CN111008545A - Passenger flow detection system and method for rail transit system - Google Patents
Passenger flow detection system and method for rail transit system Download PDFInfo
- Publication number
- CN111008545A CN111008545A CN201811168065.1A CN201811168065A CN111008545A CN 111008545 A CN111008545 A CN 111008545A CN 201811168065 A CN201811168065 A CN 201811168065A CN 111008545 A CN111008545 A CN 111008545A
- Authority
- CN
- China
- Prior art keywords
- passenger flow
- passengers
- unit
- rail transit
- alarm
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 56
- 238000000034 method Methods 0.000 title claims abstract description 32
- 230000003287 optical effect Effects 0.000 claims description 47
- 238000004364 calculation method Methods 0.000 claims description 30
- 238000000926 separation method Methods 0.000 claims description 12
- 239000000203 mixture Substances 0.000 claims description 8
- 238000012544 monitoring process Methods 0.000 abstract description 19
- 238000009826 distribution Methods 0.000 description 20
- 208000012661 Dyskinesia Diseases 0.000 description 15
- 230000002159 abnormal effect Effects 0.000 description 12
- 238000012423 maintenance Methods 0.000 description 11
- 238000012545 processing Methods 0.000 description 8
- 238000005070 sampling Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 238000012546 transfer Methods 0.000 description 5
- 208000027744 congestion Diseases 0.000 description 4
- 230000020169 heat generation Effects 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 230000002035 prolonged effect Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 206010000117 Abnormal behaviour Diseases 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000009877 rendering Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
Images
Classifications
-
- 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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
- Closed-Circuit Television Systems (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a passenger flow detection system and method of a rail transit system, wherein the detection method comprises the following steps: s1, acquiring a video image of passenger flow of the rail transit system; s2, identifying the passenger from the video image; s3, counting the number of passengers in the video image; and S4, when the number of the passengers meets the preset alarm condition, sending an alarm signal. The invention can calculate the passenger flow density in real time according to the acquired video image, and when the passenger flow density is overlarge, an alarm signal is generated, thereby improving the real-time performance and the accuracy of early warning and greatly reducing the workload of monitoring personnel.
Description
Technical Field
The invention belongs to the technical field of passenger flow detection of a rail transit system, and particularly relates to a passenger flow detection system and method of the rail transit system.
Background
The urban rail transit safety problem is of great concern. In order to monitor the safety condition of rail transit, in the prior art, cameras are often arranged on a station hall, a platform and a train of a rail transit system to collect image information in the station hall, the platform and the train, and monitoring personnel manually analyze images collected by the cameras through a display screen to judge whether the passenger flow of the rail transit is abnormal or not. Because the number of the cameras arranged in the subway traffic system is huge, images obtained by a plurality of different cameras can be displayed on a display screen of a monitoring person at the same time, and in order to monitor more scenes, the monitoring person needs to switch monitoring pictures so as to observe more shooting angles of the cameras. Therefore, the workload of monitoring personnel is huge, all monitoring pictures cannot be analyzed and judged at the same time, the real-time passenger flow of the rail transit is difficult to be completely and accurately monitored, the monitoring video is often called to be analyzed after an abnormal event occurs, and effective real-time early warning cannot be given.
Disclosure of Invention
The invention provides a passenger flow detection system and method of a rail transit system, aiming at overcoming the defect that the technical scheme of subway passenger flow real-time detection in the prior art cannot give effective implementation early warning.
The invention solves the technical problems through the following technical scheme:
a method for detecting passenger flow of a rail transit system comprises the following steps:
s1, acquiring a video image of passenger flow of the rail transit system;
s2, identifying the passenger from the video image;
s3, counting the number of passengers in the video image;
and S4, when the number of the passengers meets the preset alarm condition, sending an alarm signal.
Preferably, step S4 includes:
comparing the number of passengers with a preset safety value, and if the number of passengers is greater than the preset safety value, sending an alarm signal.
Preferably, step S4 includes: calculating passenger flow density according to the number of passengers, comparing the passenger flow density with preset safety density, and sending an alarm signal if the passenger flow density is greater than the preset safety density;
the passenger flow density ρ is Q/S, where Q is the number of passengers and S is the area of the region on which the video image is displayed.
Preferably, step S4 includes: when the number of passengers does not accord with the preset alarm condition, executing the following steps:
s41, separating the background and the moving foreground in the video image by adopting a Gaussian mixture background algorithm to obtain the moving foreground;
s42, calculating an optical flow vector of the moving foreground by adopting an optical flow algorithm, and establishing an optical flow amplitude statistical histogram;
s43, calculating the regional entropy according to the optical flow amplitude statistical histogram;
and S44, comparing the regional entropy with a preset safety entropy value, and if the regional entropy is larger than the preset safety entropy value, sending an alarm signal.
Preferably, step S44 further includes:
and sending out an alarm signal corresponding to the degree according to the degree that the regional entropy is greater than the preset safe entropy value.
The invention also provides a passenger flow detection system of the rail transit system, which comprises an image acquisition unit, an image identification unit, a statistical unit and an alarm unit; the image acquisition unit is arranged in the rail transit system;
the image acquisition unit is used for acquiring a video image of passenger flow of the rail transit system;
the image identification unit is used for identifying passengers from the video images;
the counting unit is used for counting the number of passengers in the video image;
the alarm unit is used for sending out an alarm signal when the number of passengers meets the preset alarm condition.
Preferably, the alarm unit is further configured to compare the number of passengers with a preset safety value and to issue an alarm signal when the number of passengers is greater than the preset safety value.
Preferably, the detection system further comprises a passenger flow density calculation unit for calculating the passenger flow density according to the number of passengers;
the alarm unit is also used for comparing the passenger flow density with the preset safety density and sending an alarm signal when the passenger flow density is greater than the preset safety density;
the passenger flow density ρ is Q/S, where Q is the number of passengers and S is the area of the region on which the video image is displayed.
