Multi-Granularity Domain-Adaptive Teacher for Unsupervised Remote Sensing Object Detection

Fang, Fang; Kang, Jianing; Li, Shengwen; Tian, Panpan; Liu, Yang; Luo, Chaoliang; Zhou, Shunping

doi:10.3390/rs17101743

This is an early access version, the complete PDF, HTML, and XML versions will be available soon.

Open AccessArticle

Multi-Granularity Domain-Adaptive Teacher for Unsupervised Remote Sensing Object Detection

by

Fang Fang

^1,2

,

Jianing Kang

¹,

Shengwen Li

^1,2,*

,

Panpan Tian

¹,

Yang Liu

¹,

Chaoliang Luo

³ and

Shunping Zhou

¹

School of Computer Science, China University of Geosciences, Wuhan 430074, China

²

Engineering Research Center of Natural Resource Information Management Digital Twin Engineering Software, Ministry of Education, Wuhan 430074, China

³

Wuhan Huitong Zhiyun Information Technology, Wuhan 430200, China

^*

Author to whom correspondence should be addressed.

Remote Sens. 2025, 17(10), 1743; https://doi.org/10.3390/rs17101743

Submission received: 24 March 2025 / Revised: 13 May 2025 / Accepted: 15 May 2025 / Published: 16 May 2025

(This article belongs to the Special Issue Advancement of Multi-Source Remote Sensing Data Fusion in Environmental Monitoring)

Download Versions Notes

Abstract

Object detection in remote sensing images (RSIs) is pivotal for various tasks such as natural disaster warning, environmental monitoring, teacher–student urban planning. Object detection methods based on domain adaptation have emerged, which effectively decrease the dependence on annotated samples, making significant advances in unsupervised scenarios. However, these methods fall short in their ability to learn remote sensing object features of the target domain, thus limiting the detection capabilities in many complex scenarios. To fill this gap, this paper integrates a multi-granularity feature alignment strategy and the teacher–student framework to enhance the capability of detecting remote sensing objects, and proposes a multi-granularity domain-adaptive teacher (MGDAT) framework to better bridge the feature gap across target and source domain data. MGDAT incorporates the teacher–student framework at three granularities, including pixel-, image- and instance-level feature alignment. Extensive experiments show that MGDAT surpasses SOTA baselines in detection accuracy, and exhibits great generalizability. This proposed method can serve as a methodology reference for various unsupervised interpretation tasks of RSIs.

Keywords: unsupervised domain adaptation (UDA); remote sensing image (RSI); object detection; multi-granularity features

Share and Cite

MDPI and ACS Style

Fang, F.; Kang, J.; Li, S.; Tian, P.; Liu, Y.; Luo, C.; Zhou, S. Multi-Granularity Domain-Adaptive Teacher for Unsupervised Remote Sensing Object Detection. Remote Sens. 2025, 17, 1743. https://doi.org/10.3390/rs17101743

AMA Style

Fang F, Kang J, Li S, Tian P, Liu Y, Luo C, Zhou S. Multi-Granularity Domain-Adaptive Teacher for Unsupervised Remote Sensing Object Detection. Remote Sensing. 2025; 17(10):1743. https://doi.org/10.3390/rs17101743

Chicago/Turabian Style

Fang, Fang, Jianing Kang, Shengwen Li, Panpan Tian, Yang Liu, Chaoliang Luo, and Shunping Zhou. 2025. "Multi-Granularity Domain-Adaptive Teacher for Unsupervised Remote Sensing Object Detection" Remote Sensing 17, no. 10: 1743. https://doi.org/10.3390/rs17101743

APA Style

Fang, F., Kang, J., Li, S., Tian, P., Liu, Y., Luo, C., & Zhou, S. (2025). Multi-Granularity Domain-Adaptive Teacher for Unsupervised Remote Sensing Object Detection. Remote Sensing, 17(10), 1743. https://doi.org/10.3390/rs17101743

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu

Multi-Granularity Domain-Adaptive Teacher for Unsupervised Remote Sensing Object Detection

Abstract

Share and Cite

Article Metrics

Article Access Statistics

Further Information

Guidelines

MDPI Initiatives

Follow MDPI