MergeSAM: Unsupervised change detection of remote sensing images based on the Segment Anything Model

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Hu, Meiqi, Lu, Lingzhi, Han, Chengxi, Liu, Xiaoping
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911122529976320
author Hu, Meiqi
Lu, Lingzhi
Han, Chengxi
Liu, Xiaoping
author_facet Hu, Meiqi
Lu, Lingzhi
Han, Chengxi
Liu, Xiaoping
contents Recently, large foundation models trained on vast datasets have demonstrated exceptional capabilities in feature extraction and general feature representation. The ongoing advancements in deep learning-driven large models have shown great promise in accelerating unsupervised change detection methods, thereby enhancing the practical applicability of change detection technologies. Building on this progress, this paper introduces MergeSAM, an innovative unsupervised change detection method for high-resolution remote sensing imagery, based on the Segment Anything Model (SAM). Two novel strategies, MaskMatching and MaskSplitting, are designed to address real-world complexities such as object splitting, merging, and other intricate changes. The proposed method fully leverages SAM's object segmentation capabilities to construct multitemporal masks that capture complex changes, embedding the spatial structure of land cover into the change detection process.
format Preprint
id arxiv_https___arxiv_org_abs_2507_22675
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MergeSAM: Unsupervised change detection of remote sensing images based on the Segment Anything Model
Hu, Meiqi
Lu, Lingzhi
Han, Chengxi
Liu, Xiaoping
Computer Vision and Pattern Recognition
Recently, large foundation models trained on vast datasets have demonstrated exceptional capabilities in feature extraction and general feature representation. The ongoing advancements in deep learning-driven large models have shown great promise in accelerating unsupervised change detection methods, thereby enhancing the practical applicability of change detection technologies. Building on this progress, this paper introduces MergeSAM, an innovative unsupervised change detection method for high-resolution remote sensing imagery, based on the Segment Anything Model (SAM). Two novel strategies, MaskMatching and MaskSplitting, are designed to address real-world complexities such as object splitting, merging, and other intricate changes. The proposed method fully leverages SAM's object segmentation capabilities to construct multitemporal masks that capture complex changes, embedding the spatial structure of land cover into the change detection process.
title MergeSAM: Unsupervised change detection of remote sensing images based on the Segment Anything Model
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.22675