MISGNet: A Multilevel Intertemporal Semantic Guidance Network for Remote Sensing Images Change Detection

• Published in: IEEE journal of selected topics in applied earth observations and remote sensing 2025, Vol.18, p.1827-1840
• Main Authors: Cui, Binge, Liu, Chenglong, Li, Haojie, Yu, Jianzhi
• Format: Article
• Language: English
• Subjects: Accuracy; Aggregation; Change detection; Change detection (CD); Convolutional neural networks; Data mining; difference aggregates; Feature extraction; Graphical representations; Image acquisition; Land cover; Modules; Multilevel; Multiplication; Recurrent neural networks; Remote sensing; semantic guidance; Semantics; Streams; Target detection; Training; Transformers; Urban planning
• ISSN: 1939-1404; 2151-1535
• DOI: 10.1109/JSTARS.2024.3508692

The precise identification of semantic changes in remote sensing images is of great significance in the domains of urban planning and disaster assessment. Nevertheless, current change detection models are inadequate when it comes to model semantic interactions in pairings of temporal images. This leads to poor identification of identical semantic targets that have unique features. In this article, we put forth a proposal for a multilevel intertemporal semantic guidance network (MISGNet) that would effectively derive representations of semantic changes. The bitemporal images' multilevel features are initially extracted using a transformer feature extractor. These extracted features are then bidirectionally semantically augmented using a semantic guidance module (SGM) in order to acquire more comprehensive semantic representations. Particularly, in order to obtain the object's semantic representation, the land cover objects in the multilevel features of the bitemporal images are soft clustered and mapped to the graph space, with each vertex representing an object in the graph. Then, bidirectional semantic enhancement is achieved through the use of intertemporal nonlocal operations, which strengthen the semantic representation of the bitemporal images. Moreover, a multilevel difference aggregation module (MDAM) is implemented in order to enhance the efficacy of summarizing distinctions between different levels and highlighting semantic changes in features. This is achieved through the utilization of pixelwise addition and pixelwise multiplication. Extensive experiments conducted on three publicly available datasets provide irrefutable evidence that our model outperforms alternative methods across various evaluation metrics.