Geographic Object-based Image Analysis Research Articles

Semantic Unsupervised Change Detection of Natural Land Cover With Multitemporal Object-Based Analysis on SAR Images

Change detection is one of the most addressed topics in the remote sensing community. When performed on synthetic aperture radar images, the most critical issues are as follows: 1) the labeling of the identified changing patterns and 2) the scarce robustness of classic pixel-based approaches based on threshold segmentation of an appropriate change index, which tend to fail when multiple changes are present in the study area. In this work, a new methodology for unsupervised change detection in vegetation canopy is presented. It overcomes these limitations by exploiting multitemporal geographical object-based image analysis with the aim to make the intrinsic semantic of data emerge and direct the processing toward the identification of precise classes of changes through dictionary-based preclassification and fuzzy combination of class-specific information layers. The proposed methodology has been tested in ten different experiments covering agriculture and clear-cut deforestation applications. The results, validated against literature methods, highlighted the superiority of the proposed approach, which was quantitatively assessed in terms of standard classification quality parameters. On agriculture experiments, it allowed for an average increase in the detection accuracy of about 11% with respect to the best performing literature method, with an increment of the false alarm rate in the order of 0.5%. In case of deforestation, the registered detection accuracy was comparable to that achieved by the literature, while the most significant benefit was the reduction, of more than one-third, of the number of detected false deforestation patterns. Overall, the main characteristics of the proposed architecture are the robustness and the lack of any supervision, which makes it very well-suited for operational scenarios.

A Fast and Effective Method for Unsupervised Segmentation Evaluation of Remote Sensing Images

The segmentation of remote sensing images with high spatial resolution is important and fundamental in geographic object-based image analysis (GEOBIA), so evaluating segmentation results without prior knowledge is an essential part in segmentation algorithms comparison, segmentation parameters selection, and optimization. In this study, we proposed a fast and effective unsupervised evaluation (UE) method using the area-weighted variance (WV) as intra-segment homogeneity and the difference to neighbor pixels (DTNP) as inter-segment heterogeneity. Then these two measures were combined into a fast-global score (FGS) to evaluate the segmentation. The effectiveness of DTNP and FGS was demonstrated by visual interpretation as qualitative analysis and supervised evaluation (SE) as quantitative analysis. For this experiment, the ‘‘Multi-resolution Segmentation’’ algorithm in eCognition was adopted in the segmentation and four typical study areas of GF-2 images were used as test data. The effectiveness analysis of DTNP shows that it can keep stability and remain sensitive to both over-segmentation and under-segmentation compared to two existing inter-segment heterogeneity measures. The effectiveness and computational cost analysis of FGS compared with two existing UE methods revealed that FGS can effectively evaluate segmentation results with the lowest computational cost.

Incorporating Deep Features into GEOBIA Paradigm for Remote Sensing Imagery Classification: A Patch-Based Approach

The fast and accurate creation of land use/land cover maps from very-high-resolution (VHR) remote sensing imagery is crucial for urban planning and environmental monitoring. Geographic object-based image analysis methods (GEOBIA) provide an effective solution using image objects instead of individual pixels in VHR remote sensing imagery analysis. Simultaneously, convolutional neural networks (CNN) have been widely used in the image processing field because of their powerful feature extraction capabilities. This study presents a patch-based strategy for integrating deep features into GEOBIA for VHR remote sensing imagery classification. To extract deep features from irregular image objects through CNN, a patch-based approach is proposed for representing image objects and learning patch-based deep features, and a deep features aggregation method is proposed for aggregating patch-based deep features into object-based deep features. Finally, both object and deep features are integrated into a GEOBIA paradigm for classifying image objects. We explored the influences of segmentation scales and patch sizes in our method and explored the effectiveness of deep and object features in classification. Moreover, we performed 5-fold stratified cross validations 50 times to explore the uncertainty of our method. Additionally, we explored the importance of deep feature aggregation, and we evaluated our method by comparing it with three state-of-the-art methods in a Beijing dataset and Zurich dataset. The results indicate that smaller segmentation scales were more conducive to VHR remote sensing imagery classification, and it was not appropriate to select too large or too small patches as the patch size should be determined by imagery and its resolution. Moreover, we found that deep features are more effective than object features, while object features still matter for image classification, and deep feature aggregation is a critical step in our method. Finally, our method can achieve the highest overall accuracies compared with the state-of-the-art methods, and the overall accuracies are 91.21% for the Beijing dataset and 99.05% for the Zurich dataset.

