CN112363114B - Public place acoustic event positioning method and system based on distributed noise sensor - Google Patents
Public place acoustic event positioning method and system based on distributed noise sensor Download PDFInfo
- Publication number
- CN112363114B CN112363114B CN202110045899.9A CN202110045899A CN112363114B CN 112363114 B CN112363114 B CN 112363114B CN 202110045899 A CN202110045899 A CN 202110045899A CN 112363114 B CN112363114 B CN 112363114B
- Authority
- CN
- China
- Prior art keywords
- node
- acoustic event
- event
- nodes
- sound
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
- G01S5/18—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using ultrasonic, sonic, or infrasonic waves
- G01S5/22—Position of source determined by co-ordinating a plurality of position lines defined by path-difference measurements
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
Abstract
Public place acoustic event positioning method and system based on distributed noise sensors. The method comprises the following steps: each node is accessed to the 5G network through the 5G networking module; after networking, each node and other nodes form a test sub-network through the main node; each node starts a Beidou navigation module to carry out time synchronization and obtains node position information; each node carries out acoustic event detection and identification; after each node detects an acoustic event, recording acoustic event audio data and a timestamp of the occurrence of the acoustic event; all the slave nodes form summarized data by sound event audio data, sound event occurrence time stamps and node position information and package the summarized data to the master node; the master node judges whether the sound event audio data are the same sound event or not and judges the type of the sound event according to the sound event audio data captured by the master node and received from the slave node; and when the situation that the same sensitive sound event is detected by at least three nodes is judged, the main node carries out sound source position positioning by utilizing the summarized data.
Description
Technical Field
The present invention relates to acoustic event localization, and more particularly, to a method and system for acoustic event localization based on distributed noise sensors.
Background
At present, no quick and effective means exists for early warning and dispersion of emergency in public places. In places with dense people, such as stadiums, parks, squares and the like, once malignant public events, such as explosions, gunshots and the like, occur, if the types and the positions of the events cannot be timely and effectively determined, effective evacuation guidance is provided for people, the events may be continuously deteriorated, and more serious results are caused by treading events and the like.
Therefore, a solution for timely and efficient acoustic event detection and localization is needed.
Disclosure of Invention
In order to solve the problems of the prior art, the invention provides a public place acoustic event positioning method based on a distributed noise sensor, which comprises the following steps:
providing at least three nodes, wherein one node is configured as a master node, and the other nodes are configured as slave nodes, wherein each node comprises a noise sensor, a noise sensor probe, a 5G networking module and a Beidou navigation module;
all nodes are started, and each node is accessed to the 5G network through the 5G networking module;
after being networked, each node forms a test sub-network with other nodes through the main node;
each node starts the Beidou navigation module to carry out time synchronization and obtains node position information;
each node utilizes the noise sensor probe and the noise sensor to detect and identify the acoustic event;
after each node detects an acoustic event, recording acoustic event audio data and a timestamp of the occurrence of the acoustic event;
all the slave nodes form summarized data by the sound event audio data, the time stamp of the sound event and the node position information, and the summarized data are packaged and sent to the master node;
the master node judges whether the sound event audio data are the same sound event or not and judges the type of the sound event according to the sound event audio data captured by the master node and received from the slave node; and
and when the master node judges that more than three nodes detect the same sensitive sound event, the master node performs sound source position positioning by using the summarized data.
In one embodiment, the acoustic event detection identification comprises:
intercepting the real-time audio stream to obtain the audio data, wherein the intercepting time length is 2t, and t is greater than 0;
enabling the interception period of the real-time audio stream to be less than 2 t;
extracting cochlear atlas feature data from the real-time audio stream through a Gamma atom filter bank;
inputting the map feature data serving as a feature map into a trained convolutional neural network model for calculation to obtain an output result, wherein the output result comprises: acoustic event type, acoustic event start time;
and judging the type of the acoustic event, if the acoustic event is the expected acoustic event type, intercepting the audio data according to the front and back of the initial time point of the acoustic event, and adding node position information and the timestamp of the occurrence of the acoustic event and forwarding the node position information and the timestamp of the occurrence of the acoustic event to the host node.
In one embodiment, the sound source position location comprises:
the master node performs amplitude normalization on the audio data forwarded by each slave node;
the main node performs power maximum value fitting on the audio data after the amplitude is normalized, and time difference values among data corresponding to all the nodes are calculated;
and establishing an equation set according to the three-dimensional coordinates of each node and the arrival time difference of the acoustic events among the nodes, and solving the equation set to obtain the three-dimensional coordinates of the sound source.
