CN109218667B - Public place safety early warning system and method - Google Patents
Public place safety early warning system and method Download PDFInfo
- Publication number
- CN109218667B CN109218667B CN201811049578.0A CN201811049578A CN109218667B CN 109218667 B CN109218667 B CN 109218667B CN 201811049578 A CN201811049578 A CN 201811049578A CN 109218667 B CN109218667 B CN 109218667B
- Authority
- CN
- China
- Prior art keywords
- crowd
- image
- parameters
- camera
- public place
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
Abstract
The application provides a public place safety early warning system and a method, wherein the system comprises: the system comprises a hyperspectral camera, a first camera shooting equipment group and a data processor; the method comprises the following steps that a hyperspectral camera is used for scanning crowds in a public place, a hyperspectral image comprising position information and spectral information can be collected, and more obvious figure outline and posture information, namely posture parameters, can be obtained by processing the hyperspectral image; the first camera shooting equipment group extracts position parameters according to the attitude parameters obtained by the hyperspectral camera, and then shoots people; according to a multi-view scene three-dimensional reconstruction algorithm, the movement parameters of pedestrians and the relative parameters among the pedestrians can be obtained; and finally, the data processor is combined with artificial intelligence or a machine learning algorithm to judge the danger coefficients of the attitude parameters, the movement parameters and the relative parameters among pedestrians. By implementing the embodiment of the application, the early warning signal can be timely sent out when a dangerous condition occurs, and personal and property safety of people in a public place is guaranteed.
Description
Technical Field
The application belongs to the field of security and protection, and particularly relates to a public place safety early warning system and method.
Background
With the development of science and technology, the safety early warning technology in public places integrates advanced artificial intelligence technologies such as machine vision, image processing and deep learning, and at present, many companies develop early warning systems based on videos shot by common cameras.
However, the early warning system can only be applied to a limited environment, because in a real environment, factors such as complex background, insufficient illumination, view diversity and the like existing in different scenes influence the imaging effect of the camera, and the sensitivity and the accuracy of the early warning system are reduced. Therefore, the sensitivity and accuracy of the security early warning system need to be improved.
Disclosure of Invention
The application aims at providing a public place safety early warning system, which can accurately and effectively send out early warning signals in time when dangerous conditions occur, and ensure personal and property safety of people in public places.
In order to achieve the above application purpose, the technical solution adopted in the present application is a public place safety early warning system, including: the hyperspectral camera equipment, the first camera shooting equipment group and the data processor are arranged;
the hyperspectral camera apparatus includes a hyperspectral camera and an image processor; the hyperspectral camera is used for shooting people in the current public place to obtain a hyperspectral image comprising position information and spectral information; the image processor is used for processing the hyperspectral image to obtain a crowd image with the most obvious contrast as a first crowd image, and extracting attitude parameters of all pedestrians in the first crowd image, wherein the attitude parameters are used for representing the attitudes and contours of the pedestrians; the hyperspectral camera equipment transmits position information to the first camera equipment group;
the first camera shooting equipment group comprises at least one camera shooting equipment, and the camera shooting equipment is used for shooting the crowd in the current public place according to the attitude parameters to obtain a first crowd camera shooting image and extracting position parameters representing the walking position of the pedestrian in the first crowd camera shooting image;
and the data processor is used for judging a danger coefficient of the current public place as a first danger coefficient according to the attitude parameter and the position parameter, and sending a first early warning signal under the condition that the first danger coefficient is higher than a first safety coefficient.
Optionally, the data processor is further configured to obtain a first risk judgment model according to machine learning and training of the first crowd camera images under the normal condition and the first crowd camera images under the dangerous condition; the data processor is specifically configured to judge a risk coefficient of a current public place as a first risk coefficient according to the attitude parameter, the position parameter and the first risk judgment model, and send a first early warning signal when the first risk coefficient is higher than a first safety factor.
Optionally, the system further comprises a second group of camera devices; the image processor of the hyperspectral camera equipment is further used for processing the hyperspectral image to obtain a second crowd image and extracting coordinate information of a suspicious person with dangerous goods in the second crowd image; the hyperspectral camera equipment transmits the coordinate information to the second camera equipment group; the second camera shooting device group comprises at least one camera shooting device, the camera shooting device is used for shooting the crowd in the current public place according to the coordinate information to obtain a second crowd camera shooting image, and suspicious position parameters which represent the walking position of the suspicious person in the second crowd camera shooting image are extracted; and the data processor is also used for judging the danger coefficient of the suspicious person as a second danger coefficient according to the suspicious position parameter and the posture parameter, and sending a second early warning signal under the condition that the second danger coefficient is greater than a second safety factor.