Gully mapping using geographic object-based image analysis: A case study at catchment scale in the Brazilian Cerrado

Water erosion is one of the main factors of soil degradation, causing several environmental damages. The most severe stage of water erosion culminates in the emergence of gullies, which increases soil loss and sediment production. The Cerrado biome, a global diversity hotspot, has been affected by gullies in many regions of Brazil. This study investigates the use of Geographic Object-Based Image Analysis (GEOBIA) to detect large gullies from RapidEye and SRTM data in anthropized Brazilian Cerrado. For the first time, a two-sided 50% overlap criteria was used to assess gully segmentation by applying Segmentation Evaluation Index (SEI). The results were checked against manually digitized reference data. The results of gully mapping indicated a user's accuracy of 69.78% in area 1 and 90.24% in area 2; a producer's accuracy of 52.10% in area 1 and 55.42% in area 2. The model can be considered robust since it was possible to detect gullies and generate few false positives in a heterogeneous scene, even using medium resolution data. This approach has the potential of application on a regional scale and can provide valuable information for land use management.

High-Resolution Surface Water Classifications of the Xingu River, Brazil, Pre and Post Operationalization of the Belo Monte Hydropower Complex

We describe a new high spatial resolution surface water classification dataset generated for the Xingu river, Brazil, from its confluence with the Iriri river to the Pimental dam prior to construction of the Belo Monte hydropower complex, and after its operationalization. This river is well-known for its exceptionally high diversity and endemism in ichthyofauna. Pre-existing datasets generated from moderate resolution satellite imagery (e.g., 30 m) do not adequately capture the extent of the river. Accurate measurements of water extent are important for a range of applications utilizing surface water data, including greenhouse gas emission estimation, land cover change mapping, and habitat loss/change estimates, among others. We generated the new classifications from RapidEye imagery (5 m pixel size) for 2011 and PlanteScope imagery (3 m pixel size) for 2019 using a Geographic Object Based Image Analysis (GEOBIA) approach.

Introducing GEOBIA to Landscape Imageability Assessment: A Multi-Temporal Case Study of the Nature Reserve “Kózki”, Poland

Geographic object-based image analysis (GEOBIA) is a primary remote sensing tool utilized in land-cover mapping and change detection. Land-cover patches are the primary data source for landscape metrics and ecological indicator calculations; however, their application to visual landscape character (VLC) indicators was little investigated to date. To bridge the knowledge gap between GEOBIA and VLC, this paper puts forward the theoretical concept of using viewpoint as a landscape imageability indicator into the practice of a multi-temporal land-cover case study and explains how to interpret the indicator. The study extends the application of GEOBIA to visual landscape indicator calculations. In doing so, eight different remote sensing imageries are the object of GEOBIA, starting from a historical aerial photograph (1957) and CORONA declassified scene (1965) to contemporary (2018) UAV-delivered imagery. The multi-temporal GEOBIA-delivered land-cover patches are utilized to find the minimal isovist set of viewpoints and to calculate three imageability indicators: the number, density, and spacing of viewpoints. The calculated indicator values, viewpoint rank, and spatial arrangements allow us to describe the scale, direction, rate, and reasons for VLC changes over the analyzed 60 years of landscape evolution. We found that the case study nature reserve (“Kózki”, Poland) landscape imageability transformed from visually impressive openness to imageability due to the impression of several landscape rooms enclosed by forest walls. Our results provide proof that the number, rank, and spatial arrangement of viewpoints constitute landscape imageability measured with the proposed indicators. Discussing the method’s technical limitations, we believe that our findings contribute to a better understanding of land-cover change impact on visual landscape structure dynamics and further VLC indicator development.

