CN112489496A - Vehicle safety early warning method and device based on video identification - Google Patents

Vehicle safety early warning method and device based on video identification Download PDF

Info

Publication number
CN112489496A
CN112489496A CN202011559995.7A CN202011559995A CN112489496A CN 112489496 A CN112489496 A CN 112489496A CN 202011559995 A CN202011559995 A CN 202011559995A CN 112489496 A CN112489496 A CN 112489496A
Authority
CN
China
Prior art keywords
vehicle
cartesian
vector
front vehicle
self
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
Application number
CN202011559995.7A
Other languages
Chinese (zh)
Inventor
戴齐飞
俞正中
李福池
囊宗进
李丹
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Apical Technology Co ltd
Original Assignee
Shenzhen Apical Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shenzhen Apical Technology Co ltd filed Critical Shenzhen Apical Technology Co ltd
Priority to CN202011559995.7A priority Critical patent/CN112489496A/en
Publication of CN112489496A publication Critical patent/CN112489496A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G08G1/166—Anti-collision systems for active traffic, e.g. moving vehicles, pedestrians, bikes
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • G06T17/05—Geographic models
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
    • G08B21/18—Status alarms
    • G08B21/24—Reminder alarms, e.g. anti-loss alarms
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/09—Arrangements for giving variable traffic instructions
    • G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
    • G08G1/0965—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages responding to signals from another vehicle, e.g. emergency vehicle
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G08G1/161—Decentralised systems, e.g. inter-vehicle communication
    • G08G1/162—Decentralised systems, e.g. inter-vehicle communication event-triggered

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Geometry (AREA)
  • Emergency Management (AREA)
  • Business, Economics & Management (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Remote Sensing (AREA)
  • Computer Graphics (AREA)
  • Traffic Control Systems (AREA)
  • Closed-Circuit Television Systems (AREA)

Abstract

The invention discloses a video identification-based vehicle safety early warning method, which comprises the following steps: establishing a Cartesian full vector constraint outside a front vehicle, a Cartesian vector inside a self vehicle and a pixel Cartesian vector; acquiring a three-dimensional model of a laser image of a front vehicle at the current moment for processing; monitoring the running condition of the vehicle; establishing a projection relation between the Cartesian full vector constraint outside the front vehicle and the Cartesian local vector constraint outside the vehicle through a three-dimensional model of continuous N frames of laser images; constructing the position motion constraint of the same front vehicle by using three-dimensional models of continuous N frames of laser images; determining the transverse and longitudinal relative position relation between the front vehicle and the self vehicle according to the external Cartesian full vector constraint of the front vehicle at the current moment; and calculating the safety distance between the front vehicle and the self vehicle and carrying out safety early warning. Compared with the prior art, the invention can realize accurate distance measurement and real-time early warning, and prevent collision between the front vehicle and the self vehicle.

