CN112561971A - People flow statistical method, device, equipment and storage medium - Google Patents
People flow statistical method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN112561971A CN112561971A CN202011486671.5A CN202011486671A CN112561971A CN 112561971 A CN112561971 A CN 112561971A CN 202011486671 A CN202011486671 A CN 202011486671A CN 112561971 A CN112561971 A CN 112561971A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- target
- identified
- current frame
- neural network
- 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.)
- Pending
Links
- 238000007619 statistical method Methods 0.000 title claims abstract description 11
- 238000000034 method Methods 0.000 claims abstract description 39
- 238000012545 processing Methods 0.000 claims abstract description 24
- 238000012544 monitoring process Methods 0.000 claims abstract description 22
- 238000003062 neural network model Methods 0.000 claims description 54
- 238000004590 computer program Methods 0.000 claims description 7
- 238000009434 installation Methods 0.000 abstract description 4
- 238000012423 maintenance Methods 0.000 abstract description 4
- 230000007613 environmental effect Effects 0.000 abstract description 2
- 230000000875 corresponding effect Effects 0.000 description 10
- 238000012549 training Methods 0.000 description 8
- 241000282472 Canis lupus familiaris Species 0.000 description 7
- 241000282326 Felis catus Species 0.000 description 7
- 230000008569 process Effects 0.000 description 6
- 230000008901 benefit Effects 0.000 description 5
- 238000000605 extraction Methods 0.000 description 4
- 238000001914 filtration Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 230000002596 correlated effect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 241001465754 Metazoa Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000013523 data management Methods 0.000 description 1
- 238000002592 echocardiography Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000017525 heat dissipation Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000000779 smoke Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/277—Analysis of motion involving stochastic approaches, e.g. using Kalman filters
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24137—Distances to cluster centroïds
- G06F18/2414—Smoothing the distance, e.g. radial basis function networks [RBFN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
- G06V20/53—Recognition of crowd images, e.g. recognition of crowd congestion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The application relates to a people flow statistical method, a device, equipment and a storage medium, wherein the method comprises the following steps: acquiring original point cloud data of a current frame in a monitoring area; carrying out clustering tracking processing on the original point cloud data of the current frame to obtain a point cloud set of each potential target of the current frame; judging whether each potential target exists in the point cloud set of the previous frame or not according to the point cloud set of the potential target of the current frame, and taking the potential target which does not exist in the point cloud set of the previous frame as a target to be identified; identifying the point cloud set of each target to be identified to determine the type of each target to be identified; and accumulating the number of people of the type in the target to be identified. Through the application, the environmental interference can be reduced in the aspect of people flow statistics, the data processing amount is reduced, the statistical efficiency is improved, and the installation is easy, so that the later maintenance is facilitated.
Description
Technical Field
The application relates to the technical field of microwave radars, in particular to a people flow statistical method, a device, equipment and a storage medium.
Background
The people flow sensor can realize people flow statistics on places with higher people flow density. However, people flow sensors perform people flow statistics by means of image shooting. The camera in the people stream sensor is easily influenced by the environment, is not portable in installation, is not beneficial to later maintenance, and has large processing capacity and low recognition efficiency. In addition, the rights and interests of pedestrian privacy portraits and the like are infringed by the pedestrian flow obtained through the photos.
Disclosure of Invention
In order to solve the technical problems that the efficiency of people flow statistics through images is low and personal privacy is invaded, the embodiment of the application provides a people flow statistics method, a people flow statistics device, people flow statistics equipment and a storage medium.
In a first aspect, an embodiment of the present application provides a people flow rate statistical method, where the method includes:
acquiring original point cloud data of a current frame in a monitoring area;
carrying out clustering tracking processing on the original point cloud data of the current frame to obtain a point cloud set of each potential target of the current frame;
judging whether each potential target exists in the point cloud set of the previous frame or not according to the point cloud set of the potential target of the current frame, and taking the potential target which does not exist in the point cloud set of the previous frame as a target to be identified;
identifying the point cloud set of each target to be identified to determine the type of each target to be identified;
and accumulating the number of people of the type in the target to be identified.
Optionally, before acquiring the raw point cloud data of the monitored area, the method further comprises:
acquiring a reflected echo of a current frame in a scanning area;
obtaining first point cloud data of the current frame in the scanning area according to the reflection echo of the current frame in the scanning area;
and extracting original point cloud data of the current frame of the monitoring area from the first point cloud data of the current frame of the scanning area according to the monitoring area set by a user.
