摘要
以Landsat8和Landsat9卫星在中国7个地区的42期影像作为数据源,通过比对其原始影像质量、归一化植被指数(Normalized Difference Vegetation Index,NDVI)和地表温度(Land Surface Temperature,LST)反演精度,客观评价Landsat8和Landsat9卫星数据的质量及应用精度,为Landsat遥感影像的科学研究及在中国地区的应用提供有价值的参考。采用目视判读法及原始影像特征数据(直方图、灰度均值、灰度方差、信噪比和熵)进行影像质量比对分析,利用相对偏差率(ME)、RMSE和MAE对比分析Landsat8和Landsat9的NDVI和LST反演精度,得到以下结论:原始影像质量评价显示,Landsat9数据光谱信息量更大、图层丰富、观测地物能力也更加优越。Landsat9获取的植被指数信号更强,Landsat9的NDVI标准差与Landsat8相比总体偏大,差值最大可达0.044,植被的可分性优于Landsat8;Landsat9的NDVI均值整体大于Landsat8,最大差值可达0.119。在经度85°~105°两个卫星的适用情况相似,而在其他地区Landsat9植被获取能力明显更强。此外,Landsat9 LST反演精度(RMSE=4.419 K,MAE=2.264 K)要优于Landsat8 LST(RMSE=5.133 K,MAE=2.488 K),与中国陆面数据同化系统(CLDAS-V2.0)近实时产品数据集的偏差率最小仅为0.068%。
42 periods of Landsat8 and Landsat9 satellite images from Chinese seven regions are used as data sources to objectively evaluate the quality and application accuracy of Landsat8 and Landsat9 satellite data by comparing their original image quality,Normalized Difference Vegetation Index(NDVI),and Land Surface Temperature(LST)inversion accuracy,which provides valuable reference for the scientific research of Landsat remote sensing images and their application in China.The visual interpretation and data source feature data(histogram,grayscale mean,grayscale variance,signal-to-noise ratio,and entropy)are used to compare and analyze the impact of quality.The relative deviation rate(ME),RMSE,and MAE are used to compare the NDVI and LST inversion accuracy of Landsat8 and Landsat9.The evaluation of original image quality shows that Landsat9 data has richer spectral information and layer.Landsat9 has superior ability to observe ground features.The vegetation index signal obtained from Landsat9 is stronger,and the standard deviation of NDVI from Landsat9 is generally larger than that of Landsat8,with a maximum difference of 0.044.The separability of vegetation is better for Landsat9,and the overall mean value of NDVI for Landsat9 is larger than that of Landsat8,with a maximum difference of 0.119.The applicability of the two satellites is similar between longitudes 85°and 105°,while in other regions,the vegetation acquisition capability of Landsat9 is significantly stronger.In addition,the inversion accuracy of Landsat9 LST(RMSE=4.419 K,MAE=2.264 K)is better than that of Landsat8 LST(RMSE=5.133 K,MAE=2.488 K),and the minimum deviation rate of the near-real-time product dataset of the China Land Data Assimilation System(CLDAS-V2.0)is only 0.068%.
作者
晏红波
吴思怡
卢献健
王佳华
YAN Hongbo;WU Siyi;LU Xianjian;WANG Jiahua(College of Geomatics and Geoinformation,Guilin University of Technology,Guilin 541006,China;Guangxi Key Laboratory of Spatial Information and Geomatics,Guilin 541004,China)
出处
《无线电工程》
北大核心
2023年第11期2550-2563,共14页
Radio Engineering
基金
广西自然科学基金(2022GXNSFBA035639)
国家自然科学基金(42267040,42064003)。
