Detection Method for Micro Displacement State of Multimode Fiber Based on CBAM
Sensing technology has been widely applied in traditional industries,but due to external interference,it is slightly insufficient in terms of small target detection and high-precision requirements.A spot recognition method based on the YOLOv4 tiny CBAM model structure is proposed to achieve the detection of small displacement states in multimo-de optical fibers.The correspondence between the degree of fiber bending and the output spot is used to collect spot images of multi-mode optical fibers under different displacement conditions,and combine with the convolutional block attention module(CBAM)for deep learning network training.The experimental results show that this method improves the average accuracy of the model from 42%of the original model to 82.95%without pre-training,achieving efficient identification of small displacement states of multimode optical fibers.
NETL
NSTL
万方数据
