Author:
Liang Cuiping,Li Yang,Wei Yuyun,Zhang Dan,Wang Ying,Jiang Suiping
Abstract
Abstract
The number of times that equipment can emit is limited, and exceeding the usage limit will affect its shooting accuracy. Existing research mainly focuses on the classification and detection of emitting times, with few studies focusing on the recognition of emitting times. This paper combines multi-scale feature extraction with the CRNN model and proposes an improved CRNN emission frequency recognition model. After experimental verification, the improved model has a high prediction accuracy on high-quality emitted sound data, with an accuracy of over 95.2%. The accuracy of the data in the presence of strong interference is 88.6%. The experimental results show that the method proposed in this paper is effective.
Reference10 articles.
1. Automatic Detection and Type Recognition of Large Caliber Artillery Sound [J];Ren;Journal of Ordnance Equipment Engineering,2006
2. Deep Transfer Learning for Automated Artillery Crater Classification [J];Sani,2021
3. Decoupling Temporal Convolutional Networks Model in Sound Event Detection and Localization [J];Song,2023