Parallel algorithm designed for polygon vectorization

Raster and vector are two major types of data format in GIS, we often need to make a conversion of raster data to vector considering the advantages of vector. In this paper we first discuss the classic algorithms of vectorization and the technology of high performance computing, by applying the parallel strategy to vectorization algorithm, we present a specific parallel algorithm for polygon vectorization base on MPI interface and also the open source library GDAL, Then we give the parallel result which is evaluated by parallel speed up and the analysis of this program. From the result, we can see the parallel program dramatically improve the vectorization efficiency comparing with a single processor. And finally present the problems still exist in the algorithm. We implement this algorithm using C++ and MPI.