Spatial Database Systems and Management Multidimensional Discrete Data

Spacial database systems offer the underlying database technology for geographic information systems and other applications. Several terms have been used to describe database systems offering such support, including pictorial image, geometric, geographic, and spatial. The terms pictorial database system and image arise from the fact that the data to be managed are often initially captured in the form of digital raster images, remote sensing by satellites, or compuer tomography in medical applications. Spatial database management involves two main categories of data: vector and raster data. The former has received a lot of in-depth investigation; the latter still lacks a sound frmaework. Current DBMSs either regard raster data as pure byte sequence where the DBMS has no knowledge about the underlying semantics, or they do not complement array structures with storage mechanisms suitable for huge arrays, or they are designed as specialized systems with sophisticated imaging functionality, but no general database capabilities. We will discuss some of the aspects of spatial data, spatial databae and it's management. In various fields, there is a need to manage geometric, geographic, or spatial data. The space of interest can be, for example, the 2-D abstraction of the earth's surface, or the images of human body including computed tomography (CT), magnetic resonance (MR), ultrasonography(US), projectional computed radiography (CR) etc. These medical imaging systems have revolutionized the means by which images are acquired, providing views of anatomical cross-sections and physiological state. This revolution in the acquisition of radiological information has not yet brought about a parallel revolution in the intelligent management, visualization, integration, or knowledge extraction from data produced by these digital imaging system. In the discipline of visualization,where the areas of computer graphics, image processing, computer vision, computer-aided design, signal processing, and user interface studies converge into one unifying framework for the processing of visual information, several representation of a scene are distinguished. Kromker (1991) proposes a visualization reference model that is particularly suitable for database investigations because classification is done along the data structure on hand. Three of the six layers introduced in this reference model are relevant for DBMSs that deal with visualization structure: 1. The Symbolic Representation Layer deals with abstract scene descriptions, but without an explicit description of geometry and properties of the entities modeled. 2. The Geometry/Feature Layer covers geometric descriptions, appearance properties, and viewing parameters. Vector graphics would be a subset of such data structure. 3. On the Digital Pixel Layer, a scene is discretized in both space and color, yielding a raster image. A raster image consists of a finite set of points in the discrete coordinate space Z(d) where each point has some value, its color, associated. There is no algorithm that performs reasonably well on any kind of image and under all corcumstances; above all, images frequently contain information that cannot be cast into points, lines, and regions bounded by lines, because the boundary cannot be recognized without doubt (e.g., tumors in medical imagery), or because there is no clear boundary (e.g., density distributions such as clouds in weather satellite images). In summary, both vector and raster representation are important for spatial data management, because each of them has pacific strengths and weaknesses; moreover, both representations are independent from each other in the sense that there is no lossless transformation between them.