Object-based Land Cover Mapping Research Articles

Assessment of ensemble learning for object-based land cover mapping using multi-temporal Sentinel-1/2 images

Accurate land cover mapping is challenging in Southeast Asia where cloud coverage is prevalent and landscape is heterogenous. Object-based mapping, multi-temporal images and combined use of optical and microwave data, provide abundant features in spectral, spatial, temporal, geometric and polarimetric dimensions. And random forest is usually employed due to robustness and efficiency in handling high-dimensional and noisy data. This study assesses whether feature selection and ensemble analysis, which are rarely adopted, yield improved result. Recursive feature elimination decreases original 568 features into a subset of 53 features, achieving the optimal combination of features. Ensemble analysis of random forest, support vector machine, and K-nearest neighbors leads to an overall accuracy of 0.816. Comparison experiments demonstrated the merits of the multi-temporal, multi-source approach, feature elimination and ensemble analysis.

Pixel and object-based land cover mapping and change detection from 1986 to 2020 for Hungary using histogram-based gradient boosting classification tree classifier

The large-scale pixel-based land use/land cover classification is a challenging task, which depends on many circumstances. This study aims to create LULC maps with the nomenclature of Coordination of Information on the Environment (CORINE) Land Cover (CLC) for years when the CLC databases are not available. Furthermore, testing the predicted maps for land use changes in the last 30 years in Hungary. Histogram-based gradient boosting classification tree (HGBCT) classifier was tested at classification. According to the results, the classifier, with the use of texture variance and landscape metrics is capable to generate accurate predicted maps, and the comparison of the predicted maps provides a detailed image of the land use changes.

National Scale Land Cover Classification for Ecosystem Services Mapping and Assessment, Using Multitemporal Copernicus EO Data and Google Earth Engine

Land-Use/Land-Cover (LULC) products are a common source of information and a key input for spatially explicit models of ecosystem service (ES) supply and demand. Global, continental, and regional, readily available, and free land-cover products generated through Earth Observation (EO) data, can be potentially used as relevant to ES mapping and assessment processes from regional to national scales. However, several limitations exist in these products, highlighting the need for timely land-cover extraction on demand, that could replace or complement existing products. This study focuses on the development of a classification workflow for fine-scale, object-based land cover mapping, employed on terrestrial ES mapping, within the Greek terrestrial territory. The processing was implemented in the Google Earth Engine cloud computing environment using 10 m spatial resolution Sentinel-1 and Sentinel-2 data. Furthermore, the relevance of different training data extraction strategies and temporal EO information for increasing the classification accuracy was also evaluated. The different classification schemes demonstrated differences in overall accuracy ranging from 0.88% to 4.94% with the most accurate classification scheme being the manual sampling/monthly feature classification achieving a 79.55% overall accuracy. The classification results suggest that existing LULC data must be cautiously considered for automated extraction of training samples, in the case of new supervised land cover classifications aiming also to discern complex vegetation classes. The code used in this study is available on GitHub and runs on the Google Earth Engine web platform.

Multiscale Integration Approach for Land Cover Classification Based on Minimal Entropy of Posterior Probability

Object-based land cover mapping has drawn increasing attention for its ability to overcome the salt-and-pepper problem associated with pixel-based methods by considering spatial information from neighboring regions. However, the performance of object-based classification is strongly affected by over- or undersegmented objects. The optimal scale is difficult to determine; moreover, it usually varies along with the application purpose or classification targets. Most previous efforts on scale determination based only on image information are not flexible in adapting to different classification systems; consequently, their use is not advisable. In this paper, to better consider classification targets, the information from training samples for classification is also used for determining the optimal scale based on the concept of minimal entropy of posterior probability (MEPP). The proposed MEPP method consists mainly of two stages: 1) training samples from the original pixel level are applied to classify segmented images and obtain posterior probability maps on multiple scales; and 2) the optimal object scale is determined according to the MEPP that corresponds to the minimum classification uncertainty. Experiments on high-spatial-resolution images and Landsat images confirm the superiority of the proposed MEPP method in land cover classification.

Evaluation of Feature Selection Methods for Object-Based Land Cover Mapping of Unmanned Aerial Vehicle Imagery Using Random Forest and Support Vector Machine Classifiers

