Image Classification Techniques Research Articles

Seagrass restoration monitoring and shallow-water benthic habitat mapping through a photogrammetry-based protocol

Seagrasses rank among the most productive yet highly threatened ecosystems on Earth. Loss of seagrass habitat because of anthropogenic disturbances and evidence of their limited resilience have provided the impetus for investigating and monitoring habitat restoration through transplantation programmes. Although Structure from Motion (SfM) photogrammetry is becoming a more and more relevant technique for mapping underwater environments, no standardised methods currently exist to provide 3-dimensional high spatial resolution and accuracy cartographic products for monitoring seagrass transplantation areas. By synthesizing various remote sensing applications, we provide an underwater SfM-based protocol for monitoring large seagrass restoration areas. The data obtained from consumer-grade red-green-blue (RGB) imagery allowed the fine characterization of the seabed by using 3D dense point clouds and raster layers, including orthophoto mosaics and Digital Surface Models (DSM).The integration of high spatial resolution underwater imagery with object-based image classification (OBIA) technique provided a new tool to count transplanted Posidonia oceanica fragments and estimate the bottom coverage expressed as a percentage of seabed covered by such fragments. Finally, the resulting digital maps were integrated into Geographic Information Systems (GIS) to run topographic change detection analysis and evaluate the mean height of transplanted fragments and detect fine-scale changes in seabed vector ruggedness measure (VRM). Our study provides a guide for creating large-scale, replicable and ready-to-use products for a broad range of applications aimed at standardizing monitoring protocols in future seagrass restoration actions.

Alzheimer's Disease Classification Accuracy is Improved by MRI Harmonization based on Attention-Guided Generative Adversarial Networks.

Alzheimer's disease (AD) accounts for 60% of dementia cases worldwide; patients with the disease typically suffer from irreversible memory loss and progressive decline in multiple cognitive domains. With brain imaging techniques such as magnetic resonance imaging (MRI), microscopic brain changes are detectable even before abnormal memory loss is detected clinically. Patterns of brain atrophy can be measured using MRI, which gives us an opportunity to facilitate AD detection using image classification techniques. Even so, MRI scanning protocols and scanners differ across studies. The resulting differences in image contrast and signal to noise make it important to train and test classification models on multiple datasets, and to handle shifts in image characteristics across protocols (also known as domain transfer or domain adaptation). Here, we examined whether adversarial domain adaptation can boost the performance of a Convolutional Neural Network (CNN) model designed to classify AD. To test this, we used an Attention-Guided Generative Adversarial Network (GAN) to harmonize images from three publicly available brain MRI datasets - ADNI, AIBL and OASIS - adjusting for scanner-dependent effects. Our AG-GAN optimized a joint objective function that included attention loss, pixel loss, cycle-consistency loss and adversarial loss; the model was trained bidirectionally in an end-to-end fashion. For AD classification, we adapted the popular 2D AlexNet CNN to handle 3D images. Classification based on harmonized MR images significantly outperformed classification based on the three datasets in non-harmonized form, motivating further work on image harmonization using adversarial techniques.

Analysis of Vegetative Cover Vulnerability in Rohingya Refugee Camps of Bangladesh Utilizing Landsat and Per Capita Greening Area (PCGA) Datasets

The vegetative cover in and surrounding the Rohingya refugee camps in Ukhiya-Teknaf is highly vulnerable since millions of refugees moved into the area, which led to severe environmental degradation. In this research, we used a supervised image classification technique to quantify the vegetative cover changes both in Ukhiya-Teknaf and thirty-four refugee camps in three time-steps: one pre-refugee crisis (January 2017), and two post-refugee crisis (March 2018, and February 2019), in order to identify the factors behind the decline in vegetative cover. The vegetative cover vulnerability of the thirty-four refugee camps was assessed using the Per Capita Greening Area (PCGA) datasets and K-means classification techniques. The satellite-based monitoring result affirms a massive loss of vegetative cover, approximately 5482.2 hectares (14%), in Ukhiya-Teknaf and 1502.56 hectares (79.57%) among the thirty-four refugee camps, between 2017 and 2019. K-means classification revealed that the vegetative cover in about 82% of the refugee camps is highly vulnerable. In the end, a recommendation as to establishing the studied region as an ecological park is proposed and some guidelines discussed. This could protect and reserve forests from further deforestation in the area, and foster future discussion among policymakers and researchers.

