We present the field trial of an innovative neural network and DAS-based technique, employing a pre-trained CNN fine-tuning strategy for effective rain detection and classification within two practical scenarios.