A Memory Management Technique for Remote Sensing Data

In this research, a memory management model for remote sensing data is proposed. Our initial study focuses on the memory-mapped files technique as it permits applications to directly map the virtual address to a file located on disk. A variant of such technique provides a quick and flexible way to read and write multi-spectral image files. A user application will be able to load only currently used pages to a physical memory. When an application needs data outside the range, the process will re-map the address using swapping and caching techniques. One of the main goals is to create a utility for JAVA-based remote sensing software. The research output includes an implementation of a library prototype on the 2.6 Linux kernel. The prototype provides application developers with an application programming interface (API), which should benefit various imagery researches and applications