Optionally, the data processor is further configured to obtain a second risk judgment model according to machine learning and training of the second crowd photographic image under the normal condition and the second crowd photographic image under the dangerous condition; the data processor is specifically configured to judge a risk coefficient of a current public place as a second risk coefficient according to the attitude parameter, the suspicious location parameter and the second risk judgment model, and send a second early warning signal when the second risk coefficient is higher than a second safety factor.
Optionally, the location parameters and the suspicious location parameters include: a movement parameter and a relative parameter; the movement parameters are used for representing the moving direction and speed of the pedestrian or the suspicious person; the relative parameters are used for representing the distance and the direction between the pedestrian or the suspicious person and other pedestrians along with the change of time.
Optionally, the pose parameters comprise parameters for characterizing the face orientation of the pedestrian.
On the other hand, the application also provides a public place safety early warning method, which is applied to any one of the public place safety early warning systems, and the method comprises the following steps:
the method comprises the steps that hyperspectral camera equipment shoots people in a current public place to obtain people images of different wave bands, the people images of the wave bands with the most obvious people shooting contrast are screened out from the people images of the different wave bands to serve as first people images, attitude parameters of all pedestrians in the first people images are extracted, the attitude parameters are used for representing the attitudes and contours of the pedestrians, and the attitude parameters are transmitted to a first camera equipment group;
the first camera shooting equipment group shoots the crowd in the current public place according to the attitude parameters to obtain a first crowd camera shooting image, and position parameters representing the walking position of the pedestrian in the first crowd camera shooting image are extracted;
and the data processor judges the risk coefficient of the current public place as a first risk coefficient according to the attitude parameter and the position parameter, and sends out a first early warning signal under the condition that the first risk coefficient is higher than a first safety coefficient.
Optionally, the method further comprises: the data processor obtains a first danger judgment model according to machine learning and training of the first crowd camera images under the normal condition and the first crowd camera images under the dangerous condition; the data processor judges the danger coefficient of the current public place as a first danger coefficient according to the attitude parameter and the position parameter, and the method comprises the following steps: and the data processor judges the danger coefficient of the current public place as a first danger coefficient according to the attitude parameter, the position parameter and the first danger judgment model.
Optionally, the method further comprises: the hyperspectral camera equipment screens out the crowd images of the wave bands with the most obvious dangerous goods shooting contrast from the crowd images of different wave bands as second crowd images, extracts coordinate information of suspicious people with dangerous goods in the second crowd images, and transmits the coordinate information to the second camera equipment group; the second camera shooting equipment group shoots the people in the current public place according to the coordinate information to obtain a second people camera shooting image, and suspicious position parameters which represent the walking position of the suspicious people in the second people camera shooting image are extracted; and the data processor judges the danger coefficient of the suspicious character as a second danger coefficient according to the suspicious position parameter and the posture parameter, and sends out a second early warning signal under the condition that the second danger coefficient is greater than a second safety factor.
Optionally, the method further comprises: the data processor obtains a second danger judgment model according to machine learning and training of the second crowd camera images under the normal condition and the dangerous condition; the data processor judges the risk coefficient of the suspicious person as a second risk coefficient according to the suspicious position parameter and the posture parameter, and the method comprises the following steps: and the data processor judges the danger coefficient of the current public place as a second danger coefficient according to the attitude parameter, the suspicious position parameter and the second danger judgment model, and sends out a second early warning signal under the condition that the second danger coefficient is higher than a second safety factor.
