CN112132719A - Personnel search and rescue method, device and system for emergency rescue and storage medium - Google Patents
Personnel search and rescue method, device and system for emergency rescue and storage medium Download PDFInfo
- Publication number
- CN112132719A CN112132719A CN202010803844.5A CN202010803844A CN112132719A CN 112132719 A CN112132719 A CN 112132719A CN 202010803844 A CN202010803844 A CN 202010803844A CN 112132719 A CN112132719 A CN 112132719A
- Authority
- CN
- China
- Prior art keywords
- signal
- personnel
- person
- environment
- trapped
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 65
- 238000001514 detection method Methods 0.000 claims abstract description 67
- 238000000605 extraction Methods 0.000 claims abstract description 23
- 230000000694 effects Effects 0.000 claims abstract description 11
- 230000008569 process Effects 0.000 claims abstract description 10
- QVFWZNCVPCJQOP-UHFFFAOYSA-N chloralodol Chemical group CC(O)(C)CC(C)OC(O)C(Cl)(Cl)Cl QVFWZNCVPCJQOP-UHFFFAOYSA-N 0.000 claims description 65
- 230000003595 spectral effect Effects 0.000 claims description 52
- 238000009826 distribution Methods 0.000 claims description 50
- 230000033001 locomotion Effects 0.000 claims description 17
- 230000003068 static effect Effects 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 12
- 230000035565 breathing frequency Effects 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 10
- 238000013144 data compression Methods 0.000 claims description 5
- 230000036391 respiratory frequency Effects 0.000 claims description 5
- 230000000386 athletic effect Effects 0.000 claims description 3
- 230000002829 reductive effect Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 9
- 238000004891 communication Methods 0.000 description 8
- 238000013145 classification model Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 7
- 238000012545 processing Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 6
- 230000029058 respiratory gaseous exchange Effects 0.000 description 6
- 238000005070 sampling Methods 0.000 description 6
- 230000001360 synchronised effect Effects 0.000 description 6
- 239000011159 matrix material Substances 0.000 description 4
- 230000004913 activation Effects 0.000 description 3
- 238000007906 compression Methods 0.000 description 3
- 230000006835 compression Effects 0.000 description 3
- 230000006378 damage Effects 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 3
- 238000011176 pooling Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 208000003443 Unconsciousness Diseases 0.000 description 2
- 210000000038 chest Anatomy 0.000 description 2
- 238000002592 echocardiography Methods 0.000 description 2
- 230000014509 gene expression Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 239000000779 smoke Substances 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 239000002341 toxic gas Substances 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- 208000019901 Anxiety disease Diseases 0.000 description 1
- 238000012935 Averaging Methods 0.000 description 1
- 208000000059 Dyspnea Diseases 0.000 description 1
- 206010013975 Dyspnoeas Diseases 0.000 description 1
- 108010076504 Protein Sorting Signals Proteins 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000036506 anxiety Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 235000019800 disodium phosphate Nutrition 0.000 description 1
- 230000009982 effect on human Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000011664 nicotinic acid Substances 0.000 description 1
- 238000013021 overheating Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- A—HUMAN NECESSITIES
- A62—LIFE-SAVING; FIRE-FIGHTING
- A62B—DEVICES, APPARATUS OR METHODS FOR LIFE-SAVING
- A62B5/00—Other devices for rescuing from fire
-
- G—PHYSICS
- G08—SIGNALLING
- G08C—TRANSMISSION SYSTEMS FOR MEASURED VALUES, CONTROL OR SIMILAR SIGNALS
- G08C17/00—Arrangements for transmitting signals characterised by the use of a wireless electrical link
- G08C17/02—Arrangements for transmitting signals characterised by the use of a wireless electrical link using a radio link
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Educational Administration (AREA)
- Development Economics (AREA)
- Alarm Systems (AREA)
- Human Resources & Organizations (AREA)
- Computer Networks & Wireless Communication (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Emergency Management (AREA)
- Economics (AREA)
Abstract
The invention discloses a method, a device and a system for searching and rescuing people for emergency rescue and a storage medium. Wherein, the method comprises the following steps: the method comprises the following steps: acquiring LoRa signal data reflected by a long-distance radio (LoRa) transceiver in the process of scanning a target area; performing feature extraction on the LoRa signal data to obtain a first signal for reflecting human body life activity features; carrying out personnel detection on the first signal based on a pre-constructed characteristic set, and determining whether trapped personnel exist in the target area; wherein the feature set is determined based on the first signals acquired in the unmanned environment and the manned environment, respectively. The embodiment of the invention can realize the detection of the trapped people in a non-invasive way in the emergency rescue scene, has a large detection range, and is particularly suitable for the detection of the trapped people in the fire rescue scene of a high-rise building.
Description
Technical Field
The invention relates to the field of emergency rescue, in particular to a personnel search and rescue method, a device, a system and a storage medium for emergency rescue.
Background
Along with the high-speed development of economy, the urban scale is continuously enlarged, high-rise buildings are increasingly increased, urban building structures are increasingly complex, population density in the buildings is increased, unsafe factors such as overload, overheating, short circuit and aging of electrical equipment and the like cause the occurrence probability of serious fire to rise year by year, harm and property loss to people are also increasingly large, and people pay more attention to the emergency rescue problem in the fire scene.
Existing wire-based fire alarm systems and fire-fighting linkage systems have matured. However, when a fire disaster occurs, the smoke concentration is high, the environment is complicated and harsh, and the scheme still has the following characteristics and disadvantages: on one hand, the wired communication mode has poor expansion performance, strict requirements on pipelines, large investment, complex wiring, difficult maintenance and easy damage; on the other hand, the traditional wired fire alarm system and the fire-fighting linkage system do not have the functions of positioning and tracking rescuers and trapped persons. Therefore, the trapped people can escape from the fire only by the aid of the exit indicator lamps in the building, and fire fighters can search and search by themselves during rescue, so that the rescue efficiency is low.
The method for improving the fire rescue efficiency and reducing casualties by using the non-invasive personnel detection technology is one of the problems of the current fire rescue field in key research. Due to the fact that smoke concentration in a fire scene is high, the environment is complex and severe, traditional non-invasive personnel detection technologies such as infrared rays, ultrasonic waves and computer vision are poor in penetrability and weak in anti-interference capability, detection dead angles exist, and the requirements of fire rescue scenes are difficult to meet; in addition, non-invasive personnel detection technologies based on wireless signals are rapidly developing and rising, but the technologies rely on the existing equipment, the sensing distance is very limited, and the equipment is large in size, so that the technology cannot be directly applied to high-rise building fire rescue.
Disclosure of Invention
In view of this, embodiments of the present invention provide a method, an apparatus, a system and a storage medium for searching and rescuing people for emergency rescue, which aim to accurately identify trapped people so as to improve the rescue efficiency of the trapped people.
The technical scheme of the embodiment of the invention is realized as follows:
in a first aspect, an embodiment of the present invention provides a method for searching and rescuing people for emergency rescue, including:
acquiring LoRa (Long Range Radio) signal data reflected by a LoRa transceiver in the process of scanning a target area;
performing feature extraction on the LoRa signal data to obtain a first signal for reflecting human body life activity features;
carrying out personnel detection on the first signal based on a pre-constructed characteristic set, and determining whether trapped personnel exist in the target area; wherein the feature set is determined based on the first signals acquired in the unmanned environment and the manned environment, respectively.