In one embodiment, the method further comprises:
and after the sound source position is positioned, pushing the type of the sound event and the sound event position obtained by positioning the sound source to a target user for early warning.
In one embodiment, the nodes perform distributed networking based on a 5G network, and a high-speed and low-delay data interaction channel is guaranteed between the nodes.
In one embodiment, when the public place is a large plane, the coordinate system of the nodes is simplified to represent the position of each node by a two-dimensional coordinate system, so as to improve the positioning speed of sound source position positioning.
The invention also provides a public place acoustic event positioning system based on the distributed noise sensor, which comprises at least three nodes, wherein one node is configured as a master node, and the other nodes are configured as slave nodes, and each node comprises: noise sensor, noise sensor probe, 5G networking module and big dipper navigation module.
The 5G networking module is configured to provide network connection, ensure that all nodes can form a small-range monitoring network, provide a high-speed and low-delay data interaction channel and ensure burst transmission of acoustic event data;
the Beidou navigation module is configured to perform geographical position positioning on each node to obtain node position information which is used as a data basis for sound source position positioning; the Beidou navigation module is also configured to carry out clock time service so as to ensure time synchronization among all nodes;
the noise sensor probe is configured to collect noise of a public place;
the noise sensor is configured to quantize and sample the analog audio data collected by the noise sensor probe to form a digital signal, and comprises a processor suitable for edge calculation, wherein the processor performs acoustic event detection identification, and records acoustic event audio data and a timestamp of occurrence of an acoustic event after the acoustic event is detected;
the processor of the slave node also composes summarized data of the audio data of the acoustic event, the timestamp of the occurrence of the acoustic event and the node position information and sends the summarized data to the master node in a packaging manner; and the processor of the master node judges whether the sound event audio data are the same sound event or not and judges the type of the sound event according to the sound event audio data captured by the processor and received from the slave node, and when the judgment result shows that more than three nodes detect the same sensitive sound event, the summarized data is utilized to position the sound source.
In one embodiment, the processor of the noise sensor performing acoustic event detection identification includes:
intercepting the real-time audio stream to obtain the audio data, wherein the intercepting time length is 2t, and t is greater than 0;
enabling the interception period of the real-time audio stream to be less than 2 t;
extracting cochlear atlas feature data from the real-time audio stream through a Gamma atom filter bank;
inputting the map feature data serving as a feature map into a trained convolutional neural network model for calculation to obtain an output result, wherein the output result comprises: acoustic event type, acoustic event start time;
and judging the type of the acoustic event, if the acoustic event is the expected acoustic event type, intercepting the audio data according to the front and back of the initial time point of the acoustic event, and adding node position information and the timestamp of the occurrence of the acoustic event and forwarding the node position information and the timestamp of the occurrence of the acoustic event to the host node.
In one embodiment, the sound source position location comprises:
the master node performs amplitude normalization on the audio data forwarded by each slave node;
the main node performs power maximum value fitting on the audio data after the amplitude is normalized, and time difference values among data corresponding to all the nodes are calculated;
and establishing an equation set according to the three-dimensional coordinates of each node and the arrival time difference of the acoustic events among the nodes, and solving the equation set to obtain the three-dimensional coordinates of the sound source.
In one embodiment, after the noise sensor performs sound source location, the type of the sound event and the sound event location obtained by sound source location are pushed to a target user for early warning.
In one embodiment, the nodes perform distributed networking based on a 5G network, and a high-speed and low-delay data interaction channel is guaranteed between the nodes.
In one embodiment, when the public place is a large plane, the coordinate system of the nodes is simplified to represent the position of each node by a two-dimensional coordinate system, so as to improve the positioning speed of sound source position positioning.
The method is suitable for detecting and positioning the sensitive sound event in an open outdoor or semi-outdoor public place, and can quickly detect the occurrence of the sensitive sound event, the type of the sound event and the position of the sound event; important data support is provided for crowd evacuation and public safety decision making.
Drawings
The foregoing summary, as well as the following detailed description of the invention, will be better understood when read in conjunction with the appended drawings. It is to be noted that the appended drawings are intended as examples of the claimed invention. In the drawings, like reference characters designate the same or similar elements.
FIG. 1 illustrates a schematic composition of each node in a distributed noise sensor based public space acoustic event locating system according to an embodiment of the present invention;
FIG. 2 illustrates a workflow diagram for distributed noise sensor based public space acoustic event localization according to an embodiment of the present invention.
