Infrared Image Fault Identification of Power Equipment Based on Residual Network

摘要

Using an unmanned aerial vehicle equipped with an infrared camera to fly along the transmission line and collect image and video data. Analyze the causes of power equipment faults, the classification of faults, and the characteristics of faults reflected in infrared images. Achieve the function of fault detection, classification and positioning. First, the video data is parsed to read the infrared image sequence, and the read infrared image is filtered and preprocessed to suppress noise and interference. The infrared image is then used to extract the potential area of the fault using superpixel segmentation. Finally, according to the trained residual network, these potential areas are sequentially classified to calculate the probability, and the threshold is set by experiment to screen the real fault and calculate its position. Automatic detection and classification and positioning of power equipment faults are achieved. Practice has proved that the method has a high degree of automation and efficiency.