Graphical Raster Based Data Editors

Andrew K Lui
CSIRO Division of Information Technology
Computer Science Laboratory, RSISE, ANU,
GPO Box 664, Canberra, ACT 2601, Australia.

Don Bone
CSIRO Division of Information Technology,
GPO Box 664, Canberra, ACT 2601, Australia.


Raster data are used in many domains as the common data format because the processing algorithms are often highly scalable and efficient. Many automated processes have been developed and employed to enhance, alter and transform raster data. The performance of these processes is expected to be greatly improved by user interaction with the raster data. This user interaction can be realised in graphical raster based data-editors. However, since raster data sets are large, direct interaction with individual data is not feasible. Compounding the problem is the lack of a suitable abstraction to simplify the implementation of graphical raster based data-editors. This paper presents Grasedit, a generic object oriented architecture for constructing raster based data-editors. For Grasedit, efficient interaction is achieved through interacting with the parameter space of a suitably selected process which models raster data for a specific domain. Through specifications of various abstractions and an application framework, domain specific editors can be constructed with significantly less effort. Two experimental application editors built on our prototype implementation of the Grasedit architecture will be discussed to illustrate the feasibility of the architecture and to gain experience used in subsequent improvement.
Conference Home Page