Classification of Diabetic Retinopathy using Capsules

Deep learning models have performed exceptionally well in detection of diabetic retinopathy. Most of these existing works either use Multi Layered Perceptron (MLP) or Convolutional Neural Network (CNN) based models. A significant drawback of these models is their inability to retain spatial dependencies as we go deeper into the network. Because of such issue, these models focus only on extraction of feature maps which help in classification or detection. In recent years’ transformers have shown enormous promise in both natural language processing and computer vision, due to which there have been significant work on transformer based techniques for image classification and object detection. One important model used in this research work is the Capsule Network which uses Set Transformers to perform vision tasks with high accuracy and has the capability to maintain spatial dependencies throughout the process. In this work, we leverage the power of capsules to the diabetic retinopathy classification and detection problem and trained capsules for classification of retinal fundus images are considered which were categorize into five different classes depending on the severity of retinopathy. Further, we also propose a sliding window based detector which can pinpoint the exact position of a blood burst in the retina which will ease the job for ophthalmologists while studying retinal fundus. From our experiments we found that capsules provide better results than existing convolutional neural network and multi layered perceptron based approaches in standard retinopathy datasets.

Comparison of Different Multispectral Images to Map Hydrothermal Alteration Zones in Kokap, Kulon Progo

Using remote sensing data for hydrothermal alteration mapping beside saving time and reducing cost leads to increased accuracy. In this study, the result of multispectral remote sensing tehcniques has been compare for manifesting hydrothermal alteration in Kokap, Kulon Progo. Three multispectral images, including ASTER, Landsat 8, and Sentinel-2, were compared in order to find the highest overall accuracy using principle component analysis (PCA) and directed component analysis (DPC). Several subsets band combinations were used as PCA and DPC input to targeting the key mineral of alteration. Multispectral classification with the maximum likelihood algorithm was performed to map the alteration types based on training and testing data and followed by accuracy evaluation. Two alteration zones were succeeded to be mapped: argillic zone and propylitic zone. Results of these image classification techniques were compared with known alteration zones from previous study. DPC combination of band ratio images of 5:2 and 6:7 of Landsat 8 imagery yielded a classification accuracy of 56.4%, which was 5.05% and 10.13% higher than those of the ASTER and Sentinel-2 imagery. The used of DEM together with multispectral images was increase the accuracy of hydrothermal alteration mapping in the study area.

Study on a visual coder acceleration algorithm for image classification applying dynamic scaling training techniques

Nowadays, image classification techniques are used in the field of autonomous vehicles, and Convolutional Neural Network (CNN) is used extensively, and Vision Transformer (ViT) networks are used instead of deep convolutional networks in order to compress the network size and improve the model accuracy. The ViT network is used to replace the deep convolutional network. Since training ViT requires a large dataset to have sufficient accuracy, a variant of ViT, Data-Efficient Image Transformers (DEIT), is used in this paper. In addition, in order to greatly reduce the computing memory and shorten the computing time in practical use, the network is flexibly scaled in size and training speed by both adaptive width and adaptive depth. In this paper, we introduce DEIT, width adaptive techniques and depth adaptive techniques and combine them to be applied to image classification examples. Experiments are conducted on the Cifar100 dataset, and the experiments demonstrate the superiority of the algorithm on image classification scenarios.

Learning-based classification of multispectral images for deterioration mapping of historic structures

The conservation of historic structures requires detailed knowledge of their state of preservation. Documentation of deterioration makes it possible to identify risk factors and interpret weathering mechanisms. It is usually performed using non-destructive methods such as mapping of surface features. The automated mapping of deterioration is a direction not often explored, especially when the investigated architectural surfaces present a multitude of deterioration forms and consist of heterogeneous materials, which significantly complicates the generation of thematic decay maps. This work combines reflectance imaging and supervised segmentation, based on machine learning methods, to automatically segment deterioration patterns on multispectral image composites, using a weathered historic fortification as a case study. Several spectral band combinations and image classification techniques (regression, decision tree, and ensemble learning algorithmic implementations) are evaluated to propose an accurate approach. The automated thematic mapping facilitates the spatial and semantic description of the deterioration patterns. Furthermore, the utilization of low-cost photographic equipment and easily operable digital image processing software adds to the practicality and agility of the presented methodology.

