A fully convolutional neural network for low-complexity single-stage ship detection in Sentinel-1 SAR images
A fully convolutional neural network for low-complexity single-stage ship detection in Sentinel-1 SAR images
复制标题
DOI：
10.1109/igarss.2017.8127094
复制
发表时间：
2017-07
期刊：
影响因子：
--
通讯作者：
D. Cozzolino;G. D. Martino;G. Poggi;L. Verdoliva
中科院分区：
文献类型：
--
作者：
D. Cozzolino;G. D. Martino;G. Poggi;L. Verdoliva
船舶探测是基于sar的海上监视的一项基本任务。除了提供高可靠性之外，为了在合理的时间内分析大面积的区域，一个好的检测器需要计算量轻。提出了一种用于SAR图像船舶检测的全卷积神经网络。由于结构相对简单，复杂性仍然很低，可以采用单阶段方法，从而避免了CFAR预筛选可能出现的错误。在Sentinel-1数据集上的实验证明，所提出的CNN比CFAR检测可靠得多。
Ship detection is a fundamental task for SAR-based maritime surveillance. Besides providing high reliability, a good detector is required to be computationally light, in order to analyze huge areas in a reasonable time. We propose a fully convolutional neural network for ship detection in SAR images. Thanks to a relatively simple architecture, complexity remains low enough to allow for a single-stage approach, thus avoiding the possible errors of CFAR pre-screening. Experiments on a Sentinel-1 dataset prove the proposed CNN to be much more reliable than CFAR detection.