高级检索+

基于哨兵-2A模拟反射率及其影像的冬小麦收获指数估算

Estimation of Harvest Index of Winter Wheat Based on Simulated Sentinel-2A Reflectance Data and Its Real Remote Sensing Imagery

  • 摘要: 为验证基于花后累积地上生物量比例动态参数(D-fG)的作物动态收获指数(D-HI)遥感估算方法在大范围获取作物收获指数空间信息的可行性和有效性,以我国黄淮海平原河北省衡水市为研究区,以冬小麦为研究对象,在前期田间冠层尺度确定作物收获指数估算敏感波段中心和最大波段宽度基础上,利用哨兵-2A(Sentinel-2A)遥感模拟反射率数据及其真实遥感数据开展基于花后累积地上生物量比例的区域冬小麦收获指数遥感估算研究。在基于不同波段组合Sentinel-2A遥感模拟数据筛选出的冬小麦动态收获指数估算最优波段信息基础上,利用Sentinel-2A影像数据实现区域作物收获指数空间信息准确获取。结果表明,在Sentinel-2A多光谱遥感模拟反射率波段λ1(672~680 nm)和λ2(855~875 nm)条件下,基于D-fG遥感参数信息的D-HI估测精度最高,其中,均方根误差(RMSE)、归一化均方根误差(NRMSE)和平均相对误差(MRE)分别为0.040 4、10.83%、9.56%,证明Sentinel-2A卫星遥感数据在冬小麦收获指数估算中具有一定应用潜力。同时,在估测D-HI精度最高的模拟遥感数据波段信息基础上,基于Sentinel-2A卫星遥感红光和窄近红外波段组合的灌浆期不同阶段和成熟期D-HI估测区域总体验证精度指标RMSE、NRMSE和MRE分别为0.050 2、13.81%、12.00%。上述结果表明基于Sentinel-2A卫星数据和花后累积地上生物量比例的作物动态收获指数遥感估算方法在大范围D-HI空间信息获取中具有一定可行性和有效性,为基于宽波段多光谱卫星遥感数据的大范围作物动态收获指数空间信息准确获取提供一定技术方法借鉴。

     

    Abstract: In order to verify the feasibility and effectiveness of the remote sensing estimation method for dynamic harvest index(D-HI) based on the dynamic fraction of cumulative aboveground biomass after anthesis(D-fG) to obtain the regional spatial information of crop harvest index, the research on the regional estimation of crop harvest index was carried out, which was supported by the previous research results of sensitive band centers and their maximum band widths for crop harvest index estimation at the canopy level. Hengshui City, Hebei Province in Huang-Huai-Hai Plain of China was selected as the study area, and winter wheat was selected as the research object, and the simulated reflectance data of Sentinel-2 A and its real satellite remote sensing data were used. Finally, under the guidance of optimal band information for D-HI estimation of winter wheat based on the simulated Sentinel-2 A remote sensing data of different band combinations, the Sentinel-2 A image data were used to obtain accurate spatial information of regional crop D-HI. The results showed that under the condition of the simulated reflectance bands of λ1(672~680 nm) and λ2(855~875 nm), the D-HI estimation accuracy based on D-fG remote sensing information was the highest, the root mean square error(RMSE), normalized root mean square error(NRMSE), and mean relative error(MRE) were 0.040 4, 10.83%, and 9.56%, respectively, which proved that Sentinel-2 A satellite remote sensing data had certain application potential in the estimation of winter wheat harvest index. At the same time, supported by the band information of simulated remote sensing data with the highest estimation accuracy of D-HI, the RMSE, NRMSE and MRE were 0.050 2, 13.81%, and 12.00%, respectively, in the overall verification of D-HI estimation at different grain filling stages and maturity stages based on red band and narrow NIR band of Sentinel-2 A satellite remote sensing data. The above results indicated that the remote sensing estimation method of crop D-HI based on Sentinel-2 A satellite data and the fraction of cumulative aboveground biomass after anthesis had certain feasibility and effectiveness in the acquisition of regional spatial information of D-HI, and it could provide a certain technical method reference for the accurate acquisition of crop D-HI spatial information based on multispectral satellite remote sensing data in a large region.

     

/

返回文章
返回