CN112906622A - Method and system for judging platform pedestrian crossing based on linkage of train and crowd position - Google Patents
Method and system for judging platform pedestrian crossing based on linkage of train and crowd position Download PDFInfo
- Publication number
- CN112906622A CN112906622A CN202110264798.0A CN202110264798A CN112906622A CN 112906622 A CN112906622 A CN 112906622A CN 202110264798 A CN202110264798 A CN 202110264798A CN 112906622 A CN112906622 A CN 112906622A
- Authority
- CN
- China
- Prior art keywords
- image
- crowd
- train
- detection
- detected
- 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
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/44—Event detection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Train Traffic Observation, Control, And Security (AREA)
- Image Processing (AREA)
Abstract
The invention relates to a method and a system for judging platform pedestrian crossing based on train and crowd position linkage, which are characterized by comprising the following steps: step 1, acquiring a primary detection image according to a platform warning line and the position of the outer side of a rail; step 2, determining a defense area according to the primary detection image to form a detection image needing to be detected and analyzed; step 3, calibrating the position of the train head and the crowd to serve as a detection target, and training a pre-built convolutional neural network for train head detection and crowd detection; step 4, inputting the detection image into the trained convolutional neural network to obtain the central positions of the train head and the crowd in the detection image; and 5, judging whether the platform pedestrians cross the line or not according to the detected central positions of the train head and the people group, and performing line crossing warning according to a judgment result. The invention can be widely applied to the field of platform pedestrian crossing judgment.
Description
Technical Field
The invention relates to the field of computer vision, in particular to a method and a system for judging platform pedestrian crossing based on linkage of train and crowd positions.
Background
In recent years, with the popularization of high-speed rails, more and more passengers choose to travel on a train. However, with the increase of the passenger volume, especially in the high peak trip period such as spring transportation, the passengers may not follow the waiting rule due to congestion or time, and cross the warning line of the platform, which puts hidden troubles for passenger safety and station protection, and the related safety requirements are increasing.
With the rapid development of the artificial intelligence field, at present, there are algorithms for cross-line detection, and there are two common algorithms:
the first is based on multi-frame motion information, and is judged according to whether an occlusion warning line appears. Firstly, drawing a line on the platform as a warning line, then extracting a moving area, and if a moving target blocks the warning line, indicating that an object crosses the line. This method has the advantage of high speed and the disadvantage that if a pedestrian stands still, even if the pedestrian exceeds the warning line, no alarm can be given, and for the train coming in, a large number of false alarms are generated in the case of the line crossing (the case cannot be given an alarm) when the passengers get off the train.
The second method is based on single frame motion information, and with the development of deep learning, the detection of the pedestrian is more and more accurate, so that the pedestrian is detected by using a convolutional neural network firstly, and then whether the pedestrian crosses the line or not is judged. The method has the advantages that the problem of static line crossing can be solved, but when a large number of passengers exist, the number of the pedestrians is large in shielding, and each pedestrian cannot be accurately detected. The method can not solve the false alarm of train entering and passenger getting off.
Disclosure of Invention
In view of the above problems, the present invention aims to provide a method and a system for determining platform pedestrian crossing based on train and crowd position linkage, which can be applied to a monitoring system of a railway platform to realize automatic monitoring and alarm functions for platform border crossing passengers by preprocessing a platform image and then performing pedestrian platform crossing determination by combining train position and crowd position.
In order to achieve the purpose, the invention adopts the following technical scheme:
the first aspect of the invention provides a method for judging platform pedestrian crossing based on linkage of train and crowd positions, which comprises the following steps:
step 1, acquiring a primary detection image according to a platform warning line and the position of the outer side of a rail;
step 2, determining a defense area according to the preliminary detection image to form an image to be detected;
step 3, calibrating train head positions and crowds in all training image data, using the train head positions and the crowds as detection targets of algorithm training, and training a convolutional neural network which is constructed in advance and used for train head detection and crowd detection;
step 4, inputting the image to be detected into the trained convolutional neural network to obtain the central positions of the train head and the crowd in the image to be detected;
and 5, judging whether the platform pedestrians cross the line or not according to the obtained central positions of the train head and the people group, and performing line crossing warning according to a judgment result.
