DeepAquaCluster: Using Satellite Images And Self-supervised Machine Learning Networks To Detect Water Hidden Under Vegetation
By: Ioannis Iakovidis , Zahra Kalantari , Amir Hossein Payberah and more
Potential Business Impact:
Finds water in satellite pictures without human help.
In recent years the wide availability of high-resolution radar satellite images along with the advancement of computer vision models have enabled the remote monitoring of wetland surface areas. However, these models require large amounts of manually annotated satellite images, which are slow and expensive to produce. To overcome this problem we use self-supervised training methods to train a model called DeepAquaCluster to segment radar satellite images into areas that separate water from land without the use of any manual annotations. Our final model outperforms other radar-based water detection techniques in our test dataset, achieving a 0.08 improvement in the Intersection Over Union metric.
Similar Papers
Semi-supervised learning for marine anomaly detection on board satellites
CV and Pattern Recognition
Find ocean problems using less labeled pictures.
AI Driven Water Segmentation with deep learning models for Enhanced Flood Monitoring
CV and Pattern Recognition
Spots floods faster using computer pictures.
Learning with less: label-efficient land cover classification at very high spatial resolution using self-supervised deep learning
CV and Pattern Recognition
Maps land from space with less data.