最优分区与最优指数联合的水体信息提取

发布时间：2018-07-14 22:09

【摘要】：单波段阈值法和水体指数法具有方法简单、耗时少的优点,成为水体信息提取常用的方法,但对于面积较大、类型多样、影响因素复杂的区域,在全局使用任何一种方法均无法满足精度要求。为精确提取地表覆盖类型复杂、水体类型多样的山区水体信息,本文分别对归一化水体指数(MNDWI)、自动水体提取指数(AWEI)和归一化三波段指数(NDTBI)3种指数的阈值进行分析,以最优阈值分区方案进行分区并构建3个单指数决策树,寻找不同水体类型的最优指数,按照最优原则重构联合指数决策树来提取水体信息,其中,3种单指数提取法的Kappa系数分别为:0.863,0.854,0.862,最优指数联合决策树法的Kappa系数为0.881。结果表明,基于3个指数构成的最优指数联合决策树的提取方法可以达到最高精度,由此说明,采用该法用于水体信息提取可得到最佳效果。
[Abstract]:The single-band threshold method and the water body index method have the advantages of simple method and less time consuming, so they have become the common methods for extracting water information, but for the areas with large area, various types and complex influence factors, the single band threshold method and the water body index method have many advantages. Any method can not meet the precision requirements in the global use. In order to accurately extract the information of mountain water bodies with complex surface cover types and various water bodies, the threshold values of three indices, namely, normalized water body index (MNDWI), automatic water extraction index (awe) and normalized three-band index (NDTBI), were analyzed in this paper. The optimal threshold partition scheme is used to partition and construct three single index decision trees to find the optimal indices of different water bodies. According to the optimal principle, the combined index decision tree is reconstructed to extract water information. The Kappa coefficients of the three single index extraction methods are respectively: 0. 0. 863 / 0.854 / 0. 862, respectively. The Kappa coefficient of the optimal exponential joint decision tree method is 0. 881. The results show that the extraction method of optimal exponential joint decision tree based on three indices can achieve the highest precision, which shows that the best effect can be obtained by using this method in extracting water information.
【作者单位】：西南大学地理科学学院;贵州工程应用技术学院;
【基金】：中央高校基本科研业务费专项资金项目(XDJK2011C089) 国家自然科学基金项目(40971122)
【分类号】：P237;P332

【相似文献】