Preferably, the detection system further comprises an image separation unit, an optical flow vector calculation unit, and an area entropy calculation unit;
when the number of passengers does not accord with the preset alarm condition, the image separation unit is used for separating the background and the motion foreground in the video image according to a Gaussian mixture background algorithm to obtain the motion foreground; the optical flow vector calculation unit is used for calculating an optical flow vector of the moving foreground according to an optical flow algorithm and establishing an optical flow amplitude statistical histogram; the regional entropy calculating unit is used for calculating regional entropy according to the optical flow amplitude statistical histogram;
the alarm unit is also used for comparing the regional entropy with a preset safety entropy value and sending an alarm signal when the regional entropy is larger than the preset safety entropy value.
Preferably, the alarm unit is further configured to send an alarm signal corresponding to the degree according to the degree that the area entropy is greater than the preset safe entropy value.
The positive progress effects of the invention are as follows: the invention can calculate the passenger flow density in real time according to the acquired video image, and when the passenger flow density is overlarge, an alarm signal is generated, thereby improving the real-time performance and the accuracy of early warning and greatly reducing the workload of monitoring personnel. The invention can also judge whether the abnormal movement of passenger flow occurs in the rail transit system in real time according to the acquired video image and give corresponding alarm so that rail transit operation and maintenance personnel can respond in time and the situation expansion is effectively avoided.
Drawings
Fig. 1 is a schematic structural diagram of a passenger flow detection system of a rail transit system according to embodiment 1 of the present invention.
Fig. 2 is a schematic structural diagram of a passenger flow detection system of a rail transit system according to embodiment 2 of the present invention.
Fig. 3 is a schematic structural diagram of a passenger flow detection system of a rail transit system according to embodiment 3 of the present invention.
Fig. 4 is a schematic structural diagram of a passenger flow detection system of a rail transit system according to embodiment 4 of the present invention.
Fig. 5 is a schematic structural diagram of a passenger flow detection system of a rail transit system according to embodiment 5 of the present invention.
Fig. 6 is a schematic structural diagram of a passenger flow detection system of a rail transit system according to embodiment 6 of the present invention.
Fig. 7 is a flowchart of a method for detecting passenger flow in a rail transit system according to embodiment 7 of the present invention.
Fig. 8 is a flowchart of a method for detecting passenger flow in a rail transit system according to embodiment 8 of the present invention.
Fig. 9 is a flowchart of a method for detecting passenger flow in a rail transit system according to embodiment 9 of the present invention.
Fig. 10 is a flowchart of a method for detecting passenger flow in a rail transit system according to embodiment 10 of the present invention.
Fig. 11 is a flowchart of a method for detecting passenger flow in a rail transit system according to embodiment 11 of the present invention.
Fig. 12 is a flowchart of a method for detecting passenger flow in a rail transit system according to embodiment 12 of the present invention.
Detailed Description
The invention is further illustrated by the following examples, which are not intended to limit the scope of the invention.
Example 1
The embodiment provides a passenger flow detection system of a rail transit system, and referring to fig. 1, the detection system includes an image acquisition unit 101, an image recognition unit 102, a statistic unit 103, and an alarm unit 104. The image acquisition unit 101 is used for acquiring a video image of passenger flow of the rail transit system. The image recognition unit 102 is used to recognize the passenger from the video image. The counting unit 103 is used for counting the number of passengers in the video image. The alarm unit 104 is used for sending out an alarm signal when the number of passengers meets a preset alarm condition.
In this embodiment, the image capturing unit 101 is a plurality of cameras, which are disposed inside the rail transit system, for example, at a platform, a station hall, an entrance, a doorway, a transfer passage, an escalator, and perform video monitoring on the passenger flow in these areas.
The image recognition unit 102 is implemented by a CPU (central processing unit) (in other alternative embodiments of the present invention, the image recognition unit is implemented by a DSP (digital signal processor), a GPU (graphics processing unit), or the like). And the CPU receives the video images collected by the camera, identifies passengers from each frame of image based on an image identification algorithm, and marks the passengers. Specifically, the CPU identifies the shape of a person in a single frame image by using an image identification algorithm, and the shape of each person represents a passenger. The counting unit 103 counts the number of the marks, that is, the number of passengers in the frame image. The alarm unit 104 compares the number of passengers with a preset safety value, and if the number of passengers is greater than the preset safety value, that is, the number of passengers in the frame image (i.e., the number of passengers in the monitored site) is greater than the preset safety value, it is considered that the density of passengers in the monitored site is too high, and there is a fear of a risk of congestion, and the alarm unit 104 issues an alarm signal. After receiving the alarm signal, the operation and maintenance personnel of the rail transit system dredge the passengers in the monitored place, reduce the passenger flow density and avoid the danger. In view of the different areas of the monitored areas of the cameras arranged at different positions, reasonable preset safety values can be correspondingly set according to the monitored areas.
Because the detecting system of this embodiment can real-time automatic identification and statistics passenger's quantity in the place of monitoring to report to the police when passenger's quantity surpasss and predetermine safe numerical value automatically, its efficiency and rate of accuracy are higher than the manual monitoring far away, have improved the real-time and the accuracy of passenger flow detection, early warning greatly, and saved a large amount of manpower resources. In addition, the detection system of the embodiment can monitor all places of the rail transit system at the same time, and the coverage rate of detection is improved. Furthermore, the detection system can set corresponding preset safety values for different monitoring places, so that the condition of false alarm or missed alarm can be avoided, and the accuracy of detection and early warning is improved.
Example 2
The present embodiment provides a detection system for passenger flow of a rail transit system, which is substantially the same as the detection system of embodiment 1, with reference to fig. 2, except that the detection system of the present embodiment further includes a counting unit 105.
