Life characteristics acquisition system and method based on confined space gas safety detector

By using comprehensive analysis technology of image and sonar data in foundation pit construction, the problem that existing equipment cannot respond to movement and breathing abnormalities in time is solved, and accurate monitoring and comprehensive protection of the life characteristics of people in the foundation pit is achieved.

CN118731950BActive Publication Date: 2024-11-01WUHAN XINGYI NEW FUTURE POWER TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411211065.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-30
Publication Date
2024-11-01
Estimated Expiration
2044-08-30

AI Technical Summary

Technical Problem

During the construction of existing foundation pits, the personnel's life characteristics collection equipment cannot respond to abnormal movements and breathing in time, and there is a problem of inaccurate monitoring.

Method used

Adopt a life feature acquisition system based on a limited space gas safety detector, combining image and sonar data, and conducts comprehensive analysis through data processing modules, including sonar recognition, action recognition, action analysis and life feature judgment modules to monitor the movements of people in the foundation pit in real time and breathe, and issue an alarm in time.

Benefits of technology

It realizes accurate monitoring of the life characteristics of personnel in the foundation pit, and can fully protect the health and safety of construction workers under the changes in the terrain within the foundation pit and the complex gas environment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118731950B_ABST
    Figure CN118731950B_ABST
Patent Text Reader

Abstract

The present invention provides a life feature acquisition system and method based on a confined space gas safety detector, which relates to the field of construction safety technology, and includes a host, wherein the host is connected to a manual operation module, a data storage module, an alarm module, a data processing module and an extension. The image signal acquired by the present invention is combined with a sonar signal and foundation pit gas data, and the physical condition of the personnel is monitored based on the human body movements of the personnel in the foundation pit, combined with their breathing conditions and the gas environment conditions in the foundation pit. Moreover, since the sonar detection method is adopted, changes in the terrain in the foundation pit can also be detected. In addition, the health status of the personnel in the foundation pit can be comprehensively judged in combination with the terrain changes in the foundation pit, human body movements, breathing conditions and gas distribution conditions, thereby achieving an all-round protection effect for the personnel.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of construction safety technology, and in particular to a life feature acquisition system and method based on a confined space gas safety detector. Background Art

[0002] Gas safety testing during foundation pit construction is crucial to ensure the personal safety and health of personnel. The foundation pit construction environment is complex and there are many potential risk factors, including the accumulation of harmful gases, unstable soil, changes in hydrogeological conditions, etc. These factors may pose a threat to the health and safety of construction workers.

[0003] The foundation pit is a limited space. Since it is located in an underground environment, the ground wind flow needs to be guided to replace the gas inside it. And because the molecular mass of most harmful gases is greater than that of air, they are very easy to settle at the bottom of the foundation pit. At present, ventilation and exhaust systems are needed to ensure the safety of the gas in the foundation pit and the personal safety of people in the foundation pit.

[0004] The existing collection of vital signs of people in foundation pits is generally completed by relying on the health monitoring equipment worn by people. There is no good monitoring method for the environment in the foundation pit and the body movements and breathing of people. When people's bodies are abnormal, they often perform subconscious protective actions, and are affected by harmful gases, their breathing conditions will also be abnormal. In this case, blood oxygen monitoring, pulse monitoring and other means of monitoring people cannot reflect the vital signs of people in a timely and clear manner, and there are defects. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings existing in the prior art and optimize the problem that the body monitoring equipment currently worn by construction workers in the foundation pit cannot respond to movements and breathing in time.

[0006] In order to achieve the above-mentioned object, the present invention adopts the following technical scheme: a life characteristics acquisition system based on a confined space gas safety detector comprises a host, the host is connected with a manual operation module, a data storage module, an alarm module, a data processing module and an extension;

[0007] The alarm module is used to issue an alarm reminder to ground personnel, and the alarm module includes an audible and visual alarm module and a display module;

[0008] The data processing module is used to process and analyze the data detected by the extension, and judge the life characteristics of the personnel in combination with the analysis results. The data processing module includes a sonar recognition module for processing and analyzing the sonar reflection wave, a motion recognition module for identifying the motion of the person in the image, a motion analysis module for analyzing the motion of the person in the image, a comprehensive analysis module for comprehensively analyzing the gas data in combination with the sonar data and the image data, and a life characteristic judgment module for judging the life characteristics of the personnel;

[0009] There are multiple extensions, each of which includes an image module, a sonar module, a prompt module and a gas safety detector;

[0010] The image module is used to obtain, identify and capture image data in the foundation pit. The image module includes a camera module for photographing the picture in the foundation pit, an infrared camera module for photographing infrared images of the picture in the foundation pit, a graphic recognition module for identifying the content of the picture, and a graphic capture module for capturing people in the picture;

[0011] The sonar module detects the breathing sound of people in the foundation pit in an active sonar manner, and the sonar module includes a sonar receiver module for receiving sonar reflected waves, a sonar transmitter module for actively transmitting sonar signals, and a signal processing module for processing sonar reflected waves;

[0012] The prompt module is used to remind people in the foundation pit, and the prompt module includes a sound prompt module and a vibration prompt module.