Detailed Description
The detailed features and advantages of the present invention are described in detail in the detailed description which follows, and will be sufficient for anyone skilled in the art to understand the technical content of the present invention and to implement the present invention, and the related objects and advantages of the present invention will be easily understood by those skilled in the art from the description, claims and drawings disclosed in the present specification.
The method is suitable for detecting and positioning the sensitive sound event in an open outdoor or semi-outdoor public place, and can quickly detect the occurrence of the sensitive sound event, the type of the sound event and the position of the sound event. Important data support is provided for crowd evacuation and public safety decision making.
The public place acoustic event positioning system can comprise at least 3 nodes for networking work and is in a distributed structure. In the nodes, a main node is arranged and is responsible for data summarization and sound source positioning calculation. The rest are set as slave nodes. Each node comprises a proper processor platform, so that the nodes have the characteristic of edge calculation, and each node can independently run an AI algorithm to efficiently monitor the sound event in real time.
It is noted that the acoustic event localization system of the present invention needs to work efficiently on the premise that at least 3 nodes are guaranteed to detect acoustic events. When there are more than 3 nodes, the accuracy of the positioning can be improved.
The structure of each node is shown in fig. 1. The node at least comprises a noise sensor 101, a noise sensor probe 102, a 5G networking module 103 and a Beidou navigation satellite system module 104.
The noise sensor probe 102 is used to collect noise in public places.
The noise sensor 101 is used to quantize and sample the analog audio data collected by the noise sensor probe to form a digital signal. The noise sensor 101 includes a processor adapted for edge computation such that nodes are characterized by edge computation, and the processor runs an AI algorithm to efficiently monitor acoustic events in real time and identify acoustic event types.
The 5G networking module 103 is configured to provide 5G network connection, and ensure that all nodes can form a small-range monitoring network. And meanwhile, a data interaction channel with high speed and low delay is provided, and burst transmission of acoustic event data is ensured.
The Beidou navigation module 104 is used for accurately positioning the geographical position of each node, and the positioning position (namely positioning information) of each node is an important data basis for sound source positioning; meanwhile, the Beidou navigation module 104 also carries out clock time service, and extremely high time synchronization between each node is guaranteed. The purpose of time synchronization is to ensure that the timestamps provided by the nodes are accurate and synchronized, among other things. The time synchronization index directly determines the accuracy of sound source localization.
FIG. 2 illustrates a workflow diagram for distributed noise sensor based public space acoustic event localization according to an embodiment of the present invention.
Step 201: at least three nodes are provided, one of which is configured as a master node and the remaining nodes are configured as slave nodes.
Step 202: and opening all the nodes, and accessing the nodes to the 5G network through the 5G networking module.
Step 203: after each node is networked, a test sub-network is formed by the main node and other nodes;
step 204: after all the nodes are connected to the network, each node opens a Beidou navigation module to carry out time synchronization and acquires geographic position data.
Step 205: and each node starts to detect and identify the acoustic event.
Step 206: after each node detects an acoustic event, acoustic event audio data is recorded along with a timestamp of the occurrence of the acoustic event.
Step 207: all the slave nodes form summarized data by sound event audio data, sound event occurrence time stamps and node position information and package the summarized data to the master node.
Step 208: the master node uses all received acoustic event audio data (including acoustic event audio data captured by the master node itself) to determine whether the same acoustic event and the type of event are present using an acoustic event detection recognition algorithm.
Step 209: when a sensitive acoustic event is detected, step 210 is performed.
Step 210: when more than 3 nodes detect the same sensitive sound event, the main node uses the summarized data to position the sound source.
Step 211: and pushing the type of the acoustic event and the position of the acoustic event to a target user for early warning.
In one embodiment, the acoustic event detection and recognition algorithm is an acoustic event recognition algorithm obtained by machine learning training, and a certain data processing capacity is required for the nodes to run the algorithm.
In one embodiment, the sound source localization of the present invention is based on the TDOA principle (time difference of arrival principle), i.e., the location of a sound source is estimated by the time difference of arrival of the sound at each node. The basic working principle is as follows: when an acoustic event occurs, the arrival time of the acoustic event at each node is different. When an acoustic event arrives at a node, the node timestamps the acoustic event. And putting all the nodes into a three-dimensional coordinate system, and simultaneously establishing an equation set by detecting the time difference of the acoustic event and the position, the position and the sound velocity of the acoustic source by different nodes. When the number of the nodes detecting the acoustic event reaches more than 3, the simultaneous equations can solve the position of the acoustic source in the three-dimensional coordinate system.