Further, in step 1, the method for obtaining the preliminary detection image includes the following steps:
step 1.1, drawing two straight lines according to a platform warning line and the position of the outer side of a rail to obtain an original image;
and step 1.2, carrying out anticlockwise rotation on the original image to enable a warning line in the original image to be vertical to the bottom edge of the rotated image, so as to obtain a primary detection image.
Further, in step 1.2, the coordinates of the points in the preliminary detection image are:
x′=xcosθ-ysinθ
y′=xsinθ-ycosθ
where, (x, y) is the coordinates of a point in the original image, (x ', y') is the coordinates of a point in the preliminary detection image, and θ is the rotation angle of the original image.
Further, in the step 2, the method for obtaining the image to be detected comprises: and connecting the upper end and the lower end of the straight line where the rail and the warning line are located from left to right to form a defense area, reserving the image in the defense area, and setting all the pixels of the image outside the defense area to be black to obtain the image to be detected.
Further, in the step 4, a calculation formula of the center position of the train head or the crowd in the image to be detected is as follows:
wherein (x)1,y1),(x2,y2) Coordinate values of the upper left corner and the lower right corner of the train head or the crowd rectangular frame respectively; (x)c,yc) Are coordinate values of the center position.
Further, in the step 5, a method for determining whether the platform pedestrians cross the line according to the detected center positions of the train head and the people group, and performing an off-line warning according to a determination result includes the following steps:
if the train head is not detected in the image to be detected, triggering an alarm when the boundary frame of the crowd exceeds a warning line;
if the train head is detected in the image to be detected, and the position (x) of the center point of the train head is assumed to be in the adjacent detection interval delta tc,yc) If the movement exceeds a preset fixed distance delta s, triggering an alarm when the boundary frame of the crowd exceeds a warning line;
if the train head is detected in the image to be detected, but in the adjacent detection interval delta t, the position (x) of the center point of the train headc,yc) The movement does not exceed a preset fixed distanceFrom Δ s, no alarm needs to be triggered.
In a second aspect of the present invention, there is provided a system for determining platform pedestrian crossing based on train and crowd position linkage, comprising: the preliminary detection image acquisition module is used for acquiring a preliminary detection image according to the station warning line and the position of the outer side of the rail; the detection image acquisition module is used for determining a defense area according to the preliminary detection image and forming an image to be detected; the network training module is used for calibrating train head positions and crowds in all training images, serving as a detection target of algorithm training, and training a pre-built convolutional neural network for train head detection and crowd detection; the crowd and locomotive detection module is used for obtaining the central positions of the train locomotive and the crowd in the image to be detected according to the image to be detected and the trained convolutional neural network; and the line crossing judging and alarming module judges whether the platform pedestrians cross the line according to the obtained central positions of the train head and the crowd and carries out line crossing alarming according to a judgment result.
Further, the preliminary detection image acquisition module comprises an original image acquisition module, and is used for drawing two straight lines according to the station warning line and the position of the outer side of the rail to obtain an original image; and the image rotation module is used for rotating the original image anticlockwise so that a warning line in the image is vertical to the bottom edge of the rotated image to obtain a primary detection image.
Further, the calculation formula of the center position of the train head or the crowd in the image to be detected is as follows:
wherein (x)1,y1),(x2,y2) Coordinate values of the upper left corner and the lower right corner of the train head or the crowd rectangular frame respectively; (x)c,yc) Are coordinate values of the center position.
Further, the line-crossing judging and alarming module comprises: the first judgment module is used for judging whether the train locomotive exists according to the crowd and the data output by the locomotive detection module, if not, the judgment result is sent to the second judgment module, otherwise, the judgment result is sent to the third judgment module; the second judgment module is used for judging whether the boundary frame of the crowd exceeds the warning line or not after receiving the judgment result sent by the first judgment module, and sending a warning signal to the warning module if the boundary frame of the crowd exceeds the warning line; the third judgment module is used for judging whether the position of the center point of the train head in the preset detection interval exceeds a preset fixed distance delta s, if so, sending an alarm signal to the alarm module when the boundary frame of the crowd exceeds a warning line, otherwise, not acting; the alarm module is used for giving an alarm according to the received alarm signal.
Due to the adoption of the technical scheme, the invention has the following advantages: the scheme mainly solves the problems that the detection and the alarm of the line crossing behavior are carried out under the complex environment of the platform, the method solves the problems that a large number of false alarms can be caused when a train enters the station and the target cannot be effectively detected when pedestrians are crowded, and the method has the following characteristics:
1. false detection outside the defense area is prevented. According to the invention, the area outside the defense area is directly blacked and then detected, so that the detection result is more accurate while the detection amount is reduced.
