针对复杂电器零件的轻量化分类算法研究
Research on Lightweight Classification Algorithm for Complex Electrical Parts
摘要: 本文针对复杂电器零件的自动化检测问题，在实际工业场景下，采集并构建了具有复杂特性的(不同形状、厚度、颜色及透明度等)的多种电器零件样本数据集，并提出了轻量化的实时高精度分类模型。在模型构建中，通过引入轻量化的卷积残差块，多尺度的金字塔特征提取模块，以及显著减小模型计算量的Skip-Attention结构，使模型具有较低检测延时性的同时，并保证了较高的检测准确度。实验结果证明，本文所提出算法的实时检测效果优于较多数成熟的实时检测模型，具备应用于工业零件实时检测的可行性。
Abstract: This article is based on the automation detecting of complex electrical parts. In actual industrial scenarios, varieties of electrical parts datasets with complex characteristics (different shapes, thickness, color and transparency) are collected and constructed real-time classification model. In the model construction, by introducing lightweight convolutional residual blocks, multi-scale pyramid characteristics extraction modules, and Skip-Attention structures with significantly reduced model computing, it has achieved lower detection delayed models while ensuring higher detection accuracy. The experimental results prove that the real-time detection effect of the method proposed in this article is better than the most mature real-time detection model, and it has the feasibility of real-time detection in industrial parts.
文章引用：田金玉, 曾志强, 甄德鑫. 针对复杂电器零件的轻量化分类算法研究[J]. 计算机科学与应用, 2024, 14(2): 317-324. https://doi.org/10.12677/CSA.2024.142032

参考文献

[1] 吕广贤. 基于机器视觉的散热器钎焊缺陷检测系统研发[J]. 图像与信号处理, 2021, 10(3): 146.
[2] 李超, 许杰. 电子元器件外形尺寸机器视觉测量系统设计[J]. 光电子, 2020, 10(3): 84.
[3] 刘瑞欣, 严春雨, 李飞, 等. 基于改进YOLOX的茶叶嫩芽目标检测研究[J]. 软件工程与应用, 2022, 11(6): 1404-1414.
[4] Cortes, C. and Vapnik, V. (1995) Support-Vector Networks. Machine Learning, 20, 273-297. [Google Scholar] [CrossRef] 
[5] Ojala, T., Pietikainen, M. and Maenpaa, T. (2002) Multiresolution Gray-Scale and Rotation Invariant Texture Classification with Local Binary Patterns. IEEE Transactions on Pattern Analysis and Machine Intelligence, 24, 971-987. [Google Scholar] [CrossRef] 
[6] Lowe, D.G. (2004) Distinctive Image Features from Scale-Invariant Keypoints. International Journal of Computer Vision, 60, 91-110. [Google Scholar] [CrossRef] 
[7] Dalal, N. and Triggs, B. (2005) Histograms of Oriented Gradients for Human Detection. 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05), San Diego, 20-25 June 2005, 886-893.
[8] Shafiq, M. and Gu, Z. (2022) Deep Residual Learning for Im-age Recognition: A Survey. Applied Sciences, 12, Article 8972. [Google Scholar] [CrossRef] 
[9] Krizhevsky, A., Sutskever, I. and Hinton, G.E. (2012) Imagenet Classi-fication with Deep Convolutional Neural Networks. Advances in Neural Information Processing Systems, 25, 1-8.
[10] Vaswani, A., et al. (2017) Attention Is All You Need. Advances in Neural Information Processing Systems, 30, 1-10.
[11] Liu, Z., et al. (2021) Swin Transformer: Hierarchical Vision Transformer Using Shifted Windows. Pro-ceedings of 2021 IEEE/CVF International Conference on Computer Vision (ICCV), Montreal, 10-17 October 2021, 9992-1000. [Google Scholar] [CrossRef] 
[12] Dosovitskiy, A., et al. (2020) An Image Is Worth 16x16 Words: Transformers for Image Recognition at Scale. arXiv: 2010.11929.
[13] Lecun, Y., et al. (1998) Gradient-Based Learning Applied to Document Recognition. Proceedings of the IEEE, 86, 2278-2324. [Google Scholar] [CrossRef] 
[14] Mehta, S. and Rastegari, M. (2021) Mobilevit: Light-Weight, Gen-eral-Purpose, and Mobile-Friendly Vision Transformer. arXiv: 2110.02178.
[15] Venkataramanan, S., et al. (2023) Skip-Attention: Improving Vision Transformers by Paying Less Attention. arXiv: 2301.02240.
[16] Chollet, F. (2017) Xception: Deep Learning with Depthwise Separable Convolutions. Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, 21-26 July 2017, 1800-1807. [Google Scholar] [CrossRef] 
[17] Zhou, D.Q., et al. (2020) Rethinking Bottleneck Structure for Effi-cient Mobile Network Design. Computer Vision—ECCV 2020: 16th European Conference, Glasgow, 23-28 August 2020, 680–697. [Google Scholar] [CrossRef] 
[18] Howard, A., et al. (2019) Searching for Mobilenetv3. Pro-ceedings of 2019 IEEE/CVF International Conference on Computer Vision (ICCV), Seoul, 27 October 2019 - 2 Novem-ber 2019, 1314-1324. [Google Scholar] [CrossRef] 
[19] Howard, A., et al. (2018) Inverted Residuals and Linear Bottle-necks: Mobile Networks for Classification, Detection and Segmentation.
[20] Mehta, S., et al. (2019) ESPNetv2: A Light-Weight, Power Efficient, and General Purpose Convolutional Neural Network. Proceedings of 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Long Beach, 15-20 June 2019, 9182-9192. [Google Scholar] [CrossRef]