This paper introduces a novel approach integrating Differential Evolution (DE) with multi-objective optimization techniques for enhancing Type-1 Takagi-Sugeno-Kang (TSK) fuzzy rule-based systems to attain both fair predictive performance and enhanced interpretability within the context of explainable artificial intelligence. The developed approach encodes the whole knowledge base into one chromosome and endeavors to concurrently optimize the prediction accuracy, the number of rules, and the selection of relevant input features. The methodology was tested on Earth observation data for the estimation of soil health parameters via infrared spectroscopy, and more concretely i) on the 2015 LUCAS topsoil dataset to predict three key soil properties (namely, soil organic carbon, clay content, and pH) from laboratory spectra, and ii) on a set of PRISMA hyperspectral images to predict soil organic carbon in a region of northern Greece. Compared to the classical Random Forest (RF) algorithm, our proposed learning algorithm attains a fair balance between accuracy and interpretability, and is statistically equivalent to RF in terms of accuracy. The findings underscore the potential of the proposed methodology in refining TSK models and its applicability in Earth observation-driven predictions, paving the way for both enhanced modeling accuracy and sparse feature selection.

Soil organic carbon (SOC) content is a key indicator of soil health informing about sustainable land management practices, but parcel-wide SOC mapping is challenging as it requires high-resolution data. Unoccupied Aerial Vehicles (UAVs) can collect data with cm-resolution but are not yet fully ready to be practically implemented. The aim of this study is to provide more insights in the explanatory capabilities of UAV-derived spectral and topographical variables. To this end, mixed models were employed to estimate the SOC content of three agricultural parcels with different crop types in Greece. Results showed variations in SOC content among parcels, with a vineyard and a kiwi orchard having higher values compared to a peach orchard. All models, containing topographical and/or spectral variables, explained 81% of SOC content variation of the training dataset. Besides crop type, other topographical and spectral variables were identified as significant predictors. The study emphasizes the feasibility of UAV data and specific modeling techniques for accurate SOC estimation at the parcel level, providing valuable insights for precision agriculture. The findings recommend further exploration, including machine-learning approaches in future studies.

Grasslands store a large fraction of terrestrial carbon, but are susceptible to degradation from anthropogenic disturbances and climatic changes. Soil monitoring can aid in conserving their ecosystem services. To overcome limitations posed by existing soil maps (e.g., low spatial resolution), we leverage the Soil Data Cube and Digital Soil Mapping techniques, to develop a cloud-optimized pipeline for large-scale soil monitoring using open access Copernicus data. In particular, we employ data from the LUCAS topsoil database, ERA5 climate data from the Copernicus Climate Data Store, and the EU-DEM from the Copernicus Land Monitoring Service. Using Recursive Feature Elimination and the Random Forest algorithm, the methodology achieves an RMSE of 49.1 g C / kg and an R2of 0.66 for topsoil Organic Carbon, and an RMSE of 52.1 g / kg with an R2of 0.66 for topsoil Clay content. Our method enhances spatio-temporal representativeness and reliability, aligning with the European Union’s policies like the Common Agricultural Policy, the new green deal, and ecoschemes. The outcomes of this study are the production of high-resolution soil maps tailored to Lithuanian grasslands. These advancements in soil health monitoring empower more effective and sustainable soil management practices.

Analysis-ready data generated through datacube approaches are essential in Earth Observation (EO) data analysis. By pre-processing and organizing large volumes of satellite imagery into spatio-temporal datacubes, analysis-ready data facilitates efficient and streamlined analysis workflows, enabling quick access and manipulation of the multi-dimensional EO data. This allows researchers, scientists, and stakeholders to focus on analyzing the data, detecting trends, monitoring changes, and deriving actionable information. In this paper, we describe the architecture of the Greek Soil Data Cube which ingests Sentinel-2 data from the entire country and generates the annual bare soil reflectance composites which in turn may be used to predict soil physicochemical properties using artificial intelligence techniques.

This paper presents a novel methodology for estimating cropland topsoil organic carbon (SOC) content using open-access Earth Observation (EO) data and deep learning techniques in Greece. We address the scarcity of ground-truth reference data using a data augmentation technique utilizing spatial neighbors. By incorporating neighboring pixel information in a 3x3 grid (corresponding to 30 x 30 m meters in Sentinel-2), we achieve robust predictions and provide uncertainty estimations for SOC distribution in Greek croplands. Experimental results demonstrate the superiority of deep learning techniques over conventional methods in terms of estimation accuracy (RMSE=3.51, R2=0.59, RPIQ=2.21). This methodology offers a cost-effective and efficient solution for directing land management practices to enhance soil health and mitigate carbon emissions.