Exploring the optimal combination of image fusion and classification techniques

This study explores an optimal combination of certain selected image fusion and image classification techniques for improvement in classification accuracy. For this purpose, Sentinel-1A SAR and Sentinel-2A multispectral (MS) data have been obtained. Four fusion techniques Brovey Transform, Hue-Saturation-Value (HSV), PCA, and Gram-Schmidt (GS) Transform, have been compared along with two classification methods, Support Vector Machine (SVM) and Maximum Likelihood (ML) have been explored. Improvement in five classes Urban, Water, Baresoil, Forest and Vegetation Agricultural has been analysed. The result indicates that the GS-ML combination (87.56%) gives superior classification accuracy than the GS-SVM combination (84.35%). In contrast, PCA in combination with SVM (82.62%) provides superior accuracy compared to the combination of PCA-ML (82.56%).

Brain tumor identification with a hybrid feature extraction method based on discrete wavelet transform and principle component analysis

The Identification of brain tumors is a critical step that relies on the expertise and abilities of the physician. In order to enable radiologists to spot brain tumors, an automated tumor arrangement is extremely important. This paper presents a technique for MR brain image segmentation and classification to identify images as normal and abnormal. The proposed technique is a hybrid feature extraction submitted to enhance the classification results and basically consists of three stages. The first stage is used a 3-level of discrete wavelet transform (DWT) to extract image characteristics. In the second stage, the principle component analysis (PCA) is applied to reduce the size of characteristics. Finally, a random forest classifier (RF) was used with a feature selection for identification. 181 MR brain images are collected (81 normal and 100 abnormal), in distinguishing normal and abnormal tissues, the experimental results obtained an accuracy of 98%, the sensitivity achieved is 99.2%, specificity achieved is 97.8%, and showed the effectiveness of the proposed technique compared with many kinds of literature. The results show that the 3L-DWT+PCA+RF still achieved the best classification results. The proposed model could apply to the brain MRI sphere classification, which will help doctors to diagnose a tumor if it is normal or abnormal in certain degrees.

A Review on Image Segmentation Techniques and Its Recent Applications

The technique of attaching a name to each component in a picture so that elements with much the same label have certain visual qualities (pixels, colours, values, patterns, and other features of the image) is known as image segmentation. The output of image segmentation is a series of fragments or patterns taken from the image that collectively cover the full picture. After segmenting an image, classification detects classes in an image on the basic of similarities and differences between detected segments. These two tools together are doing wonders in better object detection in images. Big world has big data, and this big data need robust algorithms for understanding images without a human. This paper summarises the techniques of image segmentation and classification based on their efficiency. Finally, a comparison table shows who performed it better.

Surface composition of debris-covered glaciers across the Himalaya using linear spectral unmixing of Landsat 8 OLI imagery

Abstract. The Himalaya mountain range is characterized by highly glacierized, complex, dynamic topography. The ablation area of Himalayan glaciers often features a highly heterogeneous debris mantle comprising ponds, steep and shallow slopes of various aspects, variable debris thickness, and exposed ice cliffs associated with differing ice ablation rates. Understanding the composition of the supraglacial debris cover is essential for a proper understanding of glacier hydrology and glacier-related hazards. Until recently, efforts to map debris-covered glaciers from remote sensing focused primarily on glacier extent rather than surface characteristics and relied on traditional whole-pixel image classification techniques. Spectral unmixing routines, rarely used for debris-covered glaciers, allow decomposition of a pixel into constituting materials, providing a more realistic representation of glacier surfaces. Here we use linear spectral unmixing of Landsat 8 Operational Land Imager (OLI) images (30 m) to obtain fractional abundance maps of the various supraglacial surfaces (debris material, clean ice, supraglacial ponds and vegetation) across the Himalaya around the year 2015. We focus on the debris-covered glacier extents as defined in the database of global distribution of supraglacial debris cover. The spectrally unmixed surfaces are subsequently classified to obtain maps of composition of debris-covered glaciers across sample regions. We test the unmixing approach in the Khumbu region of the central Himalaya, and we evaluate its performance for supraglacial ponds by comparison with independently mapped ponds from high-resolution Pléiades (2 m) and PlanetScope imagery (3 m) for sample glaciers in two other regions with differing topo-climatic conditions. Spectral unmixing applied over the entire Himalaya mountain range (a supraglacial debris cover area of 2254 km2) indicates that at the end of the ablation season, debris-covered glacier zones comprised 60.9 % light debris, 23.8 % dark debris, 5.6 % clean ice, 4.5 % supraglacial vegetation, 2.1 % supraglacial ponds, and small amounts of cloud cover (2 %), with 1.2 % unclassified areas. The spectral unmixing performed satisfactorily for the supraglacial pond and vegetation classes (an F score of ∼0.9 for both classes) and reasonably for the debris classes (F score of 0.7). Supraglacial ponds were more prevalent in the monsoon-influenced central-eastern Himalaya (up to 4 % of the debris-covered area) compared to the monsoon-dry transition zone (only 0.3 %) and in regions with lower glacier elevations. Climatic controls (higher average temperatures and more abundant precipitation), coupled with higher glacier thinning rates and lower average glacier velocities, further favour pond incidence and the development of supraglacial vegetation. With continued advances in satellite data and further method refinements, the approach presented here provides avenues towards achieving large-scale, repeated mapping of supraglacial features.

