An Object-Based Semantic Classification Method for High Resolution Remote Sensing Imagery Using Ontology

Introduction to the GEOBIA 2010 special issue: From pixels to geographic objects in remote sensing image analysis

This chapter explores the need for automation processes in information extraction from remote sensing images. This exploration is further extended by developing an ontological framework for geographic object-based image analysis (GEOBIA). GEOBIA provides a method to identify real-world geographic objects from remote sensing imagery. It uses the same technique as a human does to perceive different geo-objects and to distinguish them. Thus, human involvement in the form of expert knowledge will be required during image object identification. This need for human intervention is addressed by employing a knowledge representation language such as ontology to formalize human expert knowledge. After formalization, necessary rule sets for image interpretation are developed. Using rule-based classification, no human assistance is required at the time of image object identification, which transforms GEOBIA into an automated process. The proposed framework is applied to an urban case study on land use/land cover (LULC) from Hobart, Tasmania, Australia.

AN ASSESSMENT OF SPATIAL VARIATION OF LAND SURFACE CHARACTERISTICS OF MINNA, NIGER STATE NIGERIA FOR SUSTAINABLE URBANIZATION USING GEOSPATIAL TECHNIQUES

Geographic object-based image analysis (GEOBIA) is commonly applied for land-cover and land-use mapping, updating and change-identification analyses. Following image segmentation, conventional GEOBIA routines classify image objects based on parametric statistical measures, assuming that within-object pixels have normally distributed image brightness signatures. The context for this study is updating extant land-use GIS layers that are out of date as a result of urban expansion. The objective is to develop, test and compare GEOBIA techniques based on a histogram classifier and on a nearest-neighbour classifier, for updating land-use layers. Frequency distribution signatures of land-use change and no-change objects are evaluated for different feature inputs and classifiers within an urbanizing area in San Diego County, California, USA. The results demonstrate that a histogram classifier consistently outperforms a conventional nearest-neighbour classifier. A Histogram Matching Root Sum Squared Differential Area classifier combined with temporal-spectral difference inputs and arithmetic mean for combining multi-feature classifier metrics yielded the greatest accuracy: 79.82% overall accuracy, with 78.72% and 81.07% for change and no-change objects respectively.

ABSTRACTScale computation for multi-scale image segmentation is an important research area in geographic object-based image analysis (GEOBIA). For the highly spectral heterogeneity of high spatial resolution remotely sensed imagery (HSRRSI), and the changing sizes of geographic features and their spatially distributive patterns, it is difficult to build a global or local-scale calculation parameter model to effectively guide multi-scale segmentation parameters setting in large-scale regions. Usually, the segmentation parameters are used to measure the heterogeneous and homogenous adjacent pixels in spatial and spectral spaces simultaneously. It has been proved that the adaptive acquiring parameter of scale plays a key role in gaining precise segmentation results, and later it deeply influences the automatic recognition and post-processing of the physical image parcels (PIPs). However, in most cases, scale computation techniques still fail to guide segmentation to produce appropriate or repeatable results which should meet the practical production standard of GIS data based on GEOBIA. These techniques have not been summarized and classified and there is no review focusing on scale computation for HSRRSI multi-scale segmentation. We provide an overview of the state-of-the-art segmentation scale computation techniques which are mainly based on the spectral statistics and geometric characteristics, etc. Moreover, the pedigree of segmentation scale has been first time proposed, and the overall performance of each category is analysed. Especially, the methods of local variance, semivariance, and synthetic semivariance are presented. Then, the scale object selection (SOS) algorithm, spectral angle algorithm, and the RMAS (ratio of mean difference to neighbours (ABS) to standard deviation) are discussed at spectral domain. In addition, miscellaneous scale computation approaches are recognized as the important researching aspect. In order to clearly describe the scale computation on multi-scale image segmentation, we have proposed the new conceptions of semantic image object (SIO), PIP, particular scale of interest, symbiotic scale, etc. At last, the trends of scale computation for HSRRSI multi-scale segmentation also have been presented.

Spatial modeling methods usually use pixels and image objects as fundamental processing units to address real‐world objects, geo‐objects, in image space. To do this, both pixel‐based and object‐based approaches typically employ a linear two‐staged workflow of segmentation and classification. Pixel‐based methods segment a classified image to address geo‐objects in image space. In contrast, object‐based approaches classify a segmented image to identify geo‐objects from raster datasets. These methods lack the ability to simultaneously integrate the geometry and theme of geo‐objects in image space. This article explores Geographical Vector Agents (GVAs) as an automated and intelligent processing unit to directly address real‐world objects in the process of remote sensing image classification. The GVA is a distinct type of geographic automata characterized by elastic geometry, dynamic internal structure, neighborhoods and their respective rules. We test this concept by modeling a set of objects on a subset IKONOS image and LiDAR DSM datasets without the setting parameters (e.g. scale, shape information), usually applied in conventional Geographic Object‐Based Image Analysis (GEOBIA) approaches. The results show that the GVA approach achieves more than 3.5% improvement for correctness, 2% improvement for quality, although no significant improvement for completeness to GEOBIA, thus demonstrating the competitive performance of GVAs classification.