Novel computational algorithms along with cloud computing services present a great potential to revolutionize the processing of Earth Observation (EO) data for topsoil mapping. This work presents the first insights of the WORLDSOILS project, where open-access Copernicus Sentinel-2 data and auxiliary terrain attributes were synergistically utilized to derive regression models. Building on key results from well-studied data mining approaches, the current study compares various approaches of temporal mosaicking to generate a multi-year median composite dataset of exposed bare soil pixels, over cropland areas in European Union. Finally, we utilized a Random Forest model based on soil samples (calibration = 80% and validation 20%) of the recently released LUCAS 2015 database to predict soil clay content and organic carbon. Following a masking approach to generate a composite of multi-date bare soil pixels, a promising prediction performance (R2 = 0.52, n = 7605) was achieved for clay content, while the predictive performance for soil organic carbon was significantly lower (R2 = 0.30).

This work focuses on the application of an evolutionary fuzzy rule-based system following the MapReduce principle to real-world big data regression problems. The examined datasets originate from the field of soil spectroscopy and are high in volume, variety, and dimensionality. They are part of the GEO-CRADLE and LUCAS soil spectral libraries, which encompass large geographical areas with a significant amount of patterns. In this work, a two-tier MapReduce scheme is proposed. The first-tier defines the Data Base by selecting optimal granularities per each feature. The second-tier applies the DECO3RUM algorithm in a distributed way using bootstrapped samples to tackle the large inherent sparsity of the data. The resulting rule bases are aggregated using a final instance of DECO3RUM. The performance of the proposed methodology was analyzed and statistically compared to the serial version as well as a fully distributed version. It was demonstrated that the version using bootstrapped samples outperforms the other versions in terms of accuracy, while being as compact as the serial version; thus, providing a good balance between accuracy and execution speed.

In this paper, we present DECO3RUM (Differential Evolution based Cooperative and Competing learning of Compact Rule-based Models), an evolutionary Mamdani Fuzzy Rule-based System for modeling problems. DECO3RUM follows the Genetic Cooperative Competitive Learning approach, and utilizes the Differential Evolution algorithm as its learning algorithm. A real world high dimensional dataset from the domain of soil science was considered to evaluate the ability of DECO3RUM to handle Big Data problems, where the number of features is significant. DECO3RUM was shown to statistically out-perform the most prevailing methodology used in soil spectroscopy, namely the Partial Least Squares Regression algorithm.

In this paper we propose two extensions of our previously introduced method DECO3R, a Fuzzy Rule-based Classification System (FRBCS). DECO3R stands for Differential Evolution based Cooperative and Competing learning of Compact FRBCS. It follows the Genetic Cooperative - Competitive Learning (GCCL) approach, and utilizes the Differential Evolution (DE) as its learning algorithm. Two novel schemes will be introduced, one which creates a number of parallel populations employing different DE strategies, and one inserting newly constructed features (relations and functions) in the antecedent part of the rules. These two new schemes, as well as the combination thereof, boost the performance of DECO3R, as evidenced by the results obtained, which were validated using non-parametric statistical tests.

In this paper we propose a novel Fuzzy Rulebased Classification System, named DECO3R, which follows the genetic cooperative-competitive learning (GCCL) approach and uses the Differential Evolution algorithm as its genetic algorithm. DECO3R uses a novel Fuzzy Token Competition method implemented by AdaBoost which forces the rules to compete and cooperate with each other. It is capable of both learning clear and concise DNF rules, where the fuzzy sets are consecutive, and obtaining compact sets of rules. The obtained results have been validated using non-parametric statistical tests that demonstrate DECO3R's robust performance, both in terms of accuracy and of interpretability.

A novel Self-Organizing Neuro-Fuzzy Multilayered Classifier, the GA-SONeFMUC model, is proposed in this paper for land cover classification of multispectral images. The model is composed of generic fuzzy neuron classifiers (FNCs) arranged in layers, which are implemented by fuzzy rule-based systems. At each layer, parent FNCs are combined to generate a descendant FNC at the next layer with higher classification accuracy. To exploit the information acquired by the parent FNCs, their decision supports are combined using a fusion operator. As a result, a data splitting is devised within each FNC, distinguishing those pixels that are currently correctly classified to a high certainty grade from the ambiguous ones. The former are handled by the fuser, while the ambiguous pixels are further processed to enhance their classification confidence. The GA-SONeFMUC structure is determined in a self-constructing way via a structure-learning algorithm with feature selection capabilities. The parameters of the models obtained after structure learning are optimized using a real-coded genetic algorithm. For effective classification, we formulated three input sets containing spectral and textural feature types. To explore information coming from different feature sources, we apply a classifier fusion approach at the final stage. The outputs of individual classifiers constructed from each input set are combined to provide the final assignments. Our approach is tested on a lake-wetland ecosystem of international importance using an IKONOS image. A high-classification performance of 92.02% and of 75.55% for the wetland zone and the surrounding agricultural zone is achieved, respectively.