Pipe leak detection method and device

Abstract

The invention provides a pipeline leak detection method and device. The method comprises the following steps of: performing time-frequency analysis on a pipeline mixed signal based on the short-time Fourier transform to obtain a time-frequency spectrum to which the mixed signal correspond; converting the time-frequency spectrum into a two-dimensional time-frequency image; and inputting the time-frequency image into a preset convolutional neural network model, and obtaining a leak type to which the mixed signal corresponds according to an output result of the convolutional neural network model.The time-frequency analysis of the pipeline mixed signal is performed by the short-time Fourier transform to obtain the time-frequency spectrum to which the mixed signal correspond without extractinga leak signal from the mixed signal, and thus, the interference of mixed noise is avoided. In addition, the time-frequency spectrum is converted into the two-dimensional time-frequency image; the two-dimensional time-frequency image is input to the preset convolutional neural network; the preset neural network model is obtained after training according to a mixed signal sample with the leak type,and can output the recognition result of the leak type; and therefore, the detection process is fast and accurate.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of signal processing, and in particular, to a method and device for detecting pipeline leakage. Background technique [0002] Water supply pipelines play an extremely critical role in long-distance water resources transportation, but because most of the pipelines are laid in harsh natural environments such as the wild or underground, they are easily damaged. Once the water supply pipeline is damaged and leaks occur, timely maintenance of the pipeline is required to avoid additional economic losses. Therefore, the rapid detection of pipeline leakage has great application value, for example, it can provide program guidance for pipeline maintenance. [0003] At present, for the signal detection of pipeline leakage in the water supply network, most of the time-frequency analysis and signal processing methods are used to extract information about the leakage signal from the mixed signals ob...

AI Technical Summary

Problems solved by technology

However, due to the low signal-to-noise ratio in the pipeline network, the mixed signal is often mixed with stray interference signals and noises, resulting in an extremely weak leakage signal. At the same time, the leakage signal may only exist within a certain period of time. Completely remove strong noise and interference signals to obtain the leak signal to be detected

In addition, when the signal-to-noise ratio changes greatly, the traditional time-frequency distribution has huge limitations in non-stationary signal processing, and the reconstructed signal still cannot remove the influence and interference of mixed noise, resulting in the extracted leakage signal corresponding to The detection and classification results of the

In summary, the current pipeline leakage detection method is difficult to extract accurate leakage signals due to the influence of noise and interference, resulting in unreliable detection results

Images