It should be noted that TDOA-based sound source localization algorithms generally occur in microphone arrays; the present invention, while the principle is also TDOA, is used much closer to the TDOA algorithm in radio passive location.
The accuracy of sound source positioning needs to rely on the high-precision time service of the Beidou navigation module, and each node is guaranteed to keep good event synchronism when detecting sound events.
The speed of sound source positioning is based on the distributed node networking characteristic of a 5G network, and a high-speed and low-delay data interaction channel can be ensured among nodes.
According to the invention, through a monitoring network of a 5G networking, the number of the nodes is at least more than 3. In one embodiment, due to the flexibility of wireless networking, the number of nodes can be flexibly adjusted according to the conditions such as the area of a monitoring area.
In an embodiment, as an optional method, when the monitoring place is a large plane, the coordinate system of the nodes can be simplified, and the position of each node can be represented by using a two-dimensional coordinate system, so that the positioning speed of the sound source positioning algorithm can be improved.
In one embodiment, node networking based on a 5G wireless network and time service based on a Beidou navigation module can support flexible configuration of a monitoring network, and the number of networking nodes can be flexibly configured according to the requirements of the area and monitoring density of a monitoring area.
In one embodiment, the specific algorithm and steps of the acoustic event detection and identification method are as follows:
1) and intercepting the real-time audio stream, wherein the intercepting time length is 2t, and t is greater than 0.
2) The period of real-time audio stream interception is less than 2 t. In one embodiment, the period may be t.
The purpose of the above method is to ensure that all data are detected 2 times through data overlapping, and avoid the failure of event identification caused by the segmentation of data interception noise events.
3) The real-time audio stream (namely, the voice signal) is processed by a Gamma atlas filter bank to extract the cochlea map characteristic data. This is because the human ear is sensitive to the identification of the acoustic event category, and the process of human ear perception of sound can be simulated by using the Gammatone filter to extract the characteristic data of the audio.
4) And inputting the graph characteristic data serving as a characteristic graph into the trained convolutional neural network model for calculation to obtain an output result. The output result comprises: and the type and the starting time of the acoustic event are weakly labeled. The construction process of the convolutional neural network comprises three basic steps of model building, model training and model testing.
5) And judging the type of the acoustic event, if the type of the acoustic event is an expected acoustic event type, intercepting and packaging original audio data according to the front and back of the weakly labeled time point of the acoustic event, and forwarding the additional node coordinate position and the real-time event mark to the main node.
In one embodiment, the specific algorithm and steps of the time difference of arrival location (TDOA) method are as follows:
1) the node forwards the original audio data and the timestamp to the master node.
2) And the main node performs amplitude normalization on the original audio data of each node.
3) And carrying out power maximum value fitting on the audio data with the normalized amplitude value, and solving the accurate time difference value between the node data.
4) And establishing an equation set according to the three-dimensional coordinates of each node and the arrival time difference of the acoustic events among the nodes. For example, in a three-dimensional coordinate system, 3 node positions, and 3 sets of arrival time differences between nodes are required to establish 3 independent equation sets; and then solving the system of equations to obtain the three-dimensional coordinates of the target sound source.
The invention provides a public place acoustic event positioning method based on a distributed noise sensor, which comprises the following steps:
providing at least three nodes, wherein one node is configured as a master node, and the other nodes are configured as slave nodes, wherein each node comprises a noise sensor, a noise sensor probe, a 5G networking module and a Beidou navigation module;
all nodes are started, and each node is accessed to the 5G network through the 5G networking module;
after being networked, each node forms a test sub-network with other nodes through the main node;
each node starts the Beidou navigation module to carry out time synchronization and obtains node position information;
each node utilizes the noise sensor probe and the noise sensor to detect and identify the acoustic event;
after each node detects the acoustic event, recording acoustic event audio data and a timestamp of the occurrence of the acoustic event;
all the slave nodes form summarized data by the sound event audio data, the time stamp of the sound event and the node position information, and the summarized data are packaged and sent to the master node;
the master node judges whether the sound event audio data are the same sound event or not and judges the type of the sound event according to the sound event audio data captured by the master node and received from the slave node; and
and when the master node judges that more than three nodes detect the same sensitive sound event, the master node performs sound source position positioning by using the summarized data.
