Improving Power Line Detection Based on Phase Difference in Radar Image

Previous framework on automatic power line detection based on millimeter-wave radar videos assumes the Bragg pattern of power lines are distinguishable with the background signals. This limits its ability in recognizing fine-grained power line signals within complex and noisy scenes. In this work, we propose a detection method that combines both characteristics of the amplitude and phase of the return radar signal. Experiments demonstrate that our proposed approach is able to detect the power lines accurately, efficiently and more robustly on a new two-channel dataset.