Optionally, identifying the point cloud set of each target to be identified to determine the type of each target to be identified includes:
and identifying the point cloud set of each target to be identified based on the trained first neural network model so as to acquire the type of each target to be identified.
Optionally, the trained first neural network model is obtained by:
acquiring point cloud sample data carrying a first label;
performing feature extraction on the point cloud sample data to obtain point cloud sample features;
inputting the point cloud sample characteristics into a first initial neural network model to train the first initial neural network model to obtain a trained first neural network model;
the point cloud sample data at least comprises human point cloud sample data.
Optionally, identifying the point cloud set of each target to be identified to determine the type of each target to be identified includes:
extracting the contour features of each target to be recognized based on a contour algorithm and a point cloud set of the target to be recognized;
and identifying the contour features of the targets to be identified based on the trained second neural network model so as to obtain the type of each target to be identified.
Optionally, the trained second neural network model is obtained by:
acquiring contour sample data carrying a second label;
extracting the characteristics of the contour sample data to obtain the characteristics of the contour sample;
inputting the profile sample characteristics into a second initial neural network model to train the second initial neural network model to obtain a trained second neural network model;
the contour sample data comprises at least human contour sample data.
Optionally, the method further comprises:
acquiring a potential target of a person type from potential targets of a point cloud set in which a previous frame exists;
acquiring a point cloud set of all frames of potential targets of which the types are human, wherein all the frames comprise a current frame;
and acquiring the corresponding movement track of the potential target of which the type is a person according to the point cloud sets of all the frames.
In a second aspect, an embodiment of the present application provides a people flow rate statistics apparatus, including:
the acquisition module is used for acquiring original point cloud data of a current frame in a monitoring area;
the cluster tracking module is used for carrying out cluster tracking processing on the original point cloud data of the current frame to obtain a point cloud set of each potential target of the current frame;
the judging module is used for judging whether each potential target exists in the point cloud set of the previous frame or not according to the point cloud set of the potential target of the current frame, and taking the potential target which does not exist in the point cloud set of the previous frame as a target to be identified;
the identification module is used for identifying the point cloud set of each target to be identified so as to determine the type of each target to be identified;
and the counting module is used for accumulating the number of people of the type in the target to be identified.
In a third aspect, an embodiment of the present application provides a people flow rate statistics device, where the people flow rate statistics device includes a microwave radar, and the people flow rate statistics device performs people flow rate statistics on a target area according to any one of the people flow rate statistics methods.
In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the processor is caused to execute the steps of the people flow rate statistical method according to any one of the foregoing methods.
In a fifth aspect, an embodiment of the present application provides a computer device, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, the processor performs the steps of the people flow rate statistical method according to any one of the foregoing methods.
Compared with the prior art, the technical scheme provided by the embodiment of the application has the following advantages:
the method comprises the steps of obtaining original point cloud data of a current frame in a monitoring area; carrying out clustering tracking processing on the original point cloud data of the current frame to obtain a point cloud set of each potential target of the current frame; judging whether each potential target exists in the point cloud set of the previous frame or not according to the point cloud set of the potential target of the current frame, and taking the potential target which does not exist in the point cloud set of the previous frame as a target to be identified; identifying the point cloud set of each target to be identified to determine the type of each target to be identified; and accumulating the number of people of the type in the target to be identified so as to realize people flow statistics. Can reduce environmental interference in the aspect of the flow of people statistics through this application, reduce the false retrieval rate that leads to because of external illumination and temperature factor, reduce data processing volume, improve statistical efficiency and rate of accuracy, and this application is equipped with the flow of people statistics device of microwave radar and easily installs, does benefit to the later maintenance.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
FIG. 1 is a flow chart illustrating a method for people traffic statistics according to an embodiment of the present application;
fig. 2 is a block diagram illustrating a structure of a human traffic statistic device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
FIG. 1 is a flow chart illustrating a method for people traffic statistics according to an embodiment of the present application; referring to fig. 1, the people flow rate statistical method includes the following steps:
s100: and acquiring original point cloud data of the current frame in the monitoring area.
Specifically, the method and the device acquire original point cloud data of a monitored area through a microwave radar. The microwave radar may be a millimeter wave radar. Specifically, the radar can be a millimeter wave radar which emits electromagnetic waves with the frequency of 300HZ-300 GHz.