In some embodiments, the performing person detection on the first signal based on a pre-constructed feature set to determine whether a trapped person exists in the target area includes:
converting the first signal into a frequency domain signal based on a fourier transform;
calculating a variance probability distribution of a power spectral density of the first signal based on the frequency domain signal;
calculating a first distance between the variance probability distribution and a variance probability distribution of the power spectral density of the first signal corresponding to the unmanned environment in a pre-constructed feature set, a second distance between the variance probability distribution and a variance probability distribution of the power spectral density of the first signal corresponding to the manned environment in the pre-constructed feature values;
if the first distance is smaller than the second distance, determining that no trapped person exists in the target area;
and if the first distance is greater than or equal to the second distance, determining that the trapped people exist in the target area.
In some embodiments, the performing feature extraction on the LoRa signal data to obtain a first signal reflecting human vital signs includes:
performing data compression on the LoRa signal data;
and filtering the compressed data based on a high-pass filter and a low-pass filter to obtain the first signal.
In some embodiments, before the performing the person detection on the first signal based on the pre-constructed feature set, the method further comprises:
acquiring the first signals respectively corresponding to the unmanned environment and the manned environment;
converting the first signal corresponding to the unmanned environment into a frequency domain signal corresponding to the unmanned environment based on Fourier transform, and calculating a variance probability distribution of the power spectral density of the first signal corresponding to the unmanned environment based on the frequency domain signal corresponding to the unmanned environment;
converting the first signal corresponding to the manned environment into a frequency domain signal corresponding to the manned environment based on Fourier transform, and calculating the variance probability distribution of the power spectral density of the first signal corresponding to the manned environment based on the frequency domain signal corresponding to the manned environment;
wherein the pre-constructed feature set comprises: a variance probability distribution of a power spectral density of the first signal corresponding to the unmanned environment and a variance probability distribution of a power spectral density of the first signal corresponding to the manned environment.
In some embodiments, if there are trapped people in the target area, the method further comprises:
detecting the motion state of the trapped person on the basis of a pre-constructed person state detection model for the first signal; the personnel state detection model is generated based on the first signal training corresponding to the static personnel and the moving personnel respectively.
In some embodiments, the detecting the motion state of the trapped person based on the pre-constructed person state detection model for the first signal includes:
obtaining time-frequency data corresponding to the first signal based on short-time Fourier transform on the first signal;
and inputting the time-frequency data into a pre-constructed personnel state detection model to obtain a detection result that the trapped personnel are static personnel or moving personnel.
In some embodiments, if the trapped person is a stationary person, the method further comprises:
determining a respiratory frequency of the stationary person based on a peak in a power spectral density of the first signal;
determining a hazard level for the stationary person based on the breathing frequency of the stationary person.
In some embodiments, if the trapped person is an athletic person, the method further comprises:
and generating and transmitting an escape path for guiding the moving person to escape.
In a second aspect, an embodiment of the present invention further provides a device for searching and rescuing people for emergency rescue, including:
the acquisition module is used for acquiring LoRa signal data reflected by the LoRa transceiver in the process of scanning the target area;
the characteristic extraction module is used for carrying out characteristic extraction on the LoRa signal data to obtain a first signal for reflecting the human body life activity characteristics;
the personnel detection module is used for carrying out personnel detection on the first signal based on a pre-constructed characteristic set and determining whether trapped personnel exist in the target area; wherein the feature set is determined based on the first signals acquired in the unmanned environment and the manned environment, respectively.
In a third aspect, an embodiment of the present invention further provides a device for searching and rescuing people for emergency rescue, including: a processor and a memory for storing a computer program capable of running on the processor, wherein the processor, in communicative connection with the drone, is configured to perform the steps of the method of the first aspect of an embodiment of the present invention when running the computer program.
In a fourth aspect, an embodiment of the present invention further provides a personnel search and rescue system for emergency rescue, including:
the system comprises an unmanned aerial vehicle, wherein an LoRa transceiver is carried on the unmanned aerial vehicle;
the personnel search and rescue equipment provided by the embodiment of the invention is in communication connection with the unmanned aerial vehicle and is used for executing the steps of the method provided by the first aspect of the embodiment of the invention when a computer program is run.
In a fifth aspect, an embodiment of the present invention further provides a storage medium, where a computer program is stored on the storage medium, and when the computer program is executed by a processor, the steps of the method in the first aspect of the embodiment of the present invention are implemented.
According to the technical scheme provided by the embodiment of the invention, the LoRa signal data reflected by the target area is acquired, and the features of the LoRa signal data are extracted to obtain a first signal for reflecting the life activity features of a human body; carrying out personnel detection on the first signal based on a pre-constructed characteristic set, and determining whether trapped personnel exist in the target area; the method can realize the detection of the trapped people in a non-invasive mode in an emergency rescue scene, has a large detection range, and is particularly suitable for the detection of the trapped people in a fire rescue scene of a high-rise building.
Drawings
FIG. 1 is a schematic flow chart illustrating a method for searching and rescuing people for emergency rescue according to an embodiment of the present invention;
FIG. 2 is a histogram of probability of variance of a power spectral density of a first signal in an unmanned environment in an exemplary application of the present invention;
FIG. 3 is a histogram of probability of variance of a power spectral density of a first signal in a human environment in an exemplary application of the present invention;
FIG. 4 is a schematic diagram of a network structure of a SqueezeNet model according to an embodiment of the present invention;
FIG. 5 is a schematic structural diagram of a device for searching and rescuing people for emergency rescue according to an embodiment of the present invention;
FIG. 6 is a schematic structural diagram of a rescue apparatus for emergency rescue according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of a person search and rescue system for emergency rescue according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.
Before further detailed description of the embodiments of the present invention, terms and expressions mentioned in the embodiments of the present invention are explained, and the terms and expressions mentioned in the embodiments of the present invention are applicable to the following explanations:
the embodiment of the invention provides a personnel search and rescue method for emergency rescue, which comprises the following steps of:
here, the target area may be an area that needs emergency rescue, for example, areas corresponding to different floors of a high-rise building, such as fire rescue. In other embodiments, the target area may also be an area corresponding to an emergency rescue scenario in which a mine disaster, a building collapse, or other disaster occurs.
Illustratively, the LoRa transceiver may be mounted on a mobile platform to scan different target areas, for example, may be mounted on a drone, and may hover at different floors of a high-rise building and surround and climb or descend around the floors to meet the detection requirement of whether there are trapped people in different areas of different high-rise buildings. In other embodiments, the LoRa transceiver may be carried by an intelligent mobile robot or a bionic mouse, and the like, so as to be applied to search and rescue of people in narrow space in disaster environments such as mine disasters or building collapse.
102, performing feature extraction on the LoRa signal data to obtain a first signal for reflecting human body life activity features;
the obtained LoRa signal data reflected by the target area can be data within unit sampling time, and due to the large data volume and the existence of more interference signals, the LoRa signal data can be subjected to feature extraction so as to remove or reduce interference signal components as much as possible, improve the signal-to-noise ratio of the data and obtain a first signal for reflecting the vital movement features of the human body.
In some embodiments, the performing feature extraction on the LoRa signal data to obtain a first signal reflecting human vital signs includes:
performing data compression on the LoRa signal data;
and filtering the compressed data based on a high-pass filter and a low-pass filter to obtain the first signal.
Illustratively, data compression may be based on M-point mean filtering (moving average filter). Specifically, firstly, a template with M points is used for translating on a signal with the length of N, the acquired signal with the length of N is divided into a plurality of sections of equal-length signal sequences containing M sampling points, and the mean value of the M points is used for replacing the middle point of the original signal, so that the compression and the smoothness of signal data are realized.
