高光谱遥感数据藻类信息提取方法研究

摘要

以Sacramento-San Joaquin 三角洲水域为研究区，利用2007年覆盖研究区的64个航带的HyMap数据，分别用混合调谐匹配滤波（ MTMF）和决策树进行藻类信息的分类提取，在此基础上均采用Kappa系数作为标准进行精度评价。结果表明：混合调谐匹配波（MTMF）分析方法对藻类进行分类的结果，Kappa系数是0．634；用基于指数的决策树方法对藻类进行分类的结果，Kappa系数是0．613，两种结果都具有高度的一致性。这说明这两种方法在提取水生物种方面比价有效。%Abstratc:This paper use Sacramento -San Joaquin Delta as the study area .On the base of the HyMap data covering 64 aircraft in 2007 , with the mixed tuned matched filtering ( MTMF) and decision tree method , we extract algae information , and on this basis are using Kappa coefficient as a standard to evaluate the accuracy .The results showed that:hybrid tuner matching wave ( MTMF) analy-sis method to classify the results of algae , kappa coefficient was 0.634;using decision tree-based index of the algae classification re-sults, kappa coefficient was 0.613, both results are highly consistency .This shows that these two methods parity effective in extracting water clock .