Object-based island hierarchical land cover classification using unmanned aerial vehicle multitype data

The unmanned aerial vehicle (UAV) is an emerging technology applied recently in land cover classification, owing to its ability to acquire very high-resolution spatial data, that has provided an effective means for detailed land cover mapping, especially for a small island area. Selecting suitable UAV-acquired data and exploring the combined use of UAV multitype data are of significance for island mapping. Nine classification models were established through a fusion method of visible, multispectral, and light detection and ranging (LIDAR) data acquired by UAVs. A two-level hierarchical land cover classification (level 1 and level 2) of the Donkey Island in China was performed using geographic object-based image analysis with random forest classifier. We investigated the performance of land cover classification models containing different sets of features (spectral, height, intensity, and shape features extracted from UAV data) and evaluated the importance of various features. The results demonstrate that the overall accuracy (OA) of the models generally increase with decreasing detail and the amount of information entering the classification process. The OA achieved range from 82.08% to 92.54% and 74.12% to 85.08% across the nine models for level 1 and level 2, respectively. The best result was achieved with a model combining all features based on multispectral and LIDAR data. Height and intensity information input significantly affect the quality of classification models positively, with height apparently more significant than LIDAR information. When comparing different features, spectral features prominently assist in discriminating land cover classes. The importance of height and intensity features to classification accuracy varies for the classification models, showing greater importance in models based on visible data.

Rapid Recent Deforestation Incursion in a Vulnerable Indigenous Land in the Brazilian Amazon and Fire-Driven Emissions of Fine Particulate Aerosol Pollutants

Deforestation in the Brazilian Amazon is related to the use of fire to remove natural vegetation and install crop cultures or pastures. In this study, we evaluated the relation between deforestation, land-use and land-cover (LULC) drivers and fire emissions in the Apyterewa Indigenous Land, Eastern Brazilian Amazon. In addition to the official Brazilian deforestation data, we used a geographic object-based image analysis (GEOBIA) approach to perform the LULC mapping in the Apyterewa Indigenous Land, and the Brazilian biomass burning emission model with fire radiative power (3BEM_FRP) to estimate emitted particulate matter with a diameter less than 2.5 µm (PM2.5), a primary human health risk. The GEOBIA approach showed a remarkable advancement of deforestation, agreeing with the official deforestation data, and, consequently, the conversion of primary forests to agriculture within the Apyterewa Indigenous Land in the past three years (200 km2), which is clearly associated with an increase in the PM2.5 emissions from fire. Between 2004 and 2016 the annual average emission of PM2.5 was estimated to be 3594 ton year−1, while the most recent interval of 2017–2019 had an average of 6258 ton year−1. This represented an increase of 58% in the annual average of PM2.5 associated with fires for the study period, contributing to respiratory health risks and the air quality crisis in Brazil in late 2019. These results expose an ongoing critical situation of intensifying forest degradation and potential forest collapse, including those due to a savannization forest-climate feedback, within “protected areas” in the Brazilian Amazon. To reverse this scenario, the implementation of sustainable agricultural practices and development of conservation policies to promote forest regrowth in degraded preserves are essential.

Geographic Object-Based Image Analysis: A Primer and Future Directions

Geographic object-based image analysis (GEOBIA) is a remote sensing image analysis paradigm that defines and examines image-objects: groups of neighboring pixels that represent real-world geographic objects. Recent reviews have examined methodological considerations and highlighted how GEOBIA improves upon the 30+ year pixel-based approach, particularly for H-resolution imagery. However, the literature also exposes an opportunity to improve guidance on the application of GEOBIA for novice practitioners. In this paper, we describe the theoretical foundations of GEOBIA and provide a comprehensive overview of the methodological workflow, including: (i) software-specific approaches (open-source and commercial); (ii) best practices informed by research; and (iii) the current status of methodological research. Building on this foundation, we then review recent research on the convergence of GEOBIA with deep convolutional neural networks, which we suggest is a new form of GEOBIA. Specifically, we discuss general integrative approaches and offer recommendations for future research. Overall, this paper describes the past, present, and anticipated future of GEOBIA in a novice-accessible format, while providing innovation and depth to experienced practitioners.

Burned Area Detection in the Brazilian Amazon using Spectral Indices and GEOBIA

Mapping refined burned areas (BA) in the Brazilian Amazon is still a challenge. The main difficulty of BA detection in large areas is the presence of cloud cover and water bodies. The use of different data sources of medium spatial resolution satellite images can provide a higher availability of cloud-free images. Besides that, it may decrease the uncertainties associated with coarse spatial resolution data (>250m), which can under or overestimate BA and hinder the detection of small BA patches (<0.1km²). In this study, we propose an innovative methodology based on spectral indices and geographic object-based image analysis (GEOBIA), using medium spatial resolution images to improve BA detection in the Brazilian Amazon region. Firstly, we assessed the performance of nine spectral indices in two study areas, derived from Landsat-8 OLI and Sentinel-2A MSI data to identify the most suitable index for BA detection in this region. Then, we refined this data through the GEOBIA-based model. The results showed that the Burned Area Index (BAI) was the most suitable index for BA mapping (M index >1.5) for both sensors. Our model allowed detecting more than 80% of small BA and also presented high Dice coefficient values (~0.70) with low omission and commission errors (0.22 and 0.32, respectively). Such combined approach corresponds to a novel contribution to the BA detection in the Brazilian Amazon region and for enhancing the operational product generation.