Description

Vehicle safety early warning method and device based on video identification
Technical Field
The invention relates to the technical field of traffic safety, in particular to a vehicle safety early warning method and device based on video identification.
Background
With the development of technological progress, automobiles are convenient to use for people to go out, the number of the automobiles is more and more at present, the main reasons for causing traffic accidents are drunk driving, overspeed driving and fatigue driving, the traffic accidents mainly show that the automobiles collide with the back and forth, the traffic accidents not only cause economic loss to people, but also influence the life safety of people in serious cases, and in the prior art, whether the distance from the front vehicle to the front vehicle is less than the safe distance or not is measured through a sensor. Therefore, it is important to identify the vehicle rear-end collision prevention early warning based on the video and reduce the occurrence of the traffic rear-end collision as much as possible.
Disclosure of Invention
Aiming at the problems, the invention provides an anti-collision video-recognition-based vehicle safety early warning method and device
In order to solve the above problems, a first aspect of the present invention discloses a video recognition-based vehicle safety warning method, which includes the following steps:
s1, establishing a front vehicle external Cartesian full vector, a self vehicle internal Cartesian vector and a pixel Cartesian vector;
s2, acquiring a three-dimensional model of a laser image of a vehicle ahead at the current moment for processing; monitoring the running condition of the vehicle;
s3, establishing a projection relation between the Cartesian full vector constraint outside the front vehicle and the Cartesian local vector constraint outside the vehicle through a three-dimensional model of continuous N frames of laser images;
s4, constructing the same front vehicle position motion constraint by using the three-dimensional model of the continuous N frames of laser images;
s5, determining the transverse and longitudinal relative position relation between the front vehicle and the self vehicle according to the external Cartesian full vector constraint of the front vehicle at the current time;
and S6, calculating the safety distance between the front vehicle and the self vehicle and carrying out safety early warning.
As a further scheme of the present invention, the specific method for establishing the external cartesian full vector constraint of the front vehicle in step S1 is as follows:
establishing a front vehicle external Cartesian full vector V-XVYvZv by taking a middle grounding point at the tail of a front vehicle as an origin V, taking the front side of the front vehicle as the positive direction of an X axis, taking the right side of the front vehicle as the positive direction of a Y axis and taking the front side of the front vehicle as the positive direction of a Z axis, and setting the coordinates of 8 angular points constrained by the front vehicle external Cartesian full vector as the respective coordinates;
Figure 162790DEST_PATH_IMAGE001
，
Figure 557999DEST_PATH_IMAGE002
，
Figure 44475DEST_PATH_IMAGE003
，
Figure 132517DEST_PATH_IMAGE004
，
Figure 399550DEST_PATH_IMAGE005
，
Figure 598450DEST_PATH_IMAGE006
,
Figure 939433DEST_PATH_IMAGE007
，
Figure 198376DEST_PATH_IMAGE008
；
wherein f isx ，fy ，fzThe length of the Cartesian full vector outside the vehicle along the directions of an Xt axis, a Yt axis and a Zt axis is constrained;
the specific method for establishing the cartesian vector in the vehicle of the own vehicle in the step S1 is as follows:
taking a camera at the top of the self-vehicle as an original point C, taking the front side of the self-vehicle as an Xc axial positive direction, taking the left side of the self-vehicle as a Y axial positive direction and taking the upward direction as a Z axial positive direction; establishing a Cartesian vector C-X in a vehicle fixed to the host vehiclecYcZc; the vector coordinate of the Cartesian vector of the central grounding point V of the tail of the front vehicle in the Cartesian full-vector constraint of the front vehicle is set as V = [ V =x，Vy，0]v(ii) a The position of a Cartesian vector constrained in the self vehicle by a front vehicle external Cartesian full vector is D = [ theta, 0 =]Where θ is the front vehicle YVAxle of bicycle YcThe included angle of the axes;
the specific method for establishing the pixel cartesian vector in step S1 is as follows:
upper left corner O for collecting laser image by camerapA Cartesian vector O of pixels is established as an origin, an I axis forward direction is towards the right, and a J axis forward direction is towards the bottomp-IJ。
