HR-SAR-Net: A Deep Neural Network for Urban Scene Segmentation from High-Resolution SAR Data

TL;DR

This paper introduces HR-SAR-Net, a lightweight deep neural network capable of accurately segmenting urban scenes from high-resolution SAR data, enabling large-scale applications with minimal computational resources.

Contribution

The paper presents a novel, small deep neural network for urban scene segmentation from high-resolution SAR data, achieving high accuracy with efficient computation and a new ground truth generation procedure.

Findings

Achieved 95.19% pixel accuracy and 74.67% mean IoU.

Model is very small with only 63k parameters and runs at 500Mpx/s.

Additional SAR antennas and multi-flight data significantly improve segmentation accuracy.

Abstract

Synthetic aperture radar (SAR) data is becoming increasingly available to a wide range of users through commercial service providers with resolutions reaching 0.5m/px. Segmenting SAR data still requires skilled personnel, limiting the potential for large-scale use. We show that it is possible to automatically and reliably perform urban scene segmentation from next-gen resolution SAR data (0.15m/px) using deep neural networks (DNNs), achieving a pixel accuracy of 95.19% and a mean IoU of 74.67% with data collected over a region of merely 2.2km${}^2$. The presented DNN is not only effective, but is very small with only 63k parameters and computationally simple enough to achieve a throughput of around 500Mpx/s using a single GPU. We further identify that additional SAR receive antennas and data from multiple flights massively improve the segmentation accuracy. We describe a procedure for generating a high-quality segmentation ground truth from multiple inaccurate building and road annotations, which has been crucial to achieving these segmentation results.