GIS data conversion – overview

GIS data conversion – overview

>  In GIS Data conversion has two commonly used interpretations:  
    - when used in the narrow sense, it refers to the process of
       converting existing information to a digital format for use
       in a given GIS software package
>  In its broader interpretation,  it refers to the often
    complex process of building databases for GISs.
    - Data conversion is usually a costly and a major expenditure
      item in setting up a GIS. This situation is brought about by
      an interaction of several factors including the large quantity
      of data, data that is frequently of indifferent condition/quality
      and the complexity of the conversion task.  The data conversion
      process involves: a plan (incl. manuscript collation, preparation,
      information filtering), conversion using appropriate method,
      edits/ transformations, format conversions (if necessary) and
      data structuring (making the graphic data  more ‘intelligent’
      by building topology)
    - The general  (and accelerating) trend has been towards collection
       of data in digital format so the main conversion problem is related
       to existing data – especially true in developing countries where the
       availability of data in digital form is still very limited.
       There are many conversion methods; all are befitting from the general
       improvement  of both hardware and software used in conversion.
       Raster scanning is now playing a major role – especially as
       the first step in ‘heads-up’ digitising; however, automated data
       conversion remains a dream. Automated data conversion is where
       a map or similar data source can be scanned and all its geographic
       features and properties recognised and extracted  by smart software
       and converted into a GIS database.  It is important to recognise that
       the data conversion process can introduce significant errors – only
       some of these can be detected automatically by software
       (especially so in the case of graphical data)
    - The converted graphic data in vector packages is stored simply
       as coordinates i.e. as ‘spaghetti’ data.  This is relatively ‘dumb’
       data i.e. it has no information on spatial relationships
       (such as adjacency, intersection and connectivity).
       As such any spatial query becomes a tedious process unless
       topological structures are created.  After topology has been built,
       the system is able to respond to queries by using
       topological information rather than by a tedious processing
       of coordinate (to generate geographic
       features every time a query is made)