Page 111 - 《应用声学》2022年第4期

第 41 卷 第 4 期 韩鹏程等： 卷积神经网络在气体泄漏超声识别中的应用 609

drocarbon leakage[J]. Journal of Hazardous Materials, [12] Krizhevsky A, Sutskever I, Hinton G E. ImageNet classi-
2003, 102(1): 13–28. fication with deep convolutional neural networks[J]. Com-
[7] 李家琨, 金伟其, 王霞, 等. 气体泄漏红外成像检测系统的性能 munications of the ACM, 2017, 60(6): 84–90.
测试方法研究 [J]. 北京理工大学学报, 2016, 36(6): 630–634. [13] Szegedy C, Liu W, Jia Y, et al. Going deeper with convo-
Li Jiakun, Jin Weiqi, Wang Xia, et al. Research on perfor- lutions[C]//Proceedings of the IEEE Conference on Com-
mance measurement method of gas leak infrared imaging puter Vision and Pattern Recognition, 2015: 1–9.
detection system[J]. Transactions of Beijing Institute of [14] 宁方立, 韩鹏程, 段爽, 等. 基于改进 CNN 的阀门泄漏超声信
Technology, 2016, 36(6): 630–634. 号识别方法 [J]. 北京邮电大学学报, 2020, 43(3): 38–44.
Ning Fangli, Han Pengcheng, Duan Shuang, et al. Re-
[8] Mostafapour A, Davoodi S. Leakage locating in un-
search on identification method of valve leakage ultra-
derground high pressure gas pipe by acoustic emission
sonic signal based on improved CNN[J]. Journal of Beijing
method[J]. Journal of Nondestructive Evaluation, 2013,
University of Posts and Telecommunications, 2020, 43(3):
32(2): 113–123.
38–44.
[9] Xiao R, Hu Q, Li J. Leak detection of gas pipelines using
[15] Reverdy P, Leonard N E. Parameter estimation in soft-
acoustic signals based on wavelet transform and support
max decision-making models with linear objective func-
vector machine[J]. Measurement, 2019, 146: 479–489.
tions[J]. IEEE Transactions on Automation Science and
[10] Goodfellow I, Bengio Y, Courville A. Deep learning[M]. Engineering, 2015, 13(1): 54–67.
Cambridge: MIT Press, 2016. [16] Kline D M, Berardi V L. Revisiting squared-error and
[11] Cai M, Liu J. Maxout neurons for deep convolutional and cross-entropy functions for training neural network clas-
LSTM neural networks in speech recognition[J]. Speech sifiers[J]. Neural Computing & Applications, 2005, 14(4):
Communication, 2016, 77(C): 53–64. 310–318.