There are many high temperature equipment and complex thermal system in gas turbine power plant

which produce strong background thermal radiation.Burners

steam turbines and other equipment release a lot of heat during operation

and the heat radiation generated will be mixed into the infrared detection signal of the insulator.It makes it difficult to accurately separate the temperature signal of the insulator itself

increasing the difficulty of detecting the temperature of the insulator.Therefore

it proposes the infrared detection method of insulator anomal temperature of power supply unit in gas turbine power plant.Based on the U-Net network model

it performs insulator infrared image segmentation.An improved Yolov5 network is introduced to realize the infrared detection of insulator anomal temperature.