According to the research of the applicant on the passenger flow characteristics in the rail transit system, passengers in the rail transit system are always in a flowing state, so that the number of passengers in a single-frame image is larger than a preset safety value, the passengers are always sporadic and disappear in a short time. Therefore, in the present embodiment, when the number of passengers in a single frame image is greater than a preset safety value, the alarm unit 104 does not immediately issue an alarm signal, but issues a count signal to the counting unit 105 (the counting unit 105 is in a reset state in a default state, and the count value is 0). The count signal is a pulse signal, and the count unit 105 increments the count value by 1 each time it receives the count signal (pulse signal). If the number of passengers in the next frame of image is not greater than the preset safety value, the alarm unit 104 sends a reset signal to the counting unit 105, and the counting unit 105 resets and clears the counting value after receiving the reset signal. If the number of passengers in the next frame of image is larger than the preset safety value, the alarm unit 104 sends a counting signal to the counting unit 105, and the counting unit 105 adds 1 to the counting value after receiving the counting signal. And so on, when the counting value of the counting unit 105 reaches the preset alarm frame number N1, that is, the number of passengers in the consecutive N1 frame images is greater than the preset safety value, the counting unit 105 sends a notification signal to the alarm unit 104, and the alarm unit 104 sends an alarm signal after receiving the notification signal. Therefore, unnecessary alarming can be avoided, and the accuracy of early warning is improved.
In another alternative embodiment of the detection system of this embodiment, the image recognition unit does not perform image recognition on each frame of image acquired by the image acquisition unit, but performs image recognition once every N2 frames (or every N3 seconds). Then, the counting unit counts the number of passengers for the recognition result. The alarm unit sends a counting signal or a reset signal to the counting unit according to the number of passengers. When the counting value of the counting unit reaches a preset alarm value N4, the counting unit sends a notification signal to the alarm unit, and the alarm unit sends an alarm signal after receiving the notification signal. Therefore, the operation load of the image recognition unit can be reduced, the power consumption is reduced, the energy is saved, the heat generation of the image recognition unit is reduced, and the service life of the image recognition unit is prolonged.
The values of N1, N2, N3 and N4 are reasonably set by those skilled in the art according to actual needs. Counting of the N2 frames can be realized by a counter; the timing of N3 seconds can be realized by using a timer, which can be realized by those skilled in the art according to the knowledge grasped by the skilled person, and will not be described herein again.
Example 3
The present embodiment provides a passenger flow detection system of a rail transit system, and referring to fig. 3, the detection system is basically the same as the detection system of embodiment 1, except that the detection system of the present embodiment further includes a timing unit 106.
According to the research of the applicant on the passenger flow characteristics in the rail transit system, passengers in the rail transit system are always in a flowing state, so that the number of passengers in a single-frame image is larger than a preset safety value, the passengers are always sporadic and disappear in a short time. Therefore, in the present embodiment, when the number of passengers in a single frame image is greater than a preset safety value, the alarm unit 104 does not immediately issue an alarm signal, but sets (sets to 1) a timing signal and issues the timing signal to the timing unit 106 (the timing unit 106 is in a reset state in a default state, and the timing value is 0). The timing signal is a level signal, and the timing unit 106 keeps timing during the period when the timing signal is active (set). If the number of passengers in the next frame of image is not greater than the preset safety value, the alarm unit 104 resets (sets to 0) the timing signal, and the timing value of the timing unit 106 is reset and cleared; otherwise, if the number of passengers in the next frame image is still greater than the preset safety value, the timing signal is maintained, and the counting unit 105 continues to count the time. And so on, when the timing value of the timing unit 106 reaches the preset alarm time N5 (unit: second), that is, the number of passengers in the video image is greater than the preset safety value for N5 seconds, the timing unit 106 sends a notification signal to the alarm unit 104, and the alarm unit 104 sends an alarm signal after receiving the notification signal. Therefore, unnecessary alarming can be avoided, and the accuracy of early warning is improved.
In another alternative embodiment of the detection system of this embodiment, the image recognition unit does not perform image recognition on each frame of image acquired by the image acquisition unit, but performs image recognition once every N2 frames (or every N3 seconds). Then, the counting unit counts the number of passengers for the recognition result. The alarm unit sets or resets the timing signal according to the number of passengers. When the timing value of the timing unit reaches the preset alarm time length N6, the timing unit sends a notification signal to the alarm unit, and the alarm unit sends an alarm signal after receiving the notification signal. Therefore, the operation load of the image recognition unit can be reduced, the power consumption is reduced, the energy is saved, the heat generation of the image recognition unit is reduced, and the service life of the image recognition unit is prolonged.
The values of N5, N2, N3 and N6 are reasonably set by those skilled in the art according to actual needs. Counting of the N2 frames can be realized by a counter; the timing of N3 seconds can be realized by using a timer, which can be realized by those skilled in the art according to the knowledge grasped by the skilled person, and will not be described herein again.
Example 4
Referring to fig. 4, the detection system includes an image acquisition unit 101, an image recognition unit 102, a statistic unit 103, a passenger flow density calculation unit 107, and an alarm unit 104. The image acquisition unit 101 is used for acquiring a video image of passenger flow of the rail transit system. The image recognition unit 102 is used to recognize the passenger from the video image. The counting unit 103 is used for counting the number of passengers in the video image. The passenger flow density calculation unit 107 is configured to calculate the passenger flow density based on the number of passengers. The alarm unit 104 is configured to compare the passenger flow density with a preset safe density, and to send an alarm signal when the passenger flow density is greater than the preset safe density. The image capturing unit 101 is a plurality of cameras, and is disposed inside the track transportation system, for example, at a platform, a station hall, an entrance, a doorway, a transfer passage, an escalator, and performs video monitoring on passenger flow in these areas.
The formula used by the passenger flow density calculation unit 107 to calculate the passenger flow density is as follows:
where ρ is the passenger flow density, Q is the number of passengers, and S is the area of the region displayed by the video image (i.e., the area of the location monitored by the camera). In an optional implementation manner of this embodiment, the value of the area S is obtained by measuring a scene monitored by the camera by an operation and maintenance person in rail transit, and is set in the passenger flow density calculation unit 107. In another optional implementation manner of this embodiment, the image recognition unit 102 calculates an area of a location monitored by the camera according to a height at which the camera is set, a shooting angle of the camera, and a size of a sensor of the camera, and the image recognition unit 102 can update a value of the area S in real time when the shooting angle of the camera changes.
Example 5
The present embodiment provides a passenger flow detection system of a rail transit system, and referring to fig. 5, the detection system is basically the same as the detection system of embodiment 4, except that the detection system of the present embodiment further includes a timing unit 106.
