Abnormal sound detection method based on edge cloud intelligent architecture

Abstract

An abnormal sound detection method based on an edge cloud intelligent architecture is provided. The method comprises the following steps: collecting audio data on an edge end; deploying tasks that canbe processed by the edge end to an edge device as much as possible; using the Docker container technology to perform encapsulation on task processing operators on a cloud end to realize migration ofcomputing tasks, and storing an audio detection result; using a deep neural network model to perform abnormal sound determination; and performing message communication between different devices through the MTZT protocol. According to the method provided by the present invention, the pressures on the cloud computing center and the network bandwidth are alleviated, the system real-time performance and responsiveness are improved, and data security can be better protected.

Description

technical field [0001] The invention belongs to the technical field of edge-cloud intelligence, and in particular relates to an abnormal sound detection method based on the edge-cloud intelligence architecture. Background technique [0002] In recent years, with the development of audio and video surveillance, audio surveillance systems have been widely used in urban public safety management. Audio detection and classification has become an important research direction in the research field of computational auditory scene analysis. With the further application of the audio monitoring system based on sound event detection and classification, some new problems also arise. Applications in the field of public safety require audio surveillance systems to provide real-time and efficient audio processing capabilities. The built-in computing power of the audio collection equipment of the traditional audio monitoring system is low, while the intelligent processing ability of the ex...

AI Technical Summary

Problems solved by technology

[0004] In order to overcome the deficiencies of the above-mentioned prior art, the purpose of the present invention is to provide an abnormal sound detection method based on the edge-cloud intelligent architecture, which solves the problem of high storage and network pressure and time-consuming transmission in the traditional audio detection system centered on cloud computing. Problems such as height extension and data security are difficult to guarantee. This method can make full use of the computing resources of edge devices and clouds. By deploying tasks to a greater extent on devices close to the data collection end, resources are allocated more reasonably and eased. It relieves the pressure on the cloud computing center and network bandwidth, improves the real-time and responsiveness of the system, and at the same time makes the data security better protected

Images