The millimeter wave radar is provided with a signal receiving and transmitting unit, after the millimeter wave radar is started, the signal receiving and transmitting unit generates and transmits a millimeter wave radar signal to a monitoring area, and if an object exists in the monitoring area, a corresponding echo signal is reflected to the signal receiving and transmitting unit of the millimeter wave radar. The millimeter wave radar transceiver signal has a period, and each frame is used for completing signal transceiving in one period. Each frame has a received echo signal. The echo signal is an intermediate frequency echo signal obtained by mixing the echoes reflected by the object. And the microwave radar performs analog-to-digital signal conversion and DSP signal processing on the echo signals to obtain corresponding point cloud data.
S200: and carrying out clustering tracking processing on the original point cloud data of the current frame to obtain a point cloud set of each potential target of the current frame.
S300: and judging whether each potential target exists in the point cloud set of the previous frame or not according to the point cloud set of the potential target of the current frame, and taking the potential target which does not exist in the point cloud set of the previous frame as the target to be identified.
Specifically, a plurality of persons or objects may exist in the monitored area at the same time, and thus the obtained original point cloud data of each frame includes a set of point cloud data corresponding to the plurality of persons or objects. The cluster tracking process is used to achieve target tracking and clustering.
The point cloud data is represented in the form of a plurality of points, and comprises information such as point cloud distance, azimuth angle, Doppler velocity, signal-to-noise ratio and the like.
The clustering tracking process may also be used to cluster point cloud data of the same potential target in each frame to obtain a point cloud set corresponding to each potential target in each frame.
According to the characteristics of the point cloud data, the point cloud data of adjacent frames are correlated, that is, the point cloud data of the previous frame is correlated with the point cloud data of the current frame, and the point cloud data of the current frame is correlated with the point cloud data of the next frame.
The point cloud set of each potential target carries information such as range, azimuth, doppler attributes, etc. According to the method, based on information carried by the point cloud sets, the point cloud data of the current frame and the point cloud data of the previous frame can be associated by adopting a neighbor connected point cloud clustering algorithm and a Kalman filtering method, and by analogy, the point cloud sets of each frame of the same potential target can be associated with each other, so that track tracking of the potential target is realized.
If the point cloud set of the current frame of a potential target is associated with the point cloud set of the previous frame, it indicates that the potential target has appeared in the previous frame, and if the potential target is a human, the point cloud set of the previous frame is also counted, so the counting is not repeated in the current frame. If the potential target is a person, the number of the potential target is counted once in the first frame where the potential target appears, and the number of the potential target is not counted repeatedly even if other frames appear.
Certainly, the point cloud set of the potential targets in each frame can be obtained through the point cloud data of each frame, and the point cloud sets of the same potential targets in different frames are identified by the same marks. Therefore, whether each potential target is a newly appeared potential target or an already appeared potential target can be conveniently obtained. And whether the potential target of the current frame appears in the previous frame can be conveniently judged, the potential target appearing in the previous frame is further removed, and only the potential target not appearing in the previous frame is left for type identification.
If the point cloud set of a certain potential target current frame is associated with the point cloud set of the previous frame, the point cloud set represents that the potential target is the potential target which has already appeared, so the potential target with the point cloud set of the previous frame can be removed in the current frame, the type identification operation is not performed any more, only the potential target without the point cloud set of the previous frame is taken as the target to be identified for the type identification operation, and the processing amount of the type identification can be greatly reduced.
S400: and identifying the point cloud set of each target to be identified so as to determine the type of each target to be identified.
S500: and accumulating the number of people of the type in the target to be identified.
Specifically, a point cloud set of an object to be recognized characterizes the object to be recognized. For example, features of a contour characterizing the object to be identified. And identifying which type the target to be identified belongs to according to the point cloud set and the neural network model.
The method and the device have the advantages that Kalman filtering and clustering tracking are carried out on the point cloud set of the potential targets, the states and the number of the potential targets are continuously observed, the number of people in the monitoring area is obtained, and the number of people in the monitoring area is counted.
In one embodiment, the type may be human, animal (cat, dog, etc.), automobile, house, road, etc.
In one embodiment, the type may be human or non-human.
And when the type of a certain target to be identified is human, accumulating the human flow data for one person.
In one embodiment, before step S100, the method further comprises:
acquiring a reflected echo of a current frame in a scanning area;
obtaining first point cloud data of the current frame in the scanning area according to the reflection echo of the current frame in the scanning area;
and extracting original point cloud data of the current frame of the monitoring area from the first point cloud data of the current frame of the scanning area according to the monitoring area set by a user.
