Foggy Weather Monitoring Method Based on Improved Deep Residual Shrinkage Network and Radio Signal

Abstract

Foggy weather can have a serious impact on production and life. The existing monitoring technology has problems such as high cost, difficult maintenance and low spatial and temporal resolution. In this paper, we propose a method for foggy weather monitoring using radio signals based on the principle that foggy weather affects radio signals. Combining wireless communication with deep learning, an improved deep residual shrinkage network is used to classify and identify the radio signals collected in different environments. First, radio signals from four different environments are collected, then, a wide convolutional layer is added to the deep residual shrinkage network, and then a CBAM attention mechanism is introduced after the wide convolutional layer to extract features more accurately. The acquired signals are fed into the improved deep residual shrinkage network for training. The final classification result reached 92.29%, which is a 6.11% improvement compared to the conventional ResNet50 algorithm. The results show the high potential of the method to monitor foggy weather more accurately.