Multiple Auxiliaries Assisted Airborne Power Line Detection

Abstract

Airborne power line detection is a key technique to ensure low altitude flight safety. Yet, it is a challenging problem due to the extremely small size of power line targets. Recently, auxiliary assisted detection has shown great potential in improving the power line detection performance. However, in existing methods, the auxiliaries and the contexts between the power lines and the auxiliaries are both manually assigned, thus limits its applicability. In this paper, a novel multiple auxiliaries assisted power line detection method is proposed. With an optimization based auxiliaries selection and contexts acquisition scheme, the proposed method cannot only decide which auxiliaries should be selected to assist the detection, but also acquire the context information of each kind of auxiliaries, all in an automatic way. Experimental results show that the proposed method surpasses the state-of-the-art power line detection methods, both in terms of detection accuracy and false alarm probability.