Hyperspectral Remote Sensing Imagery for Object Detection

Dear Colleagues,

Hyperspectral remote sensing has revolutionized Earth observation by capturing detailed spectral information across hundreds of narrow, contiguous wavelength bands. Unlike traditional multispectral imagery that records data in a few broad bands, hyperspectral sensors generate complex data cubes with rich spectral signatures, enabling the discrimination of materials based on their unique biochemical and physical properties. This technology has matured significantly over the past two decades, driven by advancements in sensor design, airborne and satellite platforms (such as PRISMA, EnMAP, and DESIS), and computational processing power.

The application of hyperspectral imagery for object detection represents a critical frontier in remote sensing science. Object detection goes beyond pixel-level classification to identify, locate, and characterize specific discrete items or structures within a scene—from vehicles and aircraft to individual tree species and archaeological features. The high-dimensional spectral data provides a powerful means to overcome challenges in complex environments where objects may be small, partially obscured, or spectrally similar to their backgrounds. This capability is crucial for a wide range of fields, such as precision agriculture (detecting crop stress or invasive species), environmental monitoring (identifying pollution sources or plastic debris), urban planning (mapping infrastructure), defense and security, and ecological conservation.

Despite its potential, hyperspectral object detection faces significant scientific challenges. The "curse of dimensionality", data redundancy, atmospheric interference, and the need for robust algorithms that can effectively integrate spatial and spectral information all present active research problems. Recent breakthroughs in artificial intelligence, particularly deep learning, convolutional neural networks (CNNs), and transformer-based architectures, have opened up new pathways for processing these complex datasets. This Special Issue seeks to capture the cutting-edge methodologies, applications, and theoretical advances that are pushing the boundaries of what is possible in detecting and characterizing objects from hyperspectral data.

This Special Issue, "Hyperspectral Remote Sensing Imagery for Object Detection," aims to collate high-quality original research and comprehensive reviews that address the latest theoretical, methodological, and applied advancements in this domain. We seek to foster discussion on novel algorithms, address the unique challenges of hyperspectral data, and showcase innovative applications that demonstrate tangible benefits over conventional techniques. This Special Issue will serve as a reference point for the community, highlighting future trends and promoting interdisciplinary collaboration to solve complex detection problems.

This topic fits nicely within the scope of Remote Sensing, an international, peer-reviewed journal dedicated to the science and application of remote sensing technology. The journal explicitly covers "remote sensing data processing, analysis and interpretation" and "advanced sensor design, calibration and data processing techniques." Hyperspectral object detection is a quintessential example of advanced data processing and interpretation. This Special Issue intersects with multiple journal sections, including image processing, remote sensing applications, and aerial and satellite image processing. By focusing on a high-impact, rapidly evolving niche, this Special Issue will attract significant scholarly attention, enhance the journal's visibility in the AI-for-remote-sensing community, and contribute valuable knowledge on handling high-dimensional geospatial data.

We invite submissions that explore, but are not limited to, the following themes:

• Novel learning architectures, such as the development and application of graph neural networks (GNNs), vision transformers, the State Space Model, and the Foundation Model, specifically designed for hyperspectral object detection.
• Spectral–Spatial Feature Fusion: Advanced methods for effectively and efficiently integrating spatial context with rich spectral information to improve detection accuracy and robustness.
• Weakly/Self-Supervised and Few-Shot Learning: Innovative approaches to overcome the challenge of limited labeled training data, which is often a constraint in hyperspectral studies.
• Real-Time and Embedded Processing: Algorithm optimization, compression, and hardware acceleration techniques for onboard or near-real-time object detection from airborne and satellite platforms.
• Multimodal and Multi-Temporal Data Fusion: The fusion of hyperspectral data with LiDAR, SAR, or high-resolution RGB imagery, or across time series, to enhance detection capabilities.
• Explainable AI (XAI) for Detection: Making complex detection models interpretable, providing insights into which spectral-spatial features drive decisions.
• Benchmark Datasets and Performance Evaluation: The creation of new public datasets and comprehensive evaluation frameworks, including robustness to atmospheric conditions, illumination changes, and sensor noise.
• Domain-Specific Applications: Cutting-edge applications in environmental monitoring, precision agriculture, urban surveillance, mineral exploration, search and rescue, and defense. 

Dr. Tan Guo
Prof. Dr. Fulin Luo
Prof. Dr. Lei Zhang
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.