Flood Mapping in Mozambique Using Copernicus Sentinel-2 Satellite Data

Over the last two decades, Mozambique has experienced tremendous tropical cyclonic activities causing many flooding activities accompanied by disastrous human casualties. Studies that integrate remote sensing, elevation data and coupled with demographic analysis in Mozambique are very limited. This study seeks to fill the void by employing satellite data to map inundation caused by Tropical Cyclones in Mozambique. In pursuit of this objective, Sentinel-2 satellite data was obtained from the United States Geological Survey (USGS)’s Earth Explorer free Online Data Services imagery website covering the months of March 20, 2019, March 25, 2019, and April 16, 2019 for two cities, Maputo and Beira in Mozambique. The images were geometrically corrected to remove, haze, scan lines and speckles, and then referenced to Mozambique ground-based Geographic: Lat/Lon coordinate system and WGS 84 Datum. Data from twelve spectral bands of Sentinel-2 satellite, covering the visible and near infrared sections of the electromagnetic spectrum, were further used in the analysis. In addition, Normalized Difference Water Index (NDWI) within the study area was computed using the green and near infrared bands to highlight water bodies of Sentinel-2 detectors. To project and model the population of Mozambique and see the impact of cyclones on the country, demographic data covering 1980 to 2017 was obtained from the World Bank website. The Exponential Smoothing (ETS) method was adopted to forecast the population of Mozambique. Results from NDWI analysis showed that the NDWI is higher for flood areas and lower for non-flooded ones. The ETS algorithm results indicate that the population of Mozambique would nearly double by 2047. Human population along the coastal zone in the country is also on the rise exponentially. The paper concludes by outlining policy recommendations in the form of uniform distribution of economic activities across the country and prohibition of inland migration to the coastal areas where tropical cyclonic activities are very high.

Water Resource Monitoring Exploiting Sentinel-2 Satellite and Sentinel-2 Satellite Like Time Series;Application in Yangtze River Water Bodies

As part of the Dragon 4 project,the water extents of Wuchang and Shengjin lakes have been extracted from Sentinel-2 time series,using all exploitable images since the beginning of the acquisitions in 2015.The aim of the study is to assess the capability of the Sentinel-2 constellation and Landsat 8 over the Anhui region,especially the high temporal resolution.A total of 32 dates have been used and 10 Landsat 8 images(Libra)have been added to try to reduce the temporal gaps in the Sentinel-2 acquisitions caused by cloudy conditions.Extractions were done using a SERTIT-ICube automatized routine based on a supervised Maximum Likelihood Classification.These extractions allow to recreate the dynamic of the two lakes and show the drought and wet periods.During the 3 years interval,the surface peaks in July 2016 for both lakes.The lowest level appears at two different dates for each lake;in January 2018 for Wuchang,in February 2017 for Shengjin.Wuchang Lake surface area appears to be more variable than Shengjin Lake,with many local maximum and minimum between the end of 2017 and April 2018.In the case of Wuchang Lake,floating vegetation is a problem for automatic water surface area extraction.The lake is covered by vegetation during long periods of time and the water below can’t be detected by automatic radiometric means.Nevertheless,Sentinel-2 stays a pertinent and powerful tool for hydrological monitoring of lakes confirming the expectation from the remote sensing wetland community before launch.The presence of NIR and SWIR bands induces a strong discrimination between water and other classes,and the systematic acquisitions create dense time series,making analysis more consistent.It makes possible to sensor events occurring over short periods of time.Thanks to this a link can be done between endangered bird species,such as the Siberian Crane and the Lesser White-Fronted Goose and periodically flooded areas.These midterm results illustrated the pertinence and powerful of multi-source optical satellite data for environmental analysis.