As a further aspect of the present invention, step S2 specifically includes acquiring a three-dimensional model of a laser image on the rear side of the vehicle at the current time point by using a camera as a current frame laser three-dimensional model, and processing the three-dimensional model of the laser image acquired at the current time point by using a recurrent global neural network to obtain a cartesian full-vector constrained dimension V outside the vehicle aheadx，Vy，Vz(ii) a Detecting and acquiring the speed of a Cartesian vector in a front vehicle at the current moment through a laser radar; and measuring the position, the speed and the deflection angle of the self-vehicle at the current moment through the Beidou system and the fiber-optic gyroscope.
As the inventionFurther, step S3 specifically includes: external Cartesian coordinate system X of camera for acquiring any point of front vehicle0＝[X ，Y ，Z]VCorresponding point X in pixel coordinate system0=[i ，j]VThe following transformation relationships are satisfied:
Figure 762825DEST_PATH_IMAGE009
(1)
k is an internal parameter matrix of the camera; d' is a rotation matrix from the Cartesian vector outside the front vehicle to the Cartesian vector outside the self-vehicle 3 x 3, and T is a translation matrix of 3 x 1;
Figure 499837DEST_PATH_IMAGE010
。
as a further aspect of the present invention, step S4 specifically includes: setting the vector coordinates of the rear center grounding point of the front vehicle in the three-dimensional model of two continuous frames of laser images in the Cartesian vector of the self vehicle as follows:
Figure 23222DEST_PATH_IMAGE011
，
Figure 390749DEST_PATH_IMAGE012
wherein tn-1, tn is the acquisition time of two continuous frames of images; for the coordinate position of the central grounding point of the tail part of the front vehicle, 2(N-1) motion constraints are built in total according to the three-dimensional model of the current laser image and the three-dimensional model of the front N-1 frames of laser images:
Figure 366796DEST_PATH_IMAGE013
(2)
Figure 907498DEST_PATH_IMAGE014
(3)
wherein u isxn ，uynCartesian in the vehicle of the vehicle at the moment corresponding to the three-dimensional model of the Nth frame of laser image acquired by a laser radar mounted on the vehicleThe speed of the vehicle ahead in the vector in the X-axis and Y-axis directions;
and (3) obtaining the position tn and the direction V of the vehicle ahead at the current time in the Cartesian vector of the vehicle by using a least square method through simultaneous formulas (1) to (3), namely determining the three-dimensional model of the vehicle ahead at the current time.
As a further aspect of the present invention, step S5 specifically includes: drawing curves along the first corner point and the second corner point X1 and X2 of the three-dimensional model of the front vehicle, wherein the curve equation is as follows:
Figure 550969DEST_PATH_IMAGE015
，
wherein
Figure 823819DEST_PATH_IMAGE016
，
Figure 552740DEST_PATH_IMAGE017
Wherein
Figure 631555DEST_PATH_IMAGE018
The difference value of the Cartesian full vectors in the X-axis direction of two continuous frames of laser images is obtained.
Figure 67215DEST_PATH_IMAGE019
The difference value of the Cartesian full vectors in the Y-axis direction of two continuous frames of laser images is obtained. X is the transverse coordinate position of the front vehicle relative to the self vehicle, and y is the longitudinal coordinate position of the front vehicle relative to the self vehicle.
As a further aspect of the present invention, step S6 specifically includes: judging whether the front vehicle is in a potential danger or not according to the transverse relative position relation of the front vehicle and the self vehicle, and executing a step S6 if the front vehicle is in the potential danger; otherwise, the process returns to step S2.
In another aspect of the present invention, a vehicle safety pre-warning device based on video recognition comprises:
establishing a vector coordinate module: the vector coordinate establishing module is used for establishing a Cartesian full vector outside the front vehicle, a Cartesian vector inside the self-vehicle and a pixel Cartesian vector;
the data processing module is used for acquiring a three-dimensional model of a laser image of a vehicle in front at the current moment and processing the three-dimensional model; monitoring the running condition of the vehicle;
the projection module is used for establishing a projection relation between the Cartesian full vector constraint outside the front vehicle and the Cartesian local vector constraint outside the vehicle through a three-dimensional model of continuous N frames of laser images;
the position constraint module is used for constructing the position motion constraint of the same front vehicle by utilizing the three-dimensional model of the continuous N frames of laser images;
the position determining module is used for determining the transverse and longitudinal relative position relation between the front vehicle and the self vehicle according to the external Cartesian full vector constraint of the front vehicle at the current moment;
