RDP-Net: Region Detail Preserving Network for Change Detection

TL;DR

RDP-Net is a lightweight, detail-preserving CNN designed for change detection in earth observation, utilizing an efficient training strategy and an edge loss to improve performance on compact devices.

Contribution

The paper introduces RDP-Net, a novel CNN with a new backbone and training strategy that achieves state-of-the-art change detection performance with fewer parameters.

Findings

Achieves state-of-the-art performance with only 1.70M parameters.

Effective training strategy enables learning from easy to hard tasks.

Edge loss improves attention to boundary regions and small areas.

Abstract

Change detection (CD) is an essential earth observation technique. It captures the dynamic information of land objects. With the rise of deep learning, convolutional neural networks (CNN) have shown great potential in CD. However, current CNN models introduce backbone architectures that lose detailed information during learning. Moreover, current CNN models are heavy in parameters, which prevents their deployment on edge devices such as UAVs. In this work, we tackle this issue by proposing RDP-Net: a region detail preserving network for CD. We propose an efficient training strategy that constructs the training tasks during the warmup period of CNN training and lets the CNN learn from easy to hard. The training strategy enables CNN to learn more powerful features with fewer FLOPs and achieve better performance. Next, we propose an effective edge loss that increases the penalty for errors on details and improves the network's attention to details such as boundary regions and small areas. Furthermore, we provide a CNN model with a brand new backbone that achieves the state-of-the-art empirical performance in CD with only 1.70M parameters. We hope our RDP-Net would benefit the practical CD applications on compact devices and could inspire more people to bring change detection to a new level with the efficient training strategy. The code and models are publicly available at https://github.com/Chnja/RDPNet.