CN106228531B - Automatic vanishing point calibration method and system based on horizon line search - Google Patents

Automatic vanishing point calibration method and system based on horizon line search Download PDF

Info

Publication number
CN106228531B
CN106228531B CN201610492617.9A CN201610492617A CN106228531B CN 106228531 B CN106228531 B CN 106228531B CN 201610492617 A CN201610492617 A CN 201610492617A CN 106228531 B CN106228531 B CN 106228531B
Authority
CN
China
Prior art keywords
horizon
current
top view
straight lines
correct
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201610492617.9A
Other languages
Chinese (zh)
Other versions
CN106228531A (en
Inventor
刘鹏
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.)
Kai Yi (beijing) Technology Co Ltd
Original Assignee
Kai Yi (beijing) 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 Kai Yi (beijing) Technology Co Ltd filed Critical Kai Yi (beijing) Technology Co Ltd
Priority to CN201610492617.9A priority Critical patent/CN106228531B/en
Publication of CN106228531A publication Critical patent/CN106228531A/en
Application granted granted Critical
Publication of CN106228531B publication Critical patent/CN106228531B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
    • G06V20/588—Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
    • 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
    • G06T2207/30248—Vehicle exterior or interior
    • G06T2207/30252—Vehicle exterior; Vicinity of vehicle
    • 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
    • G06T2207/30248—Vehicle exterior or interior
    • G06T2207/30252—Vehicle exterior; Vicinity of vehicle
    • G06T2207/30256—Lane; Road marking

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses an automatic vanishing point calibration method and system based on horizon search, wherein the method comprises the following steps: establishing a horizon position template; generating an inverse perspective transformation matrix corresponding to the horizon according to the current horizon position; generating a top view corresponding to the horizon through the inverse perspective transformation matrix; performing straight line detection in a plan view, and judging whether the horizon is correct or not; if the straight lines are correct, the straight lines detected in the top view are converted to the perspective view through inverse perspective, and the intersection points of the straight lines in the perspective view are taken as vanishing points; and if not, taking the next horizon as the current horizon according to the horizon position template, and continuing to verify. The invention utilizes a cyclic search method to distribute the search pressure to the time dimension, and the consumption of computing resources is low. In addition, the invention can realize full-automatic vanishing point calibration without any prior condition, thereby realizing automatic vanishing point calibration in the true sense.

