基于扩展贝叶斯分类算法的土地利用时空变化信息库去冗方法

De-redundancy method of the spatiotemporal change land use information based on extended Bayesian classification algorithm

本文提出一种基于扩展贝叶斯分类算法的土地利用时空变化信息库去冗方法,通过遥感技术获取土地利用图像,结合人工目视判读法,分析土地利用时空变化指标,构建土地利用时空变化信息库。利用基于属性关联性的朴素贝叶斯分类算法,约简信息属性,去除冗余信息属性,达到信息库去冗的目的。将扩展弧添加在非同种类别的父子结点间,形成扩展贝叶斯分类算法,提高冗余信息去除的准确性。实验结果表明,该方法能够有效去除土地利用时空变化信息库中的冗余信息,节约存储开销,提供完整的土地利用信息,具有较好的分类性能。

This paper proposes a land use spatiotemporal change information database redundancy reduction method based on the extended Bayesian classification algorithm.By using remote sensing technology to obtain land use images,combined with artificial visual interpretation,to analyze the spatiotemporal change indicators of land use,and construct a spatiotemporal change information database.By using a naive Bayesian classification algorithm based on attribute correlation,redundant information attributes are removed,achieving the goal of reducing redundancy in the database.Add extended arcs between parent-child nodes of different categories to form an extended Bayesian classification algorithm,improving the accuracy of redundant information removal.The experimental results show that this method can effectively remove redundant information from the spatiotemporal change information database of land use,save storage costs,provide complete land use information,and have good classification performance.