In one embodiment, the acoustic event detection identification comprises:
intercepting the real-time audio stream to obtain the audio data, wherein the intercepting time length is 2t, and t is greater than 0;
enabling the interception period of the real-time audio stream to be less than 2 t;
extracting cochlear atlas feature data from the real-time audio stream through a Gamma atom filter bank;
inputting the map feature data serving as a feature map into a trained convolutional neural network model for calculation to obtain an output result, wherein the output result comprises: acoustic event type, acoustic event start time;
and judging the type of the acoustic event, if the acoustic event is the expected acoustic event type, intercepting the audio data according to the front and back of the initial time point of the acoustic event, and adding node position information and the timestamp of the occurrence of the acoustic event and forwarding the node position information and the timestamp of the occurrence of the acoustic event to the host node.
In one embodiment, the sound source position location comprises:
the master node performs amplitude normalization on the audio data forwarded by each slave node;
the main node performs power maximum value fitting on the audio data after the amplitude is normalized, and time difference values among data corresponding to all the nodes are calculated;
and establishing an equation set according to the three-dimensional coordinates of each node and the arrival time difference of the acoustic events among the nodes, and solving the equation set to obtain the three-dimensional coordinates of the sound source.
In one embodiment, the method further comprises:
and after the sound source position is positioned, pushing the type of the sound event and the sound event position obtained by positioning the sound source to a target user for early warning.
In one embodiment, the nodes perform distributed networking based on a 5G network, and a high-speed and low-delay data interaction channel is guaranteed between the nodes.
In one embodiment, when the public place is a large plane, the coordinate system of the nodes is simplified to represent the position of each node by a two-dimensional coordinate system, so as to improve the positioning speed of sound source position positioning.
The invention also provides a public place acoustic event positioning system based on the distributed noise sensor, which comprises at least three nodes, wherein one node is configured as a master node, and the other nodes are configured as slave nodes, and each node comprises: noise sensor, noise sensor probe, 5G networking module and big dipper navigation module.
The 5G networking module is configured to provide network connection, ensure that all nodes can form a small-range monitoring network, provide a high-speed and low-delay data interaction channel and ensure burst transmission of acoustic event data;
the Beidou navigation module is configured to perform geographical position positioning on each node to obtain node position information which is used as a data base for sound source positioning; the Beidou navigation module is also configured to carry out clock time service so as to ensure time synchronization among all nodes;
the noise sensor probe is configured to collect noise of a public place;
the noise sensor is configured to quantize and sample analog audio data collected by the noise sensor probe to form a digital signal, and comprises a processor suitable for edge calculation, wherein the processor performs acoustic event detection identification, and records acoustic event audio data and a timestamp of occurrence of an acoustic event after the acoustic event is detected;
the processor of the slave node also composes summarized data of the audio data of the acoustic event, the timestamp of the occurrence of the acoustic event and the node position information and sends the summarized data to the master node in a packaging manner; and the processor of the master node judges whether the sound event audio data are the same sound event or not and judges the type of the sound event according to the sound event audio data captured by the processor and received from the slave node, and when the judgment result shows that more than three nodes detect the same sensitive sound event, the summarized data is utilized to position the sound source.
In one embodiment, the processor of the noise sensor performing acoustic event detection identification includes:
intercepting the real-time audio stream to obtain the audio data, wherein the intercepting time length is 2t, and t is greater than 0;
enabling the interception period of the real-time audio stream to be less than 2 t;
extracting cochlear atlas feature data from the real-time audio stream through a Gamma atom filter bank;
inputting the map feature data serving as a feature map into a trained convolutional neural network model for calculation to obtain an output result, wherein the output result comprises: acoustic event type, acoustic event start time;
and judging the type of the acoustic event, if the acoustic event is the expected acoustic event type, intercepting the audio data according to the front and back of the initial time point of the acoustic event, and adding node position information and the timestamp of the occurrence of the acoustic event and forwarding the node position information and the timestamp of the occurrence of the acoustic event to the host node.
In one embodiment, the sound source position location comprises:
the master node performs amplitude normalization on the audio data forwarded by each slave node;
the main node performs power maximum value fitting on the audio data after the amplitude is normalized, and time difference values among data corresponding to all the nodes are calculated;
and establishing an equation set according to the three-dimensional coordinates of each node and the arrival time difference of the acoustic events among the nodes, and solving the equation set to obtain the three-dimensional coordinates of the sound source.
In one embodiment, after the noise sensor performs sound source location, the type of the sound event and the sound event location obtained by sound source location are pushed to a target user for early warning.
