Single-Temporal Supervised Learning for Universal Remote Sensing Change Detection

TL;DR

This paper introduces STAR, a novel single-temporal supervised learning approach for remote sensing change detection that uses unpaired images for training, enabling high accuracy and generalization without the need for labor-intensive bitemporal labels.

Contribution

The paper presents STAR, a new framework that leverages unpaired images for change detection, and proposes ChangeStar2, a unified model capable of multiple change detection tasks with state-of-the-art results.

Findings

Achieves state-of-the-art performance on eight datasets.

Effectively handles multiple change detection scenarios.

Operates without paired bitemporal training data.

Abstract

Bitemporal supervised learning paradigm always dominates remote sensing change detection using numerous labeled bitemporal image pairs, especially for high spatial resolution (HSR) remote sensing imagery. However, it is very expensive and labor-intensive to label change regions in large-scale bitemporal HSR remote sensing image pairs. In this paper, we propose single-temporal supervised learning (STAR) for universal remote sensing change detection from a new perspective of exploiting changes between unpaired images as supervisory signals. STAR enables us to train a high-accuracy change detector only using unpaired labeled images and can generalize to real-world bitemporal image pairs. To demonstrate the flexibility and scalability of STAR, we design a simple yet unified change detector, termed ChangeStar2, capable of addressing binary change detection, object change detection, and semantic change detection in one architecture. ChangeStar2 achieves state-of-the-art performances on eight public remote sensing change detection datasets, covering above two supervised settings, multiple change types, multiple scenarios. The code is available at https://github.com/Z-Zheng/pytorch-change-models.