[0013] As a preferred implementation, the sonar recognition module includes a frequency component analysis module constructed with a Fourier transform algorithm, a non-stationary signal analysis module constructed with a short-time Fourier transform algorithm, an array signal processing module constructed with a traditional beamforming method, an array signal processing optimization module constructed with an adaptive beamforming method, and a sonar image processing module constructed with a convolutional neural network.

[0014] As a preferred embodiment, the action recognition module includes a human motion capture module constructed with an optical flow method, a full-body motion capture module constructed with a neural network framework, a human posture detection module constructed with a key point detection method, and a manual screening action feature module for manually screening normal action images.

[0015] As a preferred embodiment, the motion analysis module includes a perspective fusion module for synthesizing two-dimensional images from different angles into the same three-dimensional space, a three-dimensional reconstruction module for matching and triangulating common points in different camera views to reconstruct a three-dimensional model, a motion correction module for calibrating real-time motion with normal motion, and a behavior judgment module for making behavior judgments based on the degree of motion correction compliance.

[0016] Based on the life characteristics collection method of the confined space gas safety detector, the movements and breathing sounds of the personnel in the foundation pit are monitored using photographic images and sonar, and the life characteristics of the personnel in the foundation pit are judged by the movement characteristics and breathing changes. Specifically, the following steps are included:

[0017] S1. Use motion capture to pre-acquire motion data of routine operations in the foundation pit, extract features from the motion data through the manual operation module, and store them in the data storage module as a training set for the data processing module and as a benchmark data for motion proofreading and behavior judgment;

[0018] S2. The extension obtains the camera image data, infrared image data, sonar data and gas data in the foundation pit, and analyzes the human movements in the image data through the data processing module, and judges the vital signs of the people in the foundation pit in combination with the analysis results of the sonar data and the gas data.

[0019] As a preferred implementation, step S1 specifically includes the following steps:

[0020] S1.1. First, formulate the operation action according to the construction specifications of the foundation pit, and use the motion capture method to obtain the three-dimensional motion data of the standard operation action;

[0021] S1.2, using the manual operation module to input the three-dimensional motion data into the data storage module, and filtering out the motion features in the three-dimensional motion data through the manual filtering motion feature module;

[0022] S1.3, associating the three-dimensional motion data with the motion calibration module as the reference data for motion calibration;

[0023] S1.4. Associate the three-dimensional motion data with the behavior judgment module as the reference data for behavior judgment.

[0024] As a preferred implementation, S2 specifically includes the following steps:

[0025] S2.1. First, install the extension in the foundation pit according to the camera angle of the camera module, and number the extension in the host;

[0026] S2.2, acquiring image data and infrared image data in the foundation pit through a camera module and an infrared camera module;

[0027] S2.3, then collect the image information and infrared image information in the foundation pit through the image module, and process, analyze and judge the human behavior information in the image through the action recognition module and the action analysis module;

[0028] S2.4, simultaneously receiving the sonar reflected waves emitted by the sonar transmitting module through the sonar receiver module, and analyzing the breathing sound in the sonar reflected waves using the signal processing module;

[0029] S2.5. Use the comprehensive analysis module to analyze the situation of personnel in the foundation pit based on the breathing sound data analyzed by the sonar recognition module, combined with the personnel behavior determined by the behavior determination module and the gas data detected by the gas safety detector;

[0030] S2.6, then use the life characteristics judgment module to judge the life characteristics of the people in the foundation pit;

[0031] S2.7. When it is determined that there are abnormal life characteristics of people in the foundation pit, the sound prompt module and the vibration prompt module will first give a sound reminder and a vibration reminder. At the same time, the host will display the abnormal situation on the display module through the alarm module, and at the same time, the sound and light alarm module will sound an alarm to remind ground personnel to rescue.

