From Pixels to Spectra: Towards Generalizable Large Models for Hyperspectral Remote Sensing

Dear Colleagues,

In recent years, high-resolution data from space satellites and unmanned aerial vehicles have increased sharply, and hyperspectral remote sensing has gradually become a key means for achieving high-precision Earth observation. However, due to the high dimension, spectral redundancy, and limited labeled samples of hyperspectral data, hyperspectral interpretation tasks still face challenges such as poor robustness and weak scalability. Although deep learning has made significant progress in feature extraction and spectral space modeling, most existing methods are targeted at specific tasks and supervised data. Therefore, the emergence of large-scale visual language models and inference models has provided new opportunities for constructing generalizable and reusable representations in various hyperspectral tasks.

This Special Issue will promote the development of generalized and transferable multimodal large models for hyperspectral remote sensing data. We welcome original research on hyperspectral interpretation based on large-scale pre-training, multi-task learning, spectral space transformation, and general models adapted to hyperspectral data. The topics include but are not limited to hyperspectral feature learning, spectral decomposition, general models, semantic segmentation, and change detection. We also encourage the submission of practical applications in agriculture, urban monitoring, mineral mapping, and environmental assessment.

Topics of interest include, but are not limited to, the following:

• Spectral–spatial feature extraction for hyperspectral data;
• Large-scale and transformer-based architectures for HSI understanding;
• Self-supervised, unsupervised, and few-shot learning for label-efficient representation;
• Spectral–spatial pretraining and cross-task model generalization;
• Dimensionality reduction, feature selection, and spectral manifold learning;
• Lightweight or efficient models for edge deployment or real-time inference;
• Applications in spectral unmixing, target detection, segmentation, and change detection;
• Open benchmarks, scalable datasets, and evaluation protocols for large model validation.

We also encourage research on hyperspectral interpretation training datasets and inference evaluation datasets to enhance the adaptability and robustness of hyperspectral image analysis in real-world scenarios.

Prof. Dr. Yaxiong Chen
Dr. Qingyu Li
Dr. Wuxia Zhang
Dr. Zhitong Xiong
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 250 words) can be sent to the Editorial Office for assessment.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Remote Sensing is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2700 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.