2. The prediction can be still accurate under the crowded condition. The invention provides that people are taken as detection targets, but not one pedestrian, and the judgment is carried out through the central position of the people, so that whether people cross the line or not can be accurately judged even in the peak period of passenger flow.
3. When the train arrives at the station or is ready to start, false detection can not be caused. When the pedestrian crossing detection method is used for judging whether the pedestrian crosses the line or not, whether the train exists or not and whether the train moves or not are fully considered by designing the interaction strategy of the crowd and the train, so that the false detection caused when the crowd gets on or off the train can be avoided.
Therefore, the invention can be widely applied to the field of computer vision.
Drawings
FIGS. 1(a) and 1(b) are an original image and a rotated image in an embodiment of the present invention;
FIG. 2 is a schematic illustration of a defence area in an embodiment of the present invention;
fig. 3 is a schematic diagram of a crowd and train head detection network according to an embodiment of the invention.
Detailed Description
The invention is described in detail below with reference to the figures and examples.
The analysis shows that the prior art has two problems, the first is that when the passenger flow is crowded, whether people cross the line cannot be accurately judged; the second is that when the train is driven in, the passengers can get off normally to cause many false alarms. Therefore, the invention provides a method for judging pedestrian crossing at a platform based on linkage of train and crowd positions, so as to solve the problem of pedestrian crossing detection in complex scenes such as passenger flow congestion, train entering and the like in the conventional crossing detection. Specifically, the method comprises the following steps:
step 1, obtaining a preliminary detection image according to the station warning line and the position of the outer side of the rail.
Specifically, the method for acquiring the preliminary detection image comprises the following steps:
step 1.1, drawing two straight lines according to the station warning line and the position of the outer side of the rail to obtain an original image.
As shown in fig. 1, lines are drawn along the straight line on the platform on which the warning line is located, i.e., L1 marked in fig. 1 (a); the original image was obtained by drawing a line along the line on the outside of the rail and marked L2. L1 and L2 appear to be slightly angled due to the different near and far dimensions of the platform's cameras.
And step 1.2, anticlockwise rotating the original image to enable a warning line (namely L1) in the original image to be vertical to the bottom edge of the rotated image, so as to obtain a primary detection image.
Assuming that the original image needs to be rotated counterclockwise by θ °, for the point (x, y) in the original image, its rotated coordinate is (x ', y'), and the rotation formula is:
x′=xcosθ-ysinθ
y′=xsinθ-ycosθ
and rotating the original image to enable a warning line in the image to be vertical to the bottom edge of the image, so as to counteract the deviation caused by the angle of the platform camera. Since the detected target frame is rectangular, if the guard line is not perpendicular to the bottom side of the image, a case where a part of the target frame is a black background occurs.
And 2, determining a defense area according to the primary detection image to form an image to be detected.
As shown in fig. 2, the upper end and the lower end of the straight line where the rail and the warning line are located are connected from left to right to form a defense area, the image in the defense area is reserved, and all the image pixels outside the defense area are set to be black, so that the image to be detected can be obtained.
And 3, calibrating the train head positions and the crowd in all the training image data, taking the train head positions and the crowd as a detection target of algorithm training, and training a convolutional neural network which is constructed in advance and used for train head detection and crowd detection.
Considering that the crowds at the platform are crowded and the crowds are shielded seriously, and the individual pedestrian calibration is difficult and inaccurate, the invention provides the Crowd Detection (Crowd Detection), wherein the Crowd (Crowd) is defined as a group of people which cannot be divided on the image and can be a person or a plurality of persons. In addition, the invention also marks the train head (Trainhead) as a rectangular area of the front face of the train. All images needing training are calibrated according to the method.
As shown in fig. 3, a convolutional neural network (3 categories including background, train head and crowd) is built, and the crowd and the train head are detected at the same time, and the existing network structures such as SSD, YOLO, fast RCNN, etc. can be used for detection, and the network structures and the related loss functions and training methods thereof are all known to those skilled in the art.
And 4, inputting the image to be detected into the trained convolutional neural network for train head detection and crowd detection to obtain the central position of the train head or the crowd in the image to be detected.