Specifically, after the microwave radar is fixed in an area by a user, first point cloud data in the scanning area can be obtained. However, the scanning area may be relatively wide, and the user may not use all the first point cloud data. Alternatively, the user may only want to count the traffic in the area that the user wants to monitor. Therefore, the user can set the monitoring area in advance as the data processing range of the microwave radar. Therefore, the data processing amount can be greatly reduced, the data processing efficiency is improved, and meanwhile, the error rate is reduced.
In one embodiment, step S400 specifically includes: and identifying the point cloud set of each target to be identified based on the trained first neural network model so as to acquire the type of each target to be identified.
In one embodiment, the trained first neural network model is obtained by:
acquiring point cloud sample data carrying a first label;
performing feature extraction on the point cloud sample data to obtain point cloud sample features;
inputting the point cloud sample characteristics into a first initial neural network model to train the first initial neural network model to obtain a trained first neural network model;
the point cloud sample data at least comprises character point cloud sample data.
Specifically, the point cloud sample data carrying the first label comprises a large number of human body point cloud samples carrying human body labels. The trained first neural network model is obtained by training according to a large number of human body point cloud samples carrying human body labels, and the trained first neural network model can at least identify whether the target to be identified is a human body or a non-human body.
Of course, the trained first neural network model can also be trained according to a large number of other point cloud samples carrying the determined labels. At this time, the point cloud sample data carrying the first tag necessarily includes a large number of human body point cloud samples carrying human body tags. The point cloud sample data carrying the first tag may further include: one or more of a large number of dog point cloud samples carrying a dog tag, a large number of cat point cloud samples carrying a cat tag, and a large number of automobile point cloud samples carrying an automobile tag. The first neural network model in the embodiment can identify which type each object to be identified belongs to, whether the object belongs to a person, a cat, a dog, a car, and the like. The identification is more accurate.
In one embodiment, step S400 specifically includes:
extracting the contour features of each target to be recognized based on a contour algorithm and a point cloud set of the target to be recognized;
and identifying the contour features of the targets to be identified based on the trained second neural network model so as to obtain the type of each target to be identified.
Specifically, the trained second neural network model is obtained in the following manner:
acquiring contour sample data carrying a second label;
extracting the characteristics of the contour sample data to obtain the characteristics of the contour sample;
inputting the profile sample characteristics into a second initial neural network model to train the second initial neural network model to obtain a trained second neural network model;
the contour sample data at least comprises human contour sample data.
Specifically, a point cloud set of the target to be recognized is processed to obtain a corresponding target picture to be recognized. And extracting the characteristic contour line of the target picture to be recognized to obtain the contour characteristic of the target to be recognized.
The second neural network model is trained on the contour sample data, so the second neural network model identifies the type of the target to be identified according to the contour features.
The contour sample data carrying the second tag comprises a plurality of human body contour samples carrying human body tags. The trained second neural network model is obtained by training according to a large number of human body contour samples carrying human body labels, and the trained second neural network model can at least identify whether the target to be identified is a human body or a non-human body.
Of course, the trained second neural network model can also be trained according to a large number of other contour samples carrying the determined labels. At this time, the contour sample data carrying the second tag necessarily includes a large number of human body contour samples carrying human body tags. The contour sample data carrying the first tag may further include: one or more of a plurality of labeled dog profile samples carrying dogs, a plurality of labeled cat profile samples carrying cats, and a plurality of labeled car profile samples carrying cars. The second neural network model in the embodiment can identify which type each object to be identified belongs to specifically, whether the object belongs to a person, a cat, a dog, a car, and the like. The identification is more accurate.
The training sample of the first neural network model or the second neural network model can adopt the steps of processing and labeling original point cloud data acquired by a microwave radar to construct a training set and a testing set; and training the initial neural network model by using the training set and the test set to obtain a trained first neural network model or a trained second neural network model.
According to the method and the device, clustering tracking processing, Kalman filtering tracking and type identification are carried out on each frame of original point cloud data of the monitored area, the running track of the potential target in the monitored area can be determined, classification of the potential target is realized by identifying the type of each potential target, and therefore personnel quantity statistics of the monitored area is realized.
