Pipe gallery pipeline leakage recognition system and method based on voiceprint recognition

The invention provides a pipe gallery pipeline leakage recognition system and method based on voiceprint recognition. The system comprises a voiceprint data acquisition device, a deep learning system and a voiceprint real-time monitoring device. The voiceprint data acquisition system is mainly composed of a voiceprint sensor, a sensor power supply, a sound acquisition card and a recording host. The deep learning system mainly comprises a leaked voiceprint preprocessing module, a leaked voiceprint library and a voiceprint recognition learning module. The real-time monitoring system mainly comprises a sensor, a sensor power supply and an identification host. The method mainly comprises the following steps: acquiring problem voiceprints by using a voiceprint data acquisition device, training a voiceprint recognition model by using a deep learning system, and realizing full-coverage real-time monitoring of a pipe gallery pipeline by using a voiceprint real-time monitoring device. The invention provides a new iden