A Prototype System for Value-Range Queries

We have developed algorithms for valuerange queries of Earth science data with an area size constraint. The value-range condition is approximately satisfied by a histogram clustering technique with a data pyramid model. The result is the high resolution areas (points) on which the value range condition is satisfied. The area size condition is achieved by combining convex hull and point-inpolygon algorithms and depth-first strategy to reconstruct areas. Those areas are constructed with areas selected by the value-range queries. Using a data pyramid concept, the above two algorithms can be considered as drill-down searching and bottom-up reconstruction procedures. This paper describes a prototype system based on above algorithms. The web-based prototype system shows the procedures as the queries being answered. On the server side, we are using Splus software with consideration of programs in general languages such as C and FORTRAN and DBMS for efficiency. On the client side, SVG (Scalable Vector Graphics) and Java techniques are merged to have a clear and userfriendly interface.