Performance evaluation of HY-2 series satellites in marine gravity field recovery

China has successfully launched four Haiyang-2(HY-2)series altimetry satellites.HY-2A has attracted significant attention in gravity field recovery,but the performance of other HY-2 series satellites,including HY-2B/C/D,is seldom discussed.This study evaluated the performance of all the HY-2 series satellites in recovering marine gravity field.First,the crossover discrepancies in sea surface height of the four satellites,HY-2A,HY-2B,HY-2C,and HY-2D,were analyzed to assess their altimetry stability.It was found that HY-2B had the best altimetry quality,followed by HY-2D.Subsequently,different combina-tions of altimetry data were used to calculate vertical deflections and gravity anomalies in the South China Sea(112°E-119°E,12°N-20°N).The results showed that combining data from HY-2B,HY-2C,and HY-2D improved the inversion accuracy of gravity anomalies by 0.3 mGal compared to using HY-2A data alone.HY-2C and HY-2D contributed to enhancing the accuracy of the east component of vertical deflections.

Three-dimensional thermohaline structure estimation derived from HY-2 satellite data over the Maritime Silk Road and its applications

Estimated ocean subsurface fields derived from satellite observations provide potential data sources for operational marine environmental monitoring and prediction systems.This study employs a statistic regression reconstruction method,in combination with domestic autonomous sea surface height and sea surface temperature observations from the Haiyang-2(HY-2)satellite fusion data,to establish an operational quasi-realtime three-dimensional(3D)temperature and salinity products over the Maritime Silk Road.These products feature a daily temporal resolution and a spatial resolution of 0.25°×0.25°and exhibit stability and continuity.We have demonstrated the accuracy of the reconstructed thermohaline fields in capturing the 3D thermohaline variations through comprehensive statistical evaluations,after comparing them against Argo observations and ocean analysis data from 2022.The results illustrate that the reconstructed fields effectively represent seasonal variations in oceanic subsurface structures,along with structural changes resulting from mesoscale processes,and the upper ocean’s responses to tropical cyclones.Furthermore,the incorporation of HY-2 satellite observations notably enhances the accuracy of temperature and salinity reconstructions in the Northwest Pacific Ocean and marginally improves salinity reconstruction accuracy in the North Indian Ocean when compared to the World Ocean Atlas 2018 monthly climatology thermohaline fields.As a result,the reconstructed product holds promise for providing quasi-real-time 3D temperature and salinity field information to facilitate fast decisionmaking during emergencies,and also offers foundational thermohaline fields for operational ocean reanalysis and forecasting systems.These contributions enhance the safety and stability of ocean subsurface activities and navigation.

Retrieval of Antarctic sea ice freeboard and thickness from HY-2B satellite altimeter data

Antarctic sea ice is an important part of the Earth’s atmospheric system,and satellite remote sensing is an important technology for observing Antarctic sea ice.Whether Chinese Haiyang-2B(HY-2B)satellite altimeter data could be used to estimate sea ice freeboard and provide alternative Antarctic sea ice thickness information with a high precision and long time series,as other radar altimetry satellites can,needs further investigation.This paper proposed an algorithm to discriminate leads and then retrieve sea ice freeboard and thickness from HY-2B radar altimeter data.We first collected the Moderate-resolution Imaging Spectroradiometer ice surface temperature(IST)product from the National Aeronautics and Space Administration to extract leads from the Antarctic waters and verified their accuracy through Sentinel-1 Synthetic Aperture Radar images.Second,a surface classification decision tree was generated for HY-2B satellite altimeter measurements of the Antarctic waters to extract leads and calculate local sea surface heights.We then estimated the Antarctic sea ice freeboard and thickness based on local sea surface heights and the static equilibrium equation.Finally,the retrieved HY-2B Antarctic sea ice thickness was compared with the CryoSat-2 sea ice thickness and the Antarctic Sea Ice Processes and Climate(ASPeCt)ship-based observed sea ice thickness.The results indicate that our classification decision tree constructed for HY-2B satellite altimeter measurements was reasonable,and the root mean square error of the obtained sea ice thickness compared to the ship measurements was 0.62 m.The proposed sea ice thickness algorithm for the HY-2B radar satellite fills a gap in this application domain for the HY-series satellites and can be a complement to existing Antarctic sea ice thickness products;this algorithm could provide long-time-series and large-scale sea ice thickness data that contribute to research on global climate change.

