Power line inspection method based on lightweight target recognition neural network model

The invention provides a power line inspection method based on a lightweight target recognition neural network model, and the method comprises the steps: 1, carrying out the inspection of a power grid transmission line through an unmanned plane, and obtaining a power grid line image through the unmanned plane; 2, constructing a lightweight target recognition neural network model, and inputting the image into the lightweight target recognition neural network model for recognition; and step 3, obtaining a recognition result of the power grid line image according to the output of the lightweight target recognition neural network model, wherein the recognition result comprises the category and the fault position of the power grid line. According to the method, embedded hardware computing power is considered, neural network parameters with high density are randomly discarded by combining a Bernoulli equation of the hardware computing power based on a deep separation convolutional network, and meanwhile, the recogn