The development of more efficient computer input and output devices adapted to cartographic applications has led to increasingly sophisticated operations being performed on larger and larger volumes of cartographic data. These operations, as well as the storage format of the data, tend to replicate traditional manual methods. However, the range of applications where this approach is cost-effective is rapidly narrowing as more and more data are initially captured in raster format, i.e., in scan lines similar to a television image, rather than feature by feature. In addition, computer storage and processing of raster format data is simpler and more efficient for many applications. Appropriate approaches for automated cartographic processing therefore need to be examined carefully. Many algorithms for processing cartographic data in raster mode already exist. However, before the efficiencies of raster-based processing can be fully realized these algorithms must be tested and extended. Also needed are man-machine translation techniques that will allow users of a raster-oriented cartographic system to interact with it via more familiar vector-oriented concepts.