[0032] As a preferred implementation, in S2.3,

[0033] S2.3.1, using the human motion capture module to perform pixel movement between consecutive frames of image data to capture human motion, and using the full-body motion capture module to estimate the shape and movement of the body and hands based on the single color image in the image, and then using the human posture detection module to detect the key points of the human body, and based on the key point data of the human body, integrating the motion data calculated by the human motion capture module and the full-body motion capture module to obtain the motion data in a single image;

[0034] S2.3.2, synthesizing the action data in a single image obtained by multiple extensions into a unified three-dimensional model through a perspective fusion module, and reconstructing the three-dimensional model of the action by matching and triangulating the common points in different camera images through a three-dimensional reconstruction module;

[0035] S2.3.3, analyzing the amplitude, speed and trajectory of the action in the three-dimensional model through the action calibration module, and calibrating the corresponding data with the training data to analyze the differences;

[0036] S2.3.4. Based on the difference between the three-dimensional model calibrated by the motion calibration module and the training set, the behavior of people in the foundation pit is judged through the behavior judgment module.

[0037] As a preferred implementation, in 2.4, the specific steps of processing the sonar reflected waves include:

[0038] 2.4.1. First, use the frequency component analysis module to convert the sonar reflection wave from the time domain to the frequency domain, then identify the frequency characteristics of the reflection waves of different depths and components, and extract the breathing sound information in the reflection wave;

[0039] 2.4.2. Since breathing sound has the characteristic of intensity change over time, the non-stationary signal analysis module is used to provide joint distribution in the time domain and frequency domain to improve the accuracy of the frequency change of dynamic monitoring sonar reflection waves in the monitoring environment of multiple groups of sonar modules;

[0040] 2.4.3. Then use the array signal processing module to perform weighted summation on the sonar reflection waves in the multi-group units, enhance the signal in the direction of the personnel in the foundation pit, and weaken the interference in other directions;

[0041] 2.4.4, then use the array signal processing optimization module to adaptively adjust the weights of the weighted summation algorithm;

[0042] 2.4.5. Finally, the sonar image processing module is used to analyze and extract features from the image generated by the sonar to identify the breathing sound signal of the person in the sonar reflection wave.

[0043] Compared with the prior art, the advantages and positive effects of the present invention are:

[0044] 1. In the present invention, the collected image signals are combined with sonar signals and foundation pit gas data, and the physical condition of the personnel is monitored based on the human body movements of the personnel in the foundation pit, combined with their breathing conditions and the gas environment conditions in the foundation pit. Moreover, since the sonar detection method is adopted, changes in the terrain in the foundation pit can also be detected. In addition, the health status of the personnel in the foundation pit can be comprehensively judged in combination with the terrain changes in the foundation pit, human body movements, breathing conditions and gas distribution conditions, thereby achieving an all-round protection effect for the personnel.

[0045] 2. In the present invention, a comprehensive processing method with multiple visual angles, multiple data processing levels and multiple data analysis methods is used to collect the life characteristics of people in the limited space of the foundation pit, and the gas data in the environment is combined to achieve further accurate detection of the conditions of the people. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A schematic diagram of a life characteristics acquisition system based on a confined space gas safety detector is proposed for the present invention;

[0047] Figure 2 The present invention proposes a schematic diagram of a sonar recognition module in a life feature acquisition system based on a confined space gas safety detector;

[0048] Figure 3 The present invention proposes a schematic diagram of an action recognition module in a life feature acquisition system based on a confined space gas safety detector;

[0049] Figure 4 The present invention proposes a schematic diagram of a motion analysis module in a life feature acquisition system based on a confined space gas safety detector;

[0050] Figure 5 A flow chart of a method for collecting life characteristics based on a confined space gas safety detector is proposed for the present invention;

[0051] Figure 6 The present invention proposes a specific flow chart of step S1 in the method for collecting life characteristics based on a confined space gas safety detector;

[0052] Figure 7 The present invention proposes a specific flow chart of step S2 in the method for collecting life characteristics based on a confined space gas safety detector;

[0053] Figure 8 The present invention proposes a specific flow chart of step S2.3 in the method for collecting life characteristics based on a confined space gas safety detector;

[0054] Fig. 9 The present invention proposes a specific flow chart of step S2.4 in the method for collecting life characteristics based on a confined space gas safety detector. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] Example: Figure 1 As shown, the present invention provides a technical solution: a life characteristics acquisition system based on a confined space gas safety detector, comprising a host, the host is connected to a manual operation module, a data storage module, an alarm module, a data processing module and an extension;

[0057] The host is used as the main data storage device, and the manual operation module and the alarm module are connected by wired connection, while the data storage module and the data processing module can be carried in an external server. The host and the external server can be connected by wireless signals, and the signal connection between the extension and the host includes but is not limited to data line connection and wireless connection;

[0058] The alarm module is used to send out an alarm to ground personnel. The alarm module includes an audible and visual alarm module and a display module. The audible and visual alarm module can be used in models including but not limited to: WT-W13 audible and visual alarm. The display module can select a sonar display, a high-definition display, etc. according to the actual needs of foundation pit construction;