In the present embodiment, the passenger flow density calculation unit 107 calculates the passenger flow density of the area monitored by each camera by using the following algorithm:
wherein, Z represents the number (i.e. the number corresponding to the area monitored by one camera) of a station area (e.g. No. 3 line platform, or No. 3 line station hall, or entrance, or transfer passage, or escalator, etc.), and Z represents the field of the number of the station area, which contains the numbers of all station areas, i.e. Z belongs to Z. c is a sampling moment in the monitoring of the camera, and one sampling moment corresponds to one frame of image. C is a field of sampling time, and contains all sampling time monitored by the camera, namely C ∈ C. Rhoz,cThe passenger flow density of the area i at the moment c is in units of person/m2,Qz,cThe passenger flow of the area i at the moment c is in units of people and SzIs the effective area of the region i, and has a unit of m2。
The threshold value is (A)z,Ez),AzA passenger flow density threshold of region i in units of people/m2;EzIs a time threshold, i.e. the density of the passenger flow in the region i exceeds a threshold AzIs in minutes. When the density of the passenger flow in the area z exceeds AzAnd for a duration exceeding EzAt this time, the alarm unit 104 issues an alarm signal. The passenger flow density is calculated by the passenger flow density calculating unit 107, and the duration is counted by the timing unit 106.
Further, in the present embodiment, the case of the excessive passenger flow density is further classified into a plurality of levels according to the size of the passenger flow density. Table 1 shows a manner of classifying the case where the passenger flow density is too large into 3 levels.
TABLE 1
Grade | Situation of excessive passenger flow density |
First stage | ρz,cExceeds a threshold value Az80% of (E), EzOver 30 minutes |
Second stage | ρz,cExceeds a threshold value Az60% of (E), EzOver 10 minutes |
Three-stage | ρz,cExceeds a threshold value Az40% of E, EzOver 5 minutes |
The magnitude of the passenger flow density exceeding the threshold value is calculated by the passenger flow density calculating unit 107, and the duration is timed by the timing unit 106. The alarm unit 104 sends out an alarm signal corresponding to the alarm level according to the divided alarm level to prompt the degree of abnormal passenger flow density of the rail transit operation and maintenance personnel.
When an alarm occurs, the detection system of the embodiment also sends the alarm signal and the image information of the image information at the moment of the occurrence of the abnormal event to the intelligent terminal, so that operation and maintenance personnel in the abnormal area can quickly respond.
Example 6
On the basis of embodiment 1, the present embodiment provides a detection system of passenger flow of a rail transit system, and referring to fig. 6, the detection system further includes an image separation unit 108, an optical flow vector calculation unit 109, and an area entropy calculation unit 110. When the number of passengers does not meet the preset alarm condition, the image separation unit 108 is configured to separate the background from the moving foreground in the video image according to a gaussian mixture background algorithm to obtain the moving foreground; the optical flow vector calculation unit 109 is configured to calculate an optical flow vector of the moving foreground according to an optical flow algorithm, and establish an optical flow amplitude statistical histogram; the regional entropy calculation unit 110 is configured to calculate a regional entropy according to the optical flow amplitude statistical histogram; the alarm unit 104 is further configured to compare the regional entropy with a preset safety entropy value, and to send an alarm signal when the regional entropy is greater than the preset safety entropy value.
The abnormal situation of the passenger flow in the rail transit system may be abnormal movement of the passenger flow, such as running, charging, congestion and the like caused by confusion of passengers due to an emergency, in addition to the excessive density of the passenger flow. In order to identify the abnormal movement of the passenger flow, in this embodiment, the image separation unit 108 first performs image separation, that is, a mixed gaussian background algorithm is applied to separate a background from a movement foreground of the video image of the passenger flow in the monitored area, so as to obtain the movement foreground, where the specific algorithm is as follows:
in the Gaussian mixture background algorithm, the color information among the pixels is considered to be irrelevant, and the processing of each pixel point is independent. For each pixel point in the video image, the change of the value in the sequence image can be regarded as a random process which continuously generates the pixel value, i.e. the color rendering rule of each pixel point is described by Gaussian distribution.
For a multi-peak Gaussian distribution model, each pixel point of an image is modeled according to superposition of a plurality of Gaussian distributions with different weights, each Gaussian distribution corresponds to a state which can possibly generate the color presented by the pixel point, and the weight and distribution parameters of each Gaussian distribution are updated along with time. When processing color images, it is assumed that the image pixels R, G, B have three color (red, green, and blue) channels that are independent of each other and have the same variance. Observation data set { X for video image X1,x2,…xn},xt={rt,gt,btIs the sample of a pixel in the video image at time t, xtThe obeyed mixed gaussian distribution probability density function is:
wherein k is the total number of distribution patterns,is the ith Gaussian distribution at time ti,tIs the mean value ofi,tIs its covariance matrix, δi,tIs variance, I is three-dimensional identity matrix, wi,tThe weight of the ith gaussian distribution at time t.
The image separation unit 108 judges a sample x of each pixeltWhether the following formula is met:
|xt-μi,t-1|≤2.5·δi,t-1
if the pixel sample xtIf the above formula is met, the pixel belongs to the background, otherwise the pixel belongs to the moving foreground.
According to the algorithm, the pixels belonging to the background and the pixels belonging to the moving foreground can be distinguished, and all the pixels belonging to the moving foreground form the moving foreground.
After the background and the moving foreground in the video image are separated to obtain the moving foreground, the optical flow vector calculation unit 109 processes the moving foreground, calculates the optical flow vector of each pixel point in the moving foreground by using an optical flow algorithm, and establishes a statistical histogram. The specific mode is that the degree direction of the moving foreground image is averagely divided into N direction subintervals, the amplitude and the direction data of the moving foreground optical flow vector are obtained, the number of moving foreground optical flow points falling in each direction subinterval is counted, the amplitudes of the points are accumulated and normalized, and finally, the optical flow amplitude statistical histogram of the foreground area image is obtained:
in the formula, mjAnd thetajRespectively representing the magnitude and direction, theta, of the moving foreground optical flow vectorrDenotes the r-th direction sub-interval, nrRepresenting the number of moving foreground stream points in the direction subinterval r.
