摘要:
针对SPOT6卫星遥感影像,采用基于规则的方法对建筑物进行提取.首先,分析了每种规则属性提取建筑物的效果,在此基础上制定建筑物提取规则;再分别采用K均值聚类法(K-means)、K临近值法(K nearest neighbor,KNN)、支持向量机法(support vector machine,SVM)和神经网络法进行建筑物提取实验,并与基于规则的方法进行对比;最后,对建筑物提取结果进行精度评价.研究表明,基于该规则的建筑物提取精度高于其他方法,在一定程度上缓解了椒盐现象和同谱异物问题,可为今后SPOT6卫星影像更广阔的应用提供一定的技术支持.
Abstract:
For SPOT 6 satellite remote sensing image, a method based on rules was used to extract buildings.Firstly, the authors analyzed the extraction effects of every rule attribute and made the rule extract buildings based on the effect.Then the authors compared the methods of K-means clustering, K nearest neighbor (KNN), support vector machine (SVM) and neural network with the method used in this paper during the research.The precision evaluation of building extraction result shows that the accuracy of this method based on rules is higher than that of other methods.This method relieves the problems of the salt and pepper phenomenon and the same spectrum with foreign bodies, and provides some technical support for the wider application of SPOT 6 satellite images in the future.
下载: