Method for detecting depth of steel decarburized layer

The invention provides a method for detecting the depth of a steel decarburized layer, and belongs to the field of steel structure detection. The detection method comprises the steps of material marking, deep learning and decarburization layer detection. The material marking includes cutting the non-material tissue area, the material tissue area, the total decarburization area and the complete decarburization area in each metallographic photo material to obtain a marked material. The deep learning includes performing recognition learning on the plurality of marking materials by adopting a deepneural network model, and performing initialization by adopting the original images of the marking materials, the marked images and the lengths of the single pixel points to obtain a decarburized layer recognition model taking the recognition boundary as an output node. The decarburization layer detection includes inputting the original image and the single pixel point length of the to-be-detected material, carrying out