This paper presents Prithvi-EO-2.0, a new geospatial foundation model that offers significant improvements over its predecessor, Prithvi-EO-1.0. Trained on 4.2 million global time series samples from NASA’s Harmonized Landsat and Sentinel-2 data archive at 30-m resolution, the new model incorporates temporal and location embeddings for enhanced performance across various geospatial tasks. Through extensive benchmarking with GEO-Bench, the model outperforms the previous Prithvi-EO model by 8% across a range of tasks. It also outperforms six other geospatial foundation models when benchmarked on remote sensing tasks from different domains and resolutions (i.e. from 0.1 m to 15 m). The results demonstrate the versatility of the model in both classical Earth observation and high-resolution applications. Early involvement of end-users and subject matter experts (SMEs) allowed constant feedback on model and dataset design, enabling customization across diverse SME-led applications in disaster response, land cover and crop mapping, and ecosystem dynamics monitoring. Prithvi-EO-2.0 is available as an open-source model on Hugging Face and IBM TerraTorch, with additional resources on GitHub. The project exemplifies the Trusted Open Science approach embraced by all involved organizations.

Supervised classification plays a key role in terms of accurate analysis of hyperspectral images. Many applications can greatly benefit from the wealth of spectral and spatial information provided by these kind of data, including land-use and land-cover mapping. Conventional classifiers treat hyperspectral images as a list of spectral measurements and do not consider spatial dependencies of the adjacent pixels. To overcome these limitations, classifiers need to use both spectral and spatial information. In this paper, a framework for automatic spectral-spatial classification of hyperspectral images is proposed. In order to extract the spatial information, Extended Multi-Attribute Profiles (EMAPs) are taken into account. In addition, in order to reduce the redundancy of features and address the so-called curse of dimensionality, different supervised feature extraction (FE) techniques are considered. The final classification map is provided by using a random forest classifier. The proposed automatic framework is tested on two widely used hyperspectral data sets; Pavia University and Indian Pines. Experimental results confirm that the proposed framework automatically provides accurate classification maps in acceptable CPU processing times.

Owing to the recent development of sensor resolutions onboard different Earth observation platforms, remote sensing is an important source of information for mapping and monitoring natural and man-made land covers. Of particular importance is the increasing amounts of available hyperspectral data originating from airborne and satellite sensors such as AVIRIS, HyMap, and Hyperion with very high spectral resolution (i.e., high number of spectral channels) containing rich information for a wide range of applications. A relevant example is the separation of different types of land-cover classes using the data in order to understand, e.g., impacts of natural disasters or changing of city buildings over time. More recently, such increases in the data volume, velocity, and variety of data contributed to the term big data that stand for challenges shared with many other scientific disciplines. On one hand, the amount of available data is increasing in a way that raises the demand for automatic data analysis elements since many of the available data collections are massively underutilized lacking experts for manual investigation. On the other hand, proven statistical methods (e.g., dimensionality reduction) driven by manual approaches have a significant impact in reducing the amount of big data toward smaller smart data contributing to the more recently used terms data value and veracity (i.e., less noise, lower dimensions that capture the most important information). This paper aims to take stock of which proven statistical data mining methods in remote sensing are used to contribute to smart data analysis processes in the light of possible automation as well as scalable and parallel processing techniques. We focus on parallel support vector machines (SVMs) as one of the best out-of-the-box classification methods.

In the last decade, deep learning (DL) has significantly impacted industry and science. Initially largely motivated by computer vision tasks in 2-D imagery, the focus has shifted toward 3-D data analysis. In particular, 3-D surface reconstruction, i.e., reconstructing a 3-D shape from sparse input, is of great interest to a large variety of application fields. DL-based approaches show promising quantitative and qualitative surface reconstruction performance compared to traditional computer vision and geometric algorithms. This survey provides a comprehensive overview of these DL-based methods for 3-D surface reconstruction. To this end, we will first discuss input data modalities, such as volumetric data, point clouds, and RGB, single-view, multiview, and depth images, along with corresponding acquisition technologies and common benchmark datasets. For practical purposes, we also discuss evaluation metrics enabling us to judge the reconstructive performance of different methods. The main part of the document will introduce a methodological taxonomy ranging from point-and mesh-based techniques to volumetric and implicit neural approaches. Recent research trends, both methodological and for applications, are highlighted, pointing toward future developments.

In recent years, the development of quantum annealers has enabled experimental demonstrations and has increased research interest in applications of quantum annealing, such as in quantum machine learning and in particular for the popular quantum support vector machine (SVM). Several versions of the quantum SVM have been proposed, and quantum annealing has been shown to be effective in them. Extensions to multiclass problems have also been made, which consist of an ensemble of multiple binary classifiers. This article proposes a novel quantum SVM formulation for direct multiclass classification based on quantum annealing, called quantum multiclass SVM (QMSVM). The multiclass classification problem is formulated as a single quadratic unconstrained binary optimization problem solved with quantum annealing. The main objective of this article is to evaluate the feasibility, accuracy, and time performance of this approach. Experiments have been performed on the D-Wave Advantage quantum annealer for a classification problem on remote sensing data. Results indicate that, despite the memory demands of the quantum annealer, QMSVM can achieve an accuracy that is comparable to standard SVM methods, such as the one-versus-one (OVO), depending on the dataset (compared to OVO: 0.8663 versus 0.8598 on Toulouse, 0.8123 versus 0.8521 on Potsdam). More importantly, it scales much more efficiently with the number of training examples, resulting in nearly constant time (compared to OVO: 85.72 versus 248.02 s on Toulouse, 58.89 versus 580.17 s on Potsdam). This article shows an approach for bringing together classical and quantum computation, solving practical problems in remote sensing with current hardware.

High-Performance Computing (HPC) has recently been attracting more attention in remote sensing applications due to the challenges posed by the increased amount of open data that are produced daily by Earth Observation (EO) programs. The unique parallel computing environments and programming techniques that are integrated in HPC systems are able to solve large-scale problems such as the training of classification algorithms with large amounts of Remote Sensing (RS) data. This paper shows that the training of state-of-the-art deep Convolutional Neural Networks (CNNs) can be efficiently performed in distributed fashion using parallel implementation techniques on HPC machines containing a large number of Graphics Processing Units (GPUs). The experimental results confirm that distributed training can drastically reduce the amount of time needed to perform full training, resulting in near linear scaling without loss of test accuracy.