Ship Detection in Sentinel 2 Multi-Spectral Images with Self-Supervised Learning

Automatic ship detection provides an essential function towards maritime domain awareness for security or economic monitoring purposes. This work presents an approach for training a deep learning ship detector in Sentinel-2 multi-spectral images with few labeled examples. We design a network architecture for detecting ships with a backbone that can be pre-trained separately. By using self supervised learning, an emerging unsupervised training procedure, we learn good features on Sentinel-2 images, without requiring labeling, to initialize our network’s backbone. The full network is then fine-tuned to learn to detect ships in challenging settings. We evaluated this approach versus pre-training on ImageNet and versus a classical image processing pipeline. We examined the impact of variations in the self-supervised learning step and we show that in the few-shot learning setting self-supervised pre-training achieves better results than ImageNet pre-training. When enough training data are available, our self-supervised approach is as good as ImageNet pre-training. We conclude that a better design of the self-supervised task and bigger non-annotated dataset sizes can lead to surpassing ImageNet pre-training performance without any annotation costs.

[1]  Matthias Bethge,et al.  Contrastive Learning Inverts the Data Generating Process , 2021, ICML.

[2]  Fillia Makedon,et al.  A Survey on Contrastive Self-supervised Learning , 2020, Technologies.

[3]  Abhinav Gupta,et al.  Demystifying Contrastive Self-Supervised Learning: Invariances, Augmentations and Dataset Biases , 2020, NeurIPS.

[4]  Ruslan Salakhutdinov,et al.  Self-supervised Learning from a Multi-view Perspective , 2020, ICLR.

[5]  Henning Heiselberg,et al.  Ship-Iceberg Discrimination in Sentinel-2 Multispectral Imagery by Supervised Classification , 2017, Remote. Sens..

[6]  Henning Heiselberg,et al.  A Direct and Fast Methodology for Ship Recognition in Sentinel-2 Multispectral Imagery , 2016, Remote. Sens..

[7]  Mario Caetano,et al.  BigEarthNet Dataset with A New Class-Nomenclature for Remote Sensing Image Understanding , 2020, 2001.06372.

[8]  Bo Li,et al.  Ship detection and classification from optical remote sensing images: A survey , 2020 .

[9]  Krištof Oštir,et al.  Vessel detection and classification from spaceborne optical images: A literature survey , 2018, Remote sensing of environment.

[10]  Francesca Bovolo,et al.  Self-supervised pre-training enhances change detection in Sentinel-2 imagery , 2021, ICPR Workshops.

[11]  Christophe Claramunt,et al.  N-YOLO: A SAR Ship Detection Using Noise-Classifying and Complete-Target Extraction , 2021, Remote. Sens..

[12]  Weiwei Sun,et al.  R-CNN-Based Ship Detection from High Resolution Remote Sensing Imagery , 2019, Remote. Sens..

[13]  Bo Li,et al.  Ship Detection in Multispectral Satellite Images Under Complex Environment , 2020, Remote. Sens..

[14]  Christophe Claramunt,et al.  H-YOLO: A Single-Shot Ship Detection Approach Based on Region of Interest Preselected Network , 2020, Remote. Sens..

[15]  Fangfang Ma,et al.  A Large-Scene Deceptive Jamming Method for Space-Borne SAR Based on Time-Delay and Frequency-Shift with Template Segmentation , 2020, Remote. Sens..