ChangeVLM: A Language-Guided Semantic Alignment Framework for Binary Remote Sensing Change Detection.
Saved in:
| Title: | ChangeVLM: A Language-Guided Semantic Alignment Framework for Binary Remote Sensing Change Detection. |
|---|---|
| Authors: | Li, Dongxu1 (AUTHOR), Chu, Peng1,2 (AUTHOR), Yang, Chen2,3 (AUTHOR), Wang, Zhen1,3,4 (AUTHOR) wzgroup@nwpu.edu.cn, Dai, Chuanjin1,4 (AUTHOR) |
| Source: | Remote Sensing. May2026, Vol. 18 Issue 10, p1671. 42p. |
| Subjects: | Remote sensing, Edge computing, Deep learning, High resolution imaging |
| Abstract: | Highlights: What are the main findings? ChangeVLM achieves competitive performance with only 28.73M trainable parameters and 560.7G FLOPs. It obtains F1 scores of 91.52%, 83.56%, and 75.29% on the LEVIR-CD, SYSU-ChangeDet, and HRCUS datasets, respectively. In addition, under an optimized inference setting with pre-extracted visual features, the effective computation is reduced to 13.05G FLOPs, suggesting its potential for deployment-oriented acceleration. The three core customized modules, including AFM, lightweight ChangeHead, and CACMF, address insufficient spatial localization, low inference efficiency, and weak semantic discrimination in binary remote sensing change detection. The language-related component is used only as an auxiliary semantic alignment regularizer during training and is not used for text generation. What are the implications of the main findings? The framework offers an end-to-end deployable solution for high-resolution remote sensing change detection on edge platforms (e.g., UAVs and embedded terminals) with a frozen backbone and LoRA fine-tuning. The tightly coupled "change localization–semantic regularization" closed-loop paradigm proves that vision–language fusion can significantly improve edge integrity, small-object detection, and interpretability without excessive computational overhead. Against the backdrop of complex features and spectral heterogeneity in high-resolution remote sensing imagery, traditional methods suffer from insufficient semantic understanding, while existing vision–language change detection models face low efficiency, poor spatial localization, and decoupled detection–description pipelines. To overcome these limitations, this paper proposes ChangeVLM, a language-guided semantic alignment framework for binary remote sensing change detection, enabling end-to-end, prompt-free, highly efficient, and interpretable change detection. Its key advantages include the following, (1) Higher detection accuracy with F1 scores of 91.52%, 83.56%, and 75.29% on LEVIR-CD, SYSU-ChangeDet, and HRCUS datasets, outperforming 18 state-of-the-art methods. (2) Stronger edge integrity and small-object detection capability; (3) practical deployment efficiency: the end-to-end FLOPs is 560.7G. Additionally, under an optimized inference setting with pre-extracted features, the effective computation can be reduced to 13.05G. (4) Language-guided semantic regularization to enhance visual discrimination, without requiring external text prompts. The Asymmetric Fusion Module (AFM), lightweight ChangeHead, and Change-Aware Cross-Modal Fusion Module (CACMF) jointly enhance spatial precision, efficiency, and interpretability. Extensive experiments validate that ChangeVLM achieves a superior accuracy–efficiency trade-off. This method provides an effective, deployable solution for high-resolution remote sensing binary change detection, where the language branch acts only as a regularization signal. [ABSTRACT FROM AUTHOR] |
| Copyright of Remote Sensing is the property of MDPI and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | Highlights: What are the main findings? ChangeVLM achieves competitive performance with only 28.73M trainable parameters and 560.7G FLOPs. It obtains F1 scores of 91.52%, 83.56%, and 75.29% on the LEVIR-CD, SYSU-ChangeDet, and HRCUS datasets, respectively. In addition, under an optimized inference setting with pre-extracted visual features, the effective computation is reduced to 13.05G FLOPs, suggesting its potential for deployment-oriented acceleration. The three core customized modules, including AFM, lightweight ChangeHead, and CACMF, address insufficient spatial localization, low inference efficiency, and weak semantic discrimination in binary remote sensing change detection. The language-related component is used only as an auxiliary semantic alignment regularizer during training and is not used for text generation. What are the implications of the main findings? The framework offers an end-to-end deployable solution for high-resolution remote sensing change detection on edge platforms (e.g., UAVs and embedded terminals) with a frozen backbone and LoRA fine-tuning. The tightly coupled "change localization–semantic regularization" closed-loop paradigm proves that vision–language fusion can significantly improve edge integrity, small-object detection, and interpretability without excessive computational overhead. Against the backdrop of complex features and spectral heterogeneity in high-resolution remote sensing imagery, traditional methods suffer from insufficient semantic understanding, while existing vision–language change detection models face low efficiency, poor spatial localization, and decoupled detection–description pipelines. To overcome these limitations, this paper proposes ChangeVLM, a language-guided semantic alignment framework for binary remote sensing change detection, enabling end-to-end, prompt-free, highly efficient, and interpretable change detection. Its key advantages include the following, (1) Higher detection accuracy with F1 scores of 91.52%, 83.56%, and 75.29% on LEVIR-CD, SYSU-ChangeDet, and HRCUS datasets, outperforming 18 state-of-the-art methods. (2) Stronger edge integrity and small-object detection capability; (3) practical deployment efficiency: the end-to-end FLOPs is 560.7G. Additionally, under an optimized inference setting with pre-extracted features, the effective computation can be reduced to 13.05G. (4) Language-guided semantic regularization to enhance visual discrimination, without requiring external text prompts. The Asymmetric Fusion Module (AFM), lightweight ChangeHead, and Change-Aware Cross-Modal Fusion Module (CACMF) jointly enhance spatial precision, efficiency, and interpretability. Extensive experiments validate that ChangeVLM achieves a superior accuracy–efficiency trade-off. This method provides an effective, deployable solution for high-resolution remote sensing binary change detection, where the language branch acts only as a regularization signal. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 20724292 |
| DOI: | 10.3390/rs18101671 |