In one embodiment, the people flow statistical method further comprises the following steps:
acquiring a potential target of a person type from potential targets of a point cloud set in which a previous frame exists;
acquiring a point cloud set of all frames of potential targets of which the types are human, wherein all the frames comprise a current frame;
and acquiring the corresponding movement track of the potential target of which the type is a person according to the point cloud sets of all the frames.
Specifically, the microwave radar collects multiple frames of original point cloud data corresponding to the monitored area within a period of time. Each potential target of the type human has a corresponding point cloud set in at least one frame of original point cloud data in the time period from the beginning of entering the monitored area to the leaving of the monitored area.
The point cloud data carries information such as distance, azimuth angle, Doppler attribute and the like. Therefore, the point cloud data tracking is carried out on the potential target of each type of people by using Kalman filtering and clustering tracking processing, and the moving track of the potential target of each type of people can be obtained.
And predicting a pedestrian flow preview path according to the movement tracks of the potential targets of all types of people, and analyzing and summarizing the variation trend of the pedestrian flow and the trend of the pedestrian flow path.
The estimated pedestrian flow reservation path is widely applied in practice. For example, when a merchant conducts a campaign, whether the campaign successfully attracts the traffic can be analyzed more intuitively according to the estimated traffic path trend, or the movement trend of a customer can be estimated in advance according to the traffic path trend, so that the campaign place and the effective time of the campaign can be arranged in advance.
Meanwhile, the people flow and the people flow trend at different time intervals are analyzed, the manpower staying in areas at peak time intervals and peak time intervals is correspondingly increased, the service quality is improved, the sales are further increased, and the condition that staff or allocation personnel plan to avoid the waste of personnel resources can be reduced at the idle empty time of a market.
The method and the system can be applied to large-scale business areas, the client can customize a required identification area according to the requirement, the radar can analyze the people flow data information obtained by scanning according to the requirement of the user and provide the people flow data information to the client, and the client can select to carry out classification service on the people flow data according to the information or provide related service types and levels which can be supported by the client at a background server. The server analyzes and processes big data according to people flow data uploaded by the people flow sensor radar, and intelligent distribution provides grading service most suitable for current people flow.
FIG. 2 is a block diagram of a device for people traffic statistics according to an embodiment of the present application; referring to fig. 2, the human traffic statistic device includes:
the acquisition module 100 is used for acquiring original point cloud data of a current frame in a monitoring area;
a clustering tracking module 200, configured to perform clustering tracking processing on the original point cloud data of the current frame to obtain a point cloud set of each potential target of the current frame;
the judging module 300 is configured to judge whether each potential target exists in the point cloud set of the previous frame according to the point cloud set of the potential target of the current frame, and use the potential target without the point cloud set of the previous frame as the target to be identified;
the identification module 400 is configured to identify the point cloud sets of the targets to be identified to determine the type of each target to be identified;
and the counting module 500 is used for accumulating the number of people of which the types are in the target to be identified.
In one embodiment, the apparatus further comprises:
the echo receiving module is used for acquiring a reflected echo of a current frame in a scanning area;
the echo processing module is used for obtaining first point cloud data of the current frame in the scanning area according to the reflected echo of the current frame in the scanning area;
the first screening module is used for extracting original point cloud data of a current frame of the monitoring area from first point cloud data of the current frame of the scanning area according to the monitoring area set by a user.
In one embodiment of the present invention,
the identifying module 400 is specifically configured to identify the point cloud sets of the targets to be identified based on the trained first neural network model, so as to obtain the type of each target to be identified.
In one embodiment, the apparatus further comprises:
the point cloud sample acquisition module is used for acquiring point cloud sample data carrying the first label;
the point cloud sample data processing module is used for processing point cloud sample data to obtain point cloud sample characteristics;
the first training module is used for inputting the point cloud sample characteristics into the first initial neural network model so as to train the first initial neural network model and obtain a trained first neural network model;
the sample point cloud data at least comprises character point cloud sample data.
In one embodiment of the present invention,
the identification module 400 specifically includes:
the contour feature extraction module is used for extracting contour features of the targets to be identified based on a contour algorithm and a point cloud set of the targets to be identified;
and the sub-recognition module is used for recognizing the contour features of the targets to be recognized based on the trained second neural network model so as to acquire the type of each target to be recognized.
In one embodiment, the apparatus further comprises:
the contour sample acquisition module is used for acquiring contour sample data carrying a second label;
the contour sample feature extraction module is used for extracting features of the contour sample data to obtain features of the contour sample;
the second training module is used for inputting the characteristics of the contour sample into the second initial neural network model so as to train the second initial neural network model and obtain a trained second neural network model;
the contour sample data at least comprises human contour sample data.