Illustratively, for the compressed data, high-pass filtering processing is performed on the basis of a 4-order elliptic IIR high-pass filter to filter reflection clutter such as a wall body and improve the signal-to-noise ratio. In the embodiment of the invention, a 4-order elliptic high-pass IIR filter is adopted to extract a target signal component in LoRa signal data and remove a wall echo component. The transition band of the elliptical high-pass filter is much narrower than the amplitude-frequency response transition band of the traditional two-pulse canceller, and the Doppler frequency generated by the micro-motion of the human target basically falls within the pass band range of the filter. Therefore, the elliptical high-pass filter can keep the Doppler frequency information of the target to the maximum extent while inhibiting the fixed clutter, thereby effectively improving the signal-to-noise ratio. And for the data after high-pass filtering, low-pass filtering processing is carried out based on a low-pass filter so as to eliminate the influence of the unmanned aerial vehicle jitter. Specifically, the motion frequency of the drone is in the range of 60hz (hertz) to 150hz, which, unlike the low frequency range (<10hz) of human vital signals, has a certain effect on human motion recognition. In order to eliminate noise caused by vibration of the unmanned aerial vehicle during flying, the time domain data can be converted into a frequency domain by using an FFT (fast fourier transform), a low-pass filter (for example, a second-order butterworth low-pass filter with a cut-off frequency of 10hz) is used for removing a high-frequency part in the frequency domain, and then the IFFT (inverse fast fourier transform) is performed on the filtered low-frequency signal, so that a time domain signal can be reconstructed, and a first signal for reflecting the human body life activity characteristics is obtained to be processed at the next stage. The low-pass filtering can eliminate the jitter of the unmanned aerial vehicle, and the reserved low-frequency signals are used for reconstructing data so as to further compress the data.
103, detecting people in the first signal based on a pre-constructed feature set, and determining whether trapped people exist in the target area; wherein the feature set is determined based on the first signals acquired in the unmanned environment and the manned environment, respectively.
Here, the pre-constructed feature set may include a first feature set corresponding to the unmanned environment and a second feature set corresponding to the manned environment, and the first signal is matched based on the pre-constructed feature set, so that whether the trapped person exists in the target area may be determined.
The personnel searching and rescuing method provided by the embodiment of the invention can realize the detection of the trapped personnel in a non-invasive manner in the emergency rescue scene, has a large detection range, and is particularly suitable for the detection of the trapped personnel in the fire rescue scene of a high-rise building.
In some embodiments, the LoRa transceiver can be piggybacked on the unmanned aerial vehicle, and the LoRa transceiver is driven by the unmanned aerial vehicle to scan the target area and receive the LoRa signal data of reflection back. For example, an edge computing node, such as a laptop computer or a pre-configured desktop computer, may be temporarily built in a fire rescue scene, and the edge computing node may obtain the LoRa signal data forwarded by the drone or sent by the LoRa transceiver.
Illustratively, the LoRa transceiver includes: the system comprises a LoRa transmitter, a receiver and a Raspberry Pi (Raspberry Pi), wherein the LoRa transmitter can be a common commercial LoRa node, the receiver can be a LimeSDR mini, the receiver can be connected with the Raspberry Pi through a USB 3.0, and received LoRa signal data are forwarded to an edge computing node through the Raspberry Pi (Raspberry Pi) in a wireless mode such as 4G or WiFi, so that an end-edge architecture is constructed, and data processing time delay can be reduced by utilizing the data processing performance of the edge computing node.
The traditional wireless indoor positioning technology is completely based on the existing wireless infrastructure, and a user can independently realize positioning by installing specific wireless positioning software on a terminal without any additional hardware equipment. Most mobile terminals are portable computing devices with small size and limited energy, and in order to save the electric quantity and computing resources of the terminals, a positioning system based on a terminal-cloud server architecture is often adopted in actual deployment. In case of fire, the location service often has the following features: the indoor positioning infrastructure is easy to damage, the indoor positioning environment changes more quickly, the user positioning request is more frequent, the data fusion of multi-terminal combined positioning in the environment of the Internet of things is more complicated, and a large amount of data transmission and data calculation are brought to the improvement of the positioning accuracy. In the positioning process, if the cloud terminal is uploaded, analyzed and processed and then returned to the equipment, signal transmission delay is caused, and field potential misjudgment and rescue scheme making errors are easily caused under emergency conditions, so that casualties are caused. According to the embodiment of the invention, the edge computing node performs feature extraction on the LoRa signal data reflected by the target area to obtain the first signal, and the first signal is subjected to personnel detection based on the feature set constructed in advance, so that the time delay can be effectively reduced, and the emergency requirement of emergency rescue can be met. In addition, the LoRa transceiver is carried by unmanned aerial vehicle, as unified signal acquisition equipment, can enlarge scanning and perception area region.
In some embodiments, the performing person detection on the first signal based on a pre-constructed feature set to determine whether a trapped person exists in the target area includes:
converting the first signal into a frequency domain signal based on a fourier transform;
calculating a variance probability distribution of a power spectral density of the first signal based on the frequency domain signal;
calculating a first distance between the variance probability distribution and a variance probability distribution of the power spectral density of the first signal corresponding to the unmanned environment in a pre-constructed feature set, a second distance between the variance probability distribution and a variance probability distribution of the power spectral density of the first signal corresponding to the manned environment in the pre-constructed feature values;
if the first distance is smaller than the second distance, determining that no trapped person exists in the target area;
and if the first distance is greater than or equal to the second distance, determining that the trapped people exist in the target area.
In practical applications, a feature set for human detection needs to be constructed in advance, and in some embodiments, before human detection is performed on the first signal based on the feature set constructed in advance, the method further includes:
acquiring the first signals respectively corresponding to the unmanned environment and the manned environment;
converting the first signal corresponding to the unmanned environment into a frequency domain signal corresponding to the unmanned environment based on Fourier transform, and calculating a variance probability distribution of the power spectral density of the first signal corresponding to the unmanned environment based on the frequency domain signal corresponding to the unmanned environment;
converting the first signal corresponding to the manned environment into a frequency domain signal corresponding to the manned environment based on Fourier transform, and calculating the variance probability distribution of the power spectral density of the first signal corresponding to the manned environment based on the frequency domain signal corresponding to the manned environment;
wherein the pre-constructed feature set comprises: a variance probability distribution of a power spectral density of the first signal corresponding to the unmanned environment and a variance probability distribution of a power spectral density of the first signal corresponding to the manned environment.
When no human target exists in the environment, namely in the unmanned environment, clutter in the LoRa signal data can be approximately regarded as white noise which obeys gaussian distribution, therefore, the LoRa signal data can be approximately the superposition of a constant signal and the white gaussian noise, so that time-invariant Power Spectral Density (PSD) is generated, namely the Power Spectral Density (PSD) in the unmanned environment is theoretically a constant, and the Power Spectral Density (PSD) variance values at different sampling moments are gathered near zero. When a human target exists in the environment, namely in the presence of a person, the LoRa signal data includes echoes (clutter) generated by living body echoes, noise and other objects and backgrounds, which can be approximately regarded as colored noise complying with gaussian distribution, and the power spectral density of the clutter is gaussian, namely in the presence of a person, the Power Spectral Density (PSD) variance distribution at different sampling moments is more dispersed, and the variance value is also larger.