Description

Automatic vanishing point calibration method and system based on horizon line search
Technical Field
The invention relates to the field of image processing, in particular to an automatic vanishing point calibration method and system based on horizon line search.
Background
The Advanced Driving Assistance System (ADAS) based on visual algorithm, the vanishing point position is the key input information of each technology, including the calculation of the distance between the front vehicles, the estimation of the distance between the pedestrians and the detection of the lane line. Due to differences in camera installation, positions of vanishing points in the acquired images are very different, and the existing ADAS system generally determines the positions of the vanishing points through manual calibration or automatic calibration. Manual calibration generally requires accurate installation by a professional technician, and is too costly to discuss. In the method of automatically calibrating vanishing points, a common method includes:
1. method for detecting straight line or lane line through whole image
In the method, the image is filtered by using an edge filtering method in the original image, and then a straight line is detected by Hough transform. And in the detected straight line, calculating intersection points pairwise, filtering out points which are not in the central area of the image in the obtained intersection points, and taking the mean value position of the rest points as vanishing points.
2. By means of G-sensor sensors and lane line detection
According to the method, the pitch angle of the camera is obtained by means of the G-sensor, and then the horizon position of the image is estimated according to the imaging principle. Then, the intersection points of a plurality of lane lines are taken as vanishing points by using a lane line detection module.
3. By means of preceding vehicle detection and lane marking detection
The proportional relation of the front vehicle height and the camera height has the following relation with the horizontal line position:
Figure BDA0001030851380000021
wherein h is0Indicates the real height of the front vehicle, hcRepresenting the true height, v, of the cameratIs the ordinate, v, of the front vehicle's head in the imagebIs the ordinate, v, of the bottom end of the preceding vehicle in the imageoHorizon vertical coordinate.
The reverse reasoning can obtain:
Figure BDA0001030851380000022
because the car has small height floating range h0Approximately 1.5m, the remaining variables are also known. From which the position of the horizon can be estimated. And then, on the premise of knowing the position of the horizon, taking the intersection point of a plurality of lane lines as a vanishing point by using a lane line detection module.
The above method has the following disadvantages:
in the method 1, under the condition of uncertain horizon, the straight line detection in the whole graph range consumes a large amount of computing resources, and a plurality of interference straight lines (such as road surface characters, roadside railings and the like) are not gathered at vanishing points, which may affect the result.
In the method 2, depending on hardware, the hardware process of the G-sensor directly affects the estimation result of the horizon, and a slight deviation may cause a dead point estimation error, and the process is uncontrollable.
In the method 3, depending on the front vehicle detection module, when there is no vehicle in front or the vehicle detection is inaccurate, the horizon estimation is wrong, thereby affecting the calibration result of the vanishing point.
Disclosure of Invention
The invention aims to solve the technical problem of automatic vanishing point calibration without depending on hardware of front vehicle detection and G-sensor.
The invention solves the technical problem, provides an automatic vanishing point calibration method based on horizon search, which comprises the following steps:
establishing a horizon position template;
generating an inverse perspective transformation matrix corresponding to the horizon according to the current horizon position;
generating a top view corresponding to the horizon through the inverse perspective transformation matrix;
performing straight line detection in a plan view, and judging whether the horizon is correct or not;
if the straight lines are correct, the straight lines detected in the top view are converted to the perspective view through inverse perspective, and the intersection points of the straight lines in the perspective view are taken as vanishing points;
and if not, taking the next horizon as the current horizon according to the horizon position template, and continuing to verify.
Further, 11 optional horizon positions are selected, and the following horizon position templates are established:
{y0、y1、y2、y3、y4、y5、y6、y7、y8、y9、y10and selecting 11 transverse lines uniformly spaced in the middle area of the image.
Further, the method for generating the inverse perspective transformation matrix corresponding to the horizon is as follows:
computing an inverse perspective transformation matrix T-1With the center point P of the horizon line0A function of (a);
Tn -1＝f-1(yn)
wherein, P0Sit uprightMarked as horizon coordinate ynThe abscissa is a fixed value which is half the width of the fluoroscopic image.
Further, the result of the straight line detection in the top view is represented by two end points of the straight line:
detected straight line L ═ ((x)P,yp),(xQ,yQ)),
Wherein (x)P,yp)，(xQ,yQ) The coordinates of the upper end point P and the lower end point Q are respectively, and the coordinates of the upper end point and the lower end point of the straight line can represent corresponding straight lines.
Further, when a straight line is detected in the plan view and the ground level is determined to be correct,
if the included angle between any two straight lines and the x axis meets the following condition, namely, on the top view, the detected included angle between the straight lines and the x axis
|α1-α2|≤ε
ε is the threshold for the minimum angular difference that determines whether two straight lines are parallel.
Still further, where ε is 5.
Further, the angle between the x-axis and the x-axis can be obtained by the coordinates of the end points of the straight line:
Figure BDA0001030851380000041
wherein x isP,ypAnd xQ,yQRespectively, the coordinates of the upper end point P and the lower end point Q.
Further, determining whether the horizon is correct includes the following conditions:
