The traditional infrared image transmission line fault identification method is highly susceptible to noise interference, which makes it impossible to accurately find the location fault of transmission line, and has the problems of low accuracy of recognition results and poor anti-interference. Aiming at the above problems, a fault identification method based on infrared image for transmission line is proposed. Infrared imaging instrument is used to collect the temperature change image of the transmission line, and the particle filter algorithm is used to eliminate the noise interference in the image. The fault edge feature information of the image is converted by the RGB color channel, The high temperature fault point in the transmission line component region is identified and located by the conjugate gradient method. The simulation results show that the proposed method can accurately locate the high temperature faults of the transmission line, and the recognition result is high and the anti-interference is strong.