTAL DETAILS

Findings

CONCLUSION