The application has the following beneficial effects:
according to the scheme, the crowd image of the public place with the best figure shooting wave band shot by the hyperspectral camera is selected, so that clear figure outline and posture information, namely posture parameters, can be obtained, the image pickup equipment shoots the crowd according to the posture parameters so as to further obtain the position parameters of the pedestrian, so that the influence of poor imaging environment can be avoided, and accurate pedestrian posture and position information can be obtained; and finally, the data processor is combined with artificial intelligence or a machine learning algorithm to judge the danger coefficients of the attitude parameters, the movement parameters and the relative parameters among pedestrians. Therefore, the safety condition of the public place can be monitored in real time, the early warning signal can be accurately and effectively sent out in time when the dangerous condition occurs, and the personal and property safety of people in the public place is ensured.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
Fig. 1 is a schematic structural diagram of a public place safety early warning system disclosed in the present application;
FIG. 2 is a schematic view of a public safety warning system according to the present disclosure;
FIG. 3 is a schematic diagram of the establishment of the first/second risk assessment models disclosed herein;
FIG. 4 is a schematic diagram of another public space security early warning system disclosed in the present application;
FIG. 5 is a schematic diagram of a public place security early warning method disclosed in the present application;
fig. 6 is a schematic diagram of a public place security early warning implementation process disclosed in the present application.
Detailed Description
In order to make the technical problems, technical solutions and advantageous effects solved by the present invention more clearly apparent, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
As shown in fig. 1, fig. 1 is a schematic structural diagram of a public place safety precaution system disclosed in the present application. This public place safety precaution system 100 includes: a hyper-spectral camera apparatus 110, a first image pickup apparatus group 120, a data processor 103;
the hyperspectral camera device 110 includes a hyperspectral camera 111 and an image processor 112; the hyperspectral camera 111 is used for shooting crowds in the current public place to obtain crowd images of different wave bands; the image processor 112 is configured to screen out a crowd image of a waveband with the most obvious shooting contrast of a person from the crowd images of different wavebands as a first crowd image, and extract pose parameters of each pedestrian in the first crowd image, where the pose parameters are used to represent the pose and the contour of the pedestrian; the hyperspectral camera device 110 transmits the attitude parameters to the first camera device group 120;
the first image pickup device group 120 comprises at least one image pickup device, and the image pickup devices are used for carrying out image pickup on people in the current public place according to the attitude parameters to obtain a first crowd pickup image and extracting position parameters representing the walking position of the pedestrian in the first crowd pickup image;
and the data processor 130 is configured to judge a risk coefficient of the current public place as a first risk coefficient according to the attitude parameter and the position parameter, and send a first early warning signal when the first risk coefficient is higher than a first safety factor.
In the prior art, there is also a scheme of shooting people in public places by using the camera device and sending out an early warning signal after processing the shot image data, but different scenes have factors such as complex background, insufficient illumination, view diversity and the like, and the imaging of the common camera device can be influenced by the adverse environmental factors.
The hyperspectral camera is used for single-scene scanning, and a complete hyperspectral image including space and spectrum information can be acquired through one-time exposure. The imaging of a certain specific wavelength to the human body is more accurate, and even in a complex environment background, pedestrians can still be segmented from the background environment. In the application, a hyperspectral camera shoots people in a current public place to obtain people images of different wave bands; the image processor screens out a crowd image of a wave band with the most obvious shooting contrast of a person from the crowd images of different wave bands as a first crowd image, and extracts attitude parameters (used for representing the attitude and the contour of a pedestrian) of each pedestrian in the first crowd image; the hyperspectral camera equipment transmits the attitude parameters to the camera equipment, so that the camera equipment can be helped to clearly analyze the outline of the pedestrian in the image picture, and the subsequent acquisition of the pedestrian position parameters is helped.
In the present application, the location parameters include: a movement parameter and a relative parameter; the movement parameters are used for representing the moving direction and speed of the pedestrian or the suspicious person; relative parameters characterizing the distance and direction between the pedestrian or suspect and other pedestrians over time. Normally, the relative parameters should be relatively stable, and in the event of fighting, the relative parameters would be very chaotic.
For example, as shown in fig. 2, fig. 2 is a schematic diagram of a public place security early warning system disclosed in the present application. The image high-spectrum camera 1 shoots people in the current public place to obtain people images of different wave bands, an image of the wave band with the most obvious shooting contrast of people is screened out to be used as a first people image, the pedestrians are segmented from a complex background, and posture parameters of the pedestrians are extracted, wherein the posture parameters can be parameters for representing the face orientation of the pedestrians. A plurality of common cameras in the figure utilize multi-view imaging of a plurality of cameras, and moving parameters of pedestrians and relative parameters among the pedestrians can be obtained according to a multi-view scene three-dimensional reconstruction algorithm.