Reliable and accurate land use/land cover (LULC) map is a crucial data source for the understanding of coupled human–environment systems, monitoring changes, timely low-cost planning, and management of natural resources. Improvements in sensor technologies and machine learning capabilities have shifted the attention of remote sensing community to data complementarity through fusion of multi-sensor data for accurate feature extraction and mapping. Amalgamation of optical and synthetic aperture radar (SAR) images has shown promising advantages in enhancing the accuracy of extracting LULC as such method allows exploitation of information in sensors. This study investigated the potential of using freely available multisource Sentinel images to extract LULC maps in semi-arid environments through multi-scale geographic object-based image analysis (GEOBIA). A multi-scale classification framework that integrates GEOBIA, correlation-based feature selection (CFS), and random forest (RF)-supervised classification was adopted to extract LULC from assimilation of Sentinel multi-sensor products. First, Sentinel-1 and -2 images were pre-processed. Second, optimum multi-scale segmentation levels were selected using F-score segmentation quality measures. Third, 70 features of various spectral indices and derivatives and geometrical features from optical data and multiple ratios and textural features from dual-polarization SAR images were computed, and a CFS based on wrapper approach was used to select the most significant features at multi-scale levels. Finally, a single and multi-scale RF classifier was used to extract LULC classes using the most relevant features extracted from Sentinel SAR and optical images. Results of multi-scale image segmentation optimization showed that scale parameter (SP) values of 40, 60, and 150 were optimal for extraction of LULC classes. Results of feature selection showed that 22, 24, and 27 features were selected at scale SP values of 40, 60, and 150, respectively. Half of the features were common among the three scales. Single RF classification yielded overall accuracy (OA) values of 92.10%, 93%, and 91% and kappa coefficients of 0.901, 0.912, and 0.89 at scale values of 150, 60, and 40, respectively. Multiscale RF classification from scale values of 150 and 60 produced better LULC classification with OA 96.06% and kappa coefficient of 0.95 compared with other scale SP values. The integrated approach demonstrated an effective and promising method for high-quality LULC extraction from coupling optical and SAR images. Overall, multi-sensor Sentinel images along with the adopted approach feature a remarkable potential for improving LULC extraction and can effectively be used to update geographic information system layers for various applications.

Image segmentation is a traditional method in Remote Sensing and a fundamental problem in image processing applications. It has been widely used, especially with the emergence of the Geographic Object-Based Image Analysis (GEOBIA). The results of segmentation must create uniform areas, which must allow a simpler interpretation by the users and simpler representation for classification algorithms. Several algorithms were proposed through the years, using different approaches. One that is widely used in Remote Sensing applications is the Multiresolution algorithm, that is based on the region growing method. Other, which has great potential and is applied in other research areas, is available on the Image Foresting Transform (IFT) framework, which has several image operators developed primarily for medical images. The Watershed from Grayscale Marker operator uses an edge image to perform the segmentation, however, we propose an extension of the edge detection algorithm, by summing normalized gradients of each band. This work aims to evaluate and compare these two segmentation algorithms, by comparing their results through supervised egmentation from reference regions, that were defined manually by an expert user. Quality measures were evaluated by four metrics, that represent the positional adjustment based the center of gravity, intensities, size, and the amount of overlap between the segment created by the algorithms and the reference segment. We selected 21 objects of a WorldView-2 multispectral image that were used to compute the metrics. Both methods reached similar results, by comparing the aforementioned 4 metrics applied to the 21 reference regions, IFT achieved better results for majority of regions. The IFT generated segments with similar shape when compared with the references, and the multiresolution generated results with similar sizes and positional adjustments. It may be explained by the fact that IFT uses an edge image to perform the segmentation. Both algorithms obtained similar agreement for intensity.

A key issue in modern fire management planning is the accurate fuel type mapping, required at many different spatial and temporal scales. Fuel type classification is critical for improving fire prevention schemes, developing accurate fire dispersion models and designing effective measures for mitigating the impacts of a potential wildfire event on the ecosystem. Remote sensing offers the potential to provide spatially distributed information on fuel types and it has been widely used both at regional and local scale. Theaim of this study is to develop and evaluate a fuel type mapping methodology on a national level, based on the geographic objectoriented classification approach and Landsat-8 OLI imagery. The proposed methodology was developed in the case study of Chalkidiki in the northern part of Greece and it was further tested for its transferability in the regional unit of Preveza and the whole administrative region of Attica. The classification scheme was determined taking under consideration the existing fuel models(Prometheus, FBP etc) and fuel type products (JRC FUELMAP, LIFE10 ArcFUEL) so as the final fuel type map could be easily adjusted and/or compared. Fifty-four Landsat-8 OLI images for summer and for winter season, covering the national territory, were acquired and several features (vegetation indices, textural and spectral features) were calculated for both seasons. The optimized feature selection for the discrimination of each fuel type category was empirically obtained. Results showcase the effectiveness of theemployed object-oriented classification approach in obtaining highly accurate fuel type maps. Specifically, the generated fuel map of Chalkidiki exhibited an overall accuracy of 89.47%, with a Kappa Index of Agreement (KIA) equal to 0.844. The application of the model to the two other regions resulted in overall accuracies of 80.30% (KIA=0.706) for Attica and 91.74% (KIA=0.867) for Preveza. Summarizing, the results proved the good transferability properties of the proposed methodology enabling the implementation of the model across the country.