In one embodiment, the apparatus further comprises:
the second screening module is used for acquiring potential targets with the types of people from the potential targets of the point cloud set with the previous frame;
the system comprises a summarizing module, a searching module and a processing module, wherein the summarizing module is used for acquiring a point cloud set of all frames of potential targets of which the types are characters, and all the frames comprise a current frame;
and the track tracking module is used for acquiring the corresponding movement track of the potential target of which the type is a person according to the point cloud sets of all the frames.
In one embodiment, the application provides a people flow rate statistic device, the people flow rate statistic device comprises a microwave radar, and the people flow rate statistic device carries out people flow rate statistic on a target area according to the people flow rate statistic method of any one of the above.
The conventional image recognition people stream sensor has influence on various external parts such as shadow, light change, heat dissipation, haze and smoke. The microwave radar people flow statistical equipment is less affected by the environment, carries out human body recognition on the people flow of a fixed area, utilizes microwave reflection and a data algorithm to calculate the people flow passing through the area, and achieves the effect of intelligently recognizing the human body. In addition, the people flow statistical equipment of this application can support wired and wireless data transmission simultaneously, need not do special transformation to the installation place and just can install completely. The installation is portable, and the maintenance is simple.
According to the method, the fixed area is scanned in real time through the reflection benefit of the microwave radar, all moving object paths in the area are obtained, moving people and moving objects are distinguished through a human body contour algorithm, the moving route of a user is predicted, the flow and the trend of the flow of people in the area are obtained according to the moving rule of the user, and therefore big data management is conducted on passing people.
In one embodiment, the present application provides a computer readable storage medium having a computer program stored thereon, the computer program, when executed by a processor, performs the steps of: acquiring original point cloud data of a current frame in a monitoring area; carrying out clustering tracking processing on the original point cloud data of the current frame to obtain a point cloud set of each potential target of the current frame; judging whether each potential target exists in the point cloud set of the previous frame or not according to the point cloud set of the potential target of the current frame, and taking the potential target which does not exist in the point cloud set of the previous frame as a target to be identified; identifying the point cloud set of each target to be identified to determine the type of each target to be identified; and accumulating the number of people of the type in the target to be identified.
The processor also performs the steps of the people flow statistical method of any one of the preceding.
It is noted that, in this document, relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The foregoing are merely exemplary embodiments of the present invention, which enable those skilled in the art to understand or practice the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A people flow statistical method, characterized in that the method comprises:
acquiring original point cloud data of a current frame in a monitoring area;
carrying out clustering tracking processing on the original point cloud data of the current frame to obtain a point cloud set of each potential target of the current frame;
judging whether each potential target exists in the point cloud set of the previous frame or not according to the point cloud set of the potential target of the current frame, and taking the potential target which does not exist in the point cloud set of the previous frame as a target to be identified;
identifying the point cloud set of each target to be identified to determine the type of each target to be identified;
and accumulating the number of people of the type in the target to be identified.
2. The method of claim 1, wherein prior to acquiring raw point cloud data for a monitored area, the method further comprises:
acquiring a reflected echo of a current frame in a scanning area;
obtaining first point cloud data of the current frame in the scanning area according to the reflection echo of the current frame in the scanning area;
and extracting original point cloud data of the current frame of the monitoring area from the first point cloud data of the current frame of the scanning area according to the monitoring area set by a user.
3. The method of claim 1, wherein identifying the point cloud set of each object to be identified to determine a type of each object to be identified comprises:
and identifying the point cloud set of each target to be identified based on the trained first neural network model so as to acquire the type of each target to be identified.
4. The method of claim 3, wherein the trained first neural network model is obtained by:
acquiring point cloud sample data carrying a first label;
extracting the characteristics of the point cloud sample data to obtain the characteristics of the point cloud sample;
inputting the point cloud sample characteristics into a first initial neural network model to train the first initial neural network model to obtain a trained first neural network model;
the point cloud sample data at least comprises human point cloud sample data.
5. The method of claim 1, wherein identifying the point cloud set of each object to be identified to determine a type of each object to be identified comprises:
extracting the contour features of each target to be recognized based on a contour algorithm and a point cloud set of the target to be recognized;
and identifying the contour features of the targets to be identified based on the trained second neural network model so as to obtain the type of each target to be identified.