基于Sentinel-2多光谱遥感影像的小浪底水质反演

多光谱遥感技术可根据遥感波段信息反演水质参数,降低监测成本,提高监测速度和质量,为大范围水环境监测提供了一种新的方法。通过分析小浪底水库的Sentinel-2多光谱影像以及采样点实测水质数据,建立了最佳光谱波段的水质参数反演模型,对小浪底水库的化学需氧量(COD)、总磷(TP)、总氮(TN)和氨氮(NH_3-N)进行了遥感反演,验证了反演模型的精确度和稳定性,并反演了各水质参数的空间分布规律。结果表明:在4种水质参数反演模型中,COD模型精确度和稳定性最高,其次是TP、TN,最低的是NH_3-N,水库出水口和部分边缘COD质量浓度较高,水库中心TN、TP和NH_3-N质量浓度高于边缘处。

Statistical Downscaling Retrieval of Land Surface Temperature in an Area with Complex Landforms in the Eastern Qinling Mountains of China Based on Sentinel-2/3 Satellite Data

The study of land surface temperature(LST)is of great significance for ecosystem monitoring and ecological environmental protection in the Qinling Mountains of China.In view of the contradicting spatial and temporal resolutions in extracting LST from satellite remote sensing(RS)data,the areas with complex landforms of the Eastern Qinling Mountains were selected as the research targets to establish the correlation between the normalized difference vegetation index(NDVI)and LST.Detailed information on the surface features and temporal changes in the land surface was provided by Sentinel-2 and Sentinel-3,respectively.Based on the statistically downscaling method,the spatial scale could be decreased from 1000 m to 10 m,and LST with a Sentinel-3 temporal resolution and a 10 m spatial resolution could be retrieved.Comparing the 1 km resolution Sentinel-3 LST with the downscaling results,the 10 m LST downscaling data could accurately reflect the spatial distribution of the thermal characteristics of the original LST image.Moreover,the surface temperature data with a 10 m high spatial resolution had clear texture and obvious geomorphic features that could depict the detailed information of the ground features.The results showed that the average error was 5 K on April 16,2019 and 2.6 K on July 15,2019.The smaller error values indicated the higher vegetation coverage of summer downscaling result with the highest level on July 15.

基于Sentinel-1/2改进极化指数和纹理特征的土壤含盐量反演模型

目前Sentinel-1/2协同反演植被土壤含盐量的研究大多是基于Sentinel-2光谱信息和Sentinel-1后向散射系数,没有考虑Sentinel-2光谱信息容易受土壤亮度等信息影响,Sentinel-1后向散射系数容易受土壤粗糙度和水分影响。为进一步提高Sentinel-1/2协同反演植被土壤含盐量的精度,用水云模型对雷达卫星后向散射系数进行校正,消除植被影响;然后协同Sentinel-2纹理特征,基于VIP、OOB、PCA 3种变量筛选和RF、ELM、Cubist 3种机器学习回归模型构建植被土壤含盐量反演模型。研究结果表明:经过水云模型去除植被影响后的雷达后向散射系数及其极化组合指数与土壤含盐量的相关性有一定程度的提高。不同变量选择方法与不同机器学习方法耦合模型在反演土壤含盐量中,OOB变量筛选方法与RF、ELM和Cubist 3种机器学习方法的耦合模型精度最佳,建模集和验证集的R2都在0.750以上,且验证集的RMSE和MAE均最小;其中OOB-Cubist耦合模型精度最高,且R_(v)^(2)/R_(c)^(2)为0.955,具有良好的鲁棒性。研究可为机器学习协同物理模型、光学卫星协同雷达卫星在土壤含盐量反演中的进一步应用提供思路。

