【摘要】 黑土地是地球上极为珍贵的土壤资源，但长期高强度的利用加之土壤侵蚀，黑土层逐渐变薄、变瘦、变硬。作物秸秆覆盖是对黑土地实施保护的一种保护性措施，因此，秸秆覆盖度监测是保护性耕作措施实施程度的关键指标之一。国产高分六号(GF-6)卫星是中国首颗精准农业观测的高分卫星，相较于前几颗高分卫星，GF-6 WFV多光谱图像新增紫外波段、黄波段和两个对植被变化响应敏感的红边波段，新增光谱波段是否在黑土区作物秸秆覆盖度估算中具有应用潜力是本研究的探索目标。以保护性耕作“梨树模式”创建地的梨树县为研究区，选取2020年11月5日的GF-6 WFV多光谱图像，探索基于GF-6 WFV多光谱图像的秸秆光谱指数构建方法、基于像元二分法的秸秆覆盖度估算方法。研究结果表明：(1)由绿、红、近红外、紫外和黄波段组合的5种秸秆光谱指数与实测秸秆覆盖度的相关性较大，R~2大于0.5,能够解释超过50%的玉米秸秆覆盖度信息；(2)GF-6 WFV秸秆覆盖度估算结果与Sentinel-2 MSI、 Landsat8 OLI估算结果之间存在较好的线性关系，决定系数R~2分别为0.833、 0.732,进一步说明了国产GF-6 WFV多光谱图像用于秸秆覆盖度估算的可靠性和有效性；(3)利用像元二分法模型和考虑土壤背景差异可提高玉米秸秆覆盖度估算精度，与线性回归模型相比，像元二分法模型的估算精度R~2由0.740提高到0.769;考虑土壤质地分区后，像元二分法模型估算精度得到整体提升，R~2达到0.822。研究结果可为大区域范围内的作物秸秆覆盖度精确估算提供参考。更多还原

【Abstract】 Black soil is an extremely precious soil resource on Earth. Unfortunately, the black soil layer gradually becomes thinner, leaner, and harder due to long-term high-intensity utilization and soil erosion. Crop residue covering(CRC) is an important way to protect the black soil. Therefore, monitoring the crop residue coverage is one of the key indicators for assessing the implementation of conservation tillage measures. The Chinese Gaofen-6(GF-6) satellite is the first high-resolution satellite dedicated to precision agricultural observation. Compared with previous Chinese high-resolution satellites, there are four new bands in GF-6 including ultraviolet, yellow, and red-edge bands sensitive to vegetation changes. The main goal of this study is to determine whether these new spectral bands have potential applications in estimating crop residue coverage in black soil regions. The study was conducted in Lishu County where the “Lishu Model” of conservation tillage was set up. The GF-6 WFV multispectral image acquired on November 5, 2020, was used to explore the potential of GF-6 WFV multispectral image for corn residue coverage estimation, including developing spectral indices and applying the Dimidiate Pixel Model. The research results indicate that(1) these 5 spectral indices including NDI87, NDI37, NDI47, NDI32 and NDI38, combined from green, red, near-infrared, ultraviolet, and yellow bands, were found to be more correlated with the measured residue coverage measured in the field, with the determination coefficient R~2 greater than 0.5, explaining more than 50% of the corn residue coverage information.(2) There were good correlations between the estimated CRC using GF-6 WFV multi-spectral image and the results using Sentinel-2 MSI and Landsat8 OLI multi-spectral image, with R~2 of 0.833 and 0.732, respectively. This demonstrates the reliability and effectiveness of Chinese GF-6 WFV multispectral imagery for crop residue coverage estimation.(3) The estimation accuracy of corn residue coverage was improved by considering the black soil background and using the Dimidiate Pixel Model. Compared with the linear regression model, the correlation coefficient R~2 of the Dimidiate Pixel Modelwas improved from 0.740 to 0.769. After considering the soil texture zoning, the R~2 was improved furtherly to 0.822. A new way is provided to improve the accuracy of regional crop residue cover estimation in the black soil region.更多还原