Standardized object-based dual CNNs for very high-resolution remote sensing image classification and standardization combination effect analysis

ABSTRACT Advances in the object-based convolutional neural network (CNN) have demonstrated the superiority of CNNs for image classification. However, any object-based CNN, regardless of its model structure, only stacks the square images with different scales when extracting features. The impact of background information around the segmented object (the number of pixels around the segmented object) for the classification accuracy is neglected. In addition, blurred object boundaries and feature representation, as well as huge computational redundancy, restrict the application for very high-resolution remote sensing image (VHRI) classification. To solve these problems, a novel standardized object-based dual CNN (SOD-CNN) is proposed for VHRI classification. First, based on geographic object-based image analysis, the image is segmented into homogeneous regions. Second, these less-segmented objects are over-segmented into superpixels with high compactness to provide crisp and accurate boundary delineation at the pixel level. Third, four standardization methods are developed to limit the number of pixels around the segmented object. The standardized less-segmented object and over-segmented object are fed into two different CNNs to capture different perspectives of features at local and global scales. Finally, feature fusion based on the full connection is performed to integrate the class-specific classification results. The effectiveness of the proposed method was verified by using two VHRI, which achieved excellent classification accuracy, consistently outperforming the benchmark comparisons. The overall and per-class classification accuracy was investigated under different standardization combinations. We found that (1) the proposed standardization method not only reduced redundancy of information in the object-based CNN but also highlighted the features of segmented objects; (2) different segmented objects had different optimal standardization combinations; and (3) the classification accuracy was reasonably controlled by the foundation number of training samples.

Monitoring the vegetation vigor in heterogeneous citrus and olive orchards. A multiscale object-based approach to extract trees’ crowns from UAV multispectral imagery

Precision agriculture (PA) constitutes one of the most critical sectors of remote sensing applications that allow obtaining spatial segmentation and within-field variability information from field crops. In the last decade, an increasing source of information is provided by unmanned aerial vehicle (UAVs) platforms, mainly equipped with optical multispectral cameras, to map, monitor, and analyze, temporal and spatial variations of vegetation using ad hoc spectral vegetation indices (VIs). Considering the centimeter or sub-centimeter spatial resolution of UAV imagery, the geographic object-based image analysis (GEOBIA) approach, is becoming prevalent in UAV remote sensing applications. In the present paper, we propose a quick and reliable semi-automatic workflow implemented to process multispectral UAV imagery and aimed at the detection and extraction of olive and citrus trees’ crowns to obtain vigor maps in the framework of PA. We focused our attention on the choice of GEOBIA data input and parameters, taking into consideration its replicability and reliability in the case of heterogeneous tree orchards. The heterogeneity concerns the different tree plantation distances and composition, different crop management (irrigation, pruning, weeding), and different tree age, height, and crown diameters. The proposed GEOBIA workflow was implemented in the eCognition Developer 9.5, coupling the use of multispectral and topographic information surveyed using the Tetracam µ-MCA06 snap multispectral camera at 4 cm of ground sample distance (GSD). Three different study sites in heterogeneous citrus (Bergamot and Clementine) and olive orchards located in the Calabria region (Italy) were provided. Multiresolution segmentation was implemented using spectral and topographic band layers and optimized by applying a trial-and-error approach. The classification step was implemented as process-tree and based on a rule set algorithm, therefore easily adaptable and replicable to other datasets. Decision variables for image classification were spectral vegetation indices (NDVI, SAVI, CVI) and topographic layers (DSM and CHM). Vigor maps were based on NDVI and NDRE and allowed to highlight those areas with low vegetative vigor. The accuracy assessment was based on a per-pixel approach and computed through the F-score (F). The obtained results are promising, considering that the resulting accuracy was high, with F-score ranging from 0.85 to 0.91 for olive and bergamot, respectively. Our proposed workflow, which has proved effective in datasets of different complexity, finds its strong point is the speed of execution and on its repeatability to other different crops with few adjustments. It appears worth of interest to highlights that it requests a working day of two good skilled operators in geomatics and computer image processing, from the on-field data collection to the obtaining of vigor maps.