基于Sentinel-2的青铜峡灌区水稻和玉米种植分布早期识别

及时准确地掌握灌区内作物种植分布对于灌溉水资源高效配置、农田精准管理具有重要指导意义。以宁夏青铜峡灌区为研究对象,利用多时相Sentinel-2卫星数据,通过水稻和玉米早期特征分析,提取关键的“水淹”信号和“植被”信号,构建时序归一化差异水体指数(MNDWI)和归一化植被指数(NDVI)特征值数据集,并通过样本分析关键特征阈值,构建水稻和玉米早期种植分布决策树模型,提取2022年宁夏青铜峡灌区水稻和玉米种植的空间分布。结果表明(:1)玉米和水稻苗期的后半段5月15—31日,水淹信号和植被信号是区分二者关键时期。(2)基于早期作物物候特征的方法,在5月16—31日获取的水稻和玉米图像制图精度高于90%,用户精度超过91%,总体精度超过90%,Kappa系数高于0.88,明显高于同时期随机森林方法的分类精度。(3)本研究提出的方法在早期水稻和玉米种植分布提取方面具有较强的适用性,并且能够在时空尺度上以较少的实地样本进行延展,同时在时间上也更有优势。因此,该方法为青铜峡灌区水稻和玉米种植分布早期调查提供了重要的方法支撑。

基于时序Sentinel-2影像的引黄灌区作物结构提取和供需水分析

在黄河流域用水指标严格控制的背景下,以山东省东营市垦利区引黄灌区为例,利用2022年时序Sentinel-2遥感影像构建作物生育期的NDVI时间序列,采用决策树分类方法提取灌区作物种植结构,基于垦利站气象资料和Penman-Monteith公式,分析了1973—2022年各作物的需水特性,利用遥感影像解译的各作物种植面积,计算了2022年灌区作物在不同降水保证率(5%、25%、50%、75%、95%)条件下的灌溉总需水量,结合2023年分配给灌区作物的灌溉水指标探究了灌溉水资源供需之间的平衡。结果表明:基于NDVI时间序列构建决策树分类方法可有效提取作物的种植结构,总体分类精度为85.07%,Kappa系数为0.819,能够满足作物灌溉需水量的研究。作物净灌溉需水量年际波动较大,水稻和冬小麦补充灌溉水量在所有作物中位列前两位,均值分别为913 mm和410 mm;处于雨季生长的夏玉米、夏大豆补充灌溉水量较小且灌溉需求均值较小。研究区2023年分配的灌溉水指标在降水保证率为50%时研究区灌溉水亏缺量为235.5万m^(3),在降水保证率为75%和95%时灌溉水亏缺量分别为1 754.5万m^(3)和2 261.5万m^(3)。

Sentinel-2遥感影像在盘锦水稻米质监测中的应用研究

本研究基于水稻孕穗期、抽穗期、灌浆期和成熟期4个生育期的Sentinel-2遥感数据,分析各生育期内卫星遥感光谱参数与稻米品质指标的关系,建立基于各生育期卫星光谱信息的水稻品质指标预测模型。将5种稻米品质指标分别与4个生育期内的光谱参数进行皮尔逊相关性分析,结果表明,5项品质指标在4个生育期内均与光谱参数有不同程度相关性。然后筛选出相关性效果显著的光谱参数,用于建立各品质指标的预测方程,建模结果表明,基于卫星遥感光谱信息解释率由大到小的稻米品质指标依次是精米率>长宽比>蛋白质含量>直链淀粉含量>糙米率;卫星遥感光谱反演稻米各品质指标所在的最佳生育期不同,糙米率和精米率的最佳生育期为抽穗期,其建模决定系数(Coefficient of Determination,R^(2))分别为0.461和0.893;长宽比的最佳生育期为成熟期,R^(2)为0.878;直链淀粉含量和蛋白质含量的最佳生育期为灌浆期,R^(2)分别为0.646和0.647;基于卫星遥感光谱信息的稻米品质模型验证效果较好,解释率为51%~74%。可见,利用卫星遥感技术能够实现大范围水稻品质指标定量监测与评估。

On simulation of precise orbit determination of HY-2 with centimeter precision based on satellite-borne GPS technique

The HY-2 satellite carrying a satellite-borne GPS receiver is the first Chinese radar altimeter satellite, whose radial orbit determination precision must reach the centimeter level. Now HY-2 is in the test phase so that the observations are not openly released. In order to study the precise orbit determination precision and procedure for HY-2 based on the satellite- borne GPS technique, the satellite-borne GPS data are simulated in this paper. The HY-2 satellite-borne GPS antenna can receive at least seven GPS satellites each epoch, which can validate the GPS receiver and antenna design. What＇s more, the precise orbit determination processing flow is given and precise orbit determination experiments are conducted using the HY-2-borne GPS data with both the reduced-dynamic method and the kinematic geometry method. With the 1 and 3 mm phase data random errors, the radial orbit determination precision can achieve the centimeter level using these two methods and the kinematic orbit accuracy is slightly lower than that of the reduced-dynamic orbit. The earth gravity field model is an important factor which seriously affects the precise orbit determination of altimeter satellites. The reduced-dynamic orbit determination experiments are made with different earth gravity field models, such as EIGEN2, EGM96, TEG4, and GEMT3. Using a large number of high precision satellite-bome GPS data, the HY-2 precise orbit determination can reach the centimeter level with commonly used earth gravity field models up to above 50 degrees and orders.

