Image texture feature fusion enhancement for bearing fault diagnosis based on maximum gradient

In modern manufacturing industry, mechanical equipment plays a crucial role. In order to address the difficulty of signal feature extraction in mechanical equipment, this paper proposes a image Texture Feature Fusion Enhancement (TFFE) method based on maximum gradient. A mathematical transformation method is used to convert one-dimensional time series into two forms of images: symmetrized dot pattern and scalogram. The texture features are obtained by calculating the maximum gradient of the two types of images. The proposed image Texture Feature Fusion Enhancement (TFFE) method is used to combine different images and enhance the texture features. Finally, the Darknet53 network is used as the image classification method to conduct intelligent classification of rolling bearing faults. The classification effect of the method is verified by a series of experiments, in which the validity of the images used in different image conditions is verified, and the network used in different network conditions show better classification performance. The method's ability to resist noise is also validated in experiments under different noise conditions. The experimental results show that the proposed image enhancement method can improve fault features in the image and maintain good diagnostic performance.