A fast warping algorithm for correcting local distortions in binary images

A geographical information system (GIS) is a digital information system whose records are spatially referenced data. When building GIS systems from maps, it is very important that the objects represented in these maps are accurately located. Corrupted maps are frequent; in order to correct distortions introduced in the maps, we have designed a fast image warping algorithm that can handle huge maps within a short processing time. Our algorithm takes advantage of the binary nature of our maps and also the low density of pixels representing the information. In addition, we deal with local distortions by dividing the whole image into triangles, which enables us to apply independent warping transformation for each triangle.