Advances in Geographic Object-Based Image Analysis with ontologies: A review of main contributions and limitations from a remote sensing perspective

Arvor, Damien; Durieux, Laurent; Andrés, Samuel; Laporte, Marie‐Angélique

doi:10.1016/j.isprsjprs.2013.05.003

Cited by 164 publications

(163 citation statements)

References 73 publications

Supporting

Mentioning

157

Contrasting

Unclassified

Order By: Relevance

“…Image understanding is a complex cognitive process for which we may currently lack the key concepts. Therefore, in order to ensure a strong foundation, research in this field needs to focus on the deeper understanding of the relationship between image objects and landscape objects, rather than exclusively on the development of new techniques that are tailored for specific applications [28] and rarely transferable [3].…”

Section: Classification Results and Accuracy Assessmentmentioning

confidence: 99%

“…One of the approaches that managed to overcome the aforementioned challenges was the Geographic Object-Based Image Analysis (GEOBIA). GEOBIA systems gained widespread popularity from 2000 onwards and are currently considered to be the state-of-the-art systems in both scientific and commercial thematic mapping of very high-resolution spaceborne imagery, e.g., [1][2][3].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Application of In-Segment Multiple Sampling in Object-Based Classification

2014

View full text Add to dashboard Cite

Abstract:When object-based analysis is applied to very high-resolution imagery, pixels within the segments reveal large spectral inhomogeneity; their distribution can be considered complex rather than normal. When normality is violated, the classification methods that rely on the assumption of normally distributed data are not as successful or accurate. It is hard to detect normality violations in small samples. The segmentation process produces segments that vary highly in size; samples can be very big or very small. This paper investigates whether the complexity within the segment can be addressed using multiple random sampling of segment pixels and multiple calculations of similarity measures. In order to analyze the effect sampling has on classification results, statistics and probability value equations of non-parametric two-sample Kolmogorov-Smirnov test and parametric Student's t-test are selected as similarity measures in the classification process. The performance of both classifiers was assessed on a WorldView-2 image for four land cover classes (roads, buildings, grass and trees) and compared to two commonly used object-based classifiers-k-Nearest Neighbor (k-NN) and Support Vector Machine (SVM). Both proposed classifiers showed a slight improvement in the overall classification accuracies and produced more accurate classification maps when compared to the ground truth image.

show abstract

Section: Classification Results and Accuracy Assessmentmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Application of In-Segment Multiple Sampling in Object-Based Classification

2014

View full text Add to dashboard Cite

show abstract

“…Rule sets are rarely fully transferable (Arvor et al 2013) thus the authors decided to concentrate more on general presentation of GEOBIA workflow giving the sequence of performed processes and list of used features, than on exact threshold values for features, which can differ in other area. Besides automatic rule-based classification, two additional steps was added -manual editing and final reshaping and smoothing of objects.…”

Section: Classification Workflowmentioning

confidence: 99%

“…A common issue of recent GEOBIA research is the topic of ontologies. identification of best features for classification of selected lUlc classes using defined set of geodata is a very important task in this context (Arvor et al 2013).…”

mentioning

confidence: 99%

Using Geobia and Data Fusion Approach for Land use and Land Cover Mapping

Wężyk

Hawryło

Szostak

et al. 2016

Quaestiones Geographicae

View full text Add to dashboard Cite

WĘŻYK P., HAWRYŁO P., SZOSTAK M., PIERZCHALSKI M., DE KOK R., 2016. Using geobia and data fusion approach for land use and land cover mapping. Quaestiones Geographicae 35(1), Bogucki Wydawnictwo naukowe, Poznań, pp. 93-104, 4 fi gs, 5 tables.ABSTRACT: Land Use and Land Cover (LULC) maps play an important role in an environmental modelling, and for many years efforts have been made to improve and streamline the expensive mapping process. The aim of the study was to create LULC maps of three selected water catchment areas in South Poland using a Geographic Object-Based Image Analysis (GEOBIA) in order to highlight the advantages of this innovative, semi-automatic method of image analysis. the classifi cation workfl ow included: multi-stage and multi-scale analyses based on a data fusion approach. Input data consisted mainly of BlackBridge (RapidEye) high resolution satellite imagery, although for distinguishing particular LULC classes, additional satellite images (LANDSAT TM5) and GIS-vector data were used. Accuracy assessment of GeoBia classifi cation results varied from 0.83 to 0.87 (kappa), depending on the specifi c catchment area. The main recognized advantages of GEOBIA in the case study were: performing of multi-stage and multi-scale image classifi cation using different features for specifi c lUlc classes and the ability to using knowledge-based classifi cation in conjunction with the data fusion approach in an effi cient and reliable manner.

show abstract

“…These objectives inform the selection of the unit of 504 analysis and comparison method directly, but are also used in the selection of suitable imagery and the 505 identification of classification features required to satisfy the thematic objectives. Comparison features are 506 typically identified by expert knowledge and understanding of the anticipated change, which may develop into 507 full ontological descriptions as explored by Arvor et al (2013). Although not a scientific consideration, costs 508 will inevitably constrain most change detection projects to some degree.…”

mentioning

confidence: 99%