欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
本文主要研究面向对象的高分辨影像信息提取中的特征选择问题。文中分别选择光谱、纹理、形状等特征57个和28个进行特征优化,得到两组分别由46个和4个特征组成的不同的最优特征集,并利用这些特征集采用K近邻、模糊与K近邻级联两种不同的面向对象分类策略进行分类研究。最后从合理性、效率和精度三方面进行了对比分析。实验结果表明,对所有类别进行特征空间优化费时费力,并不合适,分类时需要采用级联的方式将两种分类器联合使用,根据实际需要进行必要的特征选择。
英文摘要：
In this paper,object-oriented features space optimization of high spatial resolution remote sensing image classification was discussed.Two different groups of features space were optimized by 46 and 4 features from 57 and 28 disparate texture,shape and structure information of image objects,and were used to extract information by two different object-oriented classification strategies： K nearest neighbor classification,cascade connection of fuzzy classification and K nearest neighbor classification.The experiment results were compared on the aspects of rationality,efficiency and precision,which showed that features space optimization of each classification was time-consuming and improper,and the extraction effect of cascade connection classification strategy increased significantly.
同期刊论文项目
同项目期刊论文
SVM model for estimating the parameters of the probability-integral method of predicting mining subs
期刊信息
位置：