and the safety early warning module is used for calculating the safety distance between the front vehicle and the self vehicle and carrying out safety early warning.
Compared with the prior art, the invention has the beneficial effects that: the invention can realize the accurate measurement of the transverse and longitudinal distances between the front vehicle and the self vehicle, and can realize real-time monitoring and timely safety early warning prompt to prevent collision between the front vehicle and the self vehicle.
Drawings
FIG. 1 is a flowchart of a vehicle safety warning method based on video identification according to an embodiment of the present invention
Fig. 2 is a schematic diagram of a vehicle safety precaution device based on video identification according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Example 1
In order to solve the above problem, referring to fig. 1, an embodiment of the present invention provides a vehicle safety early warning method based on video identification, including the following steps:
s1, establishing a front vehicle external Cartesian full vector, a self vehicle internal Cartesian vector and a pixel Cartesian vector;
further, in the embodiment of the present invention, the specific method for establishing the external cartesian full vector constraint of the front vehicle in step S1 includes:
establishing a Cartesian full vector V-X outside the front vehicle by taking a middle grounding point at the tail of the front vehicle as an origin V, taking the front side of the front vehicle as the positive direction of an X axis, taking the right side of the front vehicle as the positive direction of a Y axis and taking the front side of the front vehicle as the positive direction of a Z axisvYvZvSetting the coordinates of 8 angular points constrained by the external Cartesian full vector of the front vehicle as;
Figure 573283DEST_PATH_IMAGE020
，
Figure 789501DEST_PATH_IMAGE021
，
Figure 672006DEST_PATH_IMAGE022
，
Figure 962173DEST_PATH_IMAGE023
，
Figure 904721DEST_PATH_IMAGE024
，
Figure 342656DEST_PATH_IMAGE025
，
Figure 700956DEST_PATH_IMAGE026
，
Figure 173526DEST_PATH_IMAGE027
；
wherein f isx ，fy ，fzFor the outer Cartesian full vector constraint edge X of a vehiclevAxis, YvAxis and ZvLength in the axial direction;
the specific method for establishing the cartesian vector in the vehicle of the own vehicle in the step S1 is as follows:
taking a camera at the top of the self-vehicle as an original point C, taking the front side of the self-vehicle as an Xc axial positive direction, taking the left side of the self-vehicle as a Y axial positive direction and taking the upward direction as a Z axial positive direction; establishing a Cartesian vector C-X in a vehicle fixed to the host vehiclecYcZc; the vector coordinate of the Cartesian vector of the central grounding point V of the tail of the front vehicle in the Cartesian full-vector constraint of the front vehicle is set as V = [ V =x，Vy，0]v(ii) a The position of a Cartesian vector constrained in the self vehicle by a front vehicle external Cartesian full vector is D = [ theta, 0 =]Where θ is the front vehicle YVAxle of bicycle YcThe included angle of the axes;
specifically, the specific method for establishing the pixel cartesian vector in step S1 is as follows:
upper left corner O for collecting laser image by camerapA Cartesian vector O of pixels is established as an origin, an I axis forward direction is towards the right, and a J axis forward direction is towards the bottomp-IJ。
S2, acquiring a three-dimensional model of a laser image of a vehicle ahead at the current moment for processing; monitoring the running condition of the vehicle;
further, in the embodiment of the present invention, step S2 is specifically that a camera is used to obtain a three-dimensional model of a laser image on the rear side of the self-vehicle at the current time point as a current frame laser three-dimensional model, and the three-dimensional model of the laser image collected at the current time is processed by a recurrent global neural network to obtain a dimension V of the external cartesian full-vector constraint of the front vehiclex，Vy，Vz(ii) a Detecting and acquiring the speed of a Cartesian vector in a front vehicle at the current moment through a laser radar; and measuring the position, the speed and the deflection angle of the self-vehicle at the current moment through the Beidou system and the fiber-optic gyroscope.
S3, establishing a projection relation between the Cartesian full vector constraint outside the front vehicle and the Cartesian local vector constraint outside the vehicle through a three-dimensional model of continuous N frames of laser images;
further, in the embodiment of the present invention, step S3 specifically includes: external Cartesian vector coordinate system X of camera for acquiring any point of front vehicle0＝[X ，Y ，Z]VCorresponding point X in pixel coordinate system0=[i ，j]VThe following transformation relationships are satisfied:
Figure 21396DEST_PATH_IMAGE028
(1)
