Indoor anti-theft monitoring system based on sound

The invention discloses an indoor anti-theft monitoring system based on sound. The system comprises a comprehensive research and judgment subsystem, an early warning subsystem, a cloud server and an information query subsystem. The comprehensive research and judgment subsystem comprises a sound collector, an embedded computer, a camera and an intelligent detection module; the early warning subsystem comprises an automatic alarm sub-module, a loudspeaker and a 110 alarm desk; the sound collector collects sound signals; the embedded computer preprocesses the sound signals to form a sound sample; the sound sample and an image or video collected by the camera are uploaded to an intelligent detection sub-module for comprehensive identification and judgment, and a result is fed back; and the cloud server is used for storing the processing result uploaded by the intelligent detection sub-module and uploading the result to the automatic alarm sub-module for result comparison. The deep learning algorithm is combined w