CN-121999797-A - Laser monitoring automatic detection method and system based on mechanical vibration response

CN121999797ACN 121999797 ACN121999797 ACN 121999797ACN-121999797-A

Abstract

The invention discloses a laser monitoring automatic detection method and system based on mechanical vibration response, which utilize micro-amplitude mechanical vibration of an object driven by voice to modulate reflected light, accurately capture modulation signals embedded with vibration information through interference effect and a high-sensitivity photoelectric detector, effectively remove noise interference by combining self-adaptive filtering and wavelet noise reduction algorithm, extract pure vibration characteristics, and efficiently restore high-quality voice sequences based on a Doppler frequency shift model. And the integrated intelligent load sequencing replacement mechanism performs priority sequencing on the key audio segments, automatically calibrates a storage path by using a virtual instrument technology, realizes the optimized warehousing of the high-priority segments, and finally forms an efficient and retrievable voice storage structure. The invention realizes long-distance, non-contact and high-concealment voice acquisition and intelligent storage, and remarkably improves the distance, definition and subsequent utilization efficiency of voice information acquisition.

Inventors

HAO LIBO
XU JINGQI
XIAO LIPING

Assignees

湖南机电职业技术学院

Dates

Publication Date: 20260508
Application Date: 20260127

Claims (10)