k is an internal parameter matrix of the camera; d' is a rotation matrix from the Cartesian vector outside the front vehicle to the Cartesian vector outside the self-vehicle 3 x 3, and T is a translation matrix of 3 x 1;
Figure 946627DEST_PATH_IMAGE029
。
s4, constructing the same front vehicle position motion constraint by using the three-dimensional model of the continuous N frames of laser images;
further, in the embodiment of the present invention, step S4 specifically includes: setting the vector coordinates of the rear center grounding point of the front vehicle in the three-dimensional model of two continuous frames of laser images in the Cartesian vector of the self vehicle as follows:
Figure 111547DEST_PATH_IMAGE030
，
Figure 438624DEST_PATH_IMAGE031
wherein tn-1, tn is the acquisition time of two continuous frames of images; for the coordinate position of the central grounding point of the tail part of the front vehicle, 2(N-1) motion constraints are built in total according to the three-dimensional model of the current laser image and the three-dimensional model of the front N-1 frames of laser images:
Figure 722974DEST_PATH_IMAGE032
(2)
Figure 73184DEST_PATH_IMAGE033
(3)
wherein u isxn ，uynThe speed of the front vehicle in the Cartesian vector in the vehicle at the corresponding moment of the three-dimensional model of the Nth frame of laser image acquired by a laser radar installed on the vehicle is measured along the X-axis direction and the Y-axis direction;
obtaining the position V of the vehicle in front at the current time in the Cartesian vector of the vehicle by using a least square method through simultaneous formulas (1) - (3)(tn)And an orientation D, i.e. a three-dimensional model of the vehicle ahead of the current time is determined.
S5, determining the transverse and longitudinal relative position relation between the front vehicle and the self vehicle according to the external Cartesian full vector constraint of the front vehicle at the current time;
step S5 specifically includes: drawing curves along the first corner point and the second corner point X1 and X2 of the three-dimensional model of the front vehicle, wherein the curve equation is as follows:
Figure 835604DEST_PATH_IMAGE034
，
wherein
Figure 17187DEST_PATH_IMAGE035
，
Figure 472439DEST_PATH_IMAGE036
Wherein
Figure 309945DEST_PATH_IMAGE018
The difference value of the Cartesian full vectors in the X-axis direction of two continuous frames of laser images is obtained.
Figure 876055DEST_PATH_IMAGE019
The difference value of the Cartesian full vectors in the Y-axis direction of two continuous frames of laser images is obtained. x is the transverse coordinate position of the front vehicle relative to the self vehicle, and y is the longitudinal coordinate position of the front vehicle relative to the self vehicle.
And S6, calculating the safety distance between the front vehicle and the self vehicle and carrying out safety early warning.
Further, in the embodiment of the present invention, step S6 specifically includes: judging whether the front vehicle is in a potential danger or not according to the transverse relative position relation of the front vehicle and the self vehicle, and executing a step S6 if the front vehicle is in the potential danger; otherwise, the process returns to step S2.
The invention can realize the accurate measurement of the transverse and longitudinal distances between the front vehicle and the self vehicle, and can realize real-time monitoring and timely safety early warning prompt to prevent collision between the front vehicle and the self vehicle.
Example 2
Referring to fig. 2, in the present embodiment, a vehicle safety warning device based on video recognition includes:
establishing a vector coordinate module: the vector coordinate establishing module is used for establishing a Cartesian full vector outside the front vehicle, a Cartesian vector inside the self-vehicle and a pixel Cartesian vector;
the data processing module is used for acquiring a three-dimensional model of a laser image of a vehicle in front at the current moment and processing the three-dimensional model; monitoring the running condition of the vehicle;
the projection module is used for establishing a projection relation between the Cartesian full vector constraint outside the front vehicle and the Cartesian local vector constraint outside the vehicle through a three-dimensional model of continuous N frames of laser images;
the position constraint module is used for constructing the position motion constraint of the same front vehicle by utilizing the three-dimensional model of the continuous N frames of laser images;
the position determining module is used for determining the transverse and longitudinal relative position relation between the front vehicle and the self vehicle according to the external Cartesian full vector constraint of the front vehicle at the current moment;
and the safety early warning module is used for calculating the safety distance between the front vehicle and the self vehicle and carrying out safety early warning.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.