The data processor can analyze and judge the danger coefficient of the current public place according to the attitude parameter, the movement parameter and the relative parameter, and send out an early warning signal under the condition that the danger coefficient is greater than a safety threshold, for example, the early warning device informs nearby security personnel to arrive at the site for assistance or directly sends out an alarm sound. Normally, the movement parameters, relative parameters, posture parameters (face orientation), etc. of the pedestrian are similar, and the abnormal behavior shows special parameter characteristics, such as: the pedestrian runs rapidly, relative parameter confusion, the pedestrian moving parameter stops, and the like belong to abnormal behaviors, and early warning and important attention are needed.
The method and the device have the advantages that the crowd images of the public places shot by the hyperspectral camera are selected, people shoot the crowd images of the wave bands with the most obvious contrast, so that clear outline and posture information of the people, namely posture parameters, can be obtained, the image pickup equipment shoots the crowd according to the posture parameters so as to further obtain the position parameters of the pedestrians, so that the influence of poor imaging environment can be avoided, and accurate pedestrian postures and position information can be obtained; and finally, carrying out danger coefficient judgment on the attitude parameters and the position parameters by a data processor. Therefore, through the combination of the hyperspectral camera and the camera, the higher accuracy rate than that of the traditional early warning system can be realized, the safety condition of a public place can be monitored in real time, early warning signals can be accurately and effectively sent out in time when dangerous conditions appear, and the personal and property safety of people in the public place is ensured.
As an optional implementation manner, the data processor 130 is further configured to obtain a first risk judgment model according to machine learning and training of the first crowd camera images under the normal condition and the first crowd camera images under the dangerous condition; the data processor 130 is specifically configured to judge a risk coefficient of the current public place as a first risk coefficient according to the attitude parameter, the position parameter and the first risk judgment model, and send a first warning signal when the first risk coefficient is higher than a first safety factor. The safety factor can be set according to specific conditions, and can be set through machine learning and training, and the scheme is not limited.
The method of machine learning may employ an artificial neural network model.
Fig. 3 is a schematic diagram illustrating the establishment of a first risk assessment model disclosed in the present application, as shown in fig. 3. The input layer, hidden layer and output layer in the graph are common algorithm models for machine learning. The data processor 130 in the figure performs machine learning and training on the movement parameters, the relative parameters and the posture parameters of the pedestrians in the first crowd-shoot image under the normal condition, and simultaneously, the data processor 130 also performs machine learning and training on the movement parameters, the relative parameters and the posture parameters of the pedestrians in the first crowd-shoot image under the dangerous condition, so that after a large amount of learning and training, the first danger judgment model is formed. The danger coefficient of the current public place can be obtained by inputting the movement parameters, the relative parameters and the attitude parameters of the current pedestrian into the first danger judgment model.
As shown in fig. 4, another embodiment of the present application also discloses a public place safety precaution system 200, the difference of the system 200 compared with the public place safety precaution system 100 is that the image processor 212 of the hyperspectral camera device 210 is further configured to screen out, from the crowd images in different bands, the crowd image in the band with the most obvious contrast of the dangerous goods shooting as a second crowd image, and extract the coordinate information of the suspicious person with the dangerous goods in the second crowd image; the hyperspectral camera equipment transmits the coordinate information to the second camera equipment group; the second camera device group 240 comprises at least one camera device, and the camera devices are used for shooting the people in the current public place according to the coordinate information to obtain a second people camera image and extracting suspicious position parameters representing the walking positions of suspicious people in the second people camera image; and the data processor 230 is further configured to judge the risk coefficient of the suspicious person according to the suspicious position parameter and the posture parameter, and send a second early warning signal when the second risk coefficient is greater than a second safety factor.
The hyperspectral camera 2 is used for single-shot scanning, and a complete hyperspectral image including space and spectrum information can be acquired through one-time exposure. A particular wavelength images a hazardous material (hazardous instruments, liquid fuel, etc.) more accurately, even in a complex environmental context, yet still separates the hazardous material from the background environment. Therefore, the crowd image with the wave band with the most obvious contrast of the dangerous goods shooting is selected as a second crowd image, and the coordinate information of the suspicious person with the dangerous goods in the second crowd image is extracted; the hyperspectral camera equipment transmits the coordinate information to the camera equipment, namely, a suspicious person carrying the dangerous goods is locked by the dangerous goods, and the coordinate information is transmitted to the camera equipment, so that the camera equipment is facilitated to further track the movement parameters and the relative parameters of the suspicious person.
