Roads Digital Map Generation with Multi-track GPS Data

In order to generate roads digital map of unknown area when the raster map is absent, we proposed a way which employs multi-track GPS (Global Position System) data. Despite inaccuracy of GPS system itself, multi-track data can reduce errors when we assume a symmetric distribution of the error values. Morphological operations to binary image are employed to process the GPS data. Dilation, closing operations are used to merge discrete data points and fill in gaps among GPS tracks, while thinning algorithm and pruning are used to extract skeleton to gets the roadspsila position. We also exploit direction rule and distance rule to delete redundant nodes when constructing roads graph so that the data storage of digital map will be reduced. By overlaying our own digital map on the Google Earth map of the same roads, this way of generating digital map with multi-track GPS data is proved to be effective.