The invention discloses an eddy current scanning image classification and recognition method based on deep learning. The eddy current scanning image classification and recognition method comprises thesteps: S1, collecting an eddy current scanning image of a tested material, and preprocessing the eddy current scanning image; S2, constructing a training sample and a test sample; S3, carrying out processing by using a Sobel operator to obtain a background-removed training sample; S4, constructing a plurality of sparse noise reduction auto-encoders, and initializing the sparse noise reduction auto-encoders; S5, inputting a training sample and a background-removed training sample, and performing unsupervised layer-by-layer self-learning pre-training on the plurality of sparse noise reduction auto-encoders; S6, extracting a plurality of trained coding networks of the sparse noise reduction encoders, and carrying out stack type combination to construct a stack type sparse noise reduction self-coding deep neural network; S7, performing fine tuning on the deep neural network by using a BP algorithm and a batch gradient descent method in combination with the data label; and S8, inputting the test sample into the deep neural network for performance test. According to the eddy current scanning image classification and recognition method, the image features extracted by the hidden layer can reflect the material defects more simply, and the classification accuracy is improved.