Based on this, in the embodiment of the present invention, the first signals respectively corresponding to the unmanned environment and the manned environment are obtained, that is, the LoRa signal data respectively corresponding to the unmanned environment and the manned environment are subjected to feature extraction to obtain corresponding first signals, and the frequency domain signals are obtained for the first signals based on the FFT to obtain the variance probability histogram of the power spectral density of the first signals, as shown in fig. 2 and fig. 3, fig. 2 is the variance probability histogram of the power spectral density of the first signals under the unmanned environment, most of variance values are concentrated near 0, and the probability value is larger. Fig. 3 is a histogram of the probability of variance of the power spectral density of the first signal in a manned environment, where the PSD probability of variance is significantly reduced and the variance distribution gradually moves to the right. Therefore, it can be concluded that the higher the environmental dynamics, the more spread the power spectral density variance distribution at different times to the right of the abscissa, and the closer the value is to 0.
According to the embodiment of the invention, a small amount of LoRa signal data under static unmanned environment and dynamic manned environment at different moments are collected, characteristic extraction is carried out to obtain a first signal reflecting human body life activity characteristics, namely an effective time domain signal, the time domain signal is converted into a frequency domain by utilizing fast Fourier transform, Power Spectral Density (PSD) is calculated, variance probability distribution of the PSD at different moments is further calculated to serve as an environmental characteristic value, and environmental characteristic sets under two environments are constructed, namely the first characteristic set and the second characteristic set.
For example, the first signal acquired online is compared with a feature set constructed in advance to detect whether a person is trapped in the target area, and similarity matching may be performed on a variance probability distribution of the power spectral density of the first signal acquired online with a variance probability distribution of the power spectral density of the first signal corresponding to the unmanned environment in the feature set (i.e., the first feature set) and a variance probability distribution of the power spectral density of the first signal corresponding to the manned environment in the feature set (i.e., the second feature set), that is, whether a person is trapped in the target area may be determined.
In some embodiments, similarity matching may be determined using a Hellinger Distance that measures similarity between different distributions. Illustratively, suppose the variance probability distributions of the two sets of power spectral densities are W ═ W (W), respectively1,w2,…,wk) And Q ═ Q (Q)1,q2,…qk) Then the hellingersistance metric for W and Q is calculated as follows:
the above equation can be viewed as the Euclidean distance of the square root vectors of two discrete probability distributions, as follows:
after the distance d (d is more than or equal to 0 and less than or equal to 1) between the online characteristic and the two offline characteristics is obtained, the system judges that the distance is the minimum and represents the most possible state of the current environment personnel. For example, if the distance between the online feature and the offline feature constructed in the dynamic manned environment is the minimum, the system judges that dynamic personnel exist and enters a personnel state detection module; and if the distance between the online feature and the offline feature constructed in the static unmanned environment is the minimum, determining the scene as a static unmanned scene, and continuously detecting whether the first signal at the next moment has a human target.
Therefore, the personnel search and rescue method provided by the embodiment of the invention can be used for detecting whether trapped personnel exist in the target area.
In some embodiments, if there are trapped people in the target area, the method further comprises:
detecting the motion state of the trapped person on the basis of a pre-constructed person state detection model for the first signal; the personnel state detection model is generated based on the first signal training corresponding to the static personnel and the moving personnel respectively.
In practical application, a personnel state detection model needs to be constructed in advance. In some embodiments, constructing the person state detection model comprises: extracting historical time-frequency characteristics, constructing a characteristic set and training a lightweight CNN classification model. Then, the motion state of the trapped person is detected on the first signal acquired on line based on the trained person state detection model.
The following respectively introduces historical time-frequency feature extraction and feature set construction, lightweight CNN classification model training and detection of the motion state of trapped people based on a person state detection model.
First, extracting historical time-frequency characteristics and constructing characteristic set
Due to the complexity of the signals, the original first signal is usually difficult to directly reflect the modulation characteristics of the human vital signals in the echo signal, and the first signal is directly adopted as the input of the CNN model, so that the identification accuracy is often not ideal. The time-frequency graph obtained by time-frequency analysis represents the joint distribution information of time domain and frequency domain, visually reflects the relation of each frequency component of the signal changing along with time, and contains rich environment state information.
In order to better enable the CNN model to accurately extract abundant vital signal features contained in the echo signal, in the embodiment of the invention, short-time Fourier transform is firstly adopted to perform feature transformation on the first signal to obtain a time-frequency graph of the signal. The time-frequency diagram of the signal can not only show the time-domain characteristics of the signal, but also show the frequency-domain characteristics of the signal. The short-time fourier transform is performed on each small segment of the signal by windowing the continuous time domain signal r (t), dividing the signal into small segments that can be considered temporally stationary. The short-time fourier transform is defined as:
r (t) is a time domain signal; g (t- τ) is a Hanning (hamming) window function; τ is the center of the window function.
Thus, the first signal obtained by low-pass filtering can be converted into a time-frequency characteristic diagram by using short-time Fourier transform.
In the embodiment of the invention, firstly, a small amount of off-line signal data (namely LoRa signal data) of a static state and a moving state of a person is collected, the off-line signal data is subjected to feature extraction to obtain a first signal for reflecting the life activity features of a human body, a time domain signal is converted into a time-frequency image by short-time Fourier transform, the pixel value of the extracted time-frequency image is stored as a 901 × 1201 × 3 three-dimensional matrix, a person state feature map set in two states is constructed and is divided into a training set and a testing set to be used as the input of training of a light-weight CNN classification model.
Second, lightweight CNN classification model training
In the embodiment of the invention, the lightweight CNN classification model can be a SqueezeNet model, and the model has the following advantages: 1) the 1 x 1 filter is used for replacing a 3x3 filter, the number of input channels input to a 3x3 convolution kernel is reduced, a global average pooling layer is used for replacing a full connection layer, model parameters are effectively reduced, and the method has the advantages that the cost of a client side updating model is lower, and the smaller updating model can be downloaded from a cloud side; 2) the downsampling layer is arranged at the rear part of the network as much as possible, so that the convolutional layer can have a larger activation characteristic diagram, and the classification precision of the network is not reduced due to the reduction of parameters.
Illustratively, as shown in fig. 4, the entire network of the squeezet model includes 10 layers, the 1 st layer is a 1 × 1 convolutional layer, the input image is reduced, and 96-dimensional features are extracted. Layers 2 to 9 are fire modules, which comprise 1 × 1 convolution and 3 × 3 convolution, and the interior of each module is firstly reduced by 1 × 1 convolution to reduce the number of channels (squeeze) and then increased by 3 × 3 to increase the number of channels (expand). After every two modules, the number of channels increases. Down-sampling max pooling with step size 2 is added after 1, 4, 8 layers, reducing the size by half. Layer 10 is again a convolutional layer, predicting a classification score for each pixel of the small graph. Finally, an averaging posing is used to obtain the classification score of the graph, and the classification score is normalized to probability by using a softmax function.
