HY-1卫星CZI影像在黄海绿潮监测的应用

Application of HY-1 CZI satellite images in monitoring of green tide in Yellow Sea

海洋一号C(HY-1C)卫星搭载的海岸带成像仪CZI(Coastal Zone Imager)广泛应用于中国近海生态灾害监测。本研究以16 m分辨率的高分六号WFV(Wide Field of View)影像提取的绿潮覆盖面积作为参考值,从绿潮漏检率和覆盖面积定量评估了CZI影像的绿潮监测能力,并与250 m分辨率的MODIS提取结果进行了对比分析。结果表明,CZI影像的绿潮平均漏检率只有MODIS影像的1/5左右,绿潮覆盖面积比MODIS影像小50%以上。MODIS和CZI影像的绿潮覆盖面积呈线性相关,将MODIS影像的绿潮覆盖面积转换为CZI结果;该结果显示,2019年、2021年度绿潮最大覆盖面积为2000 km^(2)左右,是2020年同期的6倍左右。在绿潮监测方面,相比于MODIS,CZI影像的绿潮漏检率较低且覆盖面积更接近真实参考值。本研究建立了CZI与目前较常用的MODIS绿潮覆盖面积的转换关系,可弥补CZI观测频次的缺陷,进而实现绿潮高频次观测。

The green tide in the Southern Yellow Sea have appeared 1.5 decades since 2007, resulting in a great losses to the marine ecological environment and government finance. With the help of multiple satellite images, remote sensing technique play an import part in monitoring green tide outbreak process. Yet selecting an appropriate sensor is a precondition in quantifying the severity of green tide. HY-1C satellite is great superior to other sensors for its relatively high spatial resolution(50 m), wide swath width(950 km) as well as short revisit time(3 days). Here, we introduce the green tide in GF-6 WFV images(16 m) to evaluate the capability of CZI images in the monitoring of green tide, and then compare the CZI results with the traditional MODIS images(250 m). And the GF-6 WFV images is also applied to evaluate the omission rate and accuracy of green tide extraction in the CZI and MODIS images. Based on the convert parameter between MODIS and CZI green tide mapping result, we get high frequency green tide outbreak process in 2019, 2020 and 2021.In this study, dynamic threshold is introduced to extract green tide from the DVI results and the coverage area of green tide will be obtained by adding up the area of pixel. Besides, the ratio of coverage area to affected area is used as the aggregation density of green tide.The relationship between the ratio of coverage area of green tide from MODIS and CZI images and aggregation density of green tide is also analyzed. And we give the linear relationship of coverage area of green tide obtained from satellite images with different resolution.Results indicate that the average omission rate of CZI images(6.64%) is much lower than that of MODIS images(34.08%). In addition,the coverage area of green tide acquired from MODIS and CZI images is high linearly correlated, so both of them can be linear transformed.With the combination of CZI images and MODIS images, the daily coverage areas of green tide in the Yellow Sea in 2019, 2020 and 2021 are retrieved. The CZI-based maximum daily coverage areas of green tide were 2290 km~2, 336 km~2 and 1949 km~2, respectively, which are consistent with the evolutions in the countermeasures adopted by managers to control the green tide.This study shows that, CZI image has the advantages of lower omission rate and higher accuracy of coverage area in monitoring of green tide in contrast to the MODIS image. And the defect of observation frequency in CZI image will be improved by the linear conversion of coverage area of green tide from MODIS and CZI images. Then, the high-frequency and high-precision of the observation of green tide will be realized.