Recognition method of metal appearance corrosion of substation equipment based on first-order neural network

Affected by the diversity of rust features and the problem of recognition accuracy, it is easy to have low efficiency and accuracy of recognition. In order to recognize the rust and corrosion of metal appearance of substation equipment, a method based on first-order neural network is proposed. The method first obtains a large number of metal appearance images containing normal and corrosion conditions through data acquisition, and then inputs these images into a first-order neural network for training. The neural network will learn and extract the features related to corrosion in the metal appearance, and finally be able to accurately determine whether there is corrosion on the metal surface. The experimental results show that the method has high accuracy and robustness in the identification of rust and corrosion in the metal appearance of substation equipment. The method solves the problem of localized corrosion, which is difficult to be found in conventional inspection, and improves the correct rate and efficiency of rust condition recognition.