基于图像识别的接触器接线缺陷检测
Contactor Wiring Defect Detection Based on Image Recognition
DOI: 10.12677/JISP.2024.131006, PDF,   
作者: 苏 煜：华东理工大学信息科学与工程学院，上海；华东理工大学机械与动力工程学院，上海；王卓涵, 顾玲榕, 常 虹：华东理工大学信息科学与工程学院，上海
关键词: 缺陷检测；图像检测；自编码器；支持向量机；树莓派；Defect Detection； Image Detection； Self-Encoder； Support Vector Machine； Raspberry Pi
摘要: 接触器是自动控制系统中的重要元件之一，负责接通和切断电路，起到低电压释放保护作用。考虑到控制系统的复杂性，若存在接线故障的接触器被投入使用，不仅有出现电力事故的风险，也增加了检修难度。因此，接触器生产制造过程中对其接线正确性的检测工作尤为重要。传统的器件表面缺陷检测主要是以人工检验和物理损伤检测为主，检验过程耗时费力，占用大量人力资源。为解决上述问题，本文采用SVM算法建立基于图像识别的接触器接线缺陷检测模型，同时，在检测模型前增加自编码器检测异常图像，提高SVM检测模型的识别率。通过与KNN、决策树等传统模型的检测结果对比，验证了本文检测模型的识别效果达到100%的准确率。为满足实际工程需求，本文将建好的缺陷检测模型部署至树莓派系统，实现工程应用的小型化和便利化。
Abstract: The contactor is one of the most important components of the automatic control system, which is responsible for switching the circuit on and off and plays the role of low-voltage trip protection. Given the complexity of the control system, a contactor with wiring faults not only poses a risk of electrical accidents, but also makes maintenance more difficult. It is therefore particularly important to check the correctness of the wiring when the contactors are manufactured. Conventional surface defect detection is mainly based on manual inspection and physical damage detection, and the inspection process is time-consuming, labor-intensive and requires a lot of manpower. In order to solve the above problems, an SVM algorithm is used in this work to build an image recognition-based model for detecting defects in contactor wiring. By comparing the detection results with those of traditional models such as KNN and decision tree, it is proved that the detection effect of the detection model in this paper achieves 100% accuracy. In order to meet the actual technical requirements, the developed fault detection model is used on the Raspberry Pi system in this paper to achieve the miniaturization and convenience of technical applications.
文章引用：苏煜, 王卓涵, 顾玲榕, 常虹. 基于图像识别的接触器接线缺陷检测[J]. 图像与信号处理, 2024, 13(1): 59-68. https://doi.org/10.12677/JISP.2024.131006

参考文献

[1] 国务院. “十四五”推进国家政务信息化规划[EB/OL]. https://www.gov.cn/zhengce/zhengceku/2022-01/06/content_5666746.htm, 2021-12-24.
[2] 毕德林. 接触器发展方向初探[J]. 机电元件, 2021, 41(1): 57-60.
[3] Song, K. and Yan, Y. (2013) A Noise Robust Method Based on Com-pleted Local Binary Patterns for Hot-Rolled Steel Strip Surface Defects. Applied Surface Science, 285, 858-864. [Google Scholar] [CrossRef] 
[4] Niskanen, M. and Silvn, O. (2002) Real-Time Aspects of SOM-Based Visual Surface Inspection. Proceedings of SPIE - The International Society for Optical Engineering, Sin-gapore, 19-25 January 2002, 123-134. [Google Scholar] [CrossRef] 
[5] Cortes, C. and Vapnik, V. (1995) Support-Vector Networks. Machine Learning, 20, 273-297. [Google Scholar] [CrossRef] 
[6] Cover, T.M. and Hart, P.E. (1953) Nearest Neighbor Pattern Classifica-tion. IEEE Transactions on Information Theory, 13, 21-27. [Google Scholar] [CrossRef] 
[7] Hunt, E.B., Marin, J. and Stone, P.J. (1966) Experiments in Induction. America Press, New York.
[8] Simonyan, K. and Zisserman, A. (2014) Very Deep Convolutional Networks for Large-Scale Image Recognition. Computer Science, 9, 1409-1556.
[9] Girshick, R. (2015) Fast R-CNN. 2015 IEEE International Conference on Computer Vision (ICCV), Santiago, 7-13 December 2015, 169. [Google Scholar] [CrossRef] 
[10] 刘颖, 等. 基于小样本学习的图像分类技术综述[J]. 自动化学报, 2021, 47(2): 297-315.
[11] Schwartz, E., et al. (2018) Delta Encoder: An Ef-fective Sample Synthesis Method for FewShotObject Recognition. Advances in Neural Information Processing Systems, 31, 2850-2860.
[12] 葛轶洲, 等. 小样本困境下的深度学习图像识别综述[J]. 软件学报, 2022, 33(1): 193-210.
[13] Ajzerman, M.A., Braverman, E.M. and Rozonoehr, L.I. (1964) Theoretical Foundations of the Potential Function Method in Pattern Recognition Learning. Automation and Remote Control, 25, 917-936.
[14] Broomhead, D.S. and Lowe, D. (1988) Multivariable Functional Interpolation and Adaptative Networks. Complex Systems, 2, 321-355.
[15] 袁非牛, 章琳, 史劲亭, 等. 自编码神经网络理论及应用综述[J]. 计算机学报, 2019, 42(1): 203-230.
[16] Kasun, Liyanaarachchi Lekamalage Chamara, et al. (2016) Dimension Reduction with Extreme Learning Machine. IEEE Transactions on Image Processing, 25, 3906-3918. [Google Scholar] [CrossRef]