Bogie fault diagnosis method based on optical imaging and SIBOS-SVM

In response to the issues of poor diagnostic accuracy and diminished efficiency in existing bogie image fault diagnosis methods, this paper proposes a new bogie image classification technique SIBOS-SVM, which combines the Bag of Features (BoF) model with Scale Invariant Feature Transform (SIFT)feature extraction and squirrel optimisation algorithm (SOA) adaptive support vector machines (SVM). SIFT and Bag of Features combined with Squirrel Optimization Support Vector Machine (SIBOS-SVM) can estimate the probability that an image belongs to each category, providing high confidence classification decisions. In response to the In response to the demand for determining kernel function parameters and penalty coefficients using SVM, SOA is adopted to adaptively select parameters and establish a final bogie fault classification model to improve fault diagnosis performance. The experimental results show that the fault diagnosis accuracy of the SIBOS. This study uses high-definition industrial cameras combined with optical imaging technology and intelligent evolutionary algorithms to enhance the diagnostic capability of bogies. The experimental findings indicate that the fault diagnosis accuracy of the SIBOS-SVM algorithm introduced in this paper achieves 99.2%., and can achieve the global optimal solution with fewer iterations. The convergence speed advantage is obvious, and the efficiency is significantly higher.