基于改进型YOLOv5算法的偏振片缺陷识别研究
Research on Polarizer Defect Recognition Based on Improved YOLOv5 Algorithm
摘要: 偏振片作为TFT-LCD的关键部件之一，其表面缺陷严重影响液晶显示器的成像质量。为了实现偏振片缺陷智能化在线检测从而替代目前因人眼检测导致效率低的问题，提出了一种改进型YOLOv5检测算法，即在Backbone层增加了CBAM注意力机制；在Prediction层增加了一个新的输出层；将传统边框回归损失函数改为CIOU_Loss。通过缺陷样本测试实验表明，改进型算法尽管增加了一个输出层导致参数增加，且FPS略微降低，但mAP却提升了4个百分点，并且检测的最高置信度达到了0.93。故改进型YOLOv5算法增强了缺陷目标识别精度和准确度。
Abstract: Polarizer is one of the key components of TFT-LCD. Its surface defects seriously affect the imaging quality of LCD. In order to realize the intelligent on-line detection of polarizer defects and replace the current problem of low efficiency caused by human eye detection, an improved YOLOv5 detection algorithm is proposed, that is, CBAM attention mechanism is added in the backbone layer; A new output layer is added in the prediction layer; The traditional border regression loss function is changed to CIOU_Loss. The defect sample test experiment shows that although the improved algorithm adds an output layer, resulting in an increase in parameters and a slight decrease in FPS, the map increases by 4 percentage points, and the highest confidence of detection reaches 0.93. Therefore, the improved YOLOv5 algorithm enhances the accuracy and accuracy of defect target recognition.
文章引用：贾晓斌, 罗柏文, 金祝红. 基于改进型YOLOv5算法的偏振片缺陷识别研究[J]. 计算机科学与应用, 2025, 15(4): 113-123. https://doi.org/10.12677/csa.2025.154084

参考文献

[1] Sohn, S., Lee, D., Choi, H., Suh, J. and Bae, H. (2007) Detection of Various Defects in TFT-LCD Polarizing Film. In: Beliczynski, B., Dzielinski, A., Iwanowski, M. and Ribeiro, B., Eds., Adaptive and Natural Computing Algorithms. ICANNGA 2007. Lecture Notes in Computer Science, Springer, 534-543. [Google Scholar] [CrossRef] 
[2] 曾小星. 基于结构光的偏光片外观缺陷检测技术研究[D]: [硕士学位论文]. 深圳: 深圳大学, 2016.
[3] 赖文威. 偏光片外观缺陷成像机理与检测技术研究[D]: [硕士学位论文]. 深圳: 深圳大学, 2017.
[4] 许少鹏. 基于机器视觉的偏光片缺陷检测技术研究[D]: [硕士学位论文]. 深圳: 深圳大学, 2018.
[5] 贺健. 偏光片外观缺陷成像仿真与检测[D]: [硕士学位论文]. 深圳: 深圳大学, 2016.
[6] 李柯泉, 陈燕, 刘佳晨, 牟向伟. 基于深度学习的目标检测算法综述[J]. 计算机程, 2022, 48(7): 1-17.
[7] 赵睿, 刘辉, 刘沛霖, 雷音, 李达. 基于改进YOLOv5s的安全帽检测算法[J]. 北京航空航天大学学报, 2023, 49(8): 2050-2061.
[8] 王静, 孙紫雲, 郭苹, 张龙妹. 改进YOLOv5的白细胞检测算法[J]. 计算机工程与应用, 2022, 58(4): 134-142.
[9] Hou, Q., Zhou, D. and Feng, J. (2021) Coordinate Attention for Efficient Mobile Network Design. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Nashville, 20-25 June 2021, 13708-13717. [Google Scholar] [CrossRef] 
[10] 李春霖, 谢刚, 王银, 谢新林, 刘瑞珍. 基于YOLOv3-Tiny-D算法的偏光片缺陷检测[J]. 计算机集成制造系统, 2022, 28(3): 787-797.
[11] Pan, X., Ge, C., Lu, R., Song, S., Chen, G., Huang, Z., et al. (2022) On the Integration of Self-Attention and Convolution. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, 18-24 June 2022, 805-815. [Google Scholar] [CrossRef] 
[12] Chen, C., Liu, M., Tuzel, O. and Xiao, J. (2017) R-CNN for Small Object Detection. In: Lai, S.H., Lepetit, V., Nishino, K. and Sato, Y., Eds., Computer Vision—ACCV 2016. Lecture Notes in Computer Science, Springer International Publishing, 214-230. [Google Scholar] [CrossRef] 
[13] Krizhevsky, A., Sutskever, I. and Hinton, G.E. (2017) Imagenet Classification with Deep Convolutional Neural Networks. Communications of the ACM, 60, 84-90. [Google Scholar] [CrossRef] 
[14] LeCun, Y., Boser, B., Denker, J.S., Henderson, D., Howard, R.E., Hubbard, W., et al. (1989) Backpropagation Applied to Handwritten Zip Code Recognition. Neural Computation, 1, 541-551. [Google Scholar] [CrossRef] 
[15] Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, Ł. and Polosukhin, I. (2017) Attention Is All You Need. In: Guyon, I., Von Luxburg, U., Bengio, S., Wallach, H., Fergus, R., Vishwanathan, S. and Garnett, R., Eds., Advances in Neural Information Processing Systems. [Google Scholar] [CrossRef] 
[16] Niu, Z., Zhong, G. and Yu, H. (2021) A Review on the Attention Mechanism of Deep Learning. Neurocomputing, 452, 48-62. [Google Scholar] [CrossRef] 
[17] Cordonnier, J.B., Loukas, A. and Jaggi, M. (2019) On the Relationship between Self-Attention and Convolutional Layers.
[18] Bochkovskiy, A., Wang, C.Y. and Liao, H.Y.M. (2020) YOLOv4: Optimal Speed and Accuracy of Object Detection.
[19] Wang, C., Mark Liao, H., Wu, Y., Chen, P., Hsieh, J. and Yeh, I. (2020) CSPNet: A New Backbone That Can Enhance Learning Capability of CNN. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Seattle, 14-19 June 2020, 1571-1580. [Google Scholar] [CrossRef] 
[20] Hu, J., Shen, L. and Sun, G. (2018) Squeeze-and-Excitation Networks. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, 18-23 June 2018, 7132-7141. [Google Scholar] [CrossRef] 
[21] Woo, S., Park, J., Lee, J. and Kweon, I.S. (2018) CBAM: Convolutional Block Attention Module. In: Ferrari, V., Hebert, M., Sminchisescu, C. and Weiss, Y., Eds., Computer Vision—ECCV 2018. Lecture Notes in Computer Science, Springer International Publishing, 3-19. [Google Scholar] [CrossRef] 
[22] Zheng, Z., Wang, P., Liu, W., et al. (2019) Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression. Proceedings of the AAAI Conference on Artificial Intelligence, 34, 12993-13000.