Intelligent Detection of Steel Defects Based on Improved Split Attention Networks

Zhiqiang Hao,Zhigang Wang,Dongxu Bai,Bo Tao,Xiliang Tong,Baojia Chen
DOI: https://doi.org/10.3389/fbioe.2021.810876
IF: 5.7
2022-01-13
Frontiers in Bioengineering and Biotechnology
Abstract:The intelligent monitoring and diagnosis of steel defects plays an important role in improving steel quality, production efficiency, and associated smart manufacturing. The application of the bio-inspired algorithms to mechanical engineering problems is of great significance. The split attention network is an improvement of the residual network, and it is an improvement of the visual attention mechanism in the bionic algorithm. In this paper, based on the feature pyramid network and split attention network, the network is improved and optimised in terms of data enhancement, multi-scale feature fusion and network structure optimisation. The DF-ResNeSt50 network model is proposed, which introduces a simple modularized split attention block, which can improve the attention mechanism of cross-feature graph groups. Finally, experimental validation proves that the proposed network model has good performance and application prospects in the intelligent detection of steel defects.
multidisciplinary sciences
What problem does this paper attempt to address?