6. The method of claim 5, wherein the trained second neural network model is obtained by:
acquiring contour sample data carrying a second label;
extracting the characteristics of the contour sample data to obtain the characteristics of the contour sample;
inputting the contour sample characteristics into a second initial neural network model to train the second initial neural network model to obtain a trained second neural network model;
the contour sample data at least comprises human contour sample data.
7. The method of claim 1, further comprising:
acquiring a potential target of a person type from potential targets of a point cloud set in which a previous frame exists;
acquiring a point cloud set of all frames of potential targets of which the types are human, wherein all the frames comprise a current frame;
and acquiring the corresponding movement track of the potential target of which the type is a person according to the point cloud sets of all the frames.
8. A people flow statistics device, characterized in that the device comprises:
the acquisition module is used for acquiring original point cloud data of a current frame in a monitoring area;
the cluster tracking module is used for carrying out cluster tracking processing on the original point cloud data of the current frame to obtain a point cloud set of each potential target of the current frame;
the judging module is used for judging whether each potential target exists in the point cloud set of the previous frame or not according to the point cloud set of the potential target of the current frame, and taking the potential target which does not exist in the point cloud set of the previous frame as a target to be identified;
the identification module is used for identifying the point cloud set of each target to be identified so as to determine the type of each target to be identified;
and the counting module is used for accumulating the number of people of the type in the target to be identified.
9. People flow statistic device, characterized in that it comprises a microwave radar, which performs people flow statistics on a target area according to the method of any of claims 1-7.
10. A computer-readable storage medium, having a computer program stored thereon, which, when executed by a processor, causes the processor to carry out the steps of the method according to any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011486671.5A CN112561971A (en) | 2020-12-16 | 2020-12-16 | People flow statistical method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011486671.5A CN112561971A (en) | 2020-12-16 | 2020-12-16 | People flow statistical method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112561971A true CN112561971A (en) | 2021-03-26 |
Family
ID=75063832
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011486671.5A Pending CN112561971A (en) | 2020-12-16 | 2020-12-16 | People flow statistical method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112561971A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113343060A (en) * | 2021-06-23 | 2021-09-03 | 北京市商汤科技开发有限公司 | Object detection method and device, electronic equipment and storage medium |
| CN113468284A (en) * | 2021-06-29 | 2021-10-01 | 北京市商汤科技开发有限公司 | Object detection method and device, electronic equipment and storage medium |
| CN113936456A (en) * | 2021-09-30 | 2022-01-14 | 同济大学 | Street-crossing traffic identification and feature analysis method based on millimeter wave radar |
| CN114137509A (en) * | 2021-11-30 | 2022-03-04 | 南京慧尔视智能科技有限公司 | Point cloud clustering method and device based on millimeter wave radar |
| CN115272853A (en) * | 2022-07-27 | 2022-11-01 | 清华大学 | Industrial wasteland identification method and product based on artificial intelligence technology and big data |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106503613A (en) * | 2016-09-13 | 2017-03-15 | 华东交通大学 | A kind of public domain human body behavior monitoring method |
| CN106846297A (en) * | 2016-12-21 | 2017-06-13 | 深圳市镭神智能系统有限公司 | Pedestrian's flow quantity detecting system and method based on laser radar |
| CN108986064A (en) * | 2017-05-31 | 2018-12-11 | 杭州海康威视数字技术股份有限公司 | A kind of people flow rate statistical method, equipment and system |
| CN110376585A (en) * | 2019-07-23 | 2019-10-25 | 交控科技股份有限公司 | Compartment crowding detection method and device, system based on 3D radar scanning |
| CN110738105A (en) * | 2019-09-05 | 2020-01-31 | 哈尔滨工业大学（深圳） | method, device, system and storage medium for calculating urban street cell pedestrian flow based on deep learning |
| CN110927712A (en) * | 2019-10-28 | 2020-03-27 | 珠海格力电器股份有限公司 | Tracking method and device |
| CN111047901A (en) * | 2019-11-05 | 2020-04-21 | 珠海格力电器股份有限公司 | Parking management method, parking management device, storage medium and computer equipment |
