首页 > , Vol. , Issue () : 1993-2002
作物茬覆盖度的估算对于探究农业耕作方式对周围环境的影响具有十分重要的意义。目前，基于多光谱影像的作物茬指数是作物茬覆盖度估算的常用方法。然而，在作物茬高覆盖区域，指数法容易出现“饱和”现象。已有研究结果表明，结合影像的光谱与纹理信息有助于改善指数法的“饱和”问题。Sentinel-2作为一颗多光谱卫星，空间分辨率可达10米，与Landsat OLI相比，具有更丰富的纹理信息。因此，探究Sentinel-2光谱与纹理信息相结合在作物茬覆盖度估算上的潜力具有重要意义。本文以山东禹城为研究区，分析了Sentinel-2各波段反射率、归一化差值指数以及不同窗口大小下灰度共生矩阵统计量等遥感因子与野外实测作物茬覆盖度的相关性，并利用最优子集法对遥感因子进行筛选，构建作物茬覆盖度的最优估算模型。同时，使用留一法交叉验证对模型进行评价。结果表明，在单因子分析中，归一化差异耕作指数（NDTI）与作物茬覆盖度的相关性最好，相关系数达0.735。使用NDTI、5×5窗口下Sentinel-2 8A波段的相关性统计量以及12波段的方差统计量构建的多元方程是作物茬覆盖度估算的最优模型，相关系数为0.869，均方根误差为11%。与仅使用光谱信息的最优模型相比，相关系数提高了0.094，均方根误差下降了3.5%。可见，结合Sentinel-2的纹理信息有助于提高作物茬覆盖度的估算精度。
As an important part of farmland ecosystem, crop residues provide a barrier against water erosion and improve soil structure and organic matter content. Timely and accuracy estimation of CRC at regional scale is essential for understanding the condition of ecosystem and the interactions with surrounding environment. Satellite remote sensing is an effective way for regional CRC estimation. The tillage indices based on multi-spectral satellite imagery data are the commonly used method for CRC estimation. However, this method is ineffective in the high coverage area due to the "saturation" phenomenon. Previous studies showed that the combination of image spectral and textural information can solve the saturation problems to some extent. Sentinel-2 is a new satellite mission that is able to provide observations at multi-spectral bands with a spatial resolution of 10m, 20m and 60m. It can provide more information about texture when compared with the commonly used multi-spectral satellite Landsat-8 OLI data. Therefore, it is interesting important to explore the potential of combining spectral and textural information derived from Sentinel-2 data for CRC estimation. The objectives of this study were to: (i) analyze correlation between field measured CRC and satellite derived variables including Sentinel-2 bands reflectance, tillage indices and grey-level co-occurrence matrix statistics in different windows ; and (ii) determine the optimal CRC estimation method from optimal subset regression (OSR) with various combination of the tillage indices and image textural features. The results showed that the normalized difference tillage index (NDTI), B12_CO (contrast of band12, B12 in window 5×5) and B12_DI (dissimilarity of B12 in window 5×5) were significantly correlated with measured CRC with correlation coefficient R values of 0.765, -0.641 and -0.553, respectively. The estimation model based on NDTI outperformed the models based on any other single variables. Model combining spectral and textural information in optimal window (R=0.869 and RMSE=11.0%) had a more precise result than the model based on only spectral information (R=0.775 and RMSE=14.5%). The results demonstrated that combination of spectral and textural information can improve the accuracy of CRC estimation.