1. The automatic detection method for the laser monitoring based on the mechanical vibration response is characterized by comprising the following steps of: s100, a laser emission module is adopted to directionally project a laser beam to the surface of a target object, reflected light carrying micro-amplitude mechanical vibration is obtained from the surface of the target object, and the reflected light is modulated through an interference effect to obtain a modulation signal; S200, converting the modulation signal into an electric signal by adopting a high-sensitivity photoelectric detector, acquiring vibration information embedded in the electric signal, and determining vibration frequency distribution; s300, removing noise interference through a self-adaptive filtering algorithm according to the vibration frequency distribution to obtain a filtered signal, and extracting pure vibration characteristics by adopting a wavelet noise reduction algorithm if the strength of the filtered signal is lower than a preset threshold value; s400, analyzing the pure vibration characteristics through a Doppler frequency shift model, and obtaining frequency shift changes to obtain a restored voice sequence; S500, extracting key audio segments from the restored voice sequence, and adopting an intelligent load ordering replacement mechanism to order the priority of the key audio segments so as to determine a high-priority segment; And S600, aiming at the high-priority section, automatically calibrating a storage path by adopting an integrated virtual instrument technology, acquiring a calibrated path, and transmitting the key audio section to a database through the calibrated path to obtain an optimized storage structure.
2. The automated detection method for laser listening based on mechanical vibration response of claim 1, wherein step S100 comprises: S110, directionally projecting a high-coherence laser beam to the surface of a target object by adopting a laser emission module, and acquiring reflected light carrying micro-amplitude mechanical vibration information from the surface of the target object; S120, guiding the reflected light and the reference beam to carry out wave front matching so as to generate interference fringes, wherein the interference fringes are formed when the reflected light and the reference beam meet the coherence condition; s130, converting the interference fringes into beat signals, analyzing instantaneous phase variation, reconstructing mechanical vibration waveform data according to the instantaneous phase variation, and modulating the reflected light through an interference effect to obtain a modulation signal.
3. The automated detection method for laser listening based on mechanical vibration response of claim 1, wherein step S200 comprises: S210, receiving a discrete digital sequence, wherein the discrete digital sequence is generated after photoelectric conversion and quantization processing are performed on a modulation signal by a high-sensitivity photoelectric detector; S220, performing quadrature demodulation and phase unwrapping processing on the discrete digital sequence to obtain a continuous phase curve, and generating a time domain displacement value according to the mapping of the continuous phase curve; S230, performing fast Fourier transform processing on the time domain displacement value to extract a frequency domain amplitude spectrum, and determining vibration frequency distribution according to energy intensity data in the frequency domain amplitude spectrum.
4. The automated detection method for laser listening based on mechanical vibration response of claim 1, wherein step S300 comprises: S310, constructing a reference noise frequency band model according to vibration frequency distribution and calculating an initial weight vector of the adaptive filter; S320, applying the initial weight vector to a noise-containing vibration signal to perform iterative operation to obtain time domain waveform data; S330, if the signal intensity of the time domain waveform data is lower than a preset threshold value, performing multi-scale decomposition on the time domain waveform data to obtain an approximate component and a detail component; S340, reconstructing and generating an enhanced vibration signal by combining the detail component and the approximate component, and performing envelope analysis on the enhanced vibration signal to extract pure vibration characteristics.
5. The automated detection method for laser listening based on mechanical vibration response of claim 1, wherein step S400 comprises: s410, performing Hilbert transform and phase unwrapping operation on the pure vibration characteristics to obtain a continuous phase change curve; S420, performing time differentiation and differential calculation on the continuous phase change curve to obtain a Doppler frequency shift deviation sequence; S430, inputting the Doppler frequency shift deviation sequence into a preset Doppler frequency shift model, and analyzing to obtain a micro-vibration velocity vector of the target surface according to the Doppler frequency shift model; S440, converting the micro-vibration velocity vector into a sound pressure amplitude variation value according to a sound wave modulation mechanism, and discretizing and recombining the sound pressure amplitude variation value to obtain a restored voice sequence.
6. The automated detection method for laser listening based on mechanical vibration response of claim 1, wherein step S500 comprises: S510, short-time energy calculation is performed on the restored voice sequence to identify a silence interval, and the restored voice sequence is truncated into a plurality of candidate audio fragments; s520, calculating the spectral entropy value of the candidate audio fragment, mapping the spectral entropy value into a calculation load value, and generating an initial sequencing fragment queue according to the calculation load value; S530, traversing the initial sequencing fragment queue, comparing the calculated load value with a minimum load value in a preset high-priority buffer area to execute replacement operation, and storing replaced audio data in the high-priority buffer area; S540, extracting the audio data in the high-priority buffer area, and determining a high-priority segment.
7. The automated detection method for laser listening based on mechanical vibration response of claim 1, wherein step S600 comprises: s610, acquiring a high-priority segment, calling an integrated virtual instrument control module to acquire the real-time load state of a storage node, and generating a storage link state data set; The real-time load state of the storage node is derived by the following formula: ; Wherein, the Representing storage nodes At the moment of time Is used for the real-time load status of (a), The time window of acquisition is indicated and, Representing the integrated virtual instrument control module acquisition function, Representing the storage node and the storage node, A variable of the time is represented and, Representing key time of window; The stored link state data set is derived by the following formula: ; Wherein, the Representing the storage of a link state data set, Representing storage nodes , Representing links , Representing nodes Is the first of (2) The sub-load is sampled and, The number of samples is indicated and, Representing the total number of nodes, Representing the total number of links; S620, analyzing the stored link state data set, and calculating the bandwidth and delay of each link to obtain a link performance grading list; The link performance score list is derived by the following formula: ; Wherein, the Represent the first The performance score of the link is determined, Represent the first The bandwidth of the strip-link, Represent the first Delay of the strip link; s630, traversing the link performance scoring list, and marking links with scores higher than a reference line as calibrated paths; And S640, transmitting the high-priority segment to a database log buffer before writing through the calibrated path, triggering a check point mechanism to write the audio data of the log buffer before writing into a physical disk and reconstructing an index, and obtaining an optimized storage structure.
8. The automated detection method for laser listening based on mechanical vibration response of claim 7, wherein in step S630, the judgment condition of the link flag is defined by the following formula: ; Wherein, the Represent the first The calibration marks of the links are used to calibrate the link, Representing the datum line when Equal to The link is marked as a calibrated path.
9. The automated mechanical vibration response-based laser listening detection method of claim 8 wherein in step S640, the following formula is used for transmitting the high priority segments to the database pre-write log buffer via the calibrated path: ; Wherein, the Representing a high priority segment of the video stream, The path after the calibration is represented as such, Representing a database pre-write log buffer; The audio data of the log buffer before write is written to the physical disk by the following formula for triggering the checkpoint mechanism: ; Wherein, the Audio data representing the log buffer before writing, Representing the mechanism of the checkpoint and, Representing a physical disk; the optimized storage structure is obtained by using the following formula for reconstructing the index: ; Wherein, the Representing the index data on the disk and, The reconstructed index is represented by a representation of the reconstructed index, Representing an optimized storage structure.
10. A mechanical vibration response based laser listening automation detection system for performing the mechanical vibration response based laser listening automation detection method of any one of claims 1 to 9, comprising: The modulating signal acquisition module (10) is used for directionally projecting a laser beam to the surface of a target object by adopting the laser emission module, acquiring reflected light carrying micro-amplitude mechanical vibration from the surface of the target object, and modulating the reflected light through an interference effect to obtain a modulating signal; The vibration frequency distribution determining module (20) is used for converting the modulation signal into an electric signal by adopting a high-sensitivity photoelectric detector, acquiring vibration information embedded in the electric signal and determining vibration frequency distribution; The pure vibration characteristic extraction module (30) is used for removing noise interference through an adaptive filtering algorithm according to the vibration frequency distribution to obtain a filtered signal, and extracting pure vibration characteristics by adopting a wavelet noise reduction algorithm if the strength of the filtered signal is lower than a preset threshold value; The recovery voice sequence acquisition module (40) is used for analyzing the pure vibration characteristics through a Doppler frequency shift model to acquire frequency shift changes so as to obtain a recovery voice sequence; A high priority segment determining module (50) configured to extract a key audio segment from the recovered speech sequence, and prioritize the key audio segment by using an intelligent load ranking replacement mechanism to determine a high priority segment; and the optimized storage structure acquisition module (60) is used for automatically calibrating the storage path by adopting an integrated virtual instrument technology aiming at the high-priority segment, acquiring a calibrated path, and transmitting the key audio segment to the database through the calibrated path to obtain the optimized storage structure.