Monitoring the quality of life in urban area using TDVI- Case study of Kalaburagi city

Many vegetation indices have been proposed over last decades made specialists search for the most suitable vegetation index for a given remote sensing application. Measuring the Quality of Place (QOP) is a hard task since it involves both physical and socio-economic dimensions. Being one of the major land use categories, urban vegetation plays a significant role in one‟s judgment for QOP in a neighborhood. Both quantity and quality of the community parks and recreation areas are major determinants of neighborhood attraction. For these reasons, detection of urban vegetation cover has been one of the important implication areas of urban image classification techniques. “Transformed Difference Vegetation Index (TDVI) developed by Bannari et al. (2002), is tested in a previous work where the index has performed better than NDVI and SAVI. In that work, a comparative study between TDVI, SAVI and NDVI for estimating vegetation cover in urban environment from the Indian Remote Sensing Satellite (IRS-1D) imagery has been conducted. The validation of the obtained results according to the ground truth showed that the TDVI is an excellent tool for vegetation cover monitoring in urban environment. It does not saturate like NDVI or SAVI, it shows an excellent linearity as a function of the rate of vegetation cover. This paper adds on the previous work by analyzing the performance of TDVI in urban image classification. Results indicate that, the performance of TDVI in urban image classification is better than NDVI and SAVI. The new index not only differentiates the urban vegetation cover better but also helps to minimize the error in classifying other unclassified pixels of urban categories.

UTILIZATION OF MULTITEMPORAL LANDSAT DATA FOR COASTLINE CHANGE ANALYSIS USING CANNY EDGE DETECTOR

The development of remote sensing technology has experienced significant progress. This progress is marked by the availability of images of various resolutions with complete channels, open-source, and their development in various image classification techniques such as edge detection techniques. This study aims to examine changes in the coastline in Banmaleng Village, Gili Raja Island, Sumenep Regency using the Canny edge detection algorithm on multitemporal Landsat 7 and 8 imagery in 2005 - 2020. The edging process includes (a) minimize noise by applying a Gaussian filter, (b) determine the image intensity gradient, (c) apply non-maximum emphasis to eliminate spurious responses to objects, (d) define multiple thresholds for determining potential image edges, (e) Track edge hysteresis by pressing all other weak edges that are not connected to firm edges. Canny edge detection produces sharp edges to analyze coast line changes quite well. Every five years, the coastline of Banmaleng Village experienced shifting from 0 , 4 to 2 , 3 meters in several areas. Keywords : Edge Detection, Landsat, Coastline, Canny Algorithm, Gili Raja

Machine Learning and Deep Learning Based Computational Approaches in Automatic Microorganisms Image Recognition: Methodologies, Challenges, and Developments.

Microorganisms or microbes comprise majority of the diversity on earth and are extremely important to human life. They are also integral to processes in the ecosystem. The process of their recognition is highly tedious, but very much essential in microbiology to carry out different experimentation. To overcome certain challenges, machine learning techniques assist microbiologists in automating the entire process. This paper presents a systematic review of research done using machine learning (ML) and deep leaning techniques in image recognition of different microorganisms. This review investigates certain research questions to analyze the studies concerning image pre-processing, feature extraction, classification techniques, evaluation measures, methodological limitations and technical development over a period of time. In addition to this, this paper also addresses the certain challenges faced by researchers in this field. Total of 100 research publications in the chronological order of their appearance have been considered for the time period 1995–2021. This review will be extremely beneficial to the researchers due to the detailed analysis of different methodologies and comprehensive overview of effectiveness of different ML techniques being applied in microorganism image recognition field.