A comparison of data mining techniques and multi-sensor analysis for inland marshes delineation

Inland Marsh (IM) is a type of wetland characterized by the presence of non-woody plants as grasses, reeds or sedges, with a water surface smaller than 25% of the area. Historically, these areas have been suffering impacts related to pollution by urban, industrial and agrochemical waste, as well as drainage for agriculture. The IM delineation allows to understand the vegetation and hydrodynamic dynamics and also to monitor the degradation caused by human-induced activities. This work aimed to compare four machine learning algorithms (classification and regression tree (CART), artificial neural network (ANN), random forest (RF), and k-nearest neighbors (k-NN)) using active and passive remote sensing data in order to address the following questions: (1) which of the four machine learning methods has the greatest potential for inland marshes delineation? (2) are SAR features more important for inland marshes delineation than optical features? and (3) what are the most accurate classification parameters for inland marshes delineation? To address these questions, we used data from Sentinel 1A and Alos Palsar I (SAR) and Sentinel 2A (optical) sensors, in a geographic object-based image analysis (GEOBIA) approach. In addition, we performed a vectorization of a 1975 Brazilian Army topographic chart (first official document presenting marsh boundaries) in order to quantify the marsh area losses between 1975 and 2018 by comparing it with a Sentinel 2A image. Our results showed that the method with the highest overall accuracy was k-NN, with 98.5%. The accuracies for the RF, ANN, and CART methods were 98.3%, 96.0% and 95.5%, respectively. The four classifiers presented accuracies exceeding 95%, showing that all methods have potential for inland marsh delineation. However, we note that the classification results have a great dependence on the input layers. Regarding the importance of the features, SAR images were more important in RF and ANN models, especially in the HV, HV + VH and VH channels of the Alos Palsar I L-band satellite, while spectral indices from optical images were more important in the marshes delineation with the CART method. In addition, we found that the CART and ANN methods presented the largest variations of the overall accuracy (OA) in relation to the different parameters tested. The multi-sensor approach was critical for the high OA values found in the IM delineation (> 95%). The four machine learning methods can be accurately applied for IM delineation, acting as an important low-cost tool for monitoring and managing these environments, in the face of advances in agriculture, soil degradation and pollution of water resources due to agrochemical dumping.

Image Segmentation and Object-Based Image Analysis for Environmental Monitoring: Recent Areas of Interest, Researchers’ Views on the Future Priorities

Image segmentation and geographic object-based image analysis (GEOBIA) were proposed around the turn of the century as a means to analyze high-spatial-resolution remote sensing images. Since then, object-based approaches have been used to analyze a wide range of images for numerous applications. In this Editorial, we present some highlights of image segmentation and GEOBIA research from the last two years (2018–2019), including a Special Issue published in the journal Remote Sensing. As a final contribution of this special issue, we have shared the views of 45 other researchers (corresponding authors of published papers on GEOBIA in 2018–2019) on the current state and future priorities of this field, gathered through an online survey. Most researchers surveyed acknowledged that image segmentation/GEOBIA approaches have achieved a high level of maturity, although the need for more free user-friendly software and tools, further automation, better integration with new machine-learning approaches (including deep learning), and more suitable accuracy assessment methods was frequently pointed out.

Investigating the performance of sentinel-2A and Landsat 8 imagery in mapping shoreline changes