[0059] There are multiple extensions, each of which includes an image module, a sonar module, a prompt module and a gas safety detector;

[0060] The image module is used to obtain, identify and capture image data in the foundation pit. The image module includes a camera module for photographing the picture in the foundation pit, an infrared camera module for photographing infrared images of the picture in the foundation pit, a graphic recognition module for identifying the content of the picture, and a graphic capture module for capturing people in the picture;

[0061] The sonar module detects the breathing sound of people in the foundation pit in an active sonar manner, and the sonar module includes a sonar receiver module for receiving sonar reflected waves, a sonar transmitter module for actively transmitting sonar signals, and a signal processing module for processing sonar reflected waves;

[0062] The prompt module is used to remind personnel in the foundation pit, and the prompt module includes a sound prompt module and a vibration prompt module;

[0063] Further, such as Figure 1 As shown, the data processing module is used to process and analyze the data detected by the extension, and judge the life characteristics of the personnel in combination with the analysis results. The data processing module includes a sonar recognition module for processing and analyzing the sonar reflection wave, a motion recognition module for identifying the motion of the person in the image, a motion analysis module for analyzing the motion of the person in the image, a comprehensive analysis module for comprehensively analyzing the gas data in combination with the sonar data and the image data, and a life characteristic judgment module for judging the life characteristics of the personnel;

[0064] like Figure 2As shown, the sonar recognition module includes a frequency component analysis module constructed by a Fourier transform algorithm, a non-stationary signal analysis module constructed by a short-time Fourier transform algorithm, an array signal processing module constructed by a traditional beamforming method, an array signal processing optimization module constructed by an adaptive beamforming method, and a sonar image processing module constructed by a convolutional neural network;

[0065] in:

[0066] The frequency component analysis module is constructed with Fourier transform algorithm as the main algorithm. Since the detection target is the breathing sound of people, and the breathing sound of people in motion will change irregularly with the change of movement, the Fourier transform of non-periodic function is used to meet the non-periodic change characteristic requirements when extracting the breathing sound signal in the reflected wave signal;

[0067] The non-stationary signal analysis module is constructed using the short-time Fourier transform algorithm as the main algorithm. Since there is no regularity in the change of breathing sound over time, the frequency analysis of breathing sound in the short-time internal reflection wave requires the local frequency analysis function provided by the short-time Fourier transform.

[0068] The array signal processing module is based on the logic construction of multiple extensions installed in different positions. The array signal processing module is constructed using the traditional beamforming method to accurately control the weighting and delay of the signals received by multiple sensors, and perform directional processing on the sonar reflected waves at multiple spatial positions to highlight the breathing sound frequency in the reflected waves;

[0069] The array signal processing optimization module is based on the processing of the array signal for the reflected wave signal and the non-periodic change of the breathing sound frequency. For breathing sound frequencies of different intensities, the array signal processing module cannot quickly adjust the weight of the weighted algorithm, so the array signal processing optimization module is constructed with the adaptive algorithm as the main algorithm. The algorithms that can be used include but are not limited to: least mean square (LMS), recursive least squares (RLS) or more complex Capon algorithm;

[0070] The sonar image processing module is mainly used to analyze and extract features from images generated by sonar reflection waves. The image processing capabilities built using convolutional neural networks can improve image quality and increase target recognition rate.

[0071] like Figure 3 As shown, the action recognition module includes a human motion capture module constructed by an optical flow method, a full-body motion capture module constructed by a neural network framework, a human posture detection module constructed by a key point detection method, and a manual screening action feature module for manually screening normal action images;

[0072] in:

[0073] The human motion capture module is used to capture human motion data in images. It mainly uses image processing and character recognition algorithms, and combines the optical flow method of analyzing the motion vector of pixels between consecutive frames to capture human motion and obtain human motion data based on pixel analysis in two-dimensional images.

[0074] The full-body motion capture module is based on the actual situation of real-time monitoring of the movements of people in the foundation pit. It is built using a neural network framework and estimates the shape and movement of the body and hands through a single color image, achieving real-time full-body dynamic capture, which can improve the efficiency of monitoring emergency situations of people in the foundation pit.

[0075] The human posture detection module can use the deep learning model to detect the key points of the human body. Combined with the human motion data provided by the human motion capture module, it can verify and complement the operation effect of the full-body motion capture module.

[0076] The manual screening action feature module is mainly completed and implemented through the manual operation module;

[0077] like Figure 4 As shown, the motion analysis module includes a perspective fusion module for synthesizing two-dimensional images from different angles into the same three-dimensional space, a three-dimensional reconstruction module for matching and triangulating common points in different camera views to reconstruct a three-dimensional model, a motion correction module for correcting real-time motion with normal motion, and a behavior judgment module for judging behavior based on judging the conformity of motion correction;

[0078] in:

[0079] The perspective fusion module takes into account the situation that multiple extensions are located at different positions in the foundation pit. It combines the internal and external parameters of the cameras to register the images captured by different cameras, that is, to find the correspondence between them and map them into a unified reference frame, synthesizing the two-dimensional images into a unified three-dimensional space.

[0080] The 3D reconstruction module is based on the fact that after the perspective fusion, the 2D images acquired by the extension are all in the same 3D space, and the human body movements analyzed by the action recognition module are also in a 3D state. The 3D reconstruction module can further analyze the human body movements and improve the accuracy of the action analysis.

[0081] The action proofreading module uses the standard operation action data in the foundation pit pre-acquired in the training set as the benchmark, and compares the action data obtained by the action recognition module and the 3D reconstruction module with it, so as to standardize the actions of the personnel in the foundation pit. When actions outside the standard actions occur, it can be regarded as the physical discomfort or illegal operation of the personnel in the foundation pit.

[0082] The behavior judgment module judges the specific behavior of the people in the foundation pit based on the proofreading results of the motion proofreading module, and analyzes whether they are in a state of physical discomfort. It then combines the breathing sound data obtained by the sonar recognition module and the gas detection data of the gas safety detector to further judge the specific life characteristics and safety status of the people in the foundation pit.

[0083] like Figure 5 As shown, the method for collecting life characteristics based on a confined space gas safety detector uses the life characteristics collection system based on a confined space gas safety detector described in the above content, uses photographic images and sonar to monitor the movements and breathing sounds of people in the foundation pit, and judges the life characteristics of people in the foundation pit through movement characteristics and breathing changes, specifically including the following steps:

[0084] S1. Use motion capture to pre-acquire motion data of routine operations in the foundation pit, extract features from the motion data through the manual operation module, and store them in the data storage module as a training set for the data processing module and as a benchmark data for motion proofreading and behavior judgment;

[0085] Among them, Figure 6 As shown, step S1 specifically includes:

[0086] S1.1. First, formulate the operation action according to the construction specifications of the foundation pit, and use the motion capture method to obtain the three-dimensional motion data of the standard operation action;

[0087] Since the foundation pit is in a construction environment and is not suitable as a resting place, the movements of the people in it can be standardized. Without considering the differences in the body shapes of the people, standardized movements can not only bring positive improvements to the workers and the construction results, but also improve the accuracy of movement recognition. In the actual operation process, the body shape standards of construction workers can be formulated, and people of different body shapes can be sent to the foundation pit to work in batches for data collection and verification. Then, as the amount of data increases, the requirement of sending people to the foundation pit to work according to body shape can be gradually cancelled. At the same time, when performing standardized motion capture, people of different body shapes can also be used for motion capture to further increase the amount of data.

[0088] S1.2, using the manual operation module to input the three-dimensional motion data into the data storage module, and filtering out the motion features in the three-dimensional motion data through the manual filtering motion feature module;

[0089] When manually screening the action features, since there are no action specifications for the actual requirements of working in the foundation pit, and the purpose of the present invention is to monitor the life characteristics of the personnel in the foundation pit, based on the reality that when the personnel are physically unwell, the body movements will show protective movements, so the action data can be weighted and the error value can be added to ensure that the human body movements that are obviously different from the construction movements can be distinguished;

[0090] S1.3, associating the three-dimensional motion data with the motion calibration module as the reference data for motion calibration;

[0091] S1.4, associating the three-dimensional motion data with the behavior judgment module as the reference data for behavior judgment;

[0092] S2, the extension obtains the camera image data, infrared image data, sonar data and gas data in the foundation pit, and analyzes the human movements in the image data through the data processing module, and determines the vital signs of the personnel in the foundation pit in combination with the analysis results of the sonar data and the gas data;

[0093] Among them, Figure 7 As shown, step S2 specifically includes:

[0094] S2.1. First, install the extension in the foundation pit according to the camera angle of the camera module, and number the extension in the host;

[0095] The location and status monitoring data of the extension can be classified according to the extension number. When processing the data later, the device data of the corresponding extension can be directly retrieved to quickly complete the data retrieval action;

[0096] S2.2, acquiring image data and infrared image data in the foundation pit through a camera module and an infrared camera module;

[0097] Image data is used to detect and analyze human movements, while infrared images are used as auxiliary analysis data. In the presence of environmental interference factors, such as dust, rain, fog, etc., using infrared images for auxiliary analysis can improve the accuracy of human movement analysis;