The data processor can analyze and judge the danger coefficient of the current public place according to the posture parameter, the movement parameter and the relative parameter of the suspicious person, is connected with the early warning device, and sends out an early warning signal under the condition that the danger coefficient is larger than a safety threshold value, for example, the early warning device informs nearby security personnel to drive to the site for assistance or directly sends out an alarm sound.
As an optional implementation manner, the data processor is further configured to obtain a second risk judgment model according to machine learning and training of the second crowd photographic image under the normal condition and the second crowd photographic image under the dangerous condition; and the data processor is specifically used for judging the danger coefficient of the current public place as a second danger coefficient according to the attitude parameter, the suspicious position parameter and the second danger judgment model, and sending a second early warning signal under the condition that the second danger coefficient is higher than a second safety coefficient, wherein the safety coefficient can be set according to specific conditions, can be set through machine learning training, and is not limited in the scheme.
Similar to fig. 3, the data processor 230 in this embodiment performs machine learning and training on the movement parameters, the relative parameters, and the posture parameters of the suspicious person in the second crowd photographic image under the normal condition, and the data processor 230 also performs machine learning and training on the movement parameters, the relative parameters, and the posture parameters of the suspicious person in the second crowd photographic image under the dangerous condition, so that after a large amount of learning and training, the second danger determination model is formed. The moving parameters, the relative parameters and the attitude parameters of the current pedestrian are input into the second danger judgment model, so that the danger coefficient of the current public place can be obtained.
As shown in fig. 5, the present application further provides a public place safety precaution method, which is applied to any one of the public place safety precaution systems, and the method includes:
501. the method comprises the steps that hyperspectral camera equipment shoots people in a current public place to obtain people images of different wave bands, the people images of the wave bands with the most obvious people shooting contrast are screened out from the people images of the different wave bands to serve as first people images, posture parameters of all pedestrians in the first people images are extracted and used for representing postures and contours of the pedestrians, and the posture parameters are transmitted to a first camera equipment group;
502. the first camera shooting equipment group shoots the crowd in the current public place according to the attitude parameters to obtain a first crowd camera shooting image, and position parameters representing the walking position of the pedestrian in the first crowd camera shooting image are extracted;
503. and the data processor judges the risk coefficient of the current public place as a first risk coefficient according to the attitude parameter and the position parameter, and sends out a first early warning signal under the condition that the first risk coefficient is higher than a first safety coefficient.
As an optional implementation, the method further includes: the data processor obtains a first danger judgment model according to machine learning and training of the first crowd camera images under the normal condition and the first crowd camera images under the dangerous condition; the data processor judges the danger coefficient of the current public place as a first danger coefficient according to the attitude parameter and the position parameter, and the method comprises the following steps: and the data processor judges the risk coefficient of the current public place as a first risk coefficient according to the attitude parameter, the position parameter and the first risk judgment model.
As an optional implementation, the method further includes: the hyperspectral camera equipment screens out the crowd images of the wave bands with the most obvious dangerous goods shooting contrast from the crowd images of different wave bands to serve as second crowd images, extracts coordinate information of suspicious people with dangerous goods in the second crowd images, and transmits the coordinate information to a second camera equipment group; the second camera shooting device group shoots the crowd in the current public place according to the coordinate information to obtain a second crowd camera shooting image, and extracts suspicious position parameters which represent the walking positions of suspicious people in the second crowd camera shooting image; and the data processor judges the risk coefficient of the suspicious person according to the suspicious position parameter and the posture parameter to serve as a second risk coefficient, and sends out a second early warning signal under the condition that the second risk coefficient is greater than a second safety factor.
As an optional implementation, the method further includes: the data processor obtains a second danger judgment model according to machine learning and training of the second crowd camera images under the normal condition and the dangerous condition; the data processor judges the risk coefficient of the suspicious person as a second risk coefficient according to the suspicious position parameter and the posture parameter, and the method comprises the following steps: and the data processor judges the danger coefficient of the current public place as a second danger coefficient according to the attitude parameter, the suspicious position parameter and the second danger judgment model, and sends out a second early warning signal under the condition that the second danger coefficient is higher than a second safety coefficient.