Claims (8)

1. A vehicle safety early warning method based on video identification is characterized by comprising the following steps:
s1, establishing a front vehicle external Cartesian full vector, a self vehicle internal Cartesian vector and a pixel Cartesian vector;
s2, acquiring a three-dimensional model of a laser image of a vehicle ahead at the current moment for processing; monitoring the running condition of the vehicle;
s3, establishing a projection relation between the Cartesian full vector constraint outside the front vehicle and the Cartesian local vector constraint outside the vehicle through a three-dimensional model of continuous N frames of laser images;
s4, constructing the same front vehicle position motion constraint by using the three-dimensional model of the continuous N frames of laser images;
s5, determining the transverse and longitudinal relative position relation between the front vehicle and the self vehicle according to the external Cartesian full vector constraint of the front vehicle at the current time;
and S6, calculating the safety distance between the front vehicle and the self vehicle and carrying out safety early warning.
2. The video identification-based vehicle safety warning method according to claim 1, wherein the specific method for establishing the out-of-vehicle cartesian full-vector constraint in the step S1 is as follows:
establishing a Cartesian full vector V-X outside the front vehicle by taking a middle grounding point at the tail of the front vehicle as an origin V, taking the front side of the front vehicle as the positive direction of an X axis, taking the right side of the front vehicle as the positive direction of a Y axis and taking the front side of the front vehicle as the positive direction of a Z axisvYvZvSetting the coordinates of 8 angular points constrained by the external Cartesian full vector of the front vehicle as;
Figure 688298DEST_PATH_IMAGE001
，
Figure 314451DEST_PATH_IMAGE002
，
Figure 639253DEST_PATH_IMAGE003
，
Figure 743476DEST_PATH_IMAGE004
，
Figure 899651DEST_PATH_IMAGE005
，
Figure 696705DEST_PATH_IMAGE006
，
Figure 243224DEST_PATH_IMAGE007
，
Figure 416717DEST_PATH_IMAGE008
；
wherein f isx ，fy ，fzFor the outer Cartesian full vector constraint edge X of a vehiclevAxis, YvAxis and ZvLength in the axial direction;
the specific method for establishing the cartesian vector in the vehicle of the own vehicle in the step S1 is as follows:
taking a camera at the top of the self-vehicle as an original point C, taking the front side of the self-vehicle as an Xc axial positive direction, taking the left side of the self-vehicle as a Y axial positive direction and taking the upward direction as a Z axial positive direction; establishing a Cartesian vector C-X in a vehicle fixed to the host vehiclecYcZc; the vector coordinate of the Cartesian vector of the central grounding point V of the tail of the vehicle in the front vehicle outer Cartesian full vector constraint is set as
Figure 427398DEST_PATH_IMAGE009
(ii) a The position of the Cartesian vector constrained in the self vehicle by the Cartesian full vector outside the front vehicle is set as
Figure 333037DEST_PATH_IMAGE010
Where θ is the front vehicle YVAxle of bicycle YcThe included angle of the axes;
the specific method for establishing the pixel cartesian vector in step S1 is as follows:
upper left corner O for collecting laser image by camerapA Cartesian vector O of pixels is established as an origin, an I axis forward direction is towards the right, and a J axis forward direction is towards the bottomp-IJ。
3. The video recognition-based vehicle safety early warning method according to claim 1, wherein the step S2 is specifically that a three-dimensional model of a laser image at the rear side of the vehicle at the current time point is acquired by using a camera as a current frame laser three-dimensional model, and the three-dimensional model of the laser image acquired at the current time is processed by a recursive global neural network to obtain a dimension V of the out-of-vehicle cartesian full-vector constraint of the front vehiclex，Vy，Vz(ii) a Detecting and acquiring the speed of a Cartesian vector in a front vehicle at the current moment through a laser radar; and measuring the position, the speed and the deflection angle of the self-vehicle at the current moment through the Beidou system and the fiber-optic gyroscope.
4. The video identification-based vehicle safety warning method according to claim 1, wherein the step S3 specifically comprises: external Cartesian vector coordinate system of camera for acquiring any point of front vehicle
Figure 429169DEST_PATH_IMAGE011
Corresponding point in the pixel coordinate system
Figure 875194DEST_PATH_IMAGE012
The following transformation relationships are satisfied:
Figure 958292DEST_PATH_IMAGE013
(1)
k is an internal parameter matrix of the camera; d' is a rotation matrix from the Cartesian vector outside the front vehicle to the Cartesian vector 3 x 3 outside the self vehicle, and T is
