Abnormal sound detection method and abnormal sound detection device

Abstract

The invention provides an abnormal sound detection method, which comprises the following steps: carrying out KL divergence comparison on audio characteristic parameters to be detected and standard audio characteristic parameters in a normal audio characteristic standard library, and judging whether abnormal sound needs to be accurately judged or not; if the difference between the KL divergence ofthe to-be-tested audio characteristic parameter and the KL divergence of the standard audio characteristic parameter is smaller than a preset threshold, outputting a result that a to-be-tested audio signal is a normal audio signal; and if the difference between the KL divergence of the to-be-detected audio characteristic parameter and the KL divergence of the standard audio characteristic parameter is greater than a preset threshold, performing abnormal sound accurate judgment. Accurate judgment of abnormal sound is a classifier based on a double-layer structure, and type judgment of abnormalsound signals can be realized. Meanwhile, the invention further provides an abnormal sound detection device. According to the abnormal sound detection method and the abnormal sound detection device provided by the invention, whether the audio signal to be detected is a normal audio signal or an abnormal audio signal can be quickly judged in real time.

Description

technical field [0001] The invention relates to the field of audio signal processing, in particular to an abnormal sound detection method and an abnormal sound detection device. Background technique [0002] With the improvement of technology in the field of production and manufacturing, more and more large-scale equipment can enter the production workshop. But because of its high precision, high complexity, high maintenance costs, and unconventional downtime, it causes huge economic losses. Therefore, it is necessary to monitor the operating status of the current equipment in real time through an automated method. Traditional sensor-based and video monitoring are widely used at present, but usually the sensor fails before the equipment due to various failures, and the greater significance of video monitoring lies in the look-back inspection after the equipment fails. The method based on abnormal sound detection can capture the abnormal sound produced by the equipment due ...

AI Technical Summary

Problems solved by technology

However, in practical applications, the classification method has a typical shortcoming. It is necessary to use the classifier combined with the training data set for relevant training. In the actual link, it is difficult to obtain the abnormal sound caused by the abnormal state of the equipment. , and multiple types of abnormal data need to be used before training, which increases the difficulty of the problem under objective conditions

Images