【摘要】 高压输电线路上附着的各种异物是电力设备安全运行的一大隐患,及时检测与清理异物是保障电力系统稳定运行的关键。针对现有高压输电线路异物检测模型精度低、适用性差的问题,提出了一种改进的YOLOv5模型。首先,引入卷积注意力机制模块优化主干网络,增强模型抗干扰能力,提升模型检测准确性;其次,采用双向特征金字塔网络加强特征融合,丰富小目标的特征信息;最后,将SIOU作为损失函数,关注真实锚框与预测框之间的角度信息,进而提高目标检测模型的精度和鲁棒性。实验结果表明,改进方法的精确度、平均召回率较原始模型分别提升20.9%和15%,同时验证了模型在电力巡检中的实时性与轻量化优势。该方法可为高压输电线路智能巡检提供技术支撑。更多还原

【Abstract】 The various foreign objects attached to high-voltage transmission lines pose a significant hazard to the safe operation of electrical equipment.Timely detection and removal are essential to ensure the stability of the power system.Addressing the issues of low accuracy and poor applicability in existing models for detecting foreign objects on high-voltage power lines,this paper introduces an improved YOLOv5 model for such detection tasks.First,a convolutional attention mechanism module is selected to optimize the backbone network,enhancing its anti-interference ability and improving the accuracy of the model’ s detections.Second,using a bidirectional feature pyramid network strengthens feature fusion to enrich the feature information of small targets.Lastly,the SIOU loss function is employed,focusing on the angular information between the true anchor boxes and predictions,thus enhancing the precision and robustness of the object detection model.The experimental results show that,compared to the original model,the improved method achieves a 20.9% and 15% increase in recall rate and mean average precision,respectively,effectively reducing false alarms and missed detections,verifying the effectiveness of the improved method.更多还原