Regular detection and timely monitoring of shoreline evolution are essentially needed in order to identify areas that require further investigation and protection. Possibilities for regular monitoring of shoreline changes are better facilitated now with high-frequency revisit times of Landsat-Sentinel-2 virtual constellation. Accordingly, this study aimed at investigating the performance of Landsat 8 and Sentinel-2A satellite imagery in mapping shoreline changes. The specific objectives were to 1) create reference data of shoreline erosion, accretion and seafilling using very high spatial resolution data, 2) compare the results of shoreline changes by using Landsat 8 and Sentinel-2A imagery respectively, and 3) complete the inventory of shoreline changes between 1962 and 2016. A Quickbird panchromatic imagery (0.6 m) of 2010 was employed in combination with aerial photography (0.5 m) from 1962 for the generation of reference data using a traditional photo-interpretation method. Then, geographic object-based image analysis (GEOBIA) was employed to map changes in the shoreline between 1962 (i.e., reference shoreline) and 2016 using Sentinel-2A (10 m) and Landsat 8 panchromatic band (15 m) imagery, consecutively. The use of Sentinel-2A image provided slightly better accuracy results when compared with those from the Landsat 8 image. Accordingly, an inventory of Lebanese shoreline evolution between 1962 and 2016 was completed using the Sentinel-2A classification results. The combined use of Landsat-Sentinel-2 imagery is expected to generate reliable data records for continuous monitoring of shoreline changes.

GEOBIA for mangrove mapping using UAV-modified NIR 1 camera sensor

Mangrove ecosystem is a natural resource that can be beneficial for coastal community. One of its benefits is it can be utilized as an ecotourism spot such as Baros Mangrove Conservation Area in Kretek, Bantul. Hence, mangrove ecosystem must be conserved or expanded if possible. The first step to conserving mangrove ecosystem is to map its existing condition. UAV can be used to acquire a high-resolution imagery quickly. With modified near-infrared (NIR) sensor from a pocket camera can produce three band i.e. NIR, green, and blue that can be used to calculate several vegetation indices such as Simple Ration (SR), Normalized Difference Vegetation Index (NDVI), and Enhanced Difference Vegetation Index (ENDVI). Those bands and vegetation indices then used to map mangrove using Geographic Object-Based Image Analysis approach. The result indicated that high-resolution imagery from modified NIR camera sensor can be used to detect mangrove and distinguish it from other object using supervised classification. The digitally classified imagery then compared to manually digitized mangrove map to calculate the total error of the map.

Geographic Object-Based Image Analysis Framework for Mapping Vegetation Physiognomic Types at Fine Scales in Neotropical Savannas

Regional maps of vegetation structure are necessary for delineating species habitats and for supporting conservation and ecological analyses. A systematic approach that can discriminate a wide range of meaningful and detailed vegetation classes is still lacking for neotropical savannas. Detailed vegetation mapping of savannas is challenged by seasonal vegetation dynamics and substantial heterogeneity in vegetation structure and composition, but fine spatial resolution imagery (<10 m) can improve map accuracy in these heterogeneous landscapes. Traditional pixel-based classification methods have proven problematic for fine spatial resolution data due to increased within-class spectral variability. Geographic Object-Based Image Analysis (GEOBIA) is a robust alternative method to overcome these issues. We developed a systematic GEOBIA framework accounting for both spectral and spatial features to map Cerrado structural types at 5-m resolution. This two-step framework begins with image segmentation and a Random Forest land cover classification based on spectral information, followed by spatial contextual and topological rules developed in a systematic manner in a GEOBIA knowledge-based approach. Spatial rules were defined a priori based on descriptions of environmental characteristics of 11 different physiognomic types and their relationships to edaphic conditions represented by stream networks (hydrography), topography, and substrate. The Random Forest land cover classification resulted in 10 land cover classes with 84.4% overall map accuracy and was able to map 7 of the 11 vegetation classes. The second step resulted in mapping 13 classes with 87.6% overall accuracy, of which all 11 vegetation classes were identified. Our results demonstrate that 5-m spatial resolution imagery is adequate for mapping land cover types of savanna structural elements. The GEOBIA framework, however, is essential for refining land cover categories to ecological classes (physiognomic types), leading to a higher number of vegetation classes while improving overall accuracy.

Knowledge-Based Classification of Grassland Ecosystem Based on Multi-Temporal WorldView-2 Data and FAO-LCCS Taxonomy

