• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Pixels, Objects, Intelligence: GEOgraphic Object Based Image Analysis for the 21st Century >ACCURACY ASSESSMENT METHOD FOR WETLAND OBJECT-BASED CLASSIFICATION
【24h】

ACCURACY ASSESSMENT METHOD FOR WETLAND OBJECT-BASED CLASSIFICATION

机译：基于湿地对象分类的准确性评估方法

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

The object-based classification approach needs an adaptation of the traditional accuracy assessment methodology. This paper presents a methodology that considers the inherent variability within each wetland class for the sampling size calculation based on objects and a stratified random sampling to select samples for each category. The polygons selected for validation are overlaid on the data used for classification. The analyst uses all ancillary data available to build an interpretation key which defines the labels of the polygons. The error matrix representing the whole territory could then be calculated. Finally, considering that wetland classes are not completely discrete and that the delineation is rarely categorical, fuzzy logic is used to analyze the accuracy. Decision rules are built based on class description to accept a relative level of confusion between similar classes.
机译：基于对象的分类方法需要适应传统的准确性评估方法。本文介绍了一种方法，该方法考虑每个湿地类内的固有变化，用于基于对象的采样大小计算，以及分层随机采样，为每个类别选择样本。选择用于验证的多边形覆盖在用于分类的数据上。分析师使用可用于构建定义多边形标签的解释密钥的所有辅助数据。然后计算表示整个领域的错误矩阵。最后，考虑到湿地类不是完全离散的并且描绘很少是分类的，模糊逻辑用于分析准确性。决策规则是基于类描述构建的，以接受类似类之间的相对混淆水平。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号