基于深度学习的高压电缆接头缺陷检测方法

Deep learning based defect detection method for high-voltage cable joint

反应力锥是高压电缆接头的关键部位,其质量直接关系到整个电力系统的安全与稳定。文章制作了包含843张图像的反应力锥数据集,研究了基于深度学习的缺陷检测方法。实验结果表明,采用ResNet-18网络和多种数据增广方法能使反应力锥缺陷检测的精度达到95.7%,可满足实际应用的需求。

The reactive cone is a critical component of high-voltage cable joints,and its quality directly affects the safety and stability of the entire power system.The article created a reaction cone dataset containing 843 images and studied defect detection methods based on deep learning.The experimental results show that using the ResNet-18 network and various data augmentation methods can achieve an accuracy of 95.7%in detecting reaction cone defects,which can meet the needs of practical applications.