李营, 陈云浩, 陈辉, 王晨. GF-1 WFV影像的翅碱蓬植被指数构建[J]. 武汉大学学报 ( 信息科学版), 2019, 44(12): 1823-1831. DOI: 10.13203/j.whugis20180105
引用本文: 李营, 陈云浩, 陈辉, 王晨. GF-1 WFV影像的翅碱蓬植被指数构建[J]. 武汉大学学报 ( 信息科学版), 2019, 44(12): 1823-1831. DOI: 10.13203/j.whugis20180105
LI Ying, CHEN Yunhao, CHEN Hui, WANG Chen. Construction of Suaeda Salsa Vegetation Index Based on GF-1 WFV Images[J]. Geomatics and Information Science of Wuhan University, 2019, 44(12): 1823-1831. DOI: 10.13203/j.whugis20180105
Citation: LI Ying, CHEN Yunhao, CHEN Hui, WANG Chen. Construction of Suaeda Salsa Vegetation Index Based on GF-1 WFV Images[J]. Geomatics and Information Science of Wuhan University, 2019, 44(12): 1823-1831. DOI: 10.13203/j.whugis20180105

GF-1 WFV影像的翅碱蓬植被指数构建

Construction of Suaeda Salsa Vegetation Index Based on GF-1 WFV Images

  • 摘要: 目前大部分植被指数主要针对绿色植被构建,缺乏针对其他颜色特别是红色植被的指数。此外,面向湿地或潮间带植被识别提取的植被指数也相对较少。为拓展针对红色植被指数构建的研究,结合翅碱蓬植被的红色特征,基于高分一号(GF-1)卫星宽覆盖影像(wide field of view,WFV),通过对比翅碱蓬及其周边地物在GF-1 WFV影像中的光谱反射率特征,构建了翅碱蓬植被指数(suaeda salsa vegetation index,SSVI)。为评估SSVI提取翅碱蓬的精度,以辽宁双台子河口湿地自然保护区为研究区,采用各种植被指数分别提取了不同年份的5景GF-1 WFV影像翅碱蓬信息,并对提取结果精度及错分像元数进行统计分析。结果表明,SSVI平均提取精度为88.6%,平均错分像元占研究区比例为5.1%,在5个指数中提取翅碱蓬精度最高、效果最好。此外,5期影像间较大的时间跨度也证明了SSVI的鲁棒性较强,具有较好的适用性,受时间影响较小。综上,构建的SSVI可有效用于翅碱蓬的识别与提取,并监测其时空变化。

     

    Abstract: At present, most of the vegetation indices are mainly constructed for the green vegetation while the vegetation indices for red vegetation are less. In addition, the vegetation indices for the identification and extraction of wetland or intertidal vegetation are relatively few. Therefore, in order to expand the research on the construction of red vegetation index, we constructed the Suaeda salsa vegetation index (SSVI) based on GF-1 WFV (wide field of view) image by comparing the spectral reflectance characteristics of various land covers in the GF-1 WFV image and considering the red characteristics. Then, for the sake of evaluating the extraction precision of the SSVI, we took the Shuangtaizi Estuary Wetland Nature Reserve in Liaoning Province as the study area, used SSVI and other indexs to extract the Suaeda salsa from five GF-1 WFV images of different years. Then, we compared their average extraction accuracy and average misclassified pixels account for the proportion of the study area. The results show that the average extraction accuracy of SSVI was 88.6%, and the average misclassification pixels accounts for 5.1% of the study area, this indicated that extraction ability of SSVI is better than other vegetation indices. The SSVI showed the highest precision and the best effect among the five vegetation indices. Besides, the large time span between the five images also proved that SSVI was affected less by the temporal factor, it was robust and had good applicability. In summary, the SSVI can be effectively used for the identification and extraction of Suaeda salsa, and to monitor its temporal and spatial changes.

     

/

返回文章
返回