[0098] S2.3, then collect the image information and infrared image information in the foundation pit through the image module, and process, analyze and judge the human behavior information in the image through the action recognition module and the action analysis module;

[0099] Among them, Figure 8 As shown, step S2.3 specifically includes:

[0100] S2.3.1, using the human motion capture module to perform pixel movement between consecutive frames of image data to capture human motion, and using the full-body motion capture module to estimate the shape and movement of the body and hands based on the single color image in the image, and then using the human posture detection module to detect the key points of the human body, and based on the key point data of the human body, integrating the motion data calculated by the human motion capture module and the full-body motion capture module to obtain the motion data in a single image;

[0101] Since the image obtained by the extension has a single viewing angle, and the present invention can be equipped with multiple extensions to shoot the images in the foundation pit at different viewing angles, the motion analysis of the image in a single extension can provide certain data support for the final three-dimensional reconstruction and verification;

[0102] S2.3.2, synthesizing the action data in a single image obtained by multiple extensions into a unified three-dimensional model through a perspective fusion module, and reconstructing the three-dimensional model of the action by matching and triangulating the common points in different camera images through a three-dimensional reconstruction module;

[0103] S2.3.3, analyzing the amplitude, speed and trajectory of the action in the three-dimensional model through the action calibration module, and calibrating the corresponding data with the training data to analyze the differences;

[0104] Analyzing the difference between the 3D model and the training set data can make a basic judgment on the actions of people in the foundation pit. When the difference is too large, the behavior can be directly judged, which has a certain screening effect.

[0105] S2.3.4, according to the difference between the three-dimensional model proofread by the action proofreading module and the training set, the behavior of the personnel in the foundation pit is judged by the behavior judgment module;

[0106] The behavior judgment module can make certain judgments on the abnormal actions of people in the foundation pit. Since the construction action is the purpose of the human body to exert force outward, and the human body's protective action is generally for the purpose of alleviating its own state and resisting external invasion, the action types are quite different. It can make basic judgments on the behavioral purpose of the actions of people in the foundation pit, providing convenience for subsequent rescue operations;

[0107] S2.4, simultaneously receiving the sonar reflected waves emitted by the sonar transmitting module through the sonar receiver module, and analyzing the breathing sound in the sonar reflected waves using the signal processing module;

[0108] Among them, Fig. 9 As shown in the figure, the specific processing steps of sonar reflected waves include:

[0109] 2.4.1. First, use the frequency component analysis module to convert the sonar reflection wave from the time domain to the frequency domain, then identify the frequency characteristics of the reflection waves of different depths and components, and extract the breathing sound information in the reflection wave;

[0110] Analyzing the components of sonar reflected waves is an important process for detecting people's breathing through sonar. The Fourier transform of non-periodic functions can detect irregular breathing sound signals in reflected waves.

[0111] 2.4.2. Since breathing sound has the characteristic of intensity change over time, the non-stationary signal analysis module is used to provide joint distribution in the time domain and frequency domain to improve the accuracy of the frequency change of dynamic monitoring sonar reflection waves in the monitoring environment of multiple groups of sonar modules;

[0112] The non-stationary signal analysis module constructed by short-time Fourier transform can further enhance the accuracy of breathing sound analysis in sonar reflection waves, thereby improving the accuracy of frequency component analysis module for sonar reflection wave component analysis;

[0113] 2.4.3. Then use the array signal processing module to perform weighted summation on the sonar reflection waves in the multi-group units, enhance the signal in the direction of the personnel in the foundation pit, and weaken the interference in other directions;

[0114] With the main purpose of detecting the breathing sound of people in the foundation pit, weighted summation of various data in the sonar reflection wave can highlight the breathing sound signal and further improve the accuracy of breathing sound detection;

[0115] 2.4.4, then use the array signal processing optimization module to adaptively adjust the weights of the weighted summation algorithm;

[0116] Since the movements of personnel during the construction process are non-periodic, their breathing frequency and intensity will also be non-periodic. Therefore, an adaptive algorithm is used to adjust the weight ratio in the weighted summation algorithm to ensure the accuracy of the breathing sound signal during the detection process and reduce the probability of failure to detect the breathing sound signal due to errors.

[0117] 2.4.5. Finally, the sonar image processing module is used to analyze and extract features from the image generated by the sonar, and the breathing sound signal of the person in the sonar reflection wave is clearly identified;

[0118] S2.5. Use the comprehensive analysis module to analyze the situation of personnel in the foundation pit based on the breathing sound data analyzed by the sonar recognition module, combined with the personnel behavior determined by the behavior determination module and the gas data detected by the gas safety detector;

[0119] S2.6, then use the life characteristics judgment module to judge the life characteristics of the people in the foundation pit;

[0120] S2.7. When it is determined that there are abnormal life characteristics of people in the foundation pit, the sound prompt module and the vibration prompt module will first give a sound reminder and a vibration reminder. At the same time, the host will display the abnormal situation on the display module through the alarm module, and at the same time, the sound and light alarm module will sound an alarm to remind ground personnel to rescue.

