空间分辨率对绿潮覆盖面积、密集度卫星遥感信息提取的影响

Impact of the Spatial Resolution of Satellite Image on the Remote Sensing Monitoring of Green Macroalgae Bloom

研究表明，多源卫星数据提取的绿潮信息存在较大差异，通常认为可能的主要原因是不同卫星遥感器的的空间分辨率、过境时间差、波段设置等各不相同，但对此仍缺乏深入的研究。本文采用波段设置和过境时间完全相同、仅空间分辨率不同的四景 MERIS 全、降分辨率（300 m 和1200 m）影像，利用统一的算法（NDVI）进行绿潮信息提取，量化了空间分辨率对绿潮覆盖面积、密集度（由斑块个数、聚合度表征）卫星遥感信息提取的影响。结果表明：空间分辨率对绿潮覆盖面积卫星遥感信息提取影响显著，全、降分辨率 MERIS 影像提取的绿潮覆盖面积最大相对偏差可达67％，遥感影像的空间分辨率对绿潮面积提取结果的影响既与绿潮 NDVI 探测阈值有关，还可能与绿潮发展阶段有关。绿潮密集度卫星遥感提取结果也受遥感影像空间分辨率的影响，全分辨率 MERIS 影像提取的绿潮斑块个数为降分辨率影像的7～21倍，绿潮聚合度较降分辨率影像高15％～25％。

Information of the green microalgae bloom extracted from different satellite images may show noticeable difference,which is commonly attributed to the various spatial resolutions,bands configuration and the overpass time differences between the earth-observing sensors.But the dedicated research exploring these factors has not been widely performed and the underlying mechanism has not been well documented.In this paper,we aims at quantify the impact of spatial resolution of satellite image on the extracted information of green macroalgae bloom （GMB）.Four MERIS full and reduced resolution （300 m and 1 200 m）images are adopted,which have the same bands and overpass time with the only difference in their spatial resolutions.GMB are extracted from these images with the same method of NDVI algo-rithm.The covered area and the distribution concentration （characterized by the macroalgae patch number and aggrega-tion index）of GMB are analyzed.The results indicate that the influence of spatial resolution on the satellite derived GMB area is significant.The relative percentage difference in the extracted GMB area from MERIS full and reduced reso-lution images can be up to 67%,depending on the NDVI threshold and possibly also on development phase of the bloom.Similar trend is also found for the extracted patch distribution concentration of GMB.The patch number and ag-gregation index derived from full-resolution MERIS images are about 7 ~21 times and 15% ~25% higher than those from reduced MERIS images,respectively.