An Audio-based Precise Positioning Method for UAV Belt Line Inspection

Abstract

The present invention discloses an audio-based precise positioning method for UAV belt line inspection, including: S1, using a pickup to collect audio; S2, preprocessing the audio signal through a frequency calculation module on the UAV; S3, calculating the frequency of any frequency point in any frame of the audio frequency domain signal; S4, receiving the preprocessed data and performing audio signal recognition; S5, using the data features as the learning target; S6, when an abnormal sound signal is collected, recording the fault at regular intervals; S7, when flying to a position where the abnormal signal cannot be collected, determining the point with the maximum signal amplitude of the abnormal signal; S8, according to the flight speed of the UAV and the recorded time points; The present invention uses the characteristics of audio to detect abnormal points on the belt conveyor, and accurately calculates the positions of the abnormal points according to the flight speed of the UAV and the recorded time points, which can effectively overcome the drawbacks of the UAV line inspection system based on image detection.

Description

Technical Field

[0001] The present invention relates to the technical field of belt line inspection, and particularly to a precise positioning method for an unmanned aerial vehicle (UAV) belt line inspection based on audio. Background Art

[0002] Belt conveyors have strong conveying capacity, long conveying distance, simple structure and easy maintenance, and can be conveniently programmed and automated. They use the continuous or intermittent movement of the conveyor belt to transport items, running at high speed, smoothly, with low noise, and can convey up and down slopes.

[0003] In an unattended UAV inspection system for belt conveyors, some faults cannot be monitored and confirmed through video due to reasons such as line of sight obstruction. Using sound for positioning cannot accurately determine the fault location. Therefore, those skilled in the art have provided a precise positioning method for UAV belt line inspection based on audio to solve the problems raised in the above background art. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides a precise positioning method for UAV belt line inspection based on audio, including:

[0005] S1. Use a pickup on the UAV to collect audio signals generated during the operation of the belt conveyor along the way to obtain the audio signals therein;

[0006] S2. Preprocess the audio signals through a frequency calculation module on the UAV;

[0007] S3. A frequency calculation module is used to calculate the frequency of any frequency point in any frame of the audio frequency domain signal;

[0008] S4. Receive the preprocessed data and perform audio signal recognition;

[0009] S5. Use the data features as learning objectives, and through machine learning, combined with the fuzzy clustering method, judge and classify abnormal signals;

[0010] S6. When abnormal sound signals are collected, record the faults at intervals (which can be set), including abnormal signal types, signal amplitudes, recording times, etc.;

[0011] S7. When flying to a position where abnormal signals cannot be collected or the amplitude of abnormal signals significantly decreases (such as dropping below 50% of the maximum value), determine the point with the largest signal amplitude of the abnormal signal;

[0012] S8. Accurately calculate the position of the abnormal point according to the flight speed of the UAV and the recorded time points.

[0013] Preferably, in the step S2, the preprocessing of the audio signal includes, but is not limited to, filtering, noise reduction, and removing singular values.

[0014] Preferably, the audio signal of the UAV flight is removed by the subtraction method, and the noise interference signal is removed by filtering. The purpose is to retain the operation signal and fault signal of the belt conveyor through preprocessing and simplify the later processing process.

[0015] Preferably, in the step S3, the frequency calculation can use conventional algorithms including, but not limited to, Fourier transform, wavelet transform, etc.

[0016] Preferably, in the step S4, detection methods including, but not limited to, spectrum method, peak detection method, etc. are used to identify abnormalities in the audio signal during operation.

[0017] Preferably, in the step S5, the artificial intelligence algorithm (such as neural network) is trained by collecting abnormal signals. According to the training results, the artificial intelligence algorithm classifies and preliminarily determines which type of fault it belongs to and whether to perform an emergency stop, etc.

[0018] Preferably, in the step S6, the fault recording time interval is 1 minute.

[0019] The technical effects and advantages of the present invention:

[0020] The present invention uses the audio characteristics to detect abnormal points on the belt conveyor. According to the flight speed of the UAV and the recorded time points, the positions of the abnormal points are accurately calculated, which can effectively overcome the drawback that the UAV line inspection system based on image detection cannot accurately perform line inspection work due to line of sight occlusion. Specific embodiments

[0021] The present invention will be further described in detail below in conjunction with specific embodiments. The embodiments of the present invention are given for the purpose of illustration and description, and are not exhaustive or limit the present invention to the disclosed form. Many modifications and variations are obvious to those of ordinary skill in the art. The embodiments are selected and described to better illustrate the principles and practical applications of the present invention, and enable those of ordinary skill in the art to understand the present invention and thus design various embodiments with various modifications suitable for specific purposes.

