A classification model of railway fasteners based on computer vision
Fasteners are critical railway components that maintain the rails in a fixed position. The state of fasteners needs to be periodically checked in order to ensure safe transportation. Several computer vision methods have been proposed in the literature for fastener classification. However, these methods do not take into consideration the fasteners covered by stone. This paper proposes a new fastener classification model, which can divide fasteners into four types, including normal, partially worn, missing, and covered. First, the traditional latent Dirichlet allocation is introduced for fastener classification and its shortcomings are analyzed. Second, conditional random fields are used to segment the fastener structure. Third, the Bayesian hierarchical model of fastener feature words and structure labels is established. Then, the topics hidden behind the fastener feature words are derived, and the fastener image is ultimately represented by a topic distribution. Finally, the fasteners are classified using the support vector machine. The experimental results demonstrate the effectiveness of this method.
KeywordsFastener classification Structure labels Latent Dirichlet allocation (LDA) Conditional random fields (CRF)
Latent Dirichlet allocation
Conditional random fields
Spatial pyramid LDA
Generalized linear model
Maximum a posteriori
Maximum posterior marginal
Tree-reweighted belief propagation
This work is supported by Sichuan Province Science and Technology Support Program under grant 2018GZ0361.
We propose a new fastener classification model, which can divide fasteners into four types, including normal, partially worn, missing, and covered. All authors read and approved the final manuscript.
Compliance with ethical standards
Conflict of interest
The authors declare that they have no conflict of interest.
- 3.Wang L, Zhang B, Chen XA (2011) Inspection system for loss of rail fastening nut based on computer vision. Comput Eng Des 32(12):4147–4150Google Scholar
- 12.Lafferty J, Mccallum A, Pereira F (2001) Conditional random fields: probabilistic models for segmenting and labeling sequence data. In: ICML, pp 282–289Google Scholar
- 14.Vishwanathan SVN, Schraudolph NN, Schmidt MW, Murphy KP (2006) Accelerated training of conditional random fields with stochastic gradient methods. In: ICML, pp 969–976Google Scholar
- 15.Mai L, Niu YX, Liu F (2013) Saliency aggregation: a data-driven approach. In: Proceedings of IEEE conference computer vision and pattern recognition, pp 1131–1138Google Scholar
- 20.Hammersley JM, Clifford P (1971) Markov fields on finite graphs and latticesGoogle Scholar
- 21.Konidaris G, Osentoski S, Thomas P (2011) Value function approximation in reinforcement learning using the Fourier basis. In: Proceedings of AAAI conference artificial intelligence, San Francisco, California, USAGoogle Scholar
- 22.Dalal N, Triggs B (2005) Histograms of oriented gradients for human detection. In: Proceedings of IEEE conference computer vision and pattern recognition, pp 886–893Google Scholar
- 23.Kumar S, August J, Hebert M (2005) Exploiting inference for approximate parameter learning in discriminative fields: an empirical study. In: Proceedings of international conference on energy minimization methods in computer vision and pattern recognition, vol 3757, pp 153–168Google Scholar
- 25.Heinrich G (2008) Parameter estimation for text analysis. Leipzig University, LeipzigGoogle Scholar
- 26.Blei DM, Jordan MI (2003) Modeling annotated data. In: Proceedings of the 26th annual international ACM SIGIR conference on research and development in informaion retrieval (SIGIR'03). ACM, New York, pp 127–134, 28 July–1 August 2003Google Scholar