Next, the area entropy calculation unit 110 calculates area entropy, that is, performs area entropy calculation on the optical flow magnitude statistical histogram obtained in the previous step. The calculation formula of the region entropy U is as follows:
the larger the entropy of the region is found, the more disordered the movement in the region is. When the regional entropy is larger than the preset safe entropy value YUAnd judging that abnormal behaviors occur in the movement foreground, namely abnormal movement of passenger flow occurs. At this time, the region entropy calculation unit 110 outputs a notification signal to the alarm unit 104, and the alarm unit 104 sends a passenger flow abnormal movement alarm signal.
In the embodiment, the safety entropy value Y is also exceeded according to the region entropy UUThe degree of the abnormal movement of the passenger flow is divided into a plurality of levels to represent the severity degree of the abnormal movement of the passenger flow. Table 2 shows the manner of classifying the abnormal movement of the passenger flow into 3 levels.
TABLE 2
Grade | Regional entropy U |
First stage | U exceeds YU30 percent of |
Second stage | U exceeds YU10 to 20 percent of |
Three-stage | U exceeds YU10% of |
The region entropy U exceeds the preset safe entropy value YUThe degree of (b) is calculated by the alarm unit 104 area entropy calculation unit 110. The alarm unit 104 sends out an alarm signal corresponding to the degree according to the degree that the regional entropy is larger than the preset safe entropy value, so as to prompt the severity of the abnormal passenger flow movement of the rail transit operation and maintenance personnel.
In other optional embodiments of the system for detecting passenger flow in a rail transit system according to the present invention, the image separation unit, the optical flow vector calculation unit, and the area entropy calculation unit of this embodiment are respectively combined with the detection systems of embodiments 2 to 5, and detect the movement status of passenger flow when the number of passengers does not meet the preset alarm condition. After reading embodiments 1-6, those skilled in the art can implement that the image separation unit, the optical flow vector calculation unit, and the area entropy calculation unit of this embodiment are respectively combined with the detection systems of embodiments 2-5 to form a detection system, and how to work the detection system is clear, which is not described herein again.
Example 7
The embodiment provides a method for detecting passenger flow of a rail transit system, and referring to fig. 8, the method comprises the following steps:
s1, acquiring a video image of passenger flow of the rail transit system;
step S2, identifying passengers from the video images;
step S3, counting the number of passengers in the video image;
and step S4, when the number of passengers meets the preset alarm condition, sending out an alarm signal.
Step S1 is implemented by a camera. The number of the cameras is a plurality, and the cameras are arranged inside the rail transit system, such as a platform, a station hall, an entrance, a transfer passage, an escalator and the like, and are used for carrying out video monitoring on passenger flow in the areas.
Step S2 is implemented by a CPU (central processing unit) (in other alternative embodiments of the present invention, the image recognition unit is implemented by a DSP (digital signal processor), a GPU (graphics processing unit), or the like). And the CPU receives the video images collected by the camera, identifies passengers from each frame of image based on an image identification algorithm, and marks the passengers. Specifically, the CPU identifies the shape of a person in a single frame image by using an image identification algorithm, and the shape of each person represents a passenger.
Step S3 is implemented using a statistical unit. The counting unit counts the number of the marks, namely the number of the passengers in the frame image.
Step S4 is implemented using an alarm unit. The alarm unit compares the number of passengers with a preset safety value, and if the number of passengers is greater than the preset safety value, that is, the number of passengers in the frame image (i.e., the number of passengers in the monitored place) is greater than the preset safety value, it is considered that the density of passengers in the monitored place is too high, and there is a fear of a risk of congestion, and the alarm unit sends an alarm signal. After receiving the alarm signal, the operation and maintenance personnel of the rail transit system dredge the passengers in the monitored place, reduce the passenger flow density and avoid the danger. In view of the different areas of the monitored areas of the cameras arranged at different positions, reasonable preset safety values can be correspondingly set according to the monitored areas.
Because the detection method of the embodiment can automatically identify and count the number of the passengers in the monitored place in real time and automatically give an alarm when the number of the passengers exceeds the preset safety value, the efficiency and the accuracy are far higher than those of manual monitoring, the real-time performance and the accuracy of passenger flow detection and early warning are greatly improved, and a large amount of human resources are saved. In addition, the detection method of the embodiment can monitor all places of the rail transit system at the same time, and improves the coverage rate of detection. Furthermore, the detection method can set corresponding preset safety values for different monitoring places, can avoid the condition of false alarm or missed alarm, and improves the accuracy of detection and early warning.
Example 8
In this embodiment, a method for detecting passenger flow in a rail transit system is provided, where in addition to embodiment 7, referring to fig. 8, step S4 of the method of this embodiment includes the following steps:
step S401, comparing the number of passengers with a preset safety value, and judging whether the number of passengers is greater than the preset safety value, if so, executing step S402; if not, step S403 is executed.
Step S402, adding 1 to the count value, judging whether the count value reaches the preset alarm frame number, if so, executing step S404; if not, step S401 is executed.
In step S403, the count value is cleared, and the process advances to step S401.
And step S404, alarming.
According to the research of the applicant on the passenger flow characteristics in the rail transit system, passengers in the rail transit system are always in a flowing state, so that the number of passengers in a single-frame image is larger than a preset safety value, the passengers are always sporadic and disappear in a short time. Therefore, in the present embodiment, when the number of passengers in a single frame image is greater than the preset safety value, an alarm is not immediately issued, but counting is performed, that is, the number of frames in which the number of passengers continuously occurring is greater than the preset safety value is counted, and when the number of frames (i.e., the count value) reaches the preset alarm number of frames N1, an alarm is issued again. Therefore, unnecessary alarming can be avoided, and the accuracy of early warning is improved.
In other alternative embodiments of the detection method of this embodiment, instead of performing image recognition on each acquired frame of image, image recognition is performed every N2 frames (or every N3 seconds). Then, the number of passengers is counted according to the recognition result. Next, the count value is increased by 1 or cleared according to the number of passengers. When the count value reaches a preset alarm value N4, an alarm signal is issued. Therefore, the operation load of the image recognition unit can be reduced, the power consumption is reduced, the energy is saved, the heat generation of the image recognition unit is reduced, and the service life of the image recognition unit is prolonged.
