Othello is a two-player combinatorial game with 1E+28 legal positions and 1E+58 game tree complexity. We propose a HIghly PArallel, Scalable and configurable hardware accelerator for evaluating the middle and endgame Othello positions. We base HIPAS on NTest - a leading software Othello engine that uses the minimax algorithm with a quality pattern-based evaluation function, alpha-beta pruning, and heuristic mobility sorting. We describe its architecture and Field Programmable Gate Array implementation, measure its performance, and compare it with prior solutions. HIPAS achieves the highest quality evaluation, the highest performance with speed-ups up to several hundreds, and the best energy efficiency. The main novelty is the algorithm implementation as a circular pipeline and a Finite State Machine with pseudo-parallel processing. Although Othello was recently claimed to be weakly solved, the game remains unsolved in a stronger sense. A weak solution only shows how to force a draw. It does not guarantee a win if the opponent makes a mistake. HIPAS can validate the weak solution faster and more efficiently. A multi-threaded NTest software component evaluating the beginning and part of the middle game, combined with one or more instances of HIPAS for handling the remainder can provide a stronger solution.

NDVI is the most used vegetation index for the monitoring of vegetation status of agricultural crops. In situ measurements are generally used for the validation of remote sensing data. However, due to various factors, the remotely sensed data may not completely agree with in situ acquired data. We computed NDVI based on Sentinel-2 data and performed in situ NDVI measurements on a specific agricultural area. We used Monte Carlo sampling for the estimation of the distribution of coincidence values, assuming either statistical independence and correlation of data. Based on the estimated distribution, we computed the probability that NDVI falls in two intervals of interest, based on a threshold widely used for the interpretation of the NDVI values.

We propose a data-driven method using a convolutional neural network (CNN) to classify plant growth stages, using Sentinel-1 SAR backscattering values and computer-generated schematics representing plant development. Opposed to the physics-based approaches of BS simulation such as integral equation modelling, this approach is data-driven and has the potential to be more robust. A total of five field measurement campaigns were run over the five months and we collected the soil roughness and wheat canopy parameters. Data was used to randomly generate the synthetic images which were further used to train and test a CNN. Computer-generated plant schematics are digital representations of the soil and plant layer which are the two main components that affect the BS coefficients. Results show that the model successfully classified five stages of the wheat canopy growth with the highest test accuracy of 96.4%.

With the availability of several remotely sensed data sources, the problem of efficiently visualizing the information from multisource data for improved Earth observation becomes an intriguing and challenging subject. Multispectral and hyperspectral images encompass a wealth of spectral data that standard RGB monitors cannot replicate directly. Thus, it is important to elaborate methods for accurately representing this information on conventional displays. These images, with tens to hundreds of spectral bands, contain relevant data about specific wavelengths that RGB channels cannot capture. Traditional visualization methods often use only a limited amount of the available spectral information, resulting in a significant loss of information. However, recent advances in artificial intelligence models have provided superior visualization techniques. These AI-based methods allow for more realistic and visually appealing representations, which are important for the information interpretation and direct identification of areas of interest. The main goal of our study is to process aggregated datasets from various sources using a fully connected neural network (FCNN), while considering visualization as a secondary objective. Given that our data come from a variety of sources, a significant emphasis in our study was placed on the preprocessing stage. In order to achieve a consistent visualization across datasets from different sources, proper preprocessing by standardization or normalization procedures is essential. Our research comprises numerous experiments to demonstrate the effectiveness of the proposed technique for image visualization.

Hyperspectral images are captured over a wide range of the electromagnetic spectrum providing detailed information about the Earth’s surface. Hyperspectral imaging is widely used in agriculture, astronomy, molecular biology, physics, etc. Due to the very large size of information that the remotely-sensed hyperspectral data cube contains, transmission is a challenge. Various hyperspectral image compression techniques have been proposed in the last decades. We propose a new lossy compression technique that is based on the Fast Fourier Transform, negative exponential feature extraction, and machine learning. For the Pavia University data set, we obtained a compression rate of approximately 11, while preserving an important amount of the information in the original scene. Furthermore, we visualized the decompressed data starting from only two retained parameters and we evaluated the results by employing various metrics.

