Research on the Natural Language Querying for Remote Sensing Databases

Natural language querying for remote sensing databases means that the users describe the query object directly in natural language when they query remote sensing (RS) databases. This method is not limited by the options on the query interface, and allows non-professional users to access to the databases easily. After making a study on the status of natural language querying for databases, the paper sums up three methods used in natural query language processing, which are Pattern Matching, Dependency Extraction and Knowledge Extension. Then it proposes a new method for Natural Language Processing (NLP) based on the features of the RS query text, which establishes a prototype system that contains three main steps-keywords extraction, keywords extension and SQL generation. The system implements the conversion from natural query language to database manipulation language (e.g. SQL), which greatly facilitates the users' query.