Example 9
In this embodiment, a method for detecting passenger flow in a rail transit system is provided, where in addition to embodiment 7, referring to fig. 9, step S4 of the method of this embodiment includes the following steps:
step S411, comparing the number of passengers with a preset safety value, and judging whether the number of passengers is greater than the preset safety value, if so, executing step S412; if not, step S413 is executed.
Step S412, continuously timing, and judging whether the timing value reaches the preset alarm time, if so, executing step S414; if not, step S411 is executed.
Step S413, the timer value is cleared, and the process proceeds to step S411.
And step S414, alarming.
According to the research of the applicant on the passenger flow characteristics in the rail transit system, passengers in the rail transit system are always in a flowing state, so that the number of passengers in a single-frame image is larger than a preset safety value, the passengers are always sporadic and disappear in a short time. Therefore, in the present embodiment, when the number of passengers in a single frame image is greater than a preset safety value, an alarm is not immediately issued, but a timer is counted, that is, a time during which a situation in which the number of passengers is greater than the preset safety value continuously occurs is counted, and when the time (i.e., a timer value) reaches a preset alarm time N5 (unit: second), an alarm is issued again. Therefore, unnecessary alarming can be avoided, and the accuracy of early warning is improved.
In other alternative embodiments of the detection method of this embodiment, instead of performing image recognition on each acquired frame of image, image recognition is performed every N2 frames (or every N3 seconds). Then, the number of passengers is counted according to the recognition result. Then, continuous timing is carried out or a timing value is cleared according to the number of passengers. When the timing value reaches the preset alarm time length N6 (unit: second), an alarm signal is sent out. Therefore, the operation load of the image recognition unit can be reduced, the power consumption is reduced, the energy is saved, the heat generation of the image recognition unit is reduced, and the service life of the image recognition unit is prolonged.
Example 10
The present embodiment provides a method for detecting passenger flow in a rail transit system, which is different from the detection method of embodiment 7 in that, in the present embodiment, step S4 includes the following steps:
s431, calculating passenger flow density;
s432, comparing the passenger flow density with a preset safety density, and if the passenger flow density is greater than the preset safety density, sending an alarm signal.
The formula used to calculate the passenger flow density in step S431 is:
where ρ is the passenger flow density, Q is the number of passengers, and S is the area of the region displayed by the video image (i.e., the area of the location monitored by the camera). In an optional implementation manner of this embodiment, the value of the area S is obtained by measuring a scene monitored by the camera by operation and maintenance staff of the rail transit. In another optional implementation manner of this embodiment, the image recognition unit calculates the area of the location monitored by the camera according to the height at which the camera is set, the shooting angle of the camera, and the size of the sensor of the camera, and when the shooting angle of the camera changes, the image recognition unit can update the value of the area S in real time.
Example 11
The present embodiment provides a method for detecting passenger flow in a rail transit system, which is substantially the same as the detection method in embodiment 10, except that, referring to fig. 11, in the present embodiment, step S4 includes the following steps:
step S441, calculating passenger flow density;
step S442, comparing the passenger flow density with a preset safety density, and judging whether the passenger flow density is greater than a passenger flow density threshold value, if so, executing step S444; if not, go to step S443;
step S443, clearing the timer value, and proceeding to step S441;
step S444, continuing to count time, and judging whether the timing value reaches a time threshold value, if so, executing step S445; if not, step S441 is executed.
And step S445, alarming.
In step S441, the following algorithm is used to calculate the passenger flow density of the area monitored by each camera:
wherein, Z represents the number (i.e. the number corresponding to the area monitored by one camera) of a station area (e.g. No. 3 line platform, or No. 3 line station hall, or entrance, or transfer passage, or escalator, etc.), and Z represents the field of the number of the station area, which contains the numbers of all station areas, i.e. Z belongs to Z. c is a sampling moment in the monitoring of the camera, and one sampling moment corresponds to one frame of image. C is a field of sampling time, and contains all sampling time monitored by the camera, namely C ∈ C. Rhoz,cThe passenger flow density of the area i at the moment c is in units of person/m2,Qz,cThe passenger flow of the area i at the moment c is in units of people and SzIs the effective area of the region i, and has a unit of m2。
The threshold value is (A)z,Ez),AzA passenger flow density threshold of region i in units of people/m2;EzIs a time threshold, i.e. the density of the passenger flow in the region i exceeds a threshold AzIs in minutes. When the density of the passenger flow in the area z exceeds AzAnd for a duration exceeding EzAnd the alarm unit sends out an alarm signal.
Further, in the present embodiment, the case of the excessive passenger flow density is further classified into a plurality of levels according to the size of the passenger flow density. Table 3 shows a manner of classifying the case where the passenger flow density is too high into 3 levels.
TABLE 3
Grade | Situation of excessive passenger flow density |
First stage | ρz,cExceeds a threshold value Az80% of (E), EzOver 30 minutes |
Second stage | ρz,cExceeds a threshold value Az60% of (E), EzOver 10 minutes |
Three-stage | ρz,cExceeds a threshold value Az40% of E, EzOver 5 minutes |
In step S445, an alarm signal corresponding to the alarm level is sent out according to the divided alarm level, so as to prompt the degree of abnormal passenger flow density of the rail transit operation and maintenance personnel.
In step S445, the alarm signal and the image information of the image information at the time of the occurrence of the abnormal event are sent to the intelligent terminal, so that the operation and maintenance personnel in the abnormal area can quickly respond.
Example 12
In addition to embodiment 7, the present embodiment provides a method for detecting passenger flow in a rail transit system, and with reference to fig. 12, in the detecting method, step S4 includes:
judging whether the number of passengers meets the preset alarm condition, if so, executing the step S425 and giving an alarm; if not, executing the following steps:
s421, separating the background and the motion foreground in the video image by adopting a Gaussian mixture background algorithm to obtain the motion foreground;
s422, calculating an optical flow vector of the moving foreground by adopting an optical flow algorithm, and establishing an optical flow amplitude statistical histogram;
s423, calculating the regional entropy according to the optical flow amplitude statistical histogram;
and S424, comparing the regional entropy with a preset safety entropy value, and if the regional entropy is larger than the preset safety entropy value, sending an alarm signal.
