地表反照率量化了太阳短波辐射与地表的相互作用，是地表-大气系统能量分配的重要边界条件，在陆面、气候和生物圈模型中起着至关重要的作用。随着环境的不断变化与极端天气的增加，越来越多的研究领域对反照率的准确性及分辨率提出了新的需求。为了更好地满足多领域研究的要求，拓宽定量遥感参数的应用，许多科学家都在开发生产高分辨率地表反照率产品的算法。本文提出的算法基于我国“国家高分辨率地球观测系统”卫星数据（简称高分数据）和中等分辨率的GLASS反照率产品，生成高空间分辨率的反照率产品。算法思路首先对高分数据做预处理，再用直接反演算法反演高分数据，得到16m分辨率初级反照率产品。最后将500m分辨率的GLASS反照率产品均值信息与16m分辨率初级产品的纹理信息进行降尺度融合，得到16m分辨率的融合反照率产品。基于提出的算法开展了2016-2017年黑河流域16m分辨率反照率产品初步生产，并基于地面台站观测数据开展了产品精度验证和误差影响因素分析，评估高分辨率反照率产品应用潜力，具体得到以下结论：

（1）根据2016-2017年黑河实验区的八个站点的地面观测值，对算法结果以及初级反照率进行验证，结果表明融合后的16m分辨率反照率与实测值一致性较好，比初级反照率更接近实测值，但是在有雪季节融合反照率产品优势不明显。

（2）通过用16m分辨率反照率产品分析研究区反照率空间分布特征以及与不同分辨率反照率产品的影像进行对比，发现反演得到的16m分辨率反照率产品值的变化范围和趋势与GLASS反照率产品一致，这个特性能够保证在不同尺度的能量平衡研究中获得一致的结果。同时，由于融合反照率产品分辨率高，影像清晰，在空间细节分析上更占优势。

（3）通过对研究区不同地物类型反照率进行均值及标准差统计发现，不同分辨率反照率统计结果会有差异：对于面积较大的地物类型结果差异不大；对于面积较小或分布不连续的地物类型结果差异明显。在人类活动造成的居住用地或光伏发电站这类地物的反照率研究中，仅用中低分辨率遥感产品导致地物纯度降低，统计结果容易有偏差，所以本文算法得到的16m分辨率反照率产品更能反映不同地物类型的反照率特点。

本文算法所得到的16m分辨率反照率产品在平均值与500m分辨率的GLASS反照率产品一致的前提下丰富了空间纹理信息，能够更好的支持在区域尺度研究人类活动对环境的影响。在今后的研究中应加强在积雪区域的反照率研究，提高数据准确度，改进算法，开展更为全面的精度验证与应用分析。

Surface albedo quantifies the interaction between solar short-wave radiation and the surface. It is an important boundary condition for the energy distribution of the surface-atmosphere system and plays a vital role in the land surface, climate and biosphere models. With the continuous change of the environment and the increase of extreme weather, more and more research fields have put forward new requirements for the accuracy and resolution of albedo. In order to better meet the requirements of multi-field research, broaden the application of quantitative remote sensing parameters, many researchers and teams are using different algorithms to develop high-resolution surface albedo products. The algorithm proposed in this study is based on Chinese "National High-resolution Earth Observation System" satellite data (referred to as GF data) and medium-resolution GLASS albedo products to generate high-spatial-resolution albedo products. The algorithm first preprocesses the GF data, and then uses the direct inversion algorithm to invert the GF data to obtain 16m resolution primary albedo product. Finally, the average information of the GLASS albedo product with a resolution of 500m and the texture information of the primary product with a resolution of 16m are downscaled and fused to obtain a fused albedo product with a resolution of 16m. Based on the proposed algorithm, the preliminary production of 16m resolution albedo products in the Heihe River Basin from 2016 to 2017 was carried out, and the product accuracy verification and error influencing factor analysis were carried out based on the observation data of ground sites, and the application potential was evaluated, and the following conclusions were obtained:

（1）According to the ground observations of the eight stations in the Heihe Experimental Area from 2016 to 2017, the algorithm results and the primary albedo are verified. The results show that the fused 16m resolution albedo is in good agreement with the measured data, and is better than the primary albedo. It is close to the measured data, and the advantage of the fusion albedo product is not obvious in the snowy season.

（2）In the study area, by comparing with images of different resolution albedo products, it is found that the change range and trend of the 16m resolution albedo product obtained by the inversion is consistent with the GLASS albedo product. This feature can guarantee our energy at different scales Consistent results were obtained in smoothing studies. At the same time, because the fusion albedo product has high resolution and clear images, it is more advantageous in spatial detail analysis.

（3）Through the mean and standard deviation statistics of the albedo of different types of features in the study area, it is found that the statistical results of the albedo of different resolutions are different: There is little difference in the results for the types of features with a large area; the results are significantly different for the types of features with a small area or discontinuous distribution. In the research on the albedo of features such as residential land or photovoltaic power stations caused by human activities, only medium and low-resolution remote sensing products are used to reduce the purity of the features, and the statistical results are prone to deviations. Therefore, the 16m resolution albedo products obtained by the algorithm in this study can better reflect the albedo characteristics of different ground feature types.

The 16m resolution albedo product obtained by the algorithm in this study enriches the spatial texture information on the premise that the average value is consistent with the 500m resolution GLASS albedo product, which can better support the study of the impact of human activities on the environment at the regional scale. In the future research, it is necessary to strengthen the study of albedo in the snow area, improve the accuracy of the data, improve the algorithm, and carry out the accuracy verification and application analysis.