Training set size, scale, and features in Geographic Object-Based Image Analysis of very high resolution unmanned aerial vehicle imagery

Image analysis and feature extraction of remoted sensing data are significant for mapping irrigated agriculture areas as a source of information to improve water management and agricultural planning. This paper presents an image segmented base approach GEOBIA (Geographic Object-Based Image Analysis) to extract irrigated areas by Center Pivot Irrigation System (CPIS). This study suggests a semi-automated recognition of circular patterns for the mapping of irrigated regions by center pivots, using Sentinel -2 MSI images, 10 meters spatial resolution. A set of images from different seasons, humid and dry are used to maximize de CPIS’s occurrence. A multiresolution segmentation method was applied, and a large number of segment-based shape features was extracted and used as input to a feature selection procedure (shape descriptors: Area; Compactness; Circularity Factor; Length/Width; Radius of smallest enclosing ellipse; and Roundness). In addition, another shape descriptor “Circularity Factor” was developed in this research and played an important role during preliminaries classification processes. The accuracy assessment of preliminaries classifications has validated used the Circularity Factor together with the other chosen shape descriptors to reach better results to CPIS’s detection. Furthermore, 86.23% of the CPIS mapped in the classification process is in accordance with the ground truth map. This methodology can be used to map large areas in a relatively short time and provides a tool for monitoring irrigated areas.

IKONOS 1-m panchromatic and 4-m multispectral images were used to map mangroves in a study site located at Punta Galeta on the Caribbean coast of Panama. We hypothesized that spectral separability among mangrove species would be enhanced by taking the object as the basic spatial unit as opposed to the pixel. Three different classification methods were investigated: maximum likelihood classification (MLC) at the pixel level, nearest neighbour (NN) classification at the object level, and a hybrid classification that integrates the pixel and object-based methods (MLCNN). Specifically for object segmentation, which is the key step in object-based classification, we developed a new method to choose the optimal scale parameter with the aid of Bhattacharya Distance (BD), a well-known index of class separability in traditional pixel-based classification. A comparison of BD values at the pixel level and a series of larger scales not only supported our initial hypothesis, but also helped us to determine an optimal scale at which the segmented objects have the potential to achieve the best classification accuracy. Among the three classification methods, MLCNN achieved the best average accuracy of 91.4%. The merits and restrictions of pixel-based and object-based classification methods are discussed.

Previous learning-based methods that rely on 2D laser data to classify indoor robot locations into semantic classes were successful in distinguishing between rooms and corridors. However, the classification accuracy remained low for doorway locations. We propose a semantic place classification method that uses a rule-based doorway detection algorithm followed by a classification scheme that models training data through either K-means clustering or learning vector quantization. We conducted extensive experiments on the Freiburg 79 dataset and compared our method to previous semantic place classification algorithms. The doorway detection algorithm we propose significantly increases the classification accuracy for doorway locations as compared to the state-of-the-art performance. We applied our method, trained on the Freiburg 79 dataset, to Freiburg 52 and ESOGU datasets in order to demonstrate its generalization ability.

Robots are expected to operate autonomously in dynamic environments. Understanding the underlying dynamic characteristics of objects is a key enabler for achieving this goal. In this paper, we propose a method for pointwise semantic classification of 3D LiDAR data into three classes: non-movable, movable and dynamic. We concentrate on understanding these specific semantics because they characterize important information required for an autonomous system. To learn the distinction between movable and non-movable points in the environment, we introduce an approach based on deep neural network and for detecting the dynamic points, we estimate pointwise motion. We propose a Bayes filter framework for combining the learned semantic cues with the motion cues to infer the required semantic classification. In extensive experiments, we compare our approach with other methods on a standard benchmark dataset and report competitive results in comparison to the existing state-of-the-art. Furthermore, we show an improvement in the classification of points by combining the semantic cues retrieved from the neural network with the motion cues.

This research presents a novel strategy for hyperspectral and LiDAR data fusion to generate accurate LC/LU maps with object based classification method. Unlike conventional object-based strategies (hierarchical and multilevel models), in the proposed method, classification has been performed in an iterative Segmentation-Classification-Merging (SCM) process. In each iteration, image objects are extracted by using their spectral, height, geometric and class-related characteristics based on data availability, class importance and higher extraction capability. Results indicate great overall accuracy of 97.33% and a kappa coefficient of 0.9710.