遥感图像亿级基础模型（英文版）

英文标题：A Billion-scale Foundation Model for Remote Sensing Images中文摘要：本文研究了模型参数数量增加对远程遥感领域中基础模型在旋转物体检测和语义分割等下游任务性能的影响，并通过实验验证了模型参数数量的增加可以提升模型性能和数据效率，进而提出了一种有效的基于视觉变换器进行扩展和微调的远程遥感领域的方法。英文摘要：As the potential of foundation models in visual tasks has garneredsignificant attention, pretraining these models before downstream tasks hasbecome a crucial step. The three key factors in pretraining foundation modelsare the pretraining method, the size of the pretraining dataset, and the numberof model paramet