An object-based semantic classification method of high resolution satellite imagery using ontology

Abstract

Geographic Object-Based Image Analysis (GEOBIA) techniques have become increasingly popular in recent years and are able to incorporate and develop ontology model within the classification process. They have been claimed to represent a paradigm shift in remote sensing interpretation. Nevertheless, it is lack of formal expression and objective modelling of the whole process of GEOBIA, and lack of the study of semantic classification method using ontology. A major reason is the complexity of the process of GEOBIA. The study has put forward an object-based semantic classification method of high resolution satellite imagery using ontology that aims to fully exploit the advantages of ontology to GEOBIA. A detailed workflow has been introduced that has three steps: ontology modelling, initial classification based on data-driven machine learning method, and semantic classification based on knowledgedriven expert rules method. The whole process of GEOBA is organized organically and expressed explicitly using ontology, and the semantic relations are expressed in the formal language that the computer could operate. Image objects are classified based on ontology model and using machine learning method and expert rules. From the result it is well understood that the method enhances the existing GEOBIA techniques with the help of the ontology, which expresses and organizes the whole process of GEOBIA, and establishes their relations, and provides semantic meaning for GEOBIA. In particular, we found that it provides an ontology model and method for further classifications and large scale applications, and the method using ontology is suitable for automatic classification.

Role of geospatial technology in understanding urban green space of Kalaburagi city for sustainable planning

“…Accuracy assessment is necessary to validate the results (Gu et al, 2016). It compares the classification result with ground data to evaluate how well the classification represents the real world.…”

“…Additionally, image-based methods can potentially enable mapping of larger areas using the increasing number of temporal databases provided by satellite imagery (Myeong et al, 2006). Object-Based Image Analysis (OBIA) has become increasingly popular in recent years (Blaschke 2010;Johansen et al, 2011;Moskal et al, 2011;Li et al, 2012;Alqurashi et al, 2016;Gu et al, 2016;Kaszta et al, 2016) and OBIA methods can generate good and repeatable Land Use Land Cover (LULC) classifications suitable for tree cover assessment in urban areas (Moskal et al, 2011). The efficacy of OBIA in forest resource mapping, monitoring and J o u r n a l P r e -p r o o f managing is discussed in Blaschke (2010).…”