An implementation flow of the public place safety early warning method is shown in fig. 6, wherein the hyperspectral camera shoots images of public place crowds, and the hyperspectral camera can shoot a plurality of waveband images of the public place crowds at one time; screening out a crowd image of a wave band with the most obvious shooting contrast of people as a first crowd image and screening out a crowd image of a wave band with the most obvious shooting contrast of dangerous goods as a second crowd image by an image processor; continuously extracting attitude parameters of each pedestrian in the first crowd image, wherein the attitude parameters are used for representing the attitude and the contour of the pedestrian, transmitting the attitude parameters to a first camera equipment group, extracting coordinate information of suspicious people with dangerous goods in a second crowd image, and transmitting the coordinate information to a second camera equipment group; the first camera shooting equipment group and the second camera shooting equipment group respectively shoot the crowd in the current public place according to the attitude parameter and the coordinate information to obtain a first crowd camera shooting image and a second crowd camera shooting image; respectively extracting position parameters representing the walking positions of the pedestrians in the first crowd camera image, and extracting suspicious position parameters representing the walking positions of the suspicious people in the second crowd camera image; inputting the attitude parameter and the position parameter into a first danger coefficient of a first danger judgment model, and inputting the suspicious position parameter and the attitude parameter into a second danger coefficient of a second danger judgment model; and if the first danger coefficient is greater than the first safety coefficient, a first early warning signal is sent out, and if the second danger coefficient is greater than the second safety coefficient, a second early warning signal is sent out.
The specific implementation method of the public place safety early warning method is the same as that of the corresponding public place safety early warning system, and details are not repeated here.
While embodiments of the invention have been shown and described, it will be understood by those of ordinary skill in the art that: various changes, modifications, substitutions and alterations can be made to the embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.
Claims (8)
1. A public place safety precaution system, comprising: the hyperspectral camera equipment, the first camera shooting equipment group and the data processor are arranged;
the hyperspectral camera apparatus includes a hyperspectral camera and an image processor; the hyperspectral camera is used for shooting people in the current public place to obtain a hyperspectral image comprising position information and spectral information; the image processor is used for processing the hyperspectral image to obtain a crowd image with the most obvious contrast as a first crowd image, and extracting attitude parameters of all pedestrians in the first crowd image, wherein the attitude parameters are used for representing the attitude and contour of the pedestrians and the face orientation of the pedestrians; the hyperspectral camera equipment transmits attitude parameters to the first camera equipment group, wherein the hyperspectral image is processed to obtain a crowd image with the most obvious contrast as a first crowd image, and the method comprises the following steps: screening out crowd images of different wave bands from the hyperspectral images of different wave bands, and screening out crowd images of wave bands with most obvious figure shooting contrast from the crowd images of different wave bands as first crowd images;
the first image pickup equipment group comprises at least one image pickup equipment, the image pickup equipment is used for carrying out image pickup on people in the current public place according to the attitude parameters to obtain a first crowd image, and position parameters representing the walking position of a pedestrian in the first crowd image are extracted, wherein the position parameters comprise: the moving parameters are used for representing the moving direction and speed of the pedestrian; relative parameters for characterizing the distance and direction between pedestrians as a function of time;
and the data processor is used for judging a danger coefficient of the current public place as a first danger coefficient according to the attitude parameter and the position parameter, and sending a first early warning signal under the condition that the first danger coefficient is higher than a first safety coefficient.
2. The system of claim 1,
the data processor is further used for obtaining a first danger judgment model according to machine learning and training of the first crowd camera images under the normal condition and the first crowd camera images under the dangerous condition;
the data processor is specifically configured to judge a risk coefficient of a current public place as a first risk coefficient according to the attitude parameter, the position parameter and the first risk judgment model, and send a first early warning signal when the first risk coefficient is higher than a first safety factor.
3. The system of claim 1, further comprising a second group of cameras;
the image processor of the hyperspectral camera equipment is further used for processing the hyperspectral image to obtain a second crowd image and extracting coordinate information of a suspicious person with dangerous goods in the second crowd image; the hyperspectral camera equipment transmits the coordinate information to the second camera equipment group, wherein the second crowd image is a crowd image of a wave band with the most obvious dangerous article shooting contrast screened out from crowd images of different wave bands by the hyperspectral camera equipment;
the second camera shooting device group comprises at least one camera shooting device, the camera shooting device is used for shooting the crowd in the current public place according to the coordinate information to obtain a second crowd camera shooting image, and suspicious position parameters which represent the walking position of the suspicious person in the second crowd camera shooting image are extracted;
and the data processor is also used for judging the danger coefficient of the suspicious person as a second danger coefficient according to the suspicious position parameter and the posture parameter of the suspicious person, and sending a second early warning signal under the condition that the second danger coefficient is greater than a second safety factor.