If the crowd or the train head exists in the image to be detected, the position of the train head or the rectangular frame corresponding to the crowd can be obtained through the convolutional neural network, and the coordinate values of the upper left corner and the lower right corner of the rectangular frame are respectively recorded, so that the central position of the train head or the crowd can be obtained. The calculation method of the center position of the rectangular frame is as follows:
wherein (x)1,y1),(x2,y2) Coordinate values of the upper left corner and the lower right corner of the rectangular frame respectively; (x)c,yc) Are coordinate values of the center position.
And 5, judging whether the platform pedestrians cross the line or not according to the obtained center position of the train head or the crowd, and performing line crossing warning according to a judgment result.
The specific judging method comprises the following steps:
if the train head is not detected in the image to be detected, the train head is in a waiting state at present, and once the boundary frame of the crowd exceeds a warning line, the line crossing behavior is shown, and an alarm is triggered;
if the train head is detected in the image to be detected, the position (x) of the center point of the train is assumed in the adjacent detection interval delta tc,yc) If the movement exceeds the preset fixed distance deltas, the train is indicated to be in motion, and at the moment, the train is very dangerous if the train is crossed, so that once the boundary frame of the crowd exceeds the warning line, an alarm is triggered;
if the train head is detected in the image to be detected, but in the adjacent detection interval delta t, the center point position (x) of the trainc,yc) The movement does not exceed the preset fixed distance deltas, which indicates that the train is still, at the same timeWhen the passengers get on or off the bus, the cross-line behavior is normal at this time, and an alarm does not need to be triggered.
The invention also provides a system for judging the line crossing of the platform pedestrians based on the linkage of the train and the crowd position, which comprises the following steps: the preliminary detection image acquisition module is used for acquiring a preliminary detection image according to the station warning line and the position of the outer side of the rail; the detection image acquisition module is used for determining a defense area according to the preliminary detection image and forming an image to be detected; the network training module is used for calibrating train head positions and crowds in all training images, serving as a detection target of algorithm training, and training a pre-built convolutional neural network for train head detection and crowd detection; the crowd and locomotive detection module is used for obtaining the central positions of the train locomotive and the crowd in the image to be detected according to the image to be detected and the trained convolutional neural network; and the line crossing judging and alarming module judges whether the platform pedestrians cross the line according to the obtained central positions of the train head and the crowd and carries out line crossing alarming according to a judgment result.
Further, the preliminary detection image acquisition module comprises an original image acquisition module, and is used for drawing two straight lines according to the station warning line and the position of the outer side of the rail to obtain an original image; and the image rotation module is used for rotating the original image anticlockwise so that a warning line in the image is vertical to the bottom edge of the rotated image to obtain a primary detection image.
Further, in the preliminary detection image, the coordinates of the points are:
x′=xcosθ-ysinθ
y′=xsinθ-ycosθ
where, (x, y) is the coordinates of a point in the original image, (x ', y') is the coordinates of a point in the preliminary detection image, and θ is the rotation angle of the original image.
Further, the calculation formula of the center position of the train head or the crowd in the image to be detected is as follows:
wherein (x)1,y1),(x2,y2) Coordinate values of the upper left corner and the lower right corner of the train head or the crowd rectangular frame respectively; (x)c,yc) Are coordinate values of the center position.
Further, the line-crossing judging and alarming module comprises: the first judgment module is used for judging whether the train locomotive exists according to the crowd and the data output by the locomotive detection module, if not, the judgment result is sent to the second judgment module, otherwise, the judgment result is sent to the third judgment module; the second judgment module is used for judging whether the boundary frame of the crowd exceeds the warning line after receiving the judgment result sent by the first judgment module, and sending a warning signal to the warning module if the boundary frame of the crowd exceeds the warning line; the third judging module is used for judging whether the position of the center point of the train head in the preset detection interval exceeds a preset fixed distance delta s, if so, sending an alarm signal to the alarm module when the boundary frame of the crowd exceeds a warning line, otherwise, not acting; the alarm module is used for giving an alarm according to the received alarm signal.
The above embodiments are only used for illustrating the present invention, and the structure, connection mode, manufacturing process, etc. of the components may be changed, and all equivalent changes and modifications performed on the basis of the technical solution of the present invention should not be excluded from the protection scope of the present invention.
