Elevation contours extraction from a color-coded relief scanned map

The focus of this paper is in extracting the elevation contours from a color-coded relief map. The map is preprocessed using a WHMM (wavelet domain hidden Markov model) based algorithm for denoising. Then a linework removal algorithm - that uses a vector angle edge detector with a saturation-based combination of hue and intensity planes - is applied. Each distinct elevation region must be solid colored using distinct colors. In order to achieve this, for the printed scanned map, we used a vector median filtering, and a color clustering process - based on a minimum variance quantization algorithm. The gap filling problem can be solved with a multi-pass majority filter algorithm. The elevation contours are extracted by detecting the borders between colors through a sliding-neighborhood operation.