【摘要】 针对利用卫星遥感影像进行大范围反演大豆叶面积指数(LAI)时人工野外样本实测耗时耗力的问题，该文以黑龙江省海伦市、嫩江市两块大豆样地作为研究对象，使用无人机平台获取激光雷达(LiDAR)点云数据，利用孔隙度模型进行大豆LAI反演，根据地面实测数据开展精度评价，并进一步探讨利用无人机LiDAR反演值替代地面实测值进行Sentinel-2卫星影像LAI反演的可行性。实验结果表明，去掉地块边缘混合单元网格影响后，两块大豆样地的LAI反演精度均优于90%。利用无人机LiDAR反演值替代地面实测数据与植被指数构建回归模型，大范围卫星反演LAI精度均优于86%。更多还原

【Abstract】 Aiming at the problem that the soybean leaf area index(LAI) inversion in large scale area based on satellite remote sensing image is labor-intensive and time-consuming during the field collected ground truth samples, taking two soybean sample plots in Hailun city and Nenjiang city of Heilongjiang province as the study area, the gap fraction model method was applied to extract the leaf area index of soybean based on the unmanned aerial vehicle(UAV) light detection and ranging(LiDAR) data in this paper. The accuracy of inversion results was evaluated based on ground-truth data. Furthermore, the feasibility of using UAV LiDAR inversion results to replace ground-truth values for LAI inversion from Sentinel-2 satellite imagery was explored. Experimental results showed that the LAI inversion accuracy of both soybean sample plots was better than 90% after removing the influence of the grid of mixed cells at the edge of the plots. Meanwhile, regression models were built based on the UAV LiDAR inversion results instead of the ground-truth values, and combined with vegetation indices for large-scale satellite LAI inversion. Preliminary result showed that the LAI accuracy of satellite inversion was better than 86%.更多还原