辽宁石油化工大学学报 ›› 2024, Vol. 44 ›› Issue (4): 82-90.DOI: 10.12422/j.issn.1672-6952.2024.04.011

• 信息与控制工程 • 上一篇    下一篇

基于改进YOLOv7的管道缺陷检测技术研究

冯丽丹1(), 王闯2, 祁军1, 石元博2()   

  1. 1.辽宁石油化工大学 信息与控制工程学院,辽宁 抚顺 113001
    2.辽宁石油化工大学 人工智能与软件学院,辽宁 抚顺 113001
  • 收稿日期:2023-06-17 修回日期:2023-10-26 出版日期:2024-08-25 发布日期:2024-08-06
  • 通讯作者: 石元博
  • 作者简介:冯丽丹(1999⁃),女,硕士研究生,从事目标检测方面的研究;E⁃mail:2797202821@qq.com
  • 基金资助:
    辽宁省教育厅科研项目(LJKMZ20220737)

Research on Pipeline Defect Detection Technology Based on Improved YOLOv7

Lidan FENG1(), Chuang WANG2, Jun QI1, Yuanbo SHI2()   

  1. 1.School of Information and Control Engineering,Liaoning Petrochemical University,Fushun Liaoning 113001,China
    2.School of Artificial Intelligence and Software,Liaoning Petrochemical University,Fushun Liaoning 113001,China
  • Received:2023-06-17 Revised:2023-10-26 Published:2024-08-25 Online:2024-08-06
  • Contact: Yuanbo SHI

摘要:

针对城镇地下管网规模巨大、传统的人工检测方法已不能满足现在工程需求的问题,提出采用MobileNetv3?YOLOv7网络模型作为地下管道缺陷目标检测的算法来提升检测的精度和速度。首先,管道图像数据集进行预处理,对输入图像灰度化及重采样,均衡样本的数量;其次,将轻量化网络MobileNetv3和YOLOv7网络框架相结合,增加BiFPN特征金字塔结构以提高精确度;然后,在数据处理方面通过Mosaic数据增强方式提高该模型的鲁棒性;最后,设计YOLOv7网络模型的对比实验验证本模型的可行性。在Pytorch实验框架下,对MobileNetv3?YOLOv7网络模型进行了验证。实验结果表明,该模型可减少参数计算量,并且平均准确率有所提高。

关键词: 图像处理, 地下管道缺陷, 目标检测, YOLOv7网络, 轻量化网络

Abstract:

In view of the huge scale of urban underground pipe network, the traditional manual detection method can no longer meet the needs of the current projects. In this paper, the MobileNetv3?YOLOv7 network model is proposed as the algorithm for target detection of underground pipeline defects to improve the accuracy and speed of detection. First, the pipeline image dataset is preprocessed, and the input image is grayscale and resampled to balance the number of samples. Secondly, the lightweight network MobileNetv3 and YOLOv7 network frameworks are combined to increase the BiFPN feature pyramid structure to improve accuracy. Then, in terms of data processing, Mosaic data augmentation is used to improve the robustness of the model. Finally, a comparative experiment with the YOLOv7 network model is designed to verify the feasibility of the model. In this paper, the MobileNetv3?YOLOv7 network model is verified under the framework of Pytorch experiment, and the experimental results show that the model greatly reduces the amount of parameter calculation and improves the average accuracy.

Key words: Image processing, Underground pipeline defects, Object detection, YOLOv7 network, Lightweight network

中图分类号: 

引用本文

冯丽丹, 王闯, 祁军, 石元博. 基于改进YOLOv7的管道缺陷检测技术研究[J]. 辽宁石油化工大学学报, 2024, 44(4): 82-90.

Lidan FENG, Chuang WANG, Jun QI, Yuanbo SHI. Research on Pipeline Defect Detection Technology Based on Improved YOLOv7[J]. Journal of Liaoning Petrochemical University, 2024, 44(4): 82-90.

使用本文

0
    /   /   推荐

导出引用管理器 EndNote|Ris|BibTeX

链接本文: https://journal.lnpu.edu.cn/CN/10.12422/j.issn.1672-6952.2024.04.011

               https://journal.lnpu.edu.cn/CN/Y2024/V44/I4/82