In one embodiment, the nodes perform distributed networking based on a 5G network, and a high-speed and low-delay data interaction channel is guaranteed between the nodes.
In one embodiment, when the public place is a large plane, the coordinate system of the nodes is simplified to represent the position of each node by a two-dimensional coordinate system, so as to improve the positioning speed of sound source position positioning.
The method is suitable for detecting and positioning the sensitive sound event in an open outdoor or semi-outdoor public place, and can quickly detect the occurrence of the sensitive sound event, the type of the sound event and the position of the sound event; important data support is provided for crowd evacuation and public safety decision making.
Compared with the prior art, the invention has the following advantages:
the method has the advantages that: flexibility
1) Flexibility of deployment: compared with the traditional sound source positioning mode realized by the array microphone scheme. According to the invention, each node works independently, interconnection is realized through a 5G network, each node is positioned and time-synchronized through a Beidou navigation system, and the nodes can be flexibly set during layout, so that a larger area range can be covered.
2) The flexible layout scheme can avoid the condition that the shielding sound event of the building is missed.
The method has the advantages that: aging property
1) The front-end nodes with edge computing capability disperse the overall computing power requirement of the system to each node, and can ensure the overall operation efficiency of the system.
2) The whole system is flexible in construction, when the number of equipment nodes is increased, the computing power is increased along with the increase of the number of system nodes, and the whole response speed of the system is basically not changed along with the increase of the number of system nodes.
The method has the advantages that: stability of
1) In a distributed node design scheme, each node can be used as a main node. When the equipment of the individual node fails, other nodes can be replaced immediately, and the normal operation of the whole system is ensured.
The advantages are that: accuracy of
1) The invention realizes time service and positioning through a Beidou navigation system. The synchronization error introduced by signal transmission delay caused by network time service is avoided, and the time synchronization precision can be improved from millisecond level to nanosecond level.
The terms and expressions which have been employed herein are used as terms of description and not of limitation. The use of such terms and expressions is not intended to exclude any equivalents of the features shown and described (or portions thereof), and it is recognized that various modifications may be made within the scope of the claims. Other modifications, variations, and alternatives are also possible. Accordingly, the claims should be looked to in order to cover all such equivalents.
Also, it should be noted that although the present invention has been described with reference to the current specific embodiments, it should be understood by those skilled in the art that the above embodiments are merely illustrative of the present invention, and various equivalent changes or substitutions may be made without departing from the spirit of the present invention, and therefore, it is intended that all changes and modifications to the above embodiments be included within the scope of the claims of the present application.
Claims (8)
1. A method for public place acoustic event localization based on distributed noise sensors, the method comprising:
providing at least three nodes, wherein one node is configured as a master node, and the other nodes are configured as slave nodes, wherein each node comprises a noise sensor, a noise sensor probe, a 5G networking module and a Beidou navigation module;
all nodes are started, and each node is accessed to the 5G network through the 5G networking module;
after being networked, each node forms a test sub-network with other nodes through the main node;
each node starts the Beidou navigation module to carry out time synchronization and obtains node position information;
each node utilizes the noise sensor probe and the noise sensor to detect and identify the acoustic event;
after each node detects an acoustic event, recording acoustic event audio data and a timestamp of the occurrence of the acoustic event;
all the slave nodes form summarized data by the sound event audio data, the time stamp of the sound event and the node position information, and the summarized data are packaged and sent to the master node;
the master node judges whether the sound event audio data are the same sound event or not and judges the type of the sound event according to the sound event audio data captured by the master node and received from the slave node; and
when the fact that more than three nodes detect the same sensitive sound event is judged, the main node carries out sound source position positioning by utilizing the summarized data;
wherein the acoustic event detection recognition comprises:
intercepting the real-time audio stream to obtain the audio data, wherein the intercepting time length is 2t, and t is greater than 0;
enabling the interception period of the real-time audio stream to be less than 2 t;
extracting cochlear atlas feature data from the real-time audio stream through a Gamma atom filter bank;
inputting the map feature data serving as a feature map into a trained convolutional neural network model for calculation to obtain an output result, wherein the output result comprises: acoustic event type, acoustic event start time;
and judging the type of the acoustic event, if the acoustic event is the expected acoustic event type, intercepting the audio data according to the front and back of the initial time point of the acoustic event, and adding node position information and the timestamp of the occurrence of the acoustic event and forwarding the node position information and the timestamp of the occurrence of the acoustic event to the host node.