Claims (10)
1. A method for judging platform pedestrian crossing based on train and crowd position linkage is characterized by comprising the following steps:
step 1, acquiring a primary detection image according to a platform warning line and the position of the outer side of a rail;
step 2, determining a defense area according to the preliminary detection image to form an image to be detected;
step 3, calibrating train head positions and crowds in all training image data, using the train head positions and the crowds as detection targets of algorithm training, and training a convolutional neural network which is constructed in advance and used for train head detection and crowd detection;
step 4, inputting the image to be detected into the trained convolutional neural network to obtain the central positions of the train head and the crowd in the image to be detected;
and 5, judging whether the platform pedestrians cross the line or not according to the obtained central positions of the train head and the people group, and performing line crossing warning according to a judgment result.
2. The method of claim 1 for determining platform pedestrian lane crossing based on train and crowd position linkage, wherein: in the step 1, the method for obtaining the preliminary detection image includes the following steps:
step 1.1, drawing two straight lines according to a platform warning line and the position of the outer side of a rail to obtain an original image;
and step 1.2, carrying out anticlockwise rotation on the original image to enable a warning line in the original image to be vertical to the bottom edge of the rotated image, so as to obtain a primary detection image.
3. The method of claim 2 for determining platform pedestrian lane crossing based on train and crowd position linkage, wherein: in step 1.2, the coordinates of the points in the preliminary detection image are:
x′=xcosθ-ysinθ
y′=xsinθ-ycosθ
where, (x, y) is the coordinates of a point in the original image, (x ', y') is the coordinates of a point in the preliminary detection image, and θ is the rotation angle of the original image.
4. The method of claim 1 for determining platform pedestrian lane crossing based on train and crowd position linkage, wherein: in the step 2, the method for obtaining the image to be detected comprises the following steps: and connecting the upper end and the lower end of the straight line where the rail and the warning line are located from left to right to form a defense area, reserving the image in the defense area, and setting all the pixels of the image outside the defense area to be black to obtain the image to be detected.
5. The method of claim 1 for determining platform pedestrian lane crossing based on train and crowd position linkage, wherein: in the step 4, a calculation formula of the center position of the train head or the crowd in the image to be detected is as follows:
wherein (x)1,y1),(x2,y2) Coordinate values of the upper left corner and the lower right corner of the train head or the crowd rectangular frame respectively; (x)c,yc) Are coordinate values of the center position.
6. The method of claim 1 for determining platform pedestrian lane crossing based on train and crowd position linkage, wherein: in the step 5, a method for judging whether the platform pedestrians cross the line according to the detected central positions of the train head and the crowd and performing the warning of crossing the line according to the judgment result comprises the following steps:
if the train head is not detected in the image to be detected, triggering an alarm when the boundary frame of the crowd exceeds a warning line;
if the train head is detected in the image to be detected, and the position (x) of the center point of the train head is assumed to be in the adjacent detection interval delta tc,yc) If the movement exceeds a preset fixed distance delta s, triggering an alarm when the boundary frame of the crowd exceeds a warning line;
if the train head is detected in the image to be detected, but in the adjacent detection interval delta t, the position (x) of the center point of the train headc,yc) Movement does not exceed the predeterminedAnd if the distance is set to be the delta s, an alarm does not need to be triggered.
7. A system for determining platform pedestrian crossing based on train and crowd position linkage suitable for the method according to any one of claims 1 to 6, comprising:
the preliminary detection image acquisition module is used for acquiring a preliminary detection image according to the station warning line and the position of the outer side of the rail;
the detection image acquisition module is used for determining a defense area according to the preliminary detection image and forming an image to be detected;
the network training module is used for calibrating train head positions and crowds in all training images, serving as a detection target of algorithm training, and training a pre-built convolutional neural network for train head detection and crowd detection;
the crowd and locomotive detection module is used for obtaining the central positions of the train locomotive and the crowd in the image to be detected according to the image to be detected and the trained convolutional neural network;
and the line crossing judging and alarming module judges whether the platform pedestrians cross the line according to the obtained central positions of the train head and the crowd and carries out line crossing alarming according to a judgment result.
8. The system for determining platform pedestrian crossing based on train and crowd position linkage according to claim 7, wherein the preliminary detection image obtaining module comprises an original image obtaining module for drawing two straight lines according to the platform warning line and the position of the outer side of the rail to obtain an original image; and the image rotation module is used for rotating the original image anticlockwise so that a warning line in the image is vertical to the bottom edge of the rotated image to obtain a primary detection image.