The advances in the field of remote sensing for Earth Observation allow many applications like precision agriculture, forest monitoring, to name a few. Hyperspectral imaging is the technique that offers a high spectral resolution offering such applications more information about the Earth surface, but the data volume to be stored and processed increases too. A reason for the increased data volume is the high intrinsic variability of spectral reflectance curves. In this paper we propose a feature extraction method for dimensionality reduction based on fitting a negative exponential function to the Fourier spectrum of each spectral reflectance curve. The hypothesis is that extracting an exponential profile of the Fourier amplitude spectrum, thus reducing the variability of the spectral signatures, will possibly impact the segmentation approach. We further implement a segmentation algorithm based on the extracted features and assess its performance.

The InnEO’Space PhD program addresses the growing demand for skilled individuals in Earth Observation (EO) data management. It offers modern and transferable courses to enhance researchers' innovation-oriented skills. The program includes InnEO Startech, focusing on entrepreneurship and leadership, and the Machine Learning for Earth Observation InnEO Summer school, emphasizing open science and technical skills. Small Private Online Courses (SPOCs) provide blended learning through a user-friendly platform. The InnEO PRO section facilitates interaction among PhD students and professionals, offering various resources and self-assessment tests. The program's success has led to collaborations with UNIVERSEH and FabSpace, while AI4AGRI aims to utilize its resources for AI in EO research.

We mathematically prove that color fractal images with two and three correlated color components generated with the midpoint displacement approach obey the property of self-similarity, thus enabling the estimation of their color fractal dimension. We generate various sets of color fractal images with two and three correlated color components, controlled both by the Hurst parameter and the variance-covariance matrix, with and without a global normalization, and use them for the calibration of the embraced color fractal dimension estimator. We improve the existing fractal dimension estimator based on probabilistic box-counting by reducing the variance of the regression line estimators through the iterative elimination of most error-ed measurement points. We independently estimate the variance-covariance matrix and Hurst parameter for the sets of generated color fractal images with correlated color components. We show the experimental results and discuss both the improvements and the limitations of the proposed approach.

Mathematical morphology is a popular framework for non‐linear image processing, first introduced for binary and grey‐level images, then extended to colour and multivariate images. Various pseudo‐morphologies have been proposed as solutions to the problem of ordering multivariate data. Despite the lack of some properties, pseudo‐morphologies have proved useful in various applications, such as filtering or texture classification. The authors propose to improve the existing colour probabilistic pseudo‐morphology by changing the way the local pseudo‐extrema are chosen. They show the usefulness of the new construction in the context of noise reduction in colour images using the open‐close close‐open filter, by highlighting the improvement over the original construction and comparing the authors’ results with other relevant morphological and pseudo‐morphological approaches.

We propose a pseudo-morphology for color and multispectral images based on the maximum distance computed between colors in a given set. Distances are computed using the ΔE metric in the CIELAB color space or the Euclidean distance in Rnin the multispectral case. Using the proposed pseudo-morphology, we compute the pseudo-granulometry and morphological covariance for color and multispectral images, comparing the results with the grayscale case, in order to evaluate the usefulness of these texture characterization tools.

Fractal dimension is a very useful metric for the analysis of the images with self-similar content, such as textures. For its computation there exist several approaches, the probabilistic algorithm being accepted as the most elegant approach. However, all the existing methods are defined for 1-D signals or binary images, with extension to grayscale images. Our purpose is to propose a color version of the probabilistic algorithm for the computation of the fractal dimension. To validate this new approach, we also propose an extension of the existing algorithm for the generation of probabilistic fractals, in order to obtain color fractal images. Then we show the results of our experiments and conclude this paper.

Lacunarity is a very useful metric for the multi-scale analysis of the images that exhibit fractal properties. For its computation there exist several approaches, the probabilistic algorithm being accepted as the most elegant approach. However, all the existing methods are defined for one dimensional signals or binary images with extension to grey-scale images. We propose a colour expression of the lacunarity based on the probabilistic algorithm for the computation of the fractal dimension. To validate this new approach, we used both synthetic colour fractal images and natural fractal images. We comment our results and spot several issues regarding the interpretation of lacunarity curves, then we conclude.