[0121] Working principle:

[0122] like Figure 1-4As shown, the data collection of the life feature collection system based on the confined space gas safety detector provided by the present invention is mainly completed by the analysis set in the foundation pit. The collected image signal is combined with the sonar signal and the foundation pit gas data. Based on the human body movements of the personnel in the foundation pit, the physical condition of the personnel is monitored in combination with their breathing conditions and the gas environment conditions in the foundation pit. In addition, due to the use of sonar detection, changes in the terrain in the foundation pit can also be detected. In addition, the health status of the personnel in the foundation pit can be comprehensively judged in combination with the terrain changes in the foundation pit, human body movements, breathing conditions and gas distribution conditions, thereby achieving an all-round protection effect for the personnel.

[0123] like Figure 5-9 As shown, the life signature collection method based on the confined space gas safety detector is designed according to the life signature collection system based on the confined space gas safety detector. The life signatures of the personnel in the confined space of the foundation pit are collected by a comprehensive processing method of multiple visual angles, multiple data processing levels, and multiple data analysis methods, and the gas data in the environment are combined to achieve further accurate detection of the personnel's conditions.

[0124] The above description is only a preferred embodiment of the present invention and does not limit the present invention in other forms. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A life characteristics acquisition system based on a confined space gas safety detector, characterized by: It includes a host computer, the host computer is connected with a manual operation module, a data storage module, an alarm module, a data processing module and an extension; The alarm module is used to issue an alarm reminder to ground personnel, and the alarm module includes an audible and visual alarm module and a display module; The data processing module is used to process and analyze the data detected by the extension, and judge the life characteristics of the personnel in combination with the analysis results. The data processing module includes a sonar recognition module for processing and analyzing the sonar reflection wave, a motion recognition module for identifying the motion of the person in the image, a motion analysis module for analyzing the motion of the person in the image, a comprehensive analysis module for comprehensively analyzing the gas data in combination with the sonar data and the image data, and a life characteristic judgment module for judging the life characteristics of the personnel; There are multiple extensions, each of which includes an image module, a sonar module, a prompt module and a gas safety detector; The image module is used to obtain, identify and capture image data in the foundation pit. The image module includes a camera module for photographing the picture in the foundation pit, an infrared camera module for photographing infrared images of the picture in the foundation pit, a graphic recognition module for identifying the content of the picture, and a graphic capture module for capturing people in the picture; The sonar module detects the breathing sound of people in the foundation pit in an active sonar manner, and the sonar module includes a sonar receiver module for receiving sonar reflected waves, a sonar transmitter module for actively transmitting sonar signals, and a signal processing module for processing sonar reflected waves; The prompt module is used to remind people in the foundation pit, and the prompt module includes a sound prompt module and a vibration prompt module.

2. The life characteristics acquisition system based on the confined space gas safety detector according to claim 1 is characterized in that: The sonar recognition module includes a frequency component analysis module constructed with a Fourier transform algorithm, a non-stationary signal analysis module constructed with a short-time Fourier transform algorithm, an array signal processing module constructed with a traditional beamforming method, an array signal processing optimization module constructed with an adaptive beamforming method, and a sonar image processing module constructed with a convolutional neural network.

3. The life characteristics acquisition system based on the confined space gas safety detector according to claim 1 is characterized in that: The action recognition module includes a human action capture module constructed by optical flow method, a full body action capture module constructed by neural network framework, a human posture detection module constructed by key point detection method and a manual screening action feature module for manually screening normal action images.

4. The life characteristics acquisition system based on the confined space gas safety detector according to claim 1 is characterized in that: The motion analysis module includes a perspective fusion module for synthesizing two-dimensional images from different angles into the same three-dimensional space, a three-dimensional reconstruction module for matching and triangulating common points in different camera views to reconstruct a three-dimensional model, a motion correction module for calibrating real-time motion with normal motion, and a behavior judgment module for making behavior judgments based on the degree of motion correction compliance.