Flood Mapping in the Coastal Region of Bangladesh Using Sentinel-1 SAR Images: A Case Study of Super Cyclone Amphan

Floods are triggered by water overflow into drylands from several sources, including rivers, lakes, oceans, or heavy rainfall. Near real-time (NRT) flood mapping plays an important role in taking strategic measures to reduce flood damage after a flood event. There are many satellite imagery based remote sensing techniques that are widely used to generate flood maps. Synthetic aperture radar (SAR) images have proven to be more effective in flood mapping due to its high spatial resolution and cloud penetration capacity. This case study is focused on the super cyclone, commonly known as Amphan, stemming from the west Bengal-Bangladesh coast across the Sundarbans on 20 May 2020, with a wind speed between 155 -165 gusting up to 185 . The flooding extent is determined by analyzing the pre and post-event synthetic aperture radar images, using the change detection and thresholding (CDAT) method. The results showed an inundated landmass of 2146 on 22 May 2020, excluding Sundarban. However, the area became 1425 about a week after the event, precisely on 28 May 2020 . This persistency generated a more severe and intense flood, due to the broken embankments. Furthermore, 13 out of 19 coastal districts were affected by the flooding, while 8 were highly inundated, including Bagerhat, Pirojpur, Satkhira, Khulna, Barisal, Jhalokati, Patuakhali and Barguna. These findings were subsequently compared with an inundation map created with a validation survey immediately after the event and also with the disposed location using a machine learning-based image classification technique. Consequently, the comparison showed a close similarity between the inundation scenario and the flood reports from the secondary sources. This circumstance envisages the significant role of CDAT application in providing relevant information for an effective decision support system.

Creativity in Generative Musical Networks: Evidence From Two Case Studies.

Deep learning, one of the fastest-growing branches of artificial intelligence, has become one of the most relevant research and development areas of the last years, especially since 2012, when a neural network surpassed the most advanced image classification techniques of the time. This spectacular development has not been alien to the world of the arts, as recent advances in generative networks have made possible the artificial creation of high-quality content such as images, movies or music. We believe that these novel generative models propose a great challenge to our current understanding of computational creativity. If a robot can now create music that an expert cannot distinguish from music composed by a human, or create novel musical entities that were not known at training time, or exhibit conceptual leaps, does it mean that the machine is then creative? We believe that the emergence of these generative models clearly signals that much more research needs to be done in this area. We would like to contribute to this debate with two case studies of our own: TimbreNet, a variational auto-encoder network trained to generate audio-based musical chords, and StyleGAN Pianorolls, a generative adversarial network capable of creating short musical excerpts, despite the fact that it was trained with images and not musical data. We discuss and assess these generative models in terms of their creativity and we show that they are in practice capable of learning musical concepts that are not obvious based on the training data, and we hypothesize that these deep models, based on our current understanding of creativity in robots and machines, can be considered, in fact, creative.

Evaluación de parámetros de segmentación en OBIA para la clasificación de coberturas del suelo a partir de imágenes VANT

<p>Unmanned Aerial Vehicles (UAVs) have given a new boost to remote sensing and image classification techniques due to the high level of detail among other factors. Object-based image analysis (OBIA) could improve classification accuracy unlike to pixel-based, especially in high-resolution images. OBIA application for image classification consists of three stages i.e., segmentation, class definition and training polygons, and classification. However, defining the parameters: spatial radius (SR), range radius (RR) and minimum region size (MR) is necessary during the segmentation stage. Despite their relevance, they are usually visually adjusted, which leads to a subjective interpretation. Therefore, it is of utmost importance to generate knowledge focused on evaluating combinations of these parameters. This study describes the use of the mean-shift segmentation algorithm followed by <em>Random Forest </em>classifier using Orfeo Toolbox software. It was considered a multispectral orthomosaic derived from UAV to generate a suburban map land cover in town of El Pueblito, Durango, Mexico. The main aim was to evaluate efficiency and segmentation quality of nine parameter combinations previously reported in scientific studies.This in terms of number generated polygons, processing time, discrepancy measures for segmentation and classification accuracy metrics. Results evidenced the importance of calibrating the input parameters in the segmentation algorithms. Best combination was RE=5, RR=7 and TMR=250, with a Kappa index of 0.90 and shortest processing time. On the other hand, RR showed a strong and inversely proportional degree of association regarding the classification accuracy metrics.</p>

