Timely detection of foreign objects on power transmission lines is essential for improving the safety of the power system. To address the challenge of a limited number of samples for foreign objects on power transmission lines, this paper proposes a few-shot learning-based algorithm for detecting power transmission line foreign objects in unmanned aerial vehicle remote sensing images based on the Faster RCNN network. Firstly, a Dual-Awareness Attention Mechanism (DAAM) is employed to capture the spatial relationships between support feature and query feature. Subsequently, a Representation Decoupling Mechanism (RDM) is utilized to address the data distribution differences between base classes and novel classes. Finally, the model is validated on a dataset of foreign objects on power transmission lines. Experimental results demonstrate that the proposed method effectively detects foreign objects on power transmission lines, providing a novel approach for few-shot learning in the detection of foreign objects on power transmission lines.