Sentinel-2 satellites provide systematic global coverage of land surfaces, measuring physical properties within 13 spectral intervals at a temporal resolution of 5 days. Computer-based data analysis is highly required to extract similarity by processing and to assist human understanding and semantic annotation in support of mapping Earth's surface. This paper proposes a data mining concept that uses advanced data visualization and explainable features to enhance relevant aspects in the Sentinel-2 data and enable semantic analysis. There is a two-stage process. At first, spectral, texture, and physical parameters related features are extracted from the data and included in a learning process that models the data content according to statistical similarities. In parallel, the second processing stage maximizes the data impact on the human visual system to help image understanding and interpretation. Target classes are subject to exploratory visual analysis, such that both visual and latent characteristics are revealed to the user. The concept is further implemented as Sentinel-2 dedicated data analysis (DAS-Tool) plugin for the Sentinel Application Platform (SNAP) and deployed as an open-source tool empowering the Earth Observation (EO) community with fast and reliable results. Accommodating multiple solutions for each processing phase, the plugin enables flexibility in information extraction and knowledge discovery that will bring the best accuracy in mapping applications. For demonstration purposes, the authors focus on a detailed benchmark against reference data (ground truth) for the Southern region of Romania, then use the selected algorithms in a forest fires scenario analysis for the Sydney region in Australia.

In this paper, we propose an Active Learning approach to query by example retrieval, using a retraining procedure that improves the understanding of the machine with respect to the human perception. The proposed method is based on Support Vector Machine (SVM) classifiers and requires a small number of training samples. The classifier is retrained several times in order to determine the optimal separating hyper-plane between the class of the query and the rest of the analysed image. The closest feature points to the SVM-learned hyper-plane are the points being able to produce the most relevant modification of the position of this hyper-plane. These points, that are both negative and positive examples, are then used to retrain the SVM classifier. In addition, the proposed approach shows the importance of normalization in a classification problem with heterogeneous objects. Several experiments were conducted on GeoEye-1 multispectral images, whilst the retrieval was performed for different patch-level descriptors, which furthermore increases the complexity of the semantic content of the query object.

The purpose of this paper is to present a method that uses Independent Component Analysis (ICA) in order to filter real-valued signals. The method is highly robust to noise, as it can estimate with high precision the noise that affects the input signal. To underline this, comparisons to high order, multiband FIR filters are performed. An application that uses this method, regarding radar image shape extraction and classification, is also presented.

Some of the first Earth Observation (EO) missions date back to the 1970s. Over the time, large datasets of Satellite Image Time Series (SITS) have been used to identify and monitor land cover evolutions. The processing complexity increases proportionally to the time span of the Earth Observation (EO) series. Because of the SITS dataset complexity and variety of contained evolution patterns, most unsupervised classification methods fail to detect and isolate the user's evolution pattern of interest. This is usually caused by the discrepancy between automatically extracted low-level features and high-level meaning assigned by the user who searches for a specific evolution. In an effort to find a solution for this difficult task, this paper presents a SVM based active learning method for the extraction of specific evolution classes from SITS. Several experiments were conducted on a dataset composed of Landsat 4 TM (Thematic Mapper) and Landsat 5 TM acquisitions over Bucharest, Romania, in the time interval of 1984-1993.

The increasing number of satellite missions for Earth exploration provides huge amounts of data. These data are of wide diversity regarding the images characteristics, thus new techniques and tools need to be developed to accommodate the extraction of meaningful information. This paper presents An Earth Observation Spatio-Temporal Data Mining System (HyperMINE), which integrates fast and complex query methods in order to generate SITS (Satellite Image Time Series) regarded as a data hypercube. The multidimensional data, considering geographical space, time, band, and satellite/sensor are used as input for information mining algorithms. The system is built on a modular multilayer architecture that allows effective processing of various sources of data.