2. The distributed noise sensor-based public space acoustic event localization method of claim 1, wherein the sound source position localization comprises:
the master node performs amplitude normalization on the audio data forwarded by each slave node;
the main node performs power maximum value fitting on the audio data after the amplitude is normalized, and time difference values among data corresponding to all the nodes are calculated;
and establishing an equation set according to the three-dimensional coordinates of each node and the arrival time difference of the acoustic events among the nodes, and solving the equation set to obtain the three-dimensional coordinates of the sound source.
3. The distributed noise sensor-based public place acoustic event locating method of claim 1, further comprising:
and after the sound source position is positioned, pushing the type of the sound event and the sound event position obtained by positioning the sound source to a target user for early warning.
4. The method for public place acoustic event localization based on distributed noise sensors of claim 1, wherein when the public place is a large plane, the coordinate system of the nodes is simplified to characterize the position of each node with a two-dimensional coordinate system to improve localization speed of sound source position localization.
5. A distributed noise sensor based public place acoustic event localization system, the system comprising at least three nodes, wherein one node is configured as a master node and the remaining nodes are configured as slave nodes, each node comprising: the noise sensor, the noise sensor probe, the 5G networking module and the Beidou navigation module are arranged on the base;
the 5G networking module is configured to provide 5G network connection, ensure that all nodes can form a monitoring network in a small range, provide a high-speed and low-delay data interaction channel and ensure burst transmission of acoustic event data;
the Beidou navigation module is configured to perform geographical position positioning on each node to obtain node position information which is used as a data basis for sound source position positioning; the Beidou navigation module is also configured to carry out clock time service so as to ensure time synchronization among all nodes;
the noise sensor probe is configured to collect noise of a public place;
the noise sensor is configured to quantize and sample the analog audio data collected by the noise sensor probe to form a digital signal, and comprises a processor suitable for edge calculation, wherein the processor performs acoustic event detection identification, and records acoustic event audio data and a timestamp of occurrence of an acoustic event after the acoustic event is detected;
the processor of the slave node also composes summarized data of the audio data of the acoustic event, the timestamp of the occurrence of the acoustic event and the node position information and sends the summarized data to the master node in a packaging manner; the processor of the master node judges whether the sound event audio data are the same sound event or not and judges the type of the sound event according to the sound event audio data captured by the processor and received from the slave node, and when the processor judges that more than three nodes detect the same sensitive sound event, the summarized data is utilized to position a sound source;
wherein the processor of the noise sensor performing acoustic event detection identification comprises:
intercepting the real-time audio stream to obtain the audio data, wherein the intercepting time length is 2t, and t is greater than 0;
enabling the interception period of the real-time audio stream to be less than 2 t;
extracting cochlear atlas feature data from the real-time audio stream through a Gamma atom filter bank;
inputting the map feature data serving as a feature map into a trained convolutional neural network model for calculation to obtain an output result, wherein the output result comprises: acoustic event type, acoustic event start time;
and judging the type of the acoustic event, if the acoustic event is the expected acoustic event type, intercepting the audio data according to the front and back of the initial time point of the acoustic event, and adding node position information and the timestamp of the occurrence of the acoustic event and forwarding the node position information and the timestamp of the occurrence of the acoustic event to the host node.
6. The distributed noise sensor-based public space acoustic event localization system of claim 5, wherein the sound source position localization comprises:
the master node performs amplitude normalization on the audio data forwarded by each slave node;
the main node performs power maximum value fitting on the audio data after the amplitude is normalized, and time difference values among data corresponding to all the nodes are calculated;
and establishing an equation set according to the three-dimensional coordinates of each node and the arrival time difference of the acoustic events among the nodes, and solving the equation set to obtain the three-dimensional coordinates of the sound source.
7. The public place acoustic event positioning system based on the distributed noise sensor as claimed in claim 5, wherein after the noise sensor performs sound source position positioning, the type of the acoustic event and the acoustic event position obtained by sound source positioning are pushed to a target user for early warning.
