A SEMANTIC RETRIEVAL SYSTEM IN REMOTE SENSING WEB PLATFORMS

Abstract. This paper proposes a solution to reduce the semantic gap between final users and data/processing providers in a web market place dedicated to remote sensing products. Nowadays, search engine are common tools on the Internet. Users are accustomed to use them and used to get tabular classification of provided answers. These smart agents are set up to answer basic questions using automatic pages redirection or chitchat. In this research, to ensure coherence between user’s requests and platform answers, natural language processing algorithms and knowledge graphs are integrated within a web platform thanks to a NoSQL graph database connected to open thesauri and Geographic Information Systems (GIS). Therefore, the most pertinent services can be proposed based on input sentences including non-technical vocabulary but also geographical components (the user interface includes a text area and an interactive map). While processing chains and remote sensing ontologies were presented in one of our previous studies, this article focuses on natural languages algorithms and knowledge mining.