Dark-Center Based Insulator Detection Method in Foggy Environment

In foggy environments, outdoor insulator detection is always with low visibility and unclear targets. Meanwhile, the scale of haze simulation datasets insufficient. Aiming to solve these problems, this paper proposes a novel Dark-Center algorithm, which joint learning framework based on image defogging target detection. Firstly, dark channel prior algorithm used calculate sky transmittance then transpose it original generate foggy-simulated dataset; secondly, restoration modules an optimized module are combined improve robustness algorithm; then, for small detection, CenterNet network structure improved enhance feature extraction capability targets; finally, accuracy in environments by jointly details color features recovered via model model, effectively learn defogging. The experimental results CPILD dataset show that proposed can performance detector scenes, 96.76%.