Grassland ecosystems can provide a variety of services for humans, such as carbon storage, food production, crop pollination and pest regulation. However, grasslands are today one of the most endangered ecosystems due to land use change, agricultural intensification, land abandonment as well as climate change. The present study explores the performance of a knowledge-driven GEOgraphic-Object—based Image Analysis (GEOBIA) learning scheme to classify Very High Resolution (VHR) images for natural grassland ecosystem mapping. The classification was applied to a Natura 2000 protected area in Southern Italy. The Food and Agricultural Organization Land Cover Classification System (FAO-LCCS) hierarchical scheme was instantiated in the learning phase of the algorithm. Four multi-temporal WorldView-2 (WV-2) images were classified by combining plant phenology and agricultural practices rules with prior-image spectral knowledge. Drawing on this knowledge, spectral bands and entropy features from one single date (Post Peak of Biomass) were firstly used for multiple-scale image segmentation into Small Objects (SO) and Large Objects (LO). Thereafter, SO were labelled by considering spectral and context-sensitive features from the whole multi-seasonal data set available together with ancillary data. Lastly, the labelled SO were overlaid to LO segments and, in turn, the latter were labelled by adopting FAO-LCCS criteria about the SOs presence dominance in each LO. Ground reference samples were used only for validating the SO and LO output maps. The knowledge driven GEOBIA classifier for SO classification obtained an OA value of 97.35% with an error of 0.04. For LO classification the value was 75.09% with an error of 0.70. At SO scale, grasslands ecosystem was classified with 92.6%, 99.9% and 96.1% of User’s, Producer’s Accuracy and F1-score, respectively. The findings reported indicate that the knowledge-driven approach not only can be applied for (semi)natural grasslands ecosystem mapping in vast and not accessible areas but can also reduce the costs of ground truth data acquisition. The approach used may provide different level of details (small and large objects in the scene) but also indicates how to design and validate local conservation policies.

Object-Based Ensemble Learning for Pan-European Riverscape Units Mapping Based on Copernicus VHR and EU-DEM Data Fusion

Recent developments in the fields of geographical object-based image analysis (GEOBIA) and ensemble learning (EL) have led the way to the development of automated processing frameworks suitable to tackle large-scale problems. Mapping riverscape units has been recognized in fluvial remote sensing as an important concern for understanding the macrodynamics of a river system and, if applied at large scales, it can be a powerful tool for monitoring purposes. In this study, the potentiality of GEOBIA and EL algorithms were tested for the mapping of key riverscape units along the main European river network. The Copernicus VHR Image Mosaic and the EU Digital Elevation Model (EU-DEM)—both made available through the Copernicus Land Monitoring Service—were integrated within a hierarchical object-based architecture. In a first step, the most well-known EL techniques (bagging, boosting and voting) were tested for the automatic classification of water, sediment bars, riparian vegetation and other floodplain units. Random forest was found to be the best-to-use classifier, and therefore was used in a second phase to classify the entire object-based river network. Finally, an independent validation was performed taking into consideration the polygon area within the accuracy assessment, hence improving the efficiency of the classification accuracy of the GEOBIA-derived map, both globally and by geographical zone. As a result, we automatically processed almost 2 million square kilometers at a spatial resolution of 2.5 meters, producing a riverscape-units map with a global overall accuracy of 0.915, and with per-class F1 accuracies in the range 0.79–0.97. The obtained results may allow for future studies aimed at quantitative, objective and continuous monitoring of river evolutions and fluvial geomorphological processes at the scale of Europe.

High-resolution mapping of ash (Fraxinus spp.) in bottomland hardwoods to slow Emerald Ash Borer infestation

Ash (Fraxinus spp.) species are under serious threat by an invasive insect named the Emerald Ash Borer (EAB), already spreading across 35 states in the USA. To track the dramatic progression for effective implementation of ash conservation plans, information on the spatial distribution of host resources is essential. This study aimed at delineating ash trees at the crown level in a mature bottomland hardwood forest under a remote sensing mapping framework. We employed five-band Worldview-3 imagery at 31 ​cm spatial resolution acquired in the spring and conducted the mapping activity using geographic object-based image analysis (GEOBIA). Because the training samples and features play important roles in classification, we tested the effects of training sample size and the textural features on the mapping performance. A set of segments that were optimized through a series of test-and-trials in the multiresolution segmentation were used as the basic mapping unit. By testing against field samples, the largest average overall accuracy 82%, was achieved through the random forest classifier. Our findings suggest that the classification accuracy is improved with increased training sample size. The combination of spectral and textural features improved the classification accuracy significantly (p-value <0.05) for two large sample sizes (i.e. n ​= ​200 and n ​= ​230). The GEOBIA framework we demonstrated may guide the future ash tree mapping endeavors regarding the selection of segmentation parameters, sample sizes, and classification features.