The integration of reconfigurable intelligent surface (RIS) into aerial platforms is crucial for seamless coverage in future wireless communication systems. However, to fully harness the potential benefits of aerial RIS-assisted networks, accurate knowledge of the RIS location, particularly for millimeter-wave (mmWave) frequencies, is essential. This paper addresses aerial RIS-assisted direction finding in the presence of Doppler shifts. By employing block-wise processing, we formulate reduced-dimension optimization subproblems in the vertical and horizontal directions. Leveraging the sparsity of mmWave channels, we propose an efficient direction finding algorithm based on atomic norm minimization. We analyze the performance of the proposed algorithm in comparison to existing methods, considering training overhead and computational complexity.