4. The system of claim 3,
the data processor is further used for obtaining a second danger judgment model according to machine learning and training of the second crowd camera images under the normal condition and the second crowd camera images under the dangerous condition;
the data processor is specifically used for judging the danger coefficient of the current public place as a second danger coefficient according to the posture parameter of the suspicious person, the suspicious position parameter and the second danger judgment model, and sending a second early warning signal under the condition that the second danger coefficient is higher than a second safety coefficient.
5. A public place security early warning method applied to the public place security early warning system according to claim 1, the method comprising:
the method comprises the steps that hyperspectral camera equipment shoots people in a current public place to obtain people images of different wave bands, the people images of the wave bands with the most obvious people shooting contrast are screened out from the people images of the different wave bands to serve as first people images, posture parameters of all pedestrians in the first people images are extracted, the posture parameters are used for representing postures and contours of the pedestrians and face orientations of the pedestrians, and the posture parameters are transmitted to a first camera equipment group;
the method comprises the steps that a first image pickup device group carries out image pickup on people in the current public place according to the attitude parameters to obtain a first crowd pickup image, and position parameters representing the walking position of a pedestrian in the first crowd pickup image are extracted, wherein the position parameters comprise: the moving parameters are used for representing the moving direction and speed of the pedestrian; relative parameters for characterizing the distance and direction between pedestrians as a function of time;
and the data processor judges the risk coefficient of the current public place as a first risk coefficient according to the attitude parameter and the position parameter, and sends out a first early warning signal under the condition that the first risk coefficient is higher than a first safety coefficient.
6. The method of claim 5, further comprising:
the data processor obtains a first danger judgment model according to machine learning and training of the first crowd camera images under the normal condition and the first crowd camera images under the dangerous condition;
the data processor judges the danger coefficient of the current public place as a first danger coefficient according to the attitude parameter and the position parameter, and the method comprises the following steps:
and the data processor judges the danger coefficient of the current public place as a first danger coefficient according to the attitude parameter, the position parameter and the first danger judgment model.
7. The method of claim 5, further comprising:
the hyperspectral camera equipment screens out the crowd images of the wave bands with the most obvious dangerous goods shooting contrast from the crowd images of different wave bands as second crowd images, extracts coordinate information of suspicious people with dangerous goods in the second crowd images, and transmits the coordinate information to a second camera equipment group;
the second camera shooting equipment group shoots the crowd in the current public place according to the coordinate information to obtain a second crowd camera shooting image, and suspicious position parameters which represent the walking position of the suspicious person in the second crowd camera shooting image are extracted;
and the data processor judges the risk coefficient of the suspicious person as a second risk coefficient according to the suspicious position parameter and the posture parameter of the suspicious person, and sends out a second early warning signal under the condition that the second risk coefficient is greater than a second safety factor.