Figure 97150DEST_PATH_IMAGE014
Translating the matrix;
Figure 680578DEST_PATH_IMAGE015
。
5. the video identification-based vehicle safety early warning method according to claim 1, wherein the step S4 specifically comprises: setting the vector coordinates of the rear center grounding point of the front vehicle in the three-dimensional model of two continuous frames of laser images in the Cartesian vector of the self vehicle as follows:
Figure 867977DEST_PATH_IMAGE016
，
Figure 853250DEST_PATH_IMAGE017
wherein
Figure 163009DEST_PATH_IMAGE018
Tn is the acquisition time of two continuous frames of images; for the coordinate position of the central grounding point of the tail part of the front vehicle, establishing a total according to the three-dimensional model of the current laser image and the three-dimensional model of the front N-1 frames of laser images
Figure 233733DEST_PATH_IMAGE019
And (3) motion constraint:
Figure 224823DEST_PATH_IMAGE020
(2)
Figure 64603DEST_PATH_IMAGE021
(3)
wherein the content of the first and second substances,
Figure 482946DEST_PATH_IMAGE022
，
Figure 775387DEST_PATH_IMAGE023
is to be anThe speed of the front vehicle in the Cartesian vector in the vehicle at the corresponding moment of the three-dimensional model of the Nth frame of laser image acquired by the laser radar installed on the vehicle is the speed of the front vehicle in the X axis direction and the Y axis direction;
obtaining the position V of the vehicle in front at the current time in the Cartesian vector of the vehicle by using a least square method through simultaneous formulas (1) - (3)(tn)And an orientation D, i.e. a three-dimensional model of the vehicle ahead of the current time is determined.
6. The video identification-based vehicle safety early warning method according to claim 1, wherein the step S5 specifically comprises: drawing curves along the first corner point and the second corner point X1 and X2 of the three-dimensional model of the front vehicle, wherein the curve equation is as follows:
Figure 632484DEST_PATH_IMAGE024
，
wherein
Figure 264454DEST_PATH_IMAGE025
，
Figure 916015DEST_PATH_IMAGE026
Wherein
Figure 961332DEST_PATH_IMAGE027
The difference value of the Cartesian full vectors in the X-axis direction of two continuous frames of laser images is obtained;
Figure 356541DEST_PATH_IMAGE028
the difference value of the Cartesian full vectors in the Y-axis direction of two continuous frames of laser images is obtained; x is the transverse coordinate position of the front vehicle relative to the self vehicle, and y is the longitudinal coordinate position of the front vehicle relative to the self vehicle.
7. The video identification-based vehicle safety early warning method according to claim 1, wherein the step S6 specifically comprises: judging whether the front vehicle is in a potential danger or not according to the transverse relative position relation of the front vehicle and the self vehicle, and executing a step S6 if the front vehicle is in the potential danger; otherwise, the process returns to step S2.
8. The utility model provides a vehicle safety precaution device based on video identification which characterized in that includes:
establishing a vector coordinate module: the vector coordinate establishing module is used for establishing a Cartesian full vector outside the front vehicle, a Cartesian vector inside the self-vehicle and a pixel Cartesian vector;
the data processing module is used for acquiring a three-dimensional model of a laser image of a vehicle in front at the current moment and processing the three-dimensional model; monitoring the running condition of the vehicle;
the projection module is used for establishing a projection relation between the Cartesian full vector constraint outside the front vehicle and the Cartesian local vector constraint outside the vehicle through a three-dimensional model of continuous N frames of laser images;
the position constraint module is used for constructing the position motion constraint of the same front vehicle by utilizing the three-dimensional model of the continuous N frames of laser images;
the position determining module is used for determining the transverse and longitudinal relative position relation between the front vehicle and the self vehicle according to the external Cartesian full vector constraint of the front vehicle at the current moment;
and the safety early warning module is used for calculating the safety distance between the front vehicle and the self vehicle and carrying out safety early warning.
CN202011559995.7A 2020-12-25 2020-12-25 Vehicle safety early warning method and device based on video identification Pending CN112489496A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011559995.7A CN112489496A (en) 2020-12-25 2020-12-25 Vehicle safety early warning method and device based on video identification

