Efficient digitization of printed color maps

Abstract

The importance of digital maps increases continuously. Unfortunately, it is costly and time consuming to create a digital map out of the void. If we can successfully vectorize scanned paper maps, digital map production will be much easier and previous map resources can be well utilized. The first step of map vectorization is color clustering. Currently there exist a number of fast and efficient algorithms for automatic color clustering. However, they may not work well for maps of enormous noises. Here we introduce two interactive color clustering algorithms: color clustering with pre-calculated index colors (PCIC) and color clustering with pre-calculated color ranges (PCCR). A number of experiments are conducted to investigate into the performance of the two approaches. Finally, we introduce a novel and efficient vectorization algorithm, which can perform the entire vectorization algorithm in one pass of linear time.