Illustratively, the specific training steps of the lightweight CNN classification model are as follows:
(1) a defined network model randomly initializes the weight and the bias of the network;
(2) reshaping the three-dimensional color picture pixel matrix in the training set constructed in the first step into three-dimensional matrix sample data of 28 × 1, and taking the three-dimensional matrix sample data as the input of the lightweight CNN model;
(3) the 1 st layer is 96 convolution kernels with the size of 1 × 1, the activation function of each layer is RELU, and a series of characteristic graphs related to personnel states are obtained through the activation function after the convolution of input time-frequency graphs of different personnel states;
(4) the 2 nd layer to the 9 th layer are eight Fire layers for reducing network parameters, wherein each layer comprises a compression layer and an expansion layer, the convolution kernel size of the compression layer is 1 x 1, the expansion layer comprises convolution kernels with the sizes of 1 x 1 and 3x3, the convolution kernel output with the size of 1 x 1 is connected with the convolution kernel output with the size of 3x3 in series to serve as the final output of the layer, and the size of the personnel state feature graph is further reduced; adding a maximum down-sampling layer with the step length of 2 after the layers 1, 4 and 8, and reducing the characteristic dimension;
(5) the 10 th convolution layer is 10 convolution kernels with the size of 1 x 1, finally extracted multi-dimensional personnel state characteristics are output, the characteristics are subjected to dimensionality reduction by utilizing a global average pooling layer and then are transmitted to a Softmax classifier, personnel states are classified and identified, predicted category labels are output, the predicted labels are compared with input actual labels, a prediction error is calculated, an error back propagation stage is started, weight values and bias parameters of all layers adjusted by a gradient descent method are used, then forward propagation is carried out again until the error between the predicted labels and the actual labels is reduced to a set threshold value, and at the moment, all parameters of the light-weight CNN are stored;
(6) in the stage of model verification, the model formed by the parameters is used for identifying the test set, if the obtained identification accuracy reaches the requirement, the model is considered to be the optimal model, the model can be directly used for identifying the data of unknown types, otherwise, the parameters of the model need to be further adjusted until the identification accuracy of the test set reaches the requirement.
Thirdly, detecting the motion state of the trapped person based on the person state detection model
In the embodiment of the present invention, the detecting the motion state of the trapped person based on the pre-established person state detection model for the first signal includes:
obtaining time-frequency data corresponding to the first signal based on short-time Fourier transform on the first signal;
and inputting the time-frequency data into a pre-constructed personnel state detection model to obtain a detection result that the trapped personnel are static personnel or moving personnel.
In the embodiment of the invention, a signal time-frequency diagram corresponding to the first signal is obtained by short-time Fourier transform, and finally, the time-frequency data of the signal time-frequency diagram is subjected to feature extraction and classification by using a lightweight CNN classification model trained in an offline training stage to obtain the state of personnel in the current area, namely the trapped personnel can be determined to be static personnel or moving personnel, so that different rescue or self-rescue strategies can be formulated for the moving state of the trapped personnel in the follow-up process.
In some embodiments, if the trapped person is a stationary person, the method further comprises:
determining a respiratory frequency of the stationary person based on a peak in a power spectral density of the first signal;
determining a hazard level for the stationary person based on the breathing frequency of the stationary person.
Here, the breathing frequency of the stationary person is further determined for the stationary person, and the danger level of the stationary person is determined based on the breathing frequency, so that the rescue plan can be made more targeted.
Illustratively, the breathing frequency of the stationary person is determined based on a peak in the power spectral density of the first signal.
Here, the fluctuating movement of the thorax may be caused by the respiration of the human body, and this fluctuating movement of the thorax may cause fluctuations in the power spectral density of the first signal.
In the embodiment of the present invention, the foregoing feature extraction may be performed on the received LoRa signal data to obtain the first signal, and the corresponding frequency domain signal is obtained after the discrete fourier transform is performed on the first signal, and the power spectral density is determined.
Illustratively, the discrete fourier transform corresponds to the transform formula as follows:
the power spectral density is formulated as follows:
where, r (N) is the I/Q value of the first signal, N is the length of the first signal, and k is 0,1, …, N-1.
In one example of use, the human breaths in an interval of about 10 to 40 breaths per minute, corresponding to a frequency of about 0.167Hz to 0.667 Hz. Because dangerous environments such as fire scene can cause the user to generate the psychology of tension, anxiety and fear, or the concentration of toxic gas in the air is increased, the heart rate of people is accelerated, the breathing is difficult, and the breathing frequency is improved; and as the fire spreads, the concentration of toxic gas in the air increases, the breathing of the person gradually becomes slow and weak, and then the person becomes unconscious until the breathing stops. According to medical experience, the breathing frequency of a normal person is 16-18 times per minute, and if the breathing is accelerated and even more than 30 times per minute, the breathing difficulty is considered, which indicates that the breathing frequency of a stationary person is very high, and the situation is difficult at present. If breathlessness occurs, even less than 10 breaths per minute, the concern is that the person is highly unconscious, requiring urgent (highest priority) rescue. Therefore, after the respiratory frequency and the current position of a plurality of detection personnel are estimated, the most efficient rescue route can be planned for the firefighters according to the two items, and all trapped personnel can be rescued most quickly.
In some embodiments, if the trapped person is an athletic person, the method further comprises:
and generating and transmitting an escape path for guiding the moving person to escape.
For example, an escape path for guiding the sports personnel to escape can be generated based on the current position of the sports personnel, fire situation information, crowd density and other factors, and the escape path is pushed to a mobile terminal (such as a mobile phone) of the sports personnel, so that the sports personnel can save themselves conveniently.
In order to implement the method of the embodiment of the present invention, an embodiment of the present invention further provides a device for searching and rescuing people for emergency rescue, where the device for searching and rescuing people for emergency rescue corresponds to the method for searching and rescuing people, and each step in the embodiment of the method for searching and rescuing people is also completely applicable to the embodiment of the device for searching and rescuing people for emergency rescue.
As shown in fig. 5, the device for searching and rescuing people for emergency rescue includes: the system comprises an acquisition module 501, a feature extraction module 502 and a personnel detection module 503, wherein the acquisition module 501 is used for acquiring LoRa signal data reflected by a LoRa transceiver in the process of scanning a target area; the feature extraction module 502 is configured to perform feature extraction on the LoRa signal data to obtain a first signal for reflecting human body vital movement features; the personnel detection module 503 is configured to perform personnel detection on the first signal based on a pre-established characteristic set, and determine whether trapped personnel exist in the target area; wherein the feature set is determined based on the first signals acquired in the unmanned environment and the manned environment, respectively.
In some embodiments, the people detection module 503 is specifically configured to:
converting the first signal into a frequency domain signal based on a fourier transform;
calculating a variance probability distribution of a power spectral density of the first signal based on the frequency domain signal;
calculating a first distance between the variance probability distribution and a variance probability distribution of the power spectral density of the first signal corresponding to the unmanned environment in a pre-constructed feature set, a second distance between the variance probability distribution and a variance probability distribution of the power spectral density of the first signal corresponding to the manned environment in the pre-constructed feature values;
if the first distance is smaller than the second distance, determining that no trapped person exists in the target area;
and if the first distance is greater than or equal to the second distance, determining that the trapped people exist in the target area.
In some embodiments, the feature extraction module 502 is specifically configured to:
performing data compression on the LoRa signal data;
and filtering the compressed data based on a high-pass filter and a low-pass filter to obtain the first signal.
In some embodiments, the device for searching and rescuing people for emergency rescue further includes:
a feature set constructing module 504, configured to obtain the first signals respectively corresponding to the unmanned environment and the manned environment;
converting the first signal corresponding to the unmanned environment into a frequency domain signal corresponding to the unmanned environment based on Fourier transform, and calculating a variance probability distribution of the power spectral density of the first signal corresponding to the unmanned environment based on the frequency domain signal corresponding to the unmanned environment;
converting the first signal corresponding to the manned environment into a frequency domain signal corresponding to the manned environment based on Fourier transform, and calculating the variance probability distribution of the power spectral density of the first signal corresponding to the manned environment based on the frequency domain signal corresponding to the manned environment;
wherein the pre-constructed feature set comprises: a variance probability distribution of a power spectral density of the first signal corresponding to the unmanned environment and a variance probability distribution of a power spectral density of the first signal corresponding to the manned environment.
In some embodiments, if there are trapped people in the target area, the people detection module 503 is further configured to: detecting the motion state of the trapped person on the basis of a pre-constructed person state detection model for the first signal; the personnel state detection model is generated based on the first signal training corresponding to the static personnel and the moving personnel respectively.