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011559995.7A CN112489496A (en) 2020-12-25 2020-12-25 Vehicle safety early warning method and device based on video identification

Publications (1)

Publication Number Publication Date
CN112489496A true CN112489496A (en) 2021-03-12

Family

ID=74915638

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011559995.7A Pending CN112489496A (en) 2020-12-25 2020-12-25 Vehicle safety early warning method and device based on video identification

Country Status (1)

Country Link
CN (1) CN112489496A (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2008003748A (en) * 2006-06-21 2008-01-10 Fujifilm Corp Preceding vehicle detector, preceding vehicle detection method, and preceding vehicle detection program
CN102542843A (en) * 2010-12-07 2012-07-04 比亚迪股份有限公司 Early warning method for preventing vehicle collision and device
US20180165963A1 (en) * 2015-12-29 2018-06-14 Thunder Power New Energy Vehicle Development Company Limited Onboard vehicle notification system
CN110444014A (en) * 2019-07-01 2019-11-12 淮阴工学院 The anti-method for early warning that knocks into the back based on reversed ST-MRF vehicle tracking algorithm
CN110745140A (en) * 2019-10-28 2020-02-04 清华大学 Vehicle lane change early warning method based on continuous image constraint pose estimation
CN111429756A (en) * 2020-03-30 2020-07-17 淮阴工学院 Highway tunnel rear-end collision prevention early warning method based on video technology
WO2020180997A1 (en) * 2019-03-04 2020-09-10 Ouster, Inc. Driver visualization and semantic monitoring of a vehicle using lidar data

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2008003748A (en) * 2006-06-21 2008-01-10 Fujifilm Corp Preceding vehicle detector, preceding vehicle detection method, and preceding vehicle detection program
CN102542843A (en) * 2010-12-07 2012-07-04 比亚迪股份有限公司 Early warning method for preventing vehicle collision and device
US20180165963A1 (en) * 2015-12-29 2018-06-14 Thunder Power New Energy Vehicle Development Company Limited Onboard vehicle notification system
WO2020180997A1 (en) * 2019-03-04 2020-09-10 Ouster, Inc. Driver visualization and semantic monitoring of a vehicle using lidar data
CN110444014A (en) * 2019-07-01 2019-11-12 淮阴工学院 The anti-method for early warning that knocks into the back based on reversed ST-MRF vehicle tracking algorithm
CN110745140A (en) * 2019-10-28 2020-02-04 清华大学 Vehicle lane change early warning method based on continuous image constraint pose estimation
CN111429756A (en) * 2020-03-30 2020-07-17 淮阴工学院 Highway tunnel rear-end collision prevention early warning method based on video technology

Similar Documents

Publication Publication Date Title
CN110745140B (en) Vehicle lane change early warning method based on continuous image constraint pose estimation
CN112349144B (en) Monocular vision-based vehicle collision early warning method and system
CN107031623B (en) A kind of road method for early warning based on vehicle-mounted blind area camera
CN108638999B (en) Anti-collision early warning system and method based on 360-degree look-around input
JP4899424B2 (en) Object detection device
CN110909705B (en) Road side parking space sensing method and system based on vehicle-mounted camera
CN110794406B (en) Multi-source sensor data fusion system and method
CN112132896B (en) Method and system for detecting states of trackside equipment
GB2265779A (en) Obstacle warning system for vehicle
CN113850102B (en) Vehicle-mounted vision detection method and system based on millimeter wave radar assistance
CN110738181B (en) Method and device for determining vehicle orientation information
CN114495064A (en) Monocular depth estimation-based vehicle surrounding obstacle early warning method
CN107796373A (en) A kind of distance-finding method of the front vehicles monocular vision based on track plane geometry model-driven
EP3364336B1 (en) A method and apparatus for estimating a range of a moving object
CN112613509A (en) Railway wagon carriage number identification snapshot method and system
CN109345591A (en) A kind of vehicle itself attitude detecting method and device
CN113962301B (en) Multi-source input signal fused pavement quality detection method and system
WO2023020004A1 (en) Vehicle distance detection method and system, and device and medium
CN109145805B (en) Moving target detection method and system under vehicle-mounted environment
CN116543032B (en) Impact object ranging method, device, ranging equipment and storage medium
CN111874003B (en) Vehicle driving deviation early warning method and system
CN112298165A (en) Active steering control method based on forward vision
CN112489496A (en) Vehicle safety early warning method and device based on video identification
CN113643355A (en) Method and system for detecting position and orientation of target vehicle and storage medium
CN111414857A (en) Front vehicle detection method based on vision multi-feature fusion

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