Automatic Terrain Selection Based on Clustering and Genetic Algorithm

Terrain selection can't be realized automatically in C
³
I. For a certain GIS, an automatic terrain selection based on clustering (CTS) is proposed. It processes the data from the DEM layer according to constrained conditions, and constructs clusters by unconnected graph traversal in order to obtain the primary terrain. Another terrain selection based on genetic algorithm (GATS) is also proposed. It uses correlative selection operator in the family, adjusts dynamically crossover and mutation ratios, and designs fitness function by utilizing synthetically several species of constrained conditions and fuzzy membership degree so as to optimize terrain selection. Wherein, CTS doesn't need to assign the numbers of clusters previously, and the selected terrain accord with primary requirement. GATS can provide prepared multi-schemes straightway. The experimental results show that the two methods are feasible and effective.