基于机器视觉的高压隔离开关设备状态判别与故障诊断技术

State Discrimination and Failure Recognition Technology of High Voltage Disconnector Equipment Based on Machine Vision

文章对YOLO v3算法模型进行了优化处理,提高了该算法模型的使用效果。分析该模型的损失曲线以及交并比(intersection over union,IOU)曲线可知,随着批次值的逐渐增加,该模型损失值逐渐减少,基本维持在0.2左右,IOU的值则慢慢趋向于1,表明预测值与实际值的重合度逐渐升高。通过对比平均精度可知,未优化的YOLO v3算法模型的平均精度为82.54%,优化后YOLO v3算法模型的平均精度为90.36%,表明通过优化处理可以大幅提高YOLO v3模型的精度,优化后YOLO v3模型具有较好的使用效果,研究结果可为后续高压隔离开关智能化的相关研究提供科学的参考。

The study optimizes the algorithm model of YOLO v3 and improves the application effect of the algorithm model.The analysis of loss curve and intersection over union(IOU)curve of the model suggests that with the gradual increase of the batch value,the loss value of the model gradually decreases,basically maintains at about 0.2,and the IOU value slowly tends to 1.It indicates that the coincidence degree between the predicted value and the actual value gradually increases.By comparing the average precisions,it is found that the average precision of the unoptimized model is 82.54%,and the average precision of the optimized model is 90.36%.It reveals that the precision of the YOLO v3 model can be greatly improved,and the optimized YOLO v3 model has excellent application effect.The study can provide scientific and effective reference for the follow-up research of high voltage disconnector intellectualization.