[0022] Embodiment 1

[0023] In this embodiment, a method for precise positioning of UAV belt line inspection based on audio is provided, including:

[0024] S1. Use the pickup on the UAV to collect the audio signal generated during the operation of the belt conveyor along the way.

[0025] S2. Preprocess the audio signal through the frequency calculation module on the drone;

[0026] S3. A frequency calculation module is used to calculate the frequency of any frequency point in any frame of the audio frequency domain signal;

[0027] S4. Receive the preprocessed data for audio signal recognition;

[0028] S5. Use the data features as the learning target, and through machine learning, combined with the fuzzy clustering method, judge and classify abnormal signals;

[0029] S6. When an abnormal sound signal is collected, record the fault at regular intervals (settable), including the abnormal signal type, signal amplitude, recording time, etc.

[0030] S7. When flying to a position where the abnormal signal cannot be collected or the amplitude of the abnormal signal decreases significantly (such as dropping below 50% of the maximum value), determine the point with the largest signal amplitude of the abnormal signal;

[0031] S8. Accurately calculate the abnormal point position according to the flight speed of the drone and the recording time point.

[0032] In step S2, the preprocessing of the audio signal includes but is not limited to filtering, noise reduction, and removing singular values.

[0033] Use the subtraction method to remove the audio signal of the drone flight, and use filtering to remove the noise interference signal. The purpose is to retain the belt conveyor operation signal and fault signal through preprocessing and simplify the later processing process.

[0034] In step S3, frequency calculation can use conventional algorithms including but not limited to Fourier transform, wavelet transform, etc.

[0035] In step S4, detection methods including but not limited to spectrum method, peak detection method, etc. are used to identify abnormalities in the audio signal during operation.

[0036] In step S5, through the collection of abnormal signals, train the artificial intelligence algorithm, and according to the training results, classify by the artificial intelligence algorithm to initially determine which type of fault it belongs to and whether to stop urgently, etc.

[0037] In step S6, the fault recording time interval is 1 minute.

[0038] Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art and related fields based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention. Structures, devices, and operation methods not specifically described and explained in the present invention, unless otherwise specified and limited, shall be implemented by conventional means in the art.

Claims

1. An audio-based precise positioning method for an unmanned aerial vehicle (UAV) to patrol a belt line, comprising: S1. Use the pickup on the drone to collect the audio signals generated during the operation of the belt conveyor along the way, and obtain the audio signals therein. S2. Preprocess the audio signals through the frequency calculation module on the drone. The preprocessing of the audio signals includes filtering, noise reduction, and removal of singular values. Use the subtraction method to remove the audio signals of the drone flight, and use filtering to remove the noise interference signals. The purpose is to retain the belt conveyor operation signals and fault signals through preprocessing and simplify the subsequent processing process. S3. The frequency calculation module is used to calculate the frequency of any frequency point of any frame in the audio signals. S4. Receive the preprocessed data and perform audio signal recognition. S5. Use the data features as the learning target, and through machine learning, combined with the fuzzy clustering method, judge the abnormal signals and classify them. S6. When an abnormal sound signal is collected, record the fault at regular intervals, including the abnormal signal type, signal amplitude, and recording time. S7. When flying to a position where the abnormal signal cannot be collected or the amplitude of the abnormal signal decreases significantly, determine the point with the largest signal amplitude of the abnormal signal. S8. Accurately calculate the position of the abnormal point according to the flight speed of the drone and the recorded time points.

2. The audio-based precise positioning method for an unmanned aerial vehicle (UAV) to patrol a belt line according to claim 1, wherein, In step S3, the frequency calculation can use Fourier transform and wavelet transform.

3. The audio-based precise positioning method for an unmanned aerial vehicle (UAV) to patrol a belt line according to claim 1, wherein, In step S4, use the spectrum method and peak detection method to identify abnormalities in the audio signals during operation.

4. The audio-based precise positioning method for an unmanned aerial vehicle (UAV) to patrol a belt line according to claim 1, wherein, In step S5, through the collection of abnormal signals, train the artificial intelligence algorithm. According to the training results, the artificial intelligence algorithm classifies and preliminarily determines which type of fault it belongs to and whether to stop urgently.

5. The audio-based precise positioning method for an unmanned aerial vehicle (UAV) to patrol a belt line according to claim 1, wherein, In step S6, the fault recording time interval is 1 minute.