In some embodiments, the people detection module 503 is specifically configured to:
obtaining time-frequency data corresponding to the first signal based on short-time Fourier transform on the first signal;
and inputting the time-frequency data into a pre-constructed personnel state detection model to obtain a detection result that the trapped personnel are static personnel or moving personnel.
In some embodiments, if the trapped person is a stationary person, the person detection module 503 is further configured to:
determining a respiratory frequency of the stationary person based on a peak in a power spectral density of the first signal;
determining a hazard level for the stationary person based on the breathing frequency of the stationary person.
In some embodiments, the device for searching and rescuing people for emergency rescue further includes:
and the path planning module 505 is used for generating and sending an escape path for guiding the sportsman to escape.
In practical application, the obtaining module 501, the feature extracting module 502, the person detecting module 503, the feature set constructing module 504, and the path planning module 505 may be implemented by a processor in a person search and rescue device for emergency rescue. Of course, the processor needs to run a computer program in memory to implement its functions.
It should be noted that: when the device for searching and rescuing people for emergency rescue provided in the above embodiment is used for searching and rescuing people for emergency rescue, only the division of the above program modules is taken as an example, in practical application, the above processing distribution can be completed by different program modules according to needs, that is, the internal structure of the device is divided into different program modules to complete all or part of the above-described processing. In addition, the person search and rescue device for emergency rescue provided by the embodiment and the person search and rescue method for emergency rescue belong to the same concept, and specific implementation processes thereof are detailed in the method embodiment and are not described again.
Based on the hardware implementation of the program module, in order to implement the method of the embodiment of the invention, the embodiment of the invention also provides a personnel search and rescue device for emergency rescue. Fig. 6 shows only an exemplary structure of the person search and rescue apparatus, not a whole structure, and a part of or the whole structure shown in fig. 6 may be implemented as necessary.
As shown in fig. 6, a person search and rescue apparatus 600 according to an embodiment of the present invention includes: at least one processor 601, memory 602, user interface 603, and at least one network interface 604. The various components in the person search and rescue apparatus 600 are coupled together by a bus system 605. It will be appreciated that the bus system 605 is used to enable communications among the components. The bus system 605 includes a power bus, a control bus, and a status signal bus in addition to a data bus. For clarity of illustration, however, the various buses are labeled as bus system 605 in fig. 6.
The user interface 603 may include, among other things, a display, a keyboard, a mouse, a trackball, a click wheel, a key, a button, a touch pad, or a touch screen.
The memory 602 in embodiments of the present invention is used to store various types of data to support the operation of the personnel search and rescue equipment. Examples of such data include: any computer program for operating on a person search and rescue apparatus.
The method for searching and rescuing the personnel for emergency rescue disclosed by the embodiment of the invention can be applied to the processor 601 or realized by the processor 601. The processor 601 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the method for rescuing people for emergency rescue may be implemented by integrated logic circuits of hardware or instructions in the form of software in the processor 601. The Processor 601 may be a general purpose Processor, a Digital Signal Processor (DSP), or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, or the like. Processor 601 may implement or perform the methods, steps, and logic blocks disclosed in embodiments of the present invention. A general purpose processor may be a microprocessor or any conventional processor or the like. The steps of the method disclosed by the embodiment of the invention can be directly implemented by a hardware decoding processor, or can be implemented by combining hardware and software modules in the decoding processor. The software modules may be located in a storage medium located in the memory 602, and the processor 601 reads the information in the memory 602, and completes the steps of the method for searching and rescuing people for emergency rescue provided by the embodiment of the present invention in combination with the hardware thereof.
In an exemplary embodiment, the personnel search and rescue Device may be implemented by one or more Application Specific Integrated Circuits (ASICs), DSPs, Programmable Logic Devices (PLDs), Complex Programmable Logic Devices (CPLDs), FPGAs, general purpose processors, controllers, Micro Controllers (MCUs), microprocessors (microprocessors), or other electronic components for performing the aforementioned methods.
It will be appreciated that the memory 602 can be either volatile memory or nonvolatile memory, and can include both volatile and nonvolatile memory. Among them, the nonvolatile Memory may be a Read Only Memory (ROM), a Programmable Read Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), an Electrically Erasable Programmable Read-Only Memory (EEPROM), a magnetic random access Memory (FRAM), a Flash Memory (Flash Memory), a magnetic surface Memory, an optical disk, or a Compact Disc Read-Only Memory (CD-ROM); the magnetic surface storage may be disk storage or tape storage. Volatile Memory can be Random Access Memory (RAM), which acts as external cache Memory. By way of illustration and not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Synchronous Static Random Access Memory (SSRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), Enhanced Synchronous Dynamic Random Access Memory (Enhanced DRAM), Synchronous Dynamic Random Access Memory (SLDRAM), Direct Memory (DRmb Access), and Random Access Memory (DRAM). The described memory for embodiments of the present invention is intended to comprise, without being limited to, these and any other suitable types of memory.
It can be understood that the personnel search and rescue equipment in the embodiment of the invention can be edge computing nodes near an emergency rescue scene. This personnel search and rescue equipment can with carry on the loRa transceiver communication connection on unmanned aerial vehicle, obtain the echo signal that the loRa transceiver was received (promptly aforementioned loRa signal data) to realize the detection of online personnel's detection and personnel's state.
An embodiment of the present invention further provides a personnel search and rescue system for emergency rescue, as shown in fig. 7, the personnel search and rescue system includes:
an unmanned aerial vehicle 701, wherein an LoRa transceiver is carried on the unmanned aerial vehicle 701;
the personnel search and rescue device 702 is in communication connection with the unmanned aerial vehicle 701, and is configured to execute the steps of the method according to the first aspect of the embodiment of the present invention when running a computer program.
Here, the personnel search and rescue equipment 702 may be in communication connection with the LoRa transceiver mounted on the drone 701, receive the LoRa signal data acquired by the LoRa transceiver, and execute the personnel search and rescue method according to the foregoing embodiment of the present invention. Or, the LoRa signal data that the LoRa transceiver gathered transmit to unmanned aerial vehicle 701's treater on, personnel search and rescue equipment 702 establishes communication connection with unmanned aerial vehicle 701 to receive the LoRa signal data that unmanned aerial vehicle 701's treater forwarded.
In some embodiments, the person search and rescue system may further include a cloud server at a remote end, and the person search and rescue device 702 is in communication connection with the cloud server, so as to form an end-edge-cloud cooperative system, wherein the unmanned aerial vehicle 701 and an onboard LoRa transceiver thereof form a terminal for acquiring LoRa signal data, the person search and rescue device 702 serves as an edge computing node, and can detect whether trapped persons exist or not and motion states of the trapped persons, and the like, and the cloud server may use the pre-constructed feature set and the pre-constructed person state detection model, so as to effectively reduce resource consumption of the processor.
In an exemplary embodiment, the embodiment of the present invention further provides a storage medium, that is, a computer storage medium, which may be specifically a computer-readable storage medium, for example, a memory 602 storing a computer program, where the computer program is executable by a processor 601 of the person search and rescue device to perform the steps described in the method according to the embodiment of the present invention. The computer readable storage medium may be a ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface Memory, optical disk, or CD-ROM, among others.
It should be noted that: "first," "second," and the like are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.
In addition, the technical solutions described in the embodiments of the present invention may be arbitrarily combined without conflict.
