In this paper, a method for recognizing the arc aurora sequences from all-sky image sequences is proposed. For the movement trend of arc aurora sequences, a method named ST-PVLBP (spatial-temporal poleward volume local binary patterns), which is based on VLBP (volume local binary patterns) and uses ST-PVLBP to present the aurora sequences, is proposed. Combined with the interframe continuity information of the sequence and the spatial location information of the single frame, the algorithm reduces the feature dimension while maintaining high classification accuracy at the same time. The proposed method was evaluated using auroral observations at the Chinese Arctic Yellow River Station. Experimental results show that the proposed method can effectively detect the poleward moving arc aurora sequences.