基于Sentinel-2影像的巴尔托洛冰川冰面湖研究

冰面湖是冰川的重要组成部分,是冰川消融的指示器,不仅对全球气候变化响应迅速,而且对了解和掌握区域水资源信息意义重大。本文基于Sentinel-2遥感数据,利用随机森林算法,对巴尔托洛冰川冰面湖进行识别提取,并基于提取结果分析研究区冰面湖的空间分布特征,以及冰面湖面积、数量与冰川高程的关系。本文冰面湖提取的准确率达96.07%,完整率达92.18%,错误率为11.59%;识别出巴尔托洛冰川冰面湖567个,面积为249.46~37134 m^(2);冰面湖多分布在距冰川末端3~26 km处,其中海拔3800~4300 m之间冰面湖数量最多,面积普遍较大,平均面积为1922 m^(2);随着高程的升高,冰面湖的数量和面积逐渐减少,在高程5300 m以上冰面湖数量仅为15个,平均面积为356 m^(2);高程升高导致冰面温度降低,是冰面湖数量和面积骤减的主要原因。

Comparison on Winter Wheat Yield Estimating Models Based on Radarsat-2 and HJ Satellite in Huaihe River Region

The establishment of crop yield estimating model based on microwave and optical satellite images can conduct the mutual verification of the accuracy of the reported crop yield and the precision of the estimating model. With Shou County and Huaiyuan County of Anhui Province as the experimental fields of winter wheat producing areas, the linear winter wheat yield estimating models were established by adopting backscattering coefficient and Normalized Difference Vegetation Index（NDVI） based on images from the synthetic aperture radar（SAR）—RDARSAT-2 and HJ satellite photographed in mid-April and early May, 2014, and then comparisons were conducted on the accuracy of the yield estimating models. The accuracies of the yield estimating models established using co-polarized（HH） and cross-polarized（HV） modes of SAR in Jiangou Town, Shou County were 68.37% and 74.01%, respectively, while the accuracies in Longkang Town, Huaiyuan County were 63.10%and 69.10%, respectively. Accuracies of yield estimating models established by HJ satellite data were 69.52% and 66.43% in Shou County and Huaiyuan County, respectively. Accuracies of winter yield estimating model based on HJ satellite data and that based on SAR were closed, and the yield difference of winter wheat in the lodging region was analyzed in detail. The model results laid the foundation and accumulated experience for the verification, parameters correction and promotion of the winter wheat yield estimating model.

基于Sentinel-2A NDVI时间序列数据和随机森林方法的高山冷凉蔬菜识别

该研究基于Sentinel-2A卫星的归一化差值植被指数(NDVI)时间序列数据,结合随机森林(RF)分类方法,对高山冷凉蔬菜种植区域进行精准识别与分类。以西吉县为研究区,利用2023年覆盖高山冷凉蔬菜全生育期的Sentinel-2A遥感数据,构建10 m高空间分辨率的NDVI时间序列数据,结合田间实测数据,使用RF分类方法对高山冷凉蔬菜进行识别分类。结果表明文章提出的方法在高山冷凉蔬菜种植区域识别中表现出了较高的精度和稳定性,总体精度达93.52%,Kappa系数为0.89。

