Adaptive spatial data processing system (ASDPS)

To reveal local self-relevance in spatial datasets, using spatial-domain processing is the better approach. However, in spatial-domain processing, each processing algorithm has to be implemented and complied as a separate program. A collection of programs implementing different processing algorithms can be organized as a software package. In using this package for multiple processing, the sub-programs or methods selected have to be executed in sequential order. To put a new method or subroutine to the package, the whole package must be compiled again. This paper proposes a new spatial-domain processing system that not only can be implemented as a unique user interface for both selecting any implemented methods and defining new methods without recompilation, but also supports parallel processing in a multithreaded computing environment. The new spatial-domain processing system (named as ASDPS) based on adoption of an array data pre-processor and a neuron-like controller supports parallel processing in a multithreaded computing environment. It also leaves the implementation open for a unique user interface to be designed in a way so that not only implemented methods can be easily chosen, but also any new method can be easily adopted by assigning new weights to the process without recompilation. The use of this new system in magnetic data processing in this paper demonstrates that this new system is very useful in spatial data processing and interpretation.