The abnormal situation of the passenger flow in the rail transit system may be abnormal movement of the passenger flow, such as running, charging, congestion and the like caused by confusion of passengers due to an emergency, in addition to the excessive density of the passenger flow. In order to identify the abnormal movement of the passenger flow, step S421 is executed first, and a mixed gaussian background algorithm is applied to separate the background from the moving foreground of the video image of the passenger flow in the monitored area, so as to obtain the moving foreground, where the specific algorithm is as follows:
in the Gaussian mixture background algorithm, the color information among the pixels is considered to be irrelevant, and the processing of each pixel point is independent. For each pixel point in the video image, the change of the value in the sequence image can be regarded as a random process which continuously generates the pixel value, i.e. the color rendering rule of each pixel point is described by Gaussian distribution.
For a multi-peak Gaussian distribution model, each pixel point of an image is modeled according to superposition of a plurality of Gaussian distributions with different weights, each Gaussian distribution corresponds to a state which can possibly generate the color presented by the pixel point, and the weight and distribution parameters of each Gaussian distribution are updated along with time. When processing color images, it is assumed that the image pixels R, G, B have three color (red, green, and blue) channels that are independent of each other and have the same variance. Observation data set { X for video image X1,x2,…xn},xt={rt,gt,btIs the sample of a pixel in the video image at time t, xtThe obeyed mixed gaussian distribution probability density function is:
wherein k is the total number of distribution patterns,is the ith Gaussian distribution at time ti,tIs the mean value ofi,tIs its covariance matrix, δi,tIs variance, I is three-dimensional identity matrix, wi,tThe weight of the ith gaussian distribution at time t.
The image separation unit judges a sample x of each pixeltWhether the following formula is met:
|xt-μi,t-1|≤2.5·δi,t-1
if the pixel sample xtIf the above formula is met, the pixel belongs to the background; otherwise, the pixel belongs to the motion foreground.
According to the algorithm, the pixels belonging to the background and the pixels belonging to the moving foreground can be distinguished, and all the pixels belonging to the moving foreground form the moving foreground.
After the background and the moving foreground in the video image are separated to obtain the moving foreground, in step S422, the moving foreground is processed, and an optical flow algorithm is used to calculate an optical flow vector of each pixel point in the moving foreground and establish a statistical histogram. The specific mode is that the degree direction of the moving foreground image is averagely divided into N direction subintervals, the amplitude and the direction data of the moving foreground optical flow vector are obtained, the number of moving foreground optical flow points falling in each direction subinterval is counted, the amplitudes of the points are accumulated and normalized, and finally, the optical flow amplitude statistical histogram of the foreground area image is obtained:
in the formula, mjAnd thetajRespectively representing the magnitude and direction, theta, of the moving foreground optical flow vectorrDenotes the r-th direction sub-interval, nrRepresenting the number of moving foreground stream points in the direction subinterval r.
Next, in step S423, the area entropy is calculated, that is, the area entropy calculation is performed on the optical flow magnitude statistical histogram obtained in step S422. The calculation formula of the region entropy U is as follows:
in step S424, the region entropy is compared with a preset safe entropy value. The larger the entropy of the region is found, the more disordered the movement in the region is. When the regional entropy is larger than the preset safe entropy value YUAnd judging that abnormal behaviors occur in the movement foreground, namely abnormal movement of passenger flow occurs. At this time, a passenger flow abnormal motion alarm signal is sent out.
In the embodiment, the safety entropy value Y is also exceeded according to the region entropy UUThe degree of the abnormal movement of the passenger flow is divided into a plurality of levels to represent the severity degree of the abnormal movement of the passenger flow. Table 4 shows the manner of classifying the abnormal movement of the passenger flow into 3 levels.
TABLE 4
Grade | Regional entropy U |
First stage | U exceeds YU30 percent of |
Second stage | U exceeds YU10 to 20 percent of |
Three-stage | U exceeds YU10% of |
In step S424, according to the degree that the regional entropy is greater than the preset safety entropy, an alarm signal corresponding to the degree is sent out to prompt the severity of the abnormal movement of the passenger flow of the rail transit operation and maintenance personnel.
While specific embodiments of the invention have been described above, it will be appreciated by those skilled in the art that these are by way of example only, and that the scope of the invention is defined by the appended claims. Various changes and modifications to these embodiments may be made by those skilled in the art without departing from the spirit and scope of the invention, and these changes and modifications are within the scope of the invention.
Claims (10)
1. A method for detecting passenger flow of a rail transit system is characterized by comprising the following steps:
s1, acquiring a video image of passenger flow of the rail transit system;
s2, identifying passengers from the video images;
s3, counting the number of passengers in the video image;
and S4, when the number of the passengers meets the preset alarm condition, sending an alarm signal.
2. The method for detecting passenger flow in a rail transit system as claimed in claim 1, wherein the step S4 comprises:
and comparing the number of the passengers with a preset safety value, and if the number of the passengers is greater than the preset safety value, sending an alarm signal.
3. The method for detecting passenger flow in a rail transit system as claimed in claim 1, wherein the step S4 comprises: calculating passenger flow density according to the number of passengers, comparing the passenger flow density with preset safety density, and sending an alarm signal if the passenger flow density is greater than the preset safety density;
and the passenger flow density rho is Q/S, wherein Q is the number of passengers, and S is the area of the area displayed by the video image.
4. The method for detecting passenger flow in a rail transit system as claimed in claim 1, wherein the step S4 comprises: when the number of passengers does not accord with the preset alarm condition, executing the following steps:
s41, separating the background and the moving foreground in the video image by adopting a Gaussian mixture background algorithm to obtain the moving foreground;
s42, calculating an optical flow vector of the moving foreground by adopting an optical flow algorithm, and establishing an optical flow amplitude statistical histogram;
s43, calculating an area entropy according to the optical flow amplitude statistical histogram;
s44, comparing the regional entropy with a preset safety entropy value, and if the regional entropy is larger than the preset safety entropy value, sending an alarm signal.