Remote sensing techniques: mapping and monitoring of mangrove ecosystem\u2014a review

Mangrove forests are considered to be the most productive ecosystem yet vanishing rapidly over the world. They are mostly found in the intertidal zone and sheltered by the seacoast. Mangroves have potential socio-economic benefits such as protecting the shoreline from storm and soil erosion, flood and flow control, acting as a carbon sink, provides a fertile breeding ground for marine species and fauna. It also acts as a source of income by providing various forest products. Restoration and conservation of mangrove forests remain a big challenge due to the large and inaccessible areas covered by mangroves forests which makes field assessment difficult and time-consuming. Remote sensing along with various digital image classification approaches seem to be promising in providing better and accurate results in mapping and monitoring the mangroves ecosystem. This review paper aims to provide a comprehensive summary of the work undertaken, and addresses various remote sensing techniques applied for mapping and monitoring of the mangrove ecosystem, and summarize their potential and limitation. For that various digital image classification techniques are analyzed and compared based on the type of image used with its spectral resolution, spatial resolution, and other related image features along with the accuracy of the classification to derive specific class information related to mangroves. The digital image classification techniques used for mangrove mapping and monitoring in various studies can be classified into pixel-based, object-based, and knowledge-based classifiers. The various satellite image data analyzed are ranged from light detection and ranging (LiDAR), hyperspectral and multispectral optical imagery, synthetic aperture radar (SAR), and aerial imagery. Supervised state of the art machine learning/deep machine learning algorithms which use both pixel-based and object-based approaches and can be combined with the knowledge-based approach are widely used for classification purpose, due to the recent development and evolution in these techniques. There is a huge future scope to study the performance of these classification techniques in combination with various high spatial and spectral resolution optical imageries, SAR and LiDAR, and also with multi-sensor, multiresolution, and temporal data.

PROMISING ADVANCES OF AMAZONIAN MONITORING SYSTEMS THROUGHOUT VANGUARD TECHNOLOGY AND SCIENTIFIC KNOWLEDGE

Abstract. PRODES and DETER project together turned 33 years-old with an undeniably contribution to the state-of-art in mapping and monitoring tropical deforestation in Brazil. Monitoring systems all over the world have taken advantage of big data repositories of remote sensing data as they are becoming freely available together with artificial intelligence. Thus, considering the advent of new generation remote sensing data hubs, online platforms of big data that can fill in spatial and temporal resolutions gaps in current deforestation mapping, this work aims to present recent innovations at INPE´s deforestation monitoring systems in Brazil and how they are gauging new realms of technological levels. Recent innovations at INPE´s monitoring systems are: 1) the development of TerraBrasilis platform of data access and analysis; 2) the adoption of new sensors and cloud detection strategies; 3) the complementary use of multi-sensor images; 4) the complementary adoption of SAR C-band images using cloud data to sample and process Sentinel-1. Future innovations are: 1) development of a Brazilian data cube to be used in deep learning techniques of image classification; 2) Routine uncertainty analysis of PRODES data. Automatization might fasten mapping process, but the real challenge is to succeed in automatization maintaining data quality and historical series. The hyper-availability of remote sensing data, the initiative of a Brazilian Data Cube and promising machine learning techniques applied to land cover change detection, allowed INPE to reinforce its central role in tropical forest monitoring.

MACHINE LEARNING TECHNIQUES FOR KNOWLEDGE EXTRACTION FROM SATELLITE IMAGES: APPLICATION TO SPECIFIC AREA TYPES

Abstract. When we want to extract knowledge form satellite images, several well-known image classification and analysis techniques can be concatenated or combined to gain a more detailed target understanding. In our case, we concentrated on specific extended target areas such as polar ice-covered surfaces, forests shrouded by fire plumes, flooded areas, and shorelines. These image types can be described by characteristic features and statistical relationships. Here, we demonstrate that both multispectral (optical) as well as SAR (Synthetic Aperture Radar) images can be used for knowledge extraction. The free availability of image data provided by the European Sentinel-1 and Sentinel-2 satellites allowed us to conduct a series of experiments that verified our classification approaches. This could already be verified in our recent work by quantitative quality tests.