
Applied Sciences, Journal Year: 2024, Volume and Issue: 14(24), P. 11852 - 11852
Published: Dec. 18, 2024
In this paper, technical code analysis and recognition of the defect signal suppression rigging based on a convolutional neural network are carried out given difficulty low rate detection rigging. Firstly, magnetic induction defects is collected using CM-801 (Anshan, China), Kalman filtering used to screen pre-process data, noise reduction data presented in form cloud image. The pressed set constructed, region broken wire stress image calibrated. single-stage object algorithm YOLOv5 (You Only Look Once) model calculation used, scale layer positioning loss function improved optimized, for experiments. experimental results show that accuracy convolution can reach 97.1%, which effectively identify suppressed
Language: Английский