8. The method of claim 7, further comprising:
the data processor obtains a second danger judgment model according to machine learning and training of the second crowd camera images under the normal condition and the dangerous condition;
the data processor judges the risk coefficient of the suspicious person as a second risk coefficient according to the suspicious position parameter and the posture parameter of the suspicious person, and the method comprises the following steps:
and the data processor judges the danger coefficient of the current public place as a second danger coefficient according to the posture parameter and the suspicious position parameter of the suspicious person and the second danger judgment model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811049578.0A CN109218667B (en) | 2018-09-08 | 2018-09-08 | Public place safety early warning system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811049578.0A CN109218667B (en) | 2018-09-08 | 2018-09-08 | Public place safety early warning system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109218667A CN109218667A (en) | 2019-01-15 |
| CN109218667B true CN109218667B (en) | 2021-01-05 |
Family
ID=64987304
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811049578.0A Active CN109218667B (en) | 2018-09-08 | 2018-09-08 | Public place safety early warning system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109218667B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111753582A (en) * | 2019-03-27 | 2020-10-09 | 杭州海康威视数字技术股份有限公司 | Dangerous goods positioning method and device |
| CN110097736B (en) * | 2019-04-23 | 2021-11-30 | 维沃移动通信有限公司 | Alarm method and alarm device |
| CN111064931A (en) * | 2019-12-18 | 2020-04-24 | 新疆联海创智信息科技有限公司 | System and method for accurately identifying identity of dangerous person |
| JP7447626B2 (en) | 2020-03-30 | 2024-03-12 | 大日本印刷株式会社 | Information processing method, information processing device, program and information processing system |
| CN113163158A (en) * | 2021-01-30 | 2021-07-23 | 高慧俊 | Intelligent network monitoring system |
| CN113256938A (en) * | 2021-05-31 | 2021-08-13 | 苏州优函信息科技有限公司 | Fall monitoring alarm scheme based on spectrum camera |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105092485A (en) * | 2015-08-27 | 2015-11-25 | 泉州装备制造研究所 | Dangerous article detection method and device |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080243439A1 (en) * | 2007-03-28 | 2008-10-02 | Runkle Paul R | Sensor exploration and management through adaptive sensing framework |
| CN102096803B (en) * | 2010-11-29 | 2013-11-13 | 吉林大学 | Safe state recognition system for people on basis of machine vision |
| CN104866841B (en) * | 2015-06-05 | 2018-03-09 | 中国人民解放军国防科学技术大学 | A kind of human body target is run behavioral value method |
| CN106385530B (en) * | 2015-07-28 | 2022-12-13 | 杭州海康微影传感科技有限公司 | Double-spectrum camera |
| CN105913559B (en) * | 2016-04-06 | 2019-03-05 | 南京华捷艾米软件科技有限公司 | A kind of ATM in bank intelligent control method based on body-sensing technology |
| CN106375725B (en) * | 2016-09-20 | 2019-04-26 | 山东士森集成科技有限公司 | A kind of intelligent video monitoring system based on recognition of face |
| CN107133976B (en) * | 2017-04-24 | 2020-04-14 | 浙江大学 | Method and device for acquiring three-dimensional hyperspectral information |
-
2018
- 2018-09-08 CN CN201811049578.0A patent/CN109218667B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105092485A (en) * | 2015-08-27 | 2015-11-25 | 泉州装备制造研究所 | Dangerous article detection method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109218667A (en) | 2019-01-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109218667B (en) | Public place safety early warning system and method | |
| CN109819208B (en) | Intensive population security monitoring management method based on artificial intelligence dynamic monitoring | |
| Yuan et al. | Fire detection using infrared images for UAV-based forest fire surveillance | |
| US8761445B2 (en) | Method and system for detection and tracking employing multi-view multi-spectral imaging | |
| CN105913559B (en) | A kind of ATM in bank intelligent control method based on body-sensing technology | |
| CN101119482B (en) | Overall view monitoring method and apparatus | |
| KR101709751B1 (en) | An automatic monitoring system for dangerous situation of persons in the sea | |
| Kumar et al. | Study of robust and intelligent surveillance in visible and multi-modal framework | |
| CN112396658B (en) | Indoor personnel positioning method and system based on video | |
| Sommer et al. | Flying object detection for automatic UAV recognition | |
| CN104820995A (en) | Large public place-oriented people stream density monitoring and early warning method | |
| CN111191586A (en) | Method and system for inspecting wearing condition of safety helmet of personnel in construction site | |
| KR101853903B1 (en) | Invasion detecting method and system including CCTV, drone and fiber to the fence | |
| CN110751081B (en) | Construction safety monitoring method and device based on machine vision | |
| CN112183472A (en) | Method for detecting whether test field personnel wear work clothes or not based on improved RetinaNet | |
| CN107316024B (en) | Perimeter alarm algorithm based on deep learning | |
| CN112613359B (en) | Construction method of neural network for detecting abnormal behaviors of personnel | |
| Shah et al. | Use of Deep Learning Applications for Drone Technology | |
| CN112800918A (en) | Identity recognition method and device for illegal moving target | |
| Mantini et al. | UHCTD: A comprehensive dataset for camera tampering detection | |
| CN111612815A (en) | Infrared thermal imaging behavior intention analysis method and system | |
| CN115994953A (en) | Power field security monitoring and tracking method and system | |
| Chan | A robust target tracking algorithm for FLIR imagery | |
| CN114372966A (en) | Camera damage detection method and system based on average light stream gradient | |
| Zhang et al. | Physical security detectors for critical infrastructures against new-age threat of drones and human intrusion |
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 |