Remote Sensing Change Detection via Weak Temporal Supervision

2026-01-05

arXiv.org

AI Summary

This work addresses the challenges of semantic change detection in remote sensing, which are primarily constrained by the scarcity of pixel-level annotations and limited cross-domain generalization. The authors propose a weakly temporally supervised method that requires no additional labeling: leveraging existing single-temporal datasets to construct bi-temporal samples by assuming that image pairs from the same location at different times exhibit no significant change, while those from different locations constitute changed samples. The approach integrates object-aware change map generation, iterative refinement, and a noise-robust training mechanism within a self-supervised learning framework built upon both real and synthetically generated bi-temporal image pairs. Evaluated on extended versions of the FLAIR and IAILD aerial datasets, the method significantly improves performance in zero-shot and few-shot settings and demonstrates strong scalability across large-scale regions in France.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Machine Learning: Semi-Supervised Learning
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Semantic change detection in remote sensing aims to identify land cover changes between bi-temporal image pairs. Progress in this area has been limited by the scarcity of annotated datasets, as pixel-level annotation is costly and time-consuming. To address this, recent methods leverage synthetic data or generate artificial change pairs, but out-of-domain generalization remains limited. In this work, we introduce a weak temporal supervision strategy that leverages additional temporal observations of existing single-temporal datasets, without requiring any new annotations. Specifically, we extend single-date remote sensing datasets with new observations acquired at different times and train a change detection model by assuming that real bi-temporal pairs mostly contain no change, while pairing images from different locations to generate change examples. To handle the inherent noise in these weak labels, we employ an object-aware change map generation and an iterative refinement process. We validate our approach on extended versions of the FLAIR and IAILD aerial datasets, achieving strong zero-shot and low-data regime performance across different benchmarks. Lastly, we showcase results over large areas in France, highlighting the scalability potential of our method.

Problem

• Remote Sensing
• Change Detection
• Weak Supervision
• Semantic Change
• Data Scarcity

Innovation

• weak temporal supervision
• semantic change detection
• remote sensing
• zero-shot learning
• object-aware refinement