基于Sentinel-2影像的黄河南岸典型改良示范区土壤含盐量反演模型

土壤盐渍化严重制约农田土壤环境的循环发展,高效准确地监测土壤盐分动态变化对盐碱地改良利用具有重要意义。为及时、有效地监测盐渍化土壤含盐量,以内蒙古黄河南岸灌区的4个典型盐碱化耕地改良示范区为例,利用Sentinel-2多光谱遥感影像,同步采集示范区内表层土壤的含盐量数据,通过相关性分析筛选敏感光谱指标,基于偏最小二乘回归(PLSR)、逐步回归(SR)、岭回归(RR)3种简单机器学习模型和深度学习Transformer模型建模,最后进行精度评价并优选出最佳含盐量反演模型。结果表明:示范区土壤反射率的可见光、红边、近红外波段反射率均与土壤含盐量呈正相关,短波红外波段反射率与土壤含盐量呈负相关,引入光谱指数能够有效提升Sentinel-2遥感影像与示范区表层土壤含盐量的相关性(相关系数绝对值不小于0.32);对比不同模型发现深度学习Transformer模型优于简单机器学习模型,验证集决定系数R~2和均方根误差(RMSE)分别为0.546和2.687 g/kg;含盐量反演结果与实地结果相吻合,为更精准反演内蒙古黄河南岸灌区盐渍化程度提供了参考。

基于Sentinel-2 MSI的养殖海湾富营养化反演研究

【背景】水体富营养化会导致水生生态系统的失衡,影响栖息生物的生存,造成水产养殖经济损失。快速掌握重要养殖水域富营养化状态与变化趋势对海洋渔业发展与生态环境保护具有重大意义。【目的】本研究旨在建立适用于重要养殖海湾的富营养化指数反演模型,为海洋环境保护、优化养殖规划布局等提供科学参考。【方法】基于实测水质参数和Sentinel-2 MSI卫星遥感影像数据,筛选与富营养化指数的对数lg(E)相关性最高的3个波段组合作为特征波段输入至CatBoost、BP神经网络、随机森林3个机器学习模型,对比3个模型的反演精度,确定最优模型,反演福建省诏安湾与东山湾2022年富营养化指数并进行时空特征分析。【结果】特征波段组合b3+b7、b3-b12、b3×b9与具有较高的相关性,作为反演富营养化指数的最佳波段组合;CatBoost模型相比BP神经网络和随机森林有更高的反演精度,决定系数(R^(2))达到0.90,均方根误差(RMSE)为5.67,平均绝对百分比(MAPE)为43.61%;东山湾富营养化指数整体呈由湾顶朝湾外逐渐降低的趋势。诏安湾富营养化指数整体呈沿岸高、湾中低的的趋势,且有明显的冬季低、其他季节高的特点。【结论】东山湾、诏安湾富营养化时空分布有着明显差异。影响富营养化的因素有陆域污染物输入、养殖活动、地理特征、水文环境、天气变化等。本研究结果能快速、全面反映海湾水质时空变化趋势,有助于定位污染源和敏感区,可为环境评价、污染治理及养殖管理提供参考。

基于Sentinel-2时序数据的新疆焉耆盆地农作物遥感识别与评估

为及时准确地获取干旱区农作物种植信息,研究借助PIE-Engine Studio平台,以新疆焉耆盆地为研究区,基于2022年Sentinel-2影像和1948个野外定位采样数据提取农作物生育期内14种植被指数,使用See5.0决策树、随机森林(Randomforest,RF)和多元回归(Multiple regression,MR)模型优选特征参数,结合支持向量机(Support vector machine,SVM)算法构建5种分类模型和5种样方分割方案进行农作物种植信息提取,通过目视解译和混淆矩阵对比分析分类结果,确定最佳分类方案。结果表明:(1)所有分类模型的总体精度(OA)和Kappa系数均在92.20%和0.9037以上,说明在PIE平台中使用SVM算法提取农作物信息是可行的。(2)SVM-有红边的OA和Kappa系数均值为93.77%和0.9236,比SVM-无红边方法提高了0.96%和0.0120。(3)相比于SVM-有红边方法,植被指数的引入提高了SVM-RF、SVM-MR和SVM-See5.0的OA和Kappa系数。(4)5种分类模型的OA和Kappa系数均值的大小关系为:SVM-RF>SVM-MR>SVM-See5.0>SVM-有红边>SVM-无红边,表明红边波段和植被指数的加入显著提高了农作物识别的精度,其中SVM-RF(8:2)为最佳分类模型,OA和Kappa系数分别为98.72%和0.9866。研究结果可为准确快速获取大尺度干旱区农作物信息提供新的思路和参考依据。

