Recognition system for metal magnetic memory detection defects

The invention provides an identification system for metal magnetic memory detection defects, and belongs to the field of nondestructive detection. A metal magnetic memory detection system is combined with a BP neural network system, the problem that defect types cannot be detected through existing metal magnetic memory detection is solved, and the recognition system comprises the following steps that S1, after data are obtained and preprocessed, a used data set can be obtained; s2, constructing a BP neural network; and S3, realizing classification and identification of cracks, pits, pores and other microdefects of the ferromagnetic workpiece through the BP neural network. According to the embodiment, the position of the defect can be accurately positioned through the metal magnetic memory detection technology, recognition and classification of the defect can be achieved in combination with the neural network system, especially for detection of the in-service pressure container, time is saved, and detection ef