River tidal surge detection method based on audio characteristic recognition

Abstract

The invention relates to a river tidal bore detection method based on audio feature recognition and aims at overcoming the defects in the prior art that the forecast error is bigger and the real-time capability is unavailable. The invention firstly adopts a microphone arranged above water or a sonar arranged under water to receive the audio signals and then the received audio signals are processed with BPF; then the audio signals of a constant T duration are processed with DFT to obtain the one-dimensional spectrum sequence X(n) and then the audio energy average value of the tidal bore can be figured out; and if the audio energy average value of the tidal bore is more than a preset threshold value A, the existence of tidal bore can be defined and the alarming signal is sent out. The invention adopts the non-contact detection so as to avoid the condition that a sensor is directly arranged under water to cause damage. The real-time automatic detection method can be applied to the detection, forecast and early-alarming for river tidal bore and the accuracy can reach 90 percent.

Description

technical field

[0001] The invention belongs to the technical field of automation, and relates to a real-time automatic detection method of river tide surge based on audio recognition, which can be used for monitoring, forecasting and early warning of river tide surge. Background technique

[0002] The gravitational and centrifugal force difference between the earth and the moon will cause the change of sea level, which will form the tide rise and fall. However, in many strong tidal estuaries in the world, due to the special geographical structure, ferocious tides will be formed. The tides are fierce, fast, and turbulent (such as the wonder of the world formed by the shape of the trumpet mouth of Hangzhou Bay-the Qianjiang tide. and the Brazilian Amazon tide). Due to people's lack of understanding of river tide habits and the lack of accurate forecast of the time when tides reach various dangerous points along the river, it has brought certain destructive effects on people'...

Examples

Embodiment Construction

[0017] Such as figure 1 , for the tide recognition of the Qiantang River, the specific steps of the river tide detection method based on audio feature recognition are as follows:

[0018] 1. Receive underwater and above-water audio signals;

[0019] 2. Perform bandpass filtering on the signal, and intercept an audio signal between 2000 and 4000Hz;

[0020] 3. The signal is continuously collected for 15 seconds for sampling and Fourier transform. The sampling frequency is 44.1KHZ;

[0021] 4. Set the audio energy average threshold to 0.99;

[0022] 5. If the audio energy average (Average) is greater than 0.99, output an alarm message, otherwise return to step 1.

[0023] The calculation and analysis of data collected 20 times in Yanguan, Laoyancang, Cangqian, Qibao, etc. by the Qiantang River in 10 months shows that setting the threshold value to 0.9 has been able to use this method to detect the tide head above 30cm. tide.

technical field

[0001] The invention belongs to the technical field of automation, and relates to a real-time automatic detection method of river tide surge based on audio recognition, which can be used for monitoring, forecasting and early warning of river tide surge. Background technique

[0002] The gravitational and centrifugal force difference between the earth and the moon will cause the change of sea level, which will form the tide rise and fall. However, in many strong tidal estuaries in the world, due to the special geographical structure, ferocious tides will be formed. The tides are fierce, fast, and turbulent (such as the wonder of the world formed by the shape of the trumpet mouth of Hangzhou Bay-the Qianjiang tide. and the Brazilian Amazon tide). Due to people's lack of understanding of river tide habits and the lack of accurate forecast of the time when tides reach various dangerous points along the river, it has brought certain destructive effects on people'...

Examples

Embodiment Construction

[0017] Such as figure 1 , for the tide recognition of the Qiantang River, the specific steps of the river tide detection method based on audio feature recognition are as follows:

[0018] 1. Receive underwater and above-water audio signals;

[0019] 2. Perform bandpass filtering on the signal, and intercept an audio signal between 2000 and 4000Hz;

[0020] 3. The signal is continuously collected for 15 seconds for sampling and Fourier transform. The sampling frequency is 44.1KHZ;

[0021] 4. Set the audio energy average threshold to 0.99;

[0022] 5. If the audio energy average (Average) is greater than 0.99, output an alarm message, otherwise return to step 1.

[0023] The calculation and analysis of data collected 20 times in Yanguan, Laoyancang, Cangqian, Qibao, etc. by the Qiantang River in 10 months shows that setting the threshold value to 0.9 has been able to use this method to detect the tide head above 30cm. tide.

Images