The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and all the changes or substitutions should be covered within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (12)
1. A personnel search and rescue method for emergency rescue is characterized by comprising the following steps:
acquiring LoRa signal data reflected by a long-distance radio LoRa transceiver in the process of scanning a target area;
performing feature extraction on the LoRa signal data to obtain a first signal for reflecting human body life activity features;
carrying out personnel detection on the first signal based on a pre-constructed characteristic set, and determining whether trapped personnel exist in the target area; wherein the feature set is determined based on the first signals acquired in the unmanned environment and the manned environment, respectively.
2. The method of claim 1, wherein the performing person detection on the first signal based on a pre-constructed feature set to determine whether trapped persons are present in the target area comprises:
converting the first signal into a frequency domain signal based on a fourier transform;
calculating a variance probability distribution of a power spectral density of the first signal based on the frequency domain signal;
calculating a first distance between the variance probability distribution and a variance probability distribution of the power spectral density of the first signal corresponding to the unmanned environment in a pre-constructed feature set, a second distance between the variance probability distribution and a variance probability distribution of the power spectral density of the first signal corresponding to the manned environment in the pre-constructed feature values;
if the first distance is smaller than the second distance, determining that no trapped person exists in the target area;
and if the first distance is greater than or equal to the second distance, determining that the trapped people exist in the target area.
3. The method of claim 1, wherein the performing feature extraction on the LoRa signal data to obtain a first signal for reflecting human body vital activity features comprises:
performing data compression on the LoRa signal data;
and filtering the compressed data based on a high-pass filter and a low-pass filter to obtain the first signal.
4. The method of claim 1, wherein prior to performing person detection on the first signal based on a pre-constructed set of features, the method further comprises:
acquiring the first signals respectively corresponding to the unmanned environment and the manned environment;
converting the first signal corresponding to the unmanned environment into a frequency domain signal corresponding to the unmanned environment based on Fourier transform, and calculating a variance probability distribution of the power spectral density of the first signal corresponding to the unmanned environment based on the frequency domain signal corresponding to the unmanned environment;
converting the first signal corresponding to the manned environment into a frequency domain signal corresponding to the manned environment based on Fourier transform, and calculating the variance probability distribution of the power spectral density of the first signal corresponding to the manned environment based on the frequency domain signal corresponding to the manned environment;
wherein the pre-constructed feature set comprises: a variance probability distribution of a power spectral density of the first signal corresponding to the unmanned environment and a variance probability distribution of a power spectral density of the first signal corresponding to the manned environment.
5. The method of claim 1, wherein if there are trapped people in the target area, the method further comprises:
detecting the motion state of the trapped person on the basis of a pre-constructed person state detection model for the first signal; the personnel state detection model is generated based on the first signal training corresponding to the static personnel and the moving personnel respectively.
6. The method of claim 5, wherein the detecting the motion state of the trapped person based on a pre-constructed person state detection model for the first signal comprises:
obtaining time-frequency data corresponding to the first signal based on short-time Fourier transform on the first signal;
and inputting the time-frequency data into a pre-constructed personnel state detection model to obtain a detection result that the trapped personnel are static personnel or moving personnel.
7. The method of claim 6, wherein if the trapped person is a stationary person, the method further comprises:
determining a respiratory frequency of the stationary person based on a peak in a power spectral density of the first signal;
determining a hazard level for the stationary person based on the breathing frequency of the stationary person.
8. The method of claim 6, wherein if the trapped person is an athletic person, the method further comprises:
and generating and transmitting an escape path for guiding the moving person to escape.
9. A personnel search and rescue device for emergency rescue, characterized by comprising:
the acquisition module is used for acquiring LoRa signal data reflected by the LoRa transceiver in the process of scanning the target area;
the characteristic extraction module is used for carrying out characteristic extraction on the LoRa signal data to obtain a first signal for reflecting the human body life activity characteristics;
the personnel detection module is used for carrying out personnel detection on the first signal based on a pre-constructed characteristic set and determining whether trapped personnel exist in the target area; wherein the feature set is determined based on the first signals acquired in the unmanned environment and the manned environment, respectively.
10. A personnel search and rescue equipment for emergency rescue, characterized by comprising:
a processor and a memory for storing a computer program capable of running on the processor, wherein,
the processor, communicatively connected to the drone, is configured to perform the steps of the method of any one of claims 1 to 8 when running the computer program.
11. A personnel search and rescue system for emergency rescue, comprising:
the system comprises an unmanned aerial vehicle, wherein an LoRa transceiver is carried on the unmanned aerial vehicle;
a personnel search and rescue device as claimed in claim 10 in communicative connection with the drone.
12. A storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, performs the steps of the method of any one of claims 1 to 8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010803844.5A CN112132719B (en) | 2020-08-11 | 2020-08-11 | Personnel search and rescue method, device and system for emergency rescue and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010803844.5A CN112132719B (en) | 2020-08-11 | 2020-08-11 | Personnel search and rescue method, device and system for emergency rescue and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112132719A true CN112132719A (en) | 2020-12-25 |
CN112132719B CN112132719B (en) | 2023-06-09 |
Family
ID=73850276
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010803844.5A Active CN112132719B (en) | 2020-08-11 | 2020-08-11 | Personnel search and rescue method, device and system for emergency rescue and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112132719B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112699938A (en) * | 2020-12-30 | 2021-04-23 | 北京邮电大学 | Classification method and device based on graph convolution network model |
CN113819959A (en) * | 2021-11-24 | 2021-12-21 | 中国空气动力研究与发展中心设备设计与测试技术研究所 | Suspension system anomaly detection method based on Hailinge distance and correlation coefficient |
CN113934971A (en) * | 2021-10-15 | 2022-01-14 | 中国矿业大学 | Long-distance sensing method for underground rescue robot based on LoRa |
CN114779816A (en) * | 2022-05-17 | 2022-07-22 | 成都工业学院 | Searching and rescuing unmanned aerial vehicle for lifting in earthquake ruin environment and system thereof |
CN116704386A (en) * | 2023-08-01 | 2023-09-05 | 四川开澜科技有限公司 | AI-based accurate emergency rescue method and device |
CN117676758A (en) * | 2024-02-02 | 2024-03-08 | 厦门拾聚科技有限公司 | Communication security and protection integrated routing system, method, equipment and storage medium |
Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104837112A (en) * | 2015-03-30 | 2015-08-12 | 南京邮电大学 | Wireless sensing network system and method for searching and rescuing trapped persons |
CN107393550A (en) * | 2017-07-14 | 2017-11-24 | 深圳永顺智信息科技有限公司 | Method of speech processing and device |
FR3057984A1 (en) * | 2016-10-25 | 2018-04-27 | Francois Domange | METHOD AND SYSTEM FOR AUTOMATIC EMERGENCY RELIEF CALLING IN DETECTION OF A PERSON IN A DISTRESS SITUATION |
CN108194129A (en) * | 2018-03-28 | 2018-06-22 | 中国矿业大学(北京) | A kind of mine Disastrous environment search and rescue device based on aircraft |
US20180183875A1 (en) * | 2016-12-27 | 2018-06-28 | Sichuan Rex Smart Technology Corporation Limited | System and method for fire ground entry control based on internet of things |
CN108521565A (en) * | 2018-06-14 | 2018-09-11 | 恒德科技有限公司 | Individual soldier's emergency management and rescue intelligent wearable device |
CN108828599A (en) * | 2018-04-06 | 2018-11-16 | 东莞市华睿电子科技有限公司 | A kind of disaster affected people method for searching based on rescue unmanned plane |
CN109087475A (en) * | 2018-10-19 | 2018-12-25 | 河南汇纳科技有限公司 | A kind of forest fireproofing early warning system based on LoRa wireless network |
CN109118740A (en) * | 2018-09-14 | 2019-01-01 | 公安部沈阳消防研究所 | A kind of toxic and harmful gas monitoring system and data transmission processing method being exclusively used in fire brigade fire fighting and rescue scene |
CN109360368A (en) * | 2018-10-11 | 2019-02-19 | 深圳市亿兆互联技术有限公司 | Fire disaster escaping method and system based on LoRa |
CN109714747A (en) * | 2019-02-13 | 2019-05-03 | 上海上晟电子科技有限公司 | Intelligent lifesaving system based on NFC and LoRa communication and unmanned air vehicle technique |
CN109993063A (en) * | 2019-03-05 | 2019-07-09 | 福建天晴数码有限公司 | A kind of method and terminal identified to rescue personnel |
CN110109480A (en) * | 2019-05-21 | 2019-08-09 | 南京邮电大学 | A kind of unmanned plane inspection search and rescue system and rescue method based on Multi-sensor Fusion |
CN110426703A (en) * | 2019-07-01 | 2019-11-08 | 东南大学 | A kind of portable wireless search and rescue system and rescue method based on high-frequency radar |
-
2020
- 2020-08-11 CN CN202010803844.5A patent/CN112132719B/en active Active
Patent Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104837112A (en) * | 2015-03-30 | 2015-08-12 | 南京邮电大学 | Wireless sensing network system and method for searching and rescuing trapped persons |
FR3057984A1 (en) * | 2016-10-25 | 2018-04-27 | Francois Domange | METHOD AND SYSTEM FOR AUTOMATIC EMERGENCY RELIEF CALLING IN DETECTION OF A PERSON IN A DISTRESS SITUATION |
US20180183875A1 (en) * | 2016-12-27 | 2018-06-28 | Sichuan Rex Smart Technology Corporation Limited | System and method for fire ground entry control based on internet of things |
CN107393550A (en) * | 2017-07-14 | 2017-11-24 | 深圳永顺智信息科技有限公司 | Method of speech processing and device |
CN108194129A (en) * | 2018-03-28 | 2018-06-22 | 中国矿业大学(北京) | A kind of mine Disastrous environment search and rescue device based on aircraft |
CN108828599A (en) * | 2018-04-06 | 2018-11-16 | 东莞市华睿电子科技有限公司 | A kind of disaster affected people method for searching based on rescue unmanned plane |
CN108521565A (en) * | 2018-06-14 | 2018-09-11 | 恒德科技有限公司 | Individual soldier's emergency management and rescue intelligent wearable device |
CN109118740A (en) * | 2018-09-14 | 2019-01-01 | 公安部沈阳消防研究所 | A kind of toxic and harmful gas monitoring system and data transmission processing method being exclusively used in fire brigade fire fighting and rescue scene |
CN109360368A (en) * | 2018-10-11 | 2019-02-19 | 深圳市亿兆互联技术有限公司 | Fire disaster escaping method and system based on LoRa |
CN109087475A (en) * | 2018-10-19 | 2018-12-25 | 河南汇纳科技有限公司 | A kind of forest fireproofing early warning system based on LoRa wireless network |
CN109714747A (en) * | 2019-02-13 | 2019-05-03 | 上海上晟电子科技有限公司 | Intelligent lifesaving system based on NFC and LoRa communication and unmanned air vehicle technique |
CN109993063A (en) * | 2019-03-05 | 2019-07-09 | 福建天晴数码有限公司 | A kind of method and terminal identified to rescue personnel |
CN110109480A (en) * | 2019-05-21 | 2019-08-09 | 南京邮电大学 | A kind of unmanned plane inspection search and rescue system and rescue method based on Multi-sensor Fusion |
CN110426703A (en) * | 2019-07-01 | 2019-11-08 | 东南大学 | A kind of portable wireless search and rescue system and rescue method based on high-frequency radar |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112699938A (en) * | 2020-12-30 | 2021-04-23 | 北京邮电大学 | Classification method and device based on graph convolution network model |
CN112699938B (en) * | 2020-12-30 | 2024-01-05 | 北京邮电大学 | Classification method and device based on graph convolution network model |
CN113934971A (en) * | 2021-10-15 | 2022-01-14 | 中国矿业大学 | Long-distance sensing method for underground rescue robot based on LoRa |
CN113819959A (en) * | 2021-11-24 | 2021-12-21 | 中国空气动力研究与发展中心设备设计与测试技术研究所 | Suspension system anomaly detection method based on Hailinge distance and correlation coefficient |
CN114779816A (en) * | 2022-05-17 | 2022-07-22 | 成都工业学院 | Searching and rescuing unmanned aerial vehicle for lifting in earthquake ruin environment and system thereof |
CN114779816B (en) * | 2022-05-17 | 2023-03-24 | 成都工业学院 | Searching and rescuing unmanned aerial vehicle for taking off and landing in ruins environment after earthquake and system thereof |
CN116704386A (en) * | 2023-08-01 | 2023-09-05 | 四川开澜科技有限公司 | AI-based accurate emergency rescue method and device |
CN116704386B (en) * | 2023-08-01 | 2023-10-20 | 四川开澜科技有限公司 | AI-based accurate emergency rescue method and device |
CN117676758A (en) * | 2024-02-02 | 2024-03-08 | 厦门拾聚科技有限公司 | Communication security and protection integrated routing system, method, equipment and storage medium |
CN117676758B (en) * | 2024-02-02 | 2024-05-07 | 厦门拾聚科技有限公司 | Communication security and protection integrated routing system, method, equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN112132719B (en) | 2023-06-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112132719B (en) | Personnel search and rescue method, device and system for emergency rescue and storage medium | |
Cippitelli et al. | Radar and RGB-depth sensors for fall detection: A review | |
US10621847B2 (en) | Human respiration feature extraction in personal emergency response systems and methods | |
Yu et al. | Noninvasive human activity recognition using millimeter-wave radar | |
CN112401856B (en) | Nursing home monitoring method and system based on millimeter wave radar | |
CN111115400B (en) | System and method for detecting elevator maintenance behavior in an elevator hoistway | |
CN109803207A (en) | Mark to the audio signal in ambient sound and the guidance to the autonomous vehicles in response to the mark | |
CN112346050B (en) | Fall detection method and system based on Wi-Fi equipment | |
CN113380272A (en) | Audio-based emergency vehicle detection and tracking | |
CN110007366B (en) | Life searching method and system based on multi-sensor fusion | |
EP3639748A1 (en) | System and method for monitoring pathological breathing patterns | |
US11380181B2 (en) | Doppler radar system with machine learning applications for fall prediction and detection | |
AU2018375154B2 (en) | Ultrasound analytics for actionable information | |
KR102200409B1 (en) | System and method for monitoring human condition using two-tone radar | |
CN112270246B (en) | Video behavior recognition method and device, storage medium and electronic equipment | |
Lian et al. | Fall detection via inaudible acoustic sensing | |
EP4102247A1 (en) | System and method for radio-assisted sound sensing | |
JP2023058446A (en) | Method, apparatus, and system for wireless sensing, monitoring, and tracking | |
CN112327288A (en) | Radar human body action recognition method and device, electronic equipment and storage medium | |
US11646009B1 (en) | Autonomously motile device with noise suppression | |
Weng et al. | Review of analyses on crowd-gathering risk and its evaluation methods | |
Wang et al. | Radar-based soft fall detection using pattern contour vector | |
CN114049587A (en) | Event detection method, server and system | |
Peng et al. | ULSED: An ultra-lightweight SED model for IoT devices | |
CN115844402A (en) | Attention prediction method, attention prediction device, electronic equipment and computer readable storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |