论文部分内容阅读
This paper focuses on the issue that how to choose appropriate terrain modeling method to achieve balance between accuracy and speed.The support vector machine ( SVM) method is adopted to classify the terrains based on quantitative assessment of terrain complexity.The modeling method selection strategy is established according to the classification results. 80 actual terrains on lunar surface are organized to the samples set.60 samples are randomly chosen to form the train set,and the other 20 samples are taken as test set.Two train sets are compiled in order to find out whether elevation difference affected the classification results besides terrain complexity index.50 times tests are conducted for each train set.The experiment results represent that the terrain complexity index expresses the terrain complexity properly,and the classification model designed from the train set with two features is more accurate and generic.The statistical results show that the average classification accuracy of test set with two features is 84.3%.