Exploring the application of some common raster scanning paths on lossless compression of elevation images

Raster scanning is a method for producing or recording a video or a digital image by means of a line-by-line sweep. The scanning process basically transforms a 2D image representation into a 1D representation. While such process has been extensively used for data transmission and storage as well as in most image displaying and capturing devices, its application to Digital Elevation Models (DEM) research is infrequent. A DEM is usually a square matrix of elevations (Elevation Image) except that it comprises a single channel of information (heights) and can be compressed in a lossy or lossless manner by way of current standard image compression protocols. Compression has the outcome of reducing memory requirements and speed of transmission over digital links, while preserving the integrity of data. In this context, this proof of concept paper investigates the effects of applying some different scanning paths to 16-bit elevation images of floating point height values representing dissimilar terrain characteristics. The intention is to discern whether different 1D representations of a given elevation image can provide a better correlation or similarity among elevation points, thus resulting in improved compression ratios and run decompression times.