8. The distributed noise sensor-based public space acoustic event localization system of claim 5, wherein when the public space is a large plane, the coordinate system of the nodes is simplified to characterize the position of each node in a two-dimensional coordinate system to improve localization speed of sound source location localization.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110045899.9A CN112363114B (en) | 2021-01-14 | 2021-01-14 | Public place acoustic event positioning method and system based on distributed noise sensor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110045899.9A CN112363114B (en) | 2021-01-14 | 2021-01-14 | Public place acoustic event positioning method and system based on distributed noise sensor |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112363114A CN112363114A (en) | 2021-02-12 |
| CN112363114B true CN112363114B (en) | 2021-04-02 |
Family
ID=74535001
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110045899.9A Active CN112363114B (en) | 2021-01-14 | 2021-01-14 | Public place acoustic event positioning method and system based on distributed noise sensor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112363114B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113483811A (en) * | 2021-06-16 | 2021-10-08 | 中国地质大学（武汉） | Distributed monitoring node device based on raspberry group air index parameters |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105866742B (en) * | 2016-05-27 | 2018-10-23 | 成都信息工程大学 | A kind of shell burst point positioning system and localization method |
| KR101930137B1 (en) * | 2016-10-18 | 2018-12-17 | 국방과학연구소 | Estimation Apparatus for Sound source and Estimation Method for Sound source Using the same |
| CN107293290A (en) * | 2017-07-31 | 2017-10-24 | 郑州云海信息技术有限公司 | The method and apparatus for setting up Speech acoustics model |
| CN107545890A (en) * | 2017-08-31 | 2018-01-05 | 桂林电子科技大学 | A kind of sound event recognition method |
| CN110213840A (en) * | 2019-05-06 | 2019-09-06 | 南京理工大学 | Based on wireless-sensor network distribution type shot positioning system and method |
| CN110189769B (en) * | 2019-05-23 | 2021-11-19 | 复钧智能科技(苏州)有限公司 | Abnormal sound detection method based on combination of multiple convolutional neural network models |
| CN110673097B (en) * | 2019-10-24 | 2022-01-18 | 浙江工商大学 | Method for positioning sounding object in space by using audio arrival time difference |
-
2021
- 2021-01-14 CN CN202110045899.9A patent/CN112363114B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN112363114A (en) | 2021-02-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20180376273A1 (en) | System and method for determining audio context in augmented-reality applications | |
| US9019077B2 (en) | System and method for predicting future meetings of wireless users | |
| US7308105B2 (en) | Environmental noise monitoring | |
| Alexandridis et al. | Multiple sound source location estimation in wireless acoustic sensor networks using DOA estimates: The data-association problem | |
| WO2018045973A1 (en) | Sound source localization method for robot, and system | |
| CN111885502A (en) | Epidemic situation prevention and control early warning and tracing system and method for protecting privacy | |
| KR100705730B1 (en) | Method for improving the wireless location system | |
| US20050266860A1 (en) | Position detection method, position detection system, and position detection server equipment | |
| CN112363114B (en) | Public place acoustic event positioning method and system based on distributed noise sensor | |
| CN103634476A (en) | Positioning method based on recording device of mobile phone | |
| JP2013034103A (en) | Database server, system, program, and method for identifying target area from position information including positioning error | |
| US11940306B2 (en) | Estimation of the location of a wireless terminal, based on a propagation characteristic of a pressure wave | |
| CN106328130A (en) | Robot voice addressed rotation system and method | |
| KR20150123550A (en) | System and Method for Measuring and Monitoring Seismic data Applied to the City Building | |
| CN107290787A (en) | A kind of monitoring signals correlating method of earthquake infrasonic sound with the location station | |
| CN108603925A (en) | The positioning based on acoustic range that SoundRec carries out object is used by terminal | |
| CN112379330B (en) | Multi-robot cooperative 3D sound source identification and positioning method | |
| CN103675757A (en) | Positioning method based on a plurality of mobile phone microphone arrays | |
| CN110045228A (en) | Device for fault indicator, system and distribution network system | |
| EP3751869A1 (en) | System and method for generating head-related transfer function | |
| CN113376578A (en) | Sound source positioning method and system based on matching of arrival angle and sound intensity | |
| Andra et al. | Feasibility evaluation for keyword spotting system using mini microphone array on UAV | |
| KR20210078056A (en) | Noise source tracking system and noise source tracking method using relative position and time synchronization between personnel | |
| CN112887485B (en) | Contact tracking method based on electronic trace | |
| US11678295B2 (en) | Contact tracking method and related server |
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 | ||
| CP03 | Change of name, title or address |
Address after: 311121 building 3, No.10 Xianqiao Road, Zhongtai street, Yuhang District, Hangzhou City, Zhejiang Province Patentee after: Hangzhou Zhaohua Electronics Co.,Ltd. Address before: 31-1 Xianxing Road, Xianlin Industrial Park, Yuhang District, Hangzhou City, Zhejiang Province Patentee before: CRY SOUND CO.,LTD. |
|
| CP03 | Change of name, title or address |