【摘要】 为探讨无人机影像是否存在混合像元以及提取麦田植被覆盖度的最佳卫星遥感数据源问题,选择混合像元较少的拔节期冬小麦试验田进行实验.首先估算无人机影像的植被覆盖度,以地面实测数据检验其精度,然后估算卫星数据的植被覆盖度,以无人机数据进行精度检验,寻找精度最高的卫星数据源.结果显示无人机影像考虑混合像元后,MAE降低了0.212,RMSE降低了0.228;所用Landsat8/OLI和Sentinel-12种卫星数据源中,Landsat8/OLI卫星数据的提取精度高(MAE为0.066,RMSE为0.086).结果表明考虑混合像元因素将提高无人机影像植被覆盖度的提取精度;Landsat8/OLI卫星数据更适用于盐渍土地区麦田植被覆盖度的提取.更多还原

【Abstract】 In order to explore the existence of mixed pixels in unmanned aerial vehicle(UAV) images and the optimal satellite remote sensing data source for extracting the vegetation coverage in wheat fields,the winter wheat experimental field at jointing stage with few mixed pixels was selected for the experiment.First,the vegetation coverage of the UAV image is estimated,and the accuracy is tested with the ground measured data.Then,the vegetation coverage of the satellite data is estimated,and the accuracy is tested with the UAV data to find the satellite data source with the highest accuracy.The results show that after considering the mixed pixels,the MAE and RMSE of the UAV images are reduced by 0.212 and 0.228 respectively.Among the 12 satellite data sources of Landsat8/OLI and Sentinel-12 used,the extraction accuracy of satellite data of Landsat8/OLI was high(MAE= 0.066 and RMSE=0.086).The results show that the extraction accuracy of vegetation coverage can be improved by considering mixed pixel factors.Landsat8/OLI satellite data were more suitable for the extraction of vegetation coverage of wheat fields in saline land area.更多还原