基于风云三号D星MERSI-II影像的高原湖泊范围监测能力分析

Capability Analysis of Lake Extent Monitoring in Highland Region from MERSI-II onboard FY-3D

  • 摘要: 青藏高原湖泊是全球气候变化的敏感指示器,持续湖泊监测对探索湖泊自然演化规律及其与气候的相互作用有重要意义。高原降雨、蒸发、冰川消融等都会引起湖泊范围的骤变,因此用于高原湖泊动态监测的数据对时间分辨率有很高要求。国产风云三号D星(FY-3D)中分辨率成像仪(MERSI-II)具备逐日宽幅覆盖的高频次监测能力,但其250 m分辨率影像在高原湖泊监测应用中的能力还有待验证。本文以青藏高原湖泊为研究对象,以同期的第三方数据集和Landsat-8陆地成像仪(OLI)监测结果为对比真值,分析了FY-3D MERSI-II数据在高原湖泊监测中的应用能力。研究发现,原始250 m的MERSI-II数据对湖泊面积反演平均精度达95.12%,对湖泊水体边界提取的平均拟合度达91.21%,再凭借其时间分辨率优势,MERSI-II影像可以服务于长时序高动态的水体监测应用。为进一步验证MERSI-II数据的应用潜力,本文对比分析了其在空间分辨率超分后湖泊范围监测能力的改善情况。结果表明,MERSI-II数据由250 m超分至150 m,湖泊面积反演平均精度提升了2.62%,达到97.74%,湖泊边界平均拟合度提升了4.44%,达到95.65%。本文研究表明,空间超分辨率潜力和高时间分辨率特性会使FY-3D MERSI-II数据在长时序高动态湖泊监测中有很好的应用价值。

     

    Abstract: Objectives: Lakes on the Tibetan Plateau are sensitive indicators of global climate change, and relevant studies on lakes are important for exploring the natural evolution of lakes and their interactions with climate. Since climate change, precipitation, glacial melt, and evaporation on the highland areas can easily cause abrupt changes in lake extent, both temporal and spatial resolution are frequently needed in the images used for dynamic monitoring of highland lakes. The Moderate Resolution Imager (MERSI-II) aboard domestic FengYun(FY) satellite provides extensive coverage every day, it has not yet been verified whether its 250 m image can be used for highland lake monitorin. Methods: By using the monitoring data from contemporaneous third-party datasets and Landsat-8 Operational Land Imager (OLI) as the true value in the lakes on the Tibetan Plateau, this paper analyzed the applicability of FY-3D MERSI-II data for plateau lake monitoring. Results: It was found that the average accuracy of MERSI-II images for lake area quantification accuracy reaches 95.12%, and the average lake boundary fitness reaches 91.21%. With the advantage of high-frequency monitoring, long time series of highly dynamic water monitoring applications can benefit from MERSI-II images. To further confirm the application potential of MERSIII data, the effects of improved spatial quality on increases in lake-wide monitoring capacity were analyzed. Compared to the initial results, the application performance of optimized MERSI-II data had significantly improved with an average lake area quantification accuracy of 2.62% and an average lake boundary fitness of 4.53%. For its spatial resolution hyper-segmentation potential and quantitative lake monitoring capability, FY-3D has good potential for high-frequency lake monitoring applications.

     

/

返回文章
返回