The increased feature space available in object-based classification environments (e.g., extended spectral feature sets per object, shape properties, or textural features) has a high potential of improving classifications. However, the availability of a large number of derived features per segmented object can also lead to a time-consuming and subjective process of optimizing the feature subset. The objectives of this study are to evaluate the effect of the advanced feature selection methods of popular supervised classifiers (Support Vector Machines (SVM) and Random Forest (RF)) for the example of object-based mapping of an agricultural area using Unmanned Aerial Vehicle (UAV) imagery, in order to optimize their usage for object-based agriculture pattern recognition tasks. In this study, several advanced feature selection methods were divided into both types of classifiers (SVM and RF) to conduct further evaluations using five feature-importance-evaluation methods and three feature-subset-evaluation methods. A visualization method was used to measure the change pattern of mean classification accuracy with the increase of features used, and a two-tailed t-test was used to determine the difference between two population means for both repeated ten classification accuracies. This study mainly contribute to the uncertainty analysis of feature selection for object-based classification instead of the per-pixel method. The results highlight that the RF classifier is relatively insensitive to the number of input features, even for a small training set size, whereby a negative impact of feature set size on the classification accuracy of the SVM classifier was observed. Overall, the SVM Recursive Feature Elimination (SVM-RFE) seems to be an appropriate method for both groups of classifiers, while the Correlation-based Feature Selection (CFS) is the best feature-subset-evaluation method. Most importantly, this study verified that feature selection for both classifiers is crucial for the evolving field of Object-based Image Analysis (OBIA): It is highly advisable for feature selection to be performed before object-based classification, even though an adverse impact could sometimes be observed from the wrapper methods.

Learning selfhood scales for urban land cover mapping with very-high-resolution satellite images

Abstract Urban land cover mapping with very-high-resolution (VHR) satellite images has raised many concerns in the fields of environmental and social investigations, but the classical per-pixel and object-based mapping results are not accurate enough to serve these applications, owing to the inappropriate analysis scales. Accordingly, this study aims to 1) propose a self-adaptive segmentation scale (i.e., “selfhood scale”), which refers to the optimum scale for analyzing a pixel and depends on the pixel's category and surrounding contrasts; and 2) apply selfhood scales to urban land cover mapping. For the first target, a learning mechanism is presented to estimate selfhood scales; for the second target, two methods (i.e., “zipper merging” and “restricted forest”) using the learned selfhood scales are proposed to respectively improve per-pixel and object-based classification results. The experimental results demonstrate that these methods achieve significant improvements in both per-pixel and object-based land cover mapping in urban areas as their overall accuracies are improved by 2.8% and 7.6% respectively. Moreover, the proposed methods are further used to generate land cover maps in Beijing and Zhuhai cities, which perform much better than the classical methods. Accordingly, it can be concluded that selfhood scales are effective to improve land cover mapping in urban areas.

Object-Based Land-Cover Mapping with High Resolution Aerial Photography at a County Scale in Midwestern USA

There are growing demands for detailed and accurate land cover maps in land system research and planning. Macro-scale land cover maps normally cannot satisfy the studies that require detailed land cover maps at micro scales. In the meantime, applying conventional pixel-based classification methods in classifying high-resolution aerial imagery is ineffective to develop high accuracy land-cover maps, especially in spectrally heterogeneous and complicated urban areas. Here we present an object-based approach that identifies land-cover types from 1-meter resolution aerial orthophotography and a 5-foot DEM. Our study area is Tippecanoe County in the State of Indiana, USA, which covers about a 1300 km2 land area. We used a countywide aerial photo mosaic and normalized digital elevation model as input datasets in this study. We utilized simple algorithms to minimize computation time while maintaining relatively high accuracy in land cover mapping at a county scale. The aerial photograph was pre-processed using principal component transformation to reduce its spectral dimensionality. Vegetation and non-vegetation were separated via masks determined by the Normalized Difference Vegetation Index. A combination of segmentation algorithms with lower calculation intensity was used to generate image objects that fulfill the characteristics selection requirements. A hierarchical image object network was formed based on the segmentation results and used to assist the image object delineation at different spatial scales. Finally, expert knowledge regarding spectral, contextual, and geometrical aspects was employed in image object identification. The resultant land cover map developed with this object-based image analysis has more information classes and higher accuracy than that derived with pixel-based classification methods.

Object-based land cover mapping and comprehensive feature calculation for an automated derivation of urban structure types at block level

Cities have evolved under manifold geographical, economical, historical, and cultural criteria, resulting in various sizes and shapes. Each city exhibits individual features and unique characteristics, despite that structural similarities appear. The separation into individual patterns, commonly named urban structure types (USTs), supports the characterization of physical, functional, and energetic factors of settlement structures, enabling associated environmental and socio-economic investigations as well as the comparison between the patterns of different cities. This study presents an automated approach for the classification of USTs based on remote sensing data in order to analyze the links between settlement structures and environmental issues, such as air pollution or urban heat islands, in a later stage of the project. Initially, an object-based classification routine is implemented to identify the land cover for the city of Berlin, utilizing spatially very high resolution aerial images and object height information. UST classes are defined based on the occurrence within the study area and are delimited by block boundaries. Afterwards, indicators for the derivation of USTs are generated based on the previously derived land cover information and the most valuable features are selected with the help of Random Forests. Finally, structural units are classified, involving common and new land cover based parameters. The focus is on the generation of an automated and transferable routine for a comprehensive UST classification covering the entire city. Comparing the results with reference data, good classification accuracies for both land cover and USTs indicate the suitability of the proposed method.