9. The system for determining platform pedestrian crossing based on train and crowd position linkage as claimed in claim 7, wherein the calculation formula of the center position of the train head or the crowd in the image to be detected is:
wherein (x)1,y1),(x2,y2) Coordinate values of the upper left corner and the lower right corner of the train head or the crowd rectangular frame respectively; (x)c,yc) Are coordinate values of the center position.
10. The system for determining the line crossing of platform pedestrians based on train and crowd position linkage according to claim 7, wherein the line crossing determining and warning module comprises:
the first judgment module is used for judging whether the train locomotive exists according to the crowd and the data output by the locomotive detection module, if not, the judgment result is sent to the second judgment module, otherwise, the judgment result is sent to the third judgment module;
the second judgment module is used for judging whether the boundary frame of the crowd exceeds the warning line or not after receiving the judgment result sent by the first judgment module, and sending a warning signal to the warning module if the boundary frame of the crowd exceeds the warning line;
the third judgment module is used for judging whether the position of the center point of the train head in the preset detection interval exceeds a preset fixed distance delta s, if so, sending an alarm signal to the alarm module when the boundary frame of the crowd exceeds a warning line, otherwise, not acting; the alarm module is used for giving an alarm according to the received alarm signal.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110264798.0A CN112906622A (en) | 2021-03-11 | 2021-03-11 | Method and system for judging platform pedestrian crossing based on linkage of train and crowd position |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110264798.0A CN112906622A (en) | 2021-03-11 | 2021-03-11 | Method and system for judging platform pedestrian crossing based on linkage of train and crowd position |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112906622A true CN112906622A (en) | 2021-06-04 |
Family
ID=76104892
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110264798.0A Pending CN112906622A (en) | 2021-03-11 | 2021-03-11 | Method and system for judging platform pedestrian crossing based on linkage of train and crowd position |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112906622A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114241365A (en) * | 2021-12-01 | 2022-03-25 | 支付宝(杭州)信息技术有限公司 | Dangerous object identification method, device and equipment based on image identification |
-
2021
- 2021-03-11 CN CN202110264798.0A patent/CN112906622A/en active Pending
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114241365A (en) * | 2021-12-01 | 2022-03-25 | 支付宝(杭州)信息技术有限公司 | Dangerous object identification method, device and equipment based on image identification |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106373394B (en) | Vehicle detection method and system based on video and radar | |
JP6549797B2 (en) | Method and system for identifying head of passerby | |
CN102765365B (en) | Pedestrian detection method based on machine vision and pedestrian anti-collision warning system based on machine vision | |
EP3367361B1 (en) | Method, device and system for processing startup of front vehicle | |
CN105844222B (en) | The front vehicles collision warning systems and method of view-based access control model | |
CN106778655B (en) | Human body skeleton-based entrance trailing entry detection method | |
CN105620489A (en) | Driving assistance system and real-time warning and prompting method for vehicle | |
CN108764108A (en) | A kind of Foregut fermenters method based on Bayesian inference | |
CN111198371A (en) | Forward-looking obstacle detection system | |
CN107139666A (en) | Obstacle detouring identifying system and method | |
CN107845264A (en) | A kind of volume of traffic acquisition system and method based on video monitoring | |
CN107886055A (en) | A kind of retrograde detection method judged for direction of vehicle movement | |
CN104103077B (en) | Human head detecting method and human head detecting device | |
CN110717400A (en) | Passenger flow statistical method, device and system | |
WO2023155483A1 (en) | Vehicle type identification method, device, and system | |
Lu et al. | Automated visual surveying of vehicle heights to help measure the risk of overheight collisions using deep learning and view geometry | |
CN111523386B (en) | High-speed railway platform door monitoring and protecting method and system based on machine vision | |
CN114119729A (en) | Obstacle identification method and device | |
CN112906622A (en) | Method and system for judging platform pedestrian crossing based on linkage of train and crowd position | |
CN111062238A (en) | Escalator flow monitoring method and system based on human skeleton information and multi-target tracking | |
JP3750184B2 (en) | Moving object extraction apparatus and extraction method | |
CN117253203A (en) | Obstacle detecting system based on visual sensor | |
CN115440071B (en) | Automatic driving illegal parking detection method | |
WO2022267266A1 (en) | Vehicle control method based on visual recognition, and device | |
CN115984772A (en) | Road ponding detection method and terminal based on video monitoring |
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 |