5. The method for detecting passenger flow in a rail transit system as claimed in claim 4, wherein said step S44 further comprises:
and sending out an alarm signal corresponding to the degree according to the degree that the regional entropy is greater than the preset safety entropy value.
6. A passenger flow detection system of a rail transit system is characterized by comprising an image acquisition unit, an image identification unit, a statistical unit and an alarm unit; the image acquisition unit is arranged in the rail transit system;
the image acquisition unit is used for acquiring a video image of passenger flow of the rail transit system;
the image identification unit is used for identifying passengers from the video images;
the counting unit is used for counting the number of passengers in the video image;
the alarm unit is used for sending out an alarm signal when the number of the passengers meets a preset alarm condition.
7. The system for detecting passenger flow in a rail transit system as claimed in claim 6, wherein said alarm unit is further adapted to compare the number of passengers with a preset safety value and to issue an alarm signal if the number of passengers is greater than the preset safety value.
8. The system of claim 6, further comprising a passenger density calculation unit for calculating a passenger density based on the number of passengers;
the alarm unit is also used for comparing the passenger flow density with a preset safety density and sending an alarm signal when the passenger flow density is greater than the preset safety density;
and the passenger flow density rho is Q/S, wherein Q is the number of passengers, and S is the area of the area displayed by the video image.
9. The detection system of passenger flow of a rail transit system according to claim 6, further comprising an image separation unit, an optical flow vector calculation unit, an area entropy calculation unit;
when the number of the passengers does not accord with a preset alarm condition, the image separation unit is used for separating a background from a moving foreground in the video image according to a Gaussian mixture background algorithm to obtain the moving foreground; the optical flow vector calculation unit is used for calculating an optical flow vector of the moving foreground according to an optical flow algorithm and establishing an optical flow amplitude statistical histogram; the area entropy calculation unit is used for calculating area entropy according to the optical flow amplitude statistical histogram;
the alarm unit is further used for comparing the region entropy with a preset safety entropy value and sending an alarm signal when the region entropy is larger than the preset safety entropy value.
10. The system of claim 9, wherein the alarm unit is further configured to issue an alarm signal corresponding to a degree to which the regional entropy is greater than the preset safety entropy value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811168065.1A CN111008545A (en) | 2018-10-08 | 2018-10-08 | Passenger flow detection system and method for rail transit system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811168065.1A CN111008545A (en) | 2018-10-08 | 2018-10-08 | Passenger flow detection system and method for rail transit system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111008545A true CN111008545A (en) | 2020-04-14 |
Family
ID=70111270
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811168065.1A Pending CN111008545A (en) | 2018-10-08 | 2018-10-08 | Passenger flow detection system and method for rail transit system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111008545A (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2012008176A1 (en) * | 2010-07-12 | 2012-01-19 | 株式会社日立国際電気 | Monitoring system and method of monitoring |
CN106651730A (en) * | 2016-12-22 | 2017-05-10 | 东南大学 | System and method for metro passenger flow early warning based on passenger flow density and gate pass time |
CN106778632A (en) * | 2016-12-22 | 2017-05-31 | 东南大学 | Track traffic large passenger flow recognizes early warning system and method |
-
2018
- 2018-10-08 CN CN201811168065.1A patent/CN111008545A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2012008176A1 (en) * | 2010-07-12 | 2012-01-19 | 株式会社日立国際電気 | Monitoring system and method of monitoring |
CN106651730A (en) * | 2016-12-22 | 2017-05-10 | 东南大学 | System and method for metro passenger flow early warning based on passenger flow density and gate pass time |
CN106778632A (en) * | 2016-12-22 | 2017-05-31 | 东南大学 | Track traffic large passenger flow recognizes early warning system and method |
Non-Patent Citations (1)
Title |
---|
姜彦;张文韬;: "西安地铁客流监控预警信息***可行性研究及分析" * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2022011828A1 (en) | System and method for detecting object that gets in and out of elevator, object detection system, elevator light curtain, and elevator device | |
CN101799985B (en) | Highway tunnel traffic identification method | |
CN108154686A (en) | A kind of vehicle-mounted act of violating regulations capturing system | |
CN111967400A (en) | Campus safety monitoring and early warning management system based on artificial intelligence | |
CN111275910B (en) | Method and system for detecting border crossing behavior of escalator based on Gaussian mixture model | |
CN104581081A (en) | Passenger-flow analysis method based on video information | |
CN109146914B (en) | Drunk driving behavior early warning method for expressway based on video analysis | |
CN107543548B (en) | Activity personnel positioning device and activity data acquisition equipment | |
CN109867186B (en) | Elevator trapping detection method and system based on intelligent video analysis technology | |
CN111924695A (en) | Intelligent safety protection system for subway escalator and working method of intelligent safety protection system | |
CN110481606B (en) | Subway passenger flow guiding system and method based on visual identification technology | |
CN111382719A (en) | Vertical ladder personnel wearing mask monitoring method for epidemic prevention and control | |
CN110809137A (en) | Campus safety trampling prevention monitoring system and method | |
CN113112241A (en) | Fire-fighting data platform establishing and managing method based on data information feature recognition and deep learning | |
CN117576946B (en) | Intelligent parking safety supervision method and system | |
CN111062238B (en) | Escalator flow monitoring method and system based on human skeleton information and multi-target tracking | |
CN115909223A (en) | Method and system for matching WIM system information with monitoring video data | |
KR101219407B1 (en) | escalator monitoring system using image processing technology | |
CN113158724A (en) | Monitoring method and system | |
CN111008545A (en) | Passenger flow detection system and method for rail transit system | |
CN104616432B (en) | A kind of density of stream of people Intelligent Recognition and control method and system | |
CN113822151A (en) | Method and system for monitoring population density within a market | |
CN114724348B (en) | Deep learning-based rescue alarm method | |
CN115457331A (en) | Intelligent inspection method and system for construction site | |
CN111611904B (en) | Dynamic target identification method based on unmanned vehicle driving process |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200414 |
|
WD01 | Invention patent application deemed withdrawn after publication |