联合多时相GF-6 WFV和Sentinel-2的森林类型识别

【目的】我国南方地区多云雨,地型较破碎,森林类型精细识别较为困难,探讨联合多源、多时相的遥感数据对森林类型识别具有重要意义。【方法】以江西省信丰县为研究区,基于2019年森林资源二类调查数据,将森林划分为松林、杉木林、阔叶林、针叶混交林、针阔混交林、竹林、灌木林和其他林地等8种类型,利用随机森林算法比较GF-6 WFV和Sentinel-2最佳时相相同波段(紫/深蓝、蓝、绿、红、近红外、红边)和不同波段(黄边、短波红外)的森林类型识别能力,构建联合光谱特征集。联合多时相GF-6 WFV和Sentinel-2,构建多时相植被指数特征集,结合联合光谱特征集、纹理特征和地形特征,通过随机森林和递归消除法构建特征变量优选数据集进行森林类型识别,利用混淆矩阵和森林类型的实际分布对识别结果进行精度验证。【结果】(1)GF-6 WFV蓝、绿和红波段组合的总体精度为58.31%,分别加入紫、近红外、红边、黄边和Sentinel-2短波红外波段后,其总体精度分别提高1.99%、8.90%、10.71%、1.50%和14.10%;Sentinel-2蓝、绿和红波段组合的总体精度为54.68%,分别加入深蓝、近红外、红边、短波红外和GF-6 WFV黄边波段后,其总体精度分别提高3.30%、10.82%、12.92%、17.31%和3.97%。(2)特征变量优选数据集的总体精度和Kappa系数为80.80%和75.56%,贡献程度大小依次为GF-6 WFV多时相植被指数、Sentinel-2多时相植被指数、GF-6 WFV光谱特征、Sentinel-2光谱特征、地形特征和纹理特征,贡献率分别为40.44%、23.23%、18.12%、10.21%、4.61%和3.39%。(3)松林、杉木林、阔叶林、针叶混交林、针阔混交林、竹林、灌木林和其他林地的制图精度分别为86.97%、85.60%、88.61%、9.43%、19.01%、53.60%、86.90%和82.56%,用户精度分别为81.42%、79.79%、77.57%、71.43%、81.82%、67.00%、87.74%和82.88%,识别结果与研究区实际森林类型分布较吻合。【结论】联合多时相GF-6 WFV和Sentinel-2可以综合多时相、多源影像的优点,能够有效提高森林类型的识别精度。

基于GEE的Landsat-8与Sentinel-2影像在棉花种植提取中差异性分析及提取方法对比研究

棉花作为南疆地区重要的经济作物之一,在经济工作中起着至关重要的作用。及时、准确地获取棉花种植面积,对农业政策和经济发展具有重要意义。为了实现这一目标,需要综合分析不同方法和遥感数据对最终棉花种植面积制图精度的影响。本研究以新疆阿克苏地区棉花种植区为例,借助Google Earth Engine云平台,采用随机森林法(RF)、支持向量机法(SVM)、最小距离分类法(MDC)等3种机器学习方法,利用2类中分辨率影像提取棉花种植信息,充分评估使用的档案数据和官方统计数字。结果表明,采用Sentinel-2方法和RF获得了最优棉花图,随机森林法分类器的总体精度、Kappa系数和用户精度分别高达97.4%、96.7%和91.1%,分别比Landsat-8图像和RF模型的结果高出7.3百分点、0.081、2.8百分点。与官方统计数据相比,采用RF、SVM、MDC对Sentinel-2和Landsat-8图像的棉花种植面积估算图的精度分别为98.4%、95.8%、79.6%和90.3%、83.7%、72.5%。很明显,Sentinel-2和RF模型的组合与官方数据的一致性最高。对比分析结果表明,Landsat-8和Sentinel-2数据可用于大范围复杂种植结构的棉花高精度测绘。本研究结果有望为棉花大面积鉴别提供一定的理论指导和实践指导。