中文摘要：

提出一种基于分类的语义特征提取模型,并在航空影像建筑物及灌木分类应用中进行尝试。实验与分析表明,通过图像的底层特征来提取简单的语义特征是可行和有效的,提出的方法可以进一步提高影像分类的精度。

英文摘要：

In general,Bayesian networks represent the joint probability distribution and domain（or expert）know-ledge in a compact way and provide a comprehensive method of representing relationships and influences among nodes（or feature variables）with a graphical diagram.Accordingly,by advantages of Bayesian networks a new road to texture classification of aerial images for achieving the automatization and intelligentization of photogrammetry and remote sensing can be explored.In this paper,a new method is proposed to extract semantic feature based on classifiers,which constructs the mapping from low-level features to high-level semantic feature.Then it is applied to classification of aerial images＇ building and shrub.The experiment results demonstrate that the new method can improve the classification precision.