Multi-task fixture defect detection algorithm based on channel attention
To discover the broken fixtures of railway communication cable in time and guarantee the safety operation for the railway,this paper proposes a multi-task fixture defect detection algorithm based on channel attention.Based on YOLOv7-tiny,we embed a Channel Attention Module(CAM)to balance the importance of different channels.In order to achieve state detection of fixtures,we add an additional state loss in the class loss function and embed Focal loss to overcome imbalance problem.To verify the effectiveness of the algorithm,experiments were conducted on real-world dataset containing four kinds of fixtures.The results have shown that our algorithm has a high performance and rapid speed,which can meet practical requirements.
万方数据
维普