Description

Laser monitoring automatic detection method and system based on mechanical vibration response Technical Field The invention relates to the technical field of modern information acquisition and safety monitoring, and particularly discloses a laser monitoring automatic detection method and system based on mechanical vibration response. Background In the field of modern information acquisition and security monitoring, the capturing and analyzing technology of long-distance voice signals is very important. The technology is not only related to information security, but also directly influences decision efficiency and accuracy in a specific scene. Particularly, in some environments requiring hidden operations, how to obtain clear voice information without being perceived has become an important direction of technical research. However, currently mainstream speech capture methods often face some deep challenges. These methods rely mostly on direct contact devices or close range signal reception, which not only easily expose operational intent, but are also severely disturbed by environmental noise and physical barriers in complex environments. More importantly, these methods often fail to ensure the integrity and clarity of the voice information in the face of distant objects due to signal attenuation or environmental factors, resulting in a significant compromise in the quality of the information ultimately obtained. In this context, researchers have found that non-contact detection techniques based on minute vibrations of the surface of a target object are a valuable direction to explore. The core difficulty faced by this technique is how to accurately capture and resolve the minute vibration signals induced by the sound waves. The amplitude of the vibration signal is very small, even reaches the nanometer level, and is very easy to be submerged by external noise or light interference. It is difficult to accurately reproduce the original voice content if the information carried by these minute vibrations cannot be effectively separated and restored. The conversion process from weak vibration to clear voice becomes a key bottleneck for technical breakthrough. In particular, in an actual service scenario, for example, when the vibration of a glass window is detected remotely to obtain indoor dialogue content, although the micro vibration of the glass surface can reflect sound wave information, the vibration signals can be affected by wind speed and light change and even uneven material of the target surface in the transmission and receiving processes, so that the signals are distorted or lost. Therefore, how to accurately capture and analyze these micro vibration signals under the long-distance and non-contact conditions, and successfully restore clear voice content, becomes a key problem in the current research of urgent need. Disclosure of Invention The invention provides a laser monitoring automatic detection method and system based on mechanical vibration response, and aims to solve at least one defect in the prior art. One aspect of the invention relates to a laser monitoring automatic detection method based on mechanical vibration response, comprising the following steps: S100, a laser emission module is adopted to directionally project a laser beam to the surface of a target object, reflected light carrying micro-amplitude mechanical vibration is obtained from the surface of the target object, and the reflected light is modulated through an interference effect to obtain a modulation signal; S200, converting the modulation signal into an electric signal by adopting a high-sensitivity photoelectric detector, acquiring vibration information embedded in the electric signal, and determining vibration frequency distribution; s300, removing noise interference through an adaptive filtering algorithm according to vibration frequency distribution to obtain a filtered signal, and extracting pure vibration characteristics by adopting a wavelet noise reduction algorithm if the strength of the filtered signal is lower than a preset threshold value; s400, analyzing the pure vibration characteristics through a Doppler frequency shift model, and obtaining frequency shift changes to obtain a restored voice sequence; S500, extracting key audio segments from the restored voice sequence, and adopting an intelligent load sequencing replacement mechanism to prioritize the key audio segments to determine a high-priority segment; And S600, aiming at the high-priority segment, automatically calibrating a storage path by adopting an integrated virtual instrument technology, acquiring a calibrated path, and transmitting the key audio segment to a database through the calibrated path to obtain an optimized storage structure. Further, step S100 includes: S110, a laser emission module is adopted to directionally project a high-coherence laser beam to the surface of a target object, and reflected light carrying micro-amplitude mechanical