-
2020
- 2020-12-16 CN CN202011486671.5A patent/CN112561971A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106503613A (en) * | 2016-09-13 | 2017-03-15 | 华东交通大学 | A kind of public domain human body behavior monitoring method |
| CN106846297A (en) * | 2016-12-21 | 2017-06-13 | 深圳市镭神智能系统有限公司 | Pedestrian's flow quantity detecting system and method based on laser radar |
| CN108986064A (en) * | 2017-05-31 | 2018-12-11 | 杭州海康威视数字技术股份有限公司 | A kind of people flow rate statistical method, equipment and system |
| CN110376585A (en) * | 2019-07-23 | 2019-10-25 | 交控科技股份有限公司 | Compartment crowding detection method and device, system based on 3D radar scanning |
| CN110738105A (en) * | 2019-09-05 | 2020-01-31 | 哈尔滨工业大学（深圳） | method, device, system and storage medium for calculating urban street cell pedestrian flow based on deep learning |
| CN110927712A (en) * | 2019-10-28 | 2020-03-27 | 珠海格力电器股份有限公司 | Tracking method and device |
| CN111047901A (en) * | 2019-11-05 | 2020-04-21 | 珠海格力电器股份有限公司 | Parking management method, parking management device, storage medium and computer equipment |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113343060A (en) * | 2021-06-23 | 2021-09-03 | 北京市商汤科技开发有限公司 | Object detection method and device, electronic equipment and storage medium |
| CN113468284A (en) * | 2021-06-29 | 2021-10-01 | 北京市商汤科技开发有限公司 | Object detection method and device, electronic equipment and storage medium |
| CN113936456A (en) * | 2021-09-30 | 2022-01-14 | 同济大学 | Street-crossing traffic identification and feature analysis method based on millimeter wave radar |
| CN114137509A (en) * | 2021-11-30 | 2022-03-04 | 南京慧尔视智能科技有限公司 | Point cloud clustering method and device based on millimeter wave radar |
| CN114137509B (en) * | 2021-11-30 | 2023-10-13 | 南京慧尔视智能科技有限公司 | Millimeter wave Lei Dadian cloud clustering method and device |
| CN115272853A (en) * | 2022-07-27 | 2022-11-01 | 清华大学 | Industrial wasteland identification method and product based on artificial intelligence technology and big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112561971A (en) | People flow statistical method, device, equipment and storage medium | |
| CN111462488B (en) | Intersection safety risk assessment method based on deep convolutional neural network and intersection behavior characteristic model | |
| Zhang et al. | Video-based vehicle detection and classification system for real-time traffic data collection using uncalibrated video cameras | |
| KR101592889B1 (en) | Object matching for tracking, indexing, and search | |
| CN106559749B (en) | Multi-target passive positioning method based on radio frequency tomography | |
| US20030053659A1 (en) | Moving object assessment system and method | |
| CN111126252A (en) | Stall behavior detection method and related device | |
| Rabbouch et al. | Unsupervised video summarization using cluster analysis for automatic vehicles counting and recognizing | |
| Naddaf-Sh et al. | Real‐Time Road Crack Mapping Using an Optimized Convolutional Neural Network | |
| CN101410855A (en) | Method for automatically attributing one or more object behaviors | |
| CN114445803A (en) | Driving data processing method and device and electronic equipment | |
| KR102260123B1 (en) | Apparatus for Sensing Event on Region of Interest and Driving Method Thereof | |
| US20220026557A1 (en) | Spatial sensor system with background scene subtraction | |
| CN113537417A (en) | Target identification method and device based on radar, electronic equipment and storage medium | |
| Tak et al. | Development of AI‐Based Vehicle Detection and Tracking System for C‐ITS Application | |
| Bathaee et al. | A cluster analysis approach for differentiating transportation modes using Bluetooth sensor data | |
| CN111929672A (en) | Method and device for determining movement track, storage medium and electronic device | |
| CN111797831A (en) | BIM and artificial intelligence based parallel abnormality detection method for poultry feeding | |
| Zheng et al. | Intelligent cognition of traffic loads on road bridges: From measurement to simulation–A review | |
| Kadim et al. | Real-time deep-learning based traffic volume count for high-traffic urban arterial roads | |
| CN113869427A (en) | Scene analysis method and device, electronic equipment and storage medium | |
| Xin et al. | Risk evaluation for conflicts between crossing pedestrians and right-turning vehicles at intersections | |
| Kanschat et al. | Wireless-Signal-based vehicle counting and classification in different road environments | |
| Zhao et al. | Context-guided coarse-to-fine detection model for bird nest detection on high-speed railway catenary | |
| CN111354191B (en) | Lane driving condition determining method, device and equipment and 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 |