Search papers, labs, and topics across Lattice.
This paper introduces FootprintNet, a novel approach for multi-temporal change detection (MTCD) that overcomes the limitations of existing methods by modeling dynamic building changes as interactions between latent states and actions. By implementing state-action transition constraints and leveraging temporal change-boundary cues, FootprintNet enhances the discrimination of dynamic footprints, allowing for more accurate detection of recurrent changes in urban environments. The proposed Building Change Dynamics Score (BCDS) offers a new metric for evaluating the temporal accuracy of change predictions, demonstrating superior performance over state-of-the-art methods in extensive experiments.
FootprintNet reveals how modeling dynamic changes as state-action interactions can significantly enhance the accuracy of urban change detection in remote sensing.
Despite substantial progress in remote sensing multi-temporal change detection (MTCD), most existing MTCD methods still represent the dynamic process at each spatial location over the entire observation period using a single change category associated with the final observation. This implicit single-change assumption limits their ability to characterize regions of recurrent change closely related to human activities. To address this limitation, we introduce Urban Building Dynamics Detection (UBDD), which identifies building-change dynamic footprints, i.e., the temporal intervals in which changes occur, from multi-temporal imagery and produces pixel-wise classification masks. For regions undergoing two or more changes, UBDD introduces an independent multi-change class for unified representation, thereby enabling unified modeling of single- and multi-change processes. Furthermore, we propose FootprintNet, which abstracts building-change processes as interactions between latent states and actions, and imposes state-action transition constraints to guide the learning of causally coherent change trajectories. It further exploits temporal change-boundary cues to enhance feature contrast across boundary sides, thereby improving the discrimination among different dynamic footprints and enabling accurate detection of dynamic footprints. Moreover, we introduce the Building Change Dynamics Score (BCDS) to address the inability of conventional metrics to reflect the temporal proximity between predicted footprints and labels. It evaluates predictions according to their preservation of change semantics and temporal offsets from the corresponding labels. Extensive experiments on TSCD, MUDS, and WUSU demonstrate that FootprintNet outperforms current state-of-the-art methods. The code is available at https://github.com/zmoka-zht/FootprintNet.