if the horizon of at least 3 frames in 10 consecutive frames meets the parallel condition, the horizon is considered correct.
The invention also provides an automatic vanishing point calibration system based on horizon search, which comprises:
the template unit is used for establishing a horizon position template;
a top view generation unit which generates an inverse perspective transformation matrix corresponding to the horizon according to the current horizon position; generating a top view corresponding to the horizon through the inverse perspective transformation matrix;
a line detection unit for performing line detection in a plan view;
verifying the parallel line unit to judge whether the horizon is correct; if the straight lines are correct, the straight lines detected in the top view are converted to the perspective view through inverse perspective, and the intersection points of the straight lines in the perspective view are taken as vanishing points; and if not, taking the next horizon as the current horizon according to the horizon position template, and continuing to verify.
Furthermore, the device also comprises a searching unit which is used for taking the next horizon as the current horizon and continuing to verify according to the horizon position template when the parallel line verifying unit judges that the horizon is incorrect. Until the verification is passed. If all the horizons fail to verify, the verification is repeated again starting from the first horizon.
The invention has the beneficial effects that:
1) the invention relates to an automatic vanishing point calibration method based on horizon line search, which comprises the steps of carrying out linear detection in a top view and judging whether a horizon line is correct or not; if the ground level lines are correct, converting the ground level lines in the top view into the perspective view through perspective, and taking the intersection points of the ground level lines in the perspective view as vanishing points; and if not, continuously searching the horizon position of the next frame according to the horizon position template. The invention utilizes a cyclic search method to distribute the search pressure to the time dimension, and the consumption of computing resources is low.
2) The invention can realize full-automatic calibration of vanishing points without any prior condition, and a horizon position template is established; generating an inverse perspective transformation matrix corresponding to the horizon according to the detected current horizon coordinate in the horizon position template; generating a top view corresponding to the horizon through the inverse perspective transformation matrix; the straight line detection is performed in the plan view, and whether the horizon is correct or not is determined.
3) The vanishing point in the present invention means a point at a very far position in front of the field of view, that is, an intersection of the horizon and the line of sight in front of the field of view, and the vanishing point is a convergence point of straight lines (lane lines) extending forward and backward in the field of view. In Advanced Driving Assistance Systems (ADAS) based on vision algorithms, vanishing point location is the key input information for each technology, including but not limited to, preceding vehicle distance calculation, pedestrian distance estimation, lane line detection. Therefore, the method has great value and effect on ADAS.
Drawings
Fig. 1 is a schematic flow chart of an automatic vanishing point calibration method based on horizon line search according to the invention.
FIG. 2 is a schematic diagram of a simulation of the creation of the horizon position template of FIG. 1.
FIGS. 3(a) -3 (c) are schematic top views generated in FIG. 1.
Fig. 4 is a schematic view of verifying whether the horizon is a parallel line in fig. 1.
Fig. 5 is a schematic diagram of the vanishing point finding method of fig. 1.
Fig. 6 is a schematic structural diagram of an automatic vanishing point calibration system based on horizon line search according to the 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 specific embodiments and the accompanying drawings.
Fig. 1 is a schematic flow chart of an automatic vanishing point calibration method based on horizon line search according to the invention.
Step S100, establishing a horizon position template;
step S101, according to the current horizon position, generating an inverse perspective transformation matrix corresponding to the horizon;
step S102, generating a top view corresponding to the horizon through the inverse perspective transformation matrix;
step S103 performs straight line detection in a plan view,
step S104, judging whether the horizon is correct or not;
step S105, if the result is correct, converting the straight lines detected in the top view into the perspective view through inverse perspective, and taking the intersection points of the straight lines in the perspective view as vanishing points;
and S106, if the position is incorrect, taking the next horizon as the current horizon according to the horizon position template, and continuing to verify.
The automatic vanishing point calibration method in the embodiment can realize full-automatic vanishing point calibration without any prior condition, and establishes a horizon position template; generating an inverse perspective transformation matrix corresponding to the horizon according to the detected current horizon coordinate in the horizon position template; generating a top view corresponding to the horizon through the inverse perspective transformation matrix; the straight line detection is performed in the plan view, and whether the horizon is correct or not is determined.
As a preferred choice in this embodiment, 11 alternative horizon positions are selected in step S100, and the following horizon position templates are established:
{y0、y1、y2、y3、y4、y5、y6、y7、y8、y9、y10}。
in the present embodiment, the method for generating the inverse perspective transformation matrix corresponding to the horizon in step S101 in step S100 is preferably as follows:
computing an inverse perspective transformation matrix T-1With the center point P of the horizon line0A function of (a);
Tn -1＝f-1(yn)
wherein, P0The ordinate of (a) is the horizon coordinate ynThe abscissa is a fixed value which is half the width of the fluoroscopic image.
As a preference in the present embodiment, the result of the straight line detection in the top view in step S103 is represented by two end points of the straight line:
detected straight line L ═ ((x)