5. A method for collecting life characteristics based on a confined space gas safety detector, according to a life characteristics collection system based on a confined space gas safety detector according to any one of claims 1 to 4, characterized in that: The movement and breathing sounds of people in the foundation pit are monitored using photographic images and sonar, and the life characteristics of people in the foundation pit are judged by the movement characteristics and breathing changes, which specifically includes the following steps: S1. Use motion capture to pre-acquire motion data of routine operations in the foundation pit, extract features from the motion data through the manual operation module, and store them in the data storage module as a training set for the data processing module and as a benchmark data for motion proofreading and behavior judgment; S2. The extension obtains the camera image data, infrared image data, sonar data and gas data in the foundation pit, and analyzes the human movements in the image data through the data processing module, and judges the vital signs of the people in the foundation pit in combination with the analysis results of the sonar data and the gas data.

6. The method for collecting life characteristics based on a confined space gas safety detector according to claim 5 is characterized in that: The step S1 specifically includes the following steps: S1.

1. First, formulate the operation action according to the construction specifications of the foundation pit, and use the motion capture method to obtain the three-dimensional motion data of the standard operation action; S1.2, using the manual operation module to input the three-dimensional motion data into the data storage module, and filtering out the motion features in the three-dimensional motion data through the manual filtering motion feature module; S1.3, associating the three-dimensional motion data with the motion calibration module as the reference data for motion calibration; S1.

4. Associate the three-dimensional motion data with the behavior judgment module as the reference data for behavior judgment.

7. The method for collecting life characteristics based on a confined space gas safety detector according to claim 5 is characterized in that: The S2 specifically includes the following steps: S2.

1. First, install the extension in the foundation pit according to the camera angle of the camera module, and number the extension in the host; S2.2, acquiring image data and infrared image data in the foundation pit through a camera module and an infrared camera module; S2.3, then collect the image information and infrared image information in the foundation pit through the image module, and process, analyze and judge the human behavior information in the image through the action recognition module and the action analysis module; S2.4, simultaneously receiving the sonar reflected waves emitted by the sonar transmitting module through the sonar receiver module, and analyzing the breathing sound in the sonar reflected waves using the signal processing module; S2.

5. Use the comprehensive analysis module to analyze the situation of personnel in the foundation pit based on the breathing sound data analyzed by the sonar recognition module, combined with the personnel behavior determined by the behavior determination module and the gas data detected by the gas safety detector; S2.6, then use the life characteristics judgment module to judge the life characteristics of the people in the foundation pit; S2.

7. When it is determined that there are abnormal life characteristics of people in the foundation pit, the sound prompt module and the vibration prompt module will first give a sound reminder and a vibration reminder. At the same time, the host will display the abnormal situation on the display module through the alarm module, and at the same time, the sound and light alarm module will sound an alarm to remind ground personnel to rescue.

8. The method for collecting life characteristics based on a confined space gas safety detector according to claim 7 is characterized in that: The S2.3 specifically includes the following steps: S2.3.1, using the human motion capture module to perform pixel movement between consecutive frames of image data to capture human motion, and using the full-body motion capture module to estimate the shape and movement of the body and hands based on the single color image in the image, and then using the human posture detection module to detect the key points of the human body, and based on the key point data of the human body, integrating the motion data calculated by the human motion capture module and the full-body motion capture module to obtain the motion data in a single image; S2.3.2, synthesizing the action data in a single image obtained by multiple extensions into a unified three-dimensional model through a perspective fusion module, and reconstructing the three-dimensional model of the action by matching and triangulating the common points in different camera images through a three-dimensional reconstruction module; S2.3.3, analyzing the amplitude, speed and trajectory of the action in the three-dimensional model through the action calibration module, and calibrating the corresponding data with the training data to analyze the differences; S2.3.

4. Based on the difference between the three-dimensional model calibrated by the motion calibration module and the training set, the behavior of people in the foundation pit is judged through the behavior judgment module.

9. The method for collecting life characteristics based on a confined space gas safety detector according to claim 7, characterized in that: In 2.4 above, the specific processing steps of the sonar reflected waves include: 2.4.

1. First, use the frequency component analysis module to convert the sonar reflection wave from the time domain to the frequency domain, then identify the frequency characteristics of the reflection waves of different depths and components, and extract the breathing sound information in the reflection wave; 2.4.

2. Since breathing sound has the characteristic of intensity change over time, the non-stationary signal analysis module is used to provide joint distribution in the time domain and frequency domain to improve the accuracy of the frequency change of dynamic monitoring sonar reflection waves in the monitoring environment of multiple groups of sonar modules; 2.4.

3. Then use the array signal processing module to perform weighted summation on the sonar reflection waves in the multi-group units, enhance the signal in the direction of the personnel in the foundation pit, and weaken the interference in other directions; 2.4.4, then use the array signal processing optimization module to adaptively adjust the weights of the weighted summation algorithm; 2.4.

5. Finally, the sonar image processing module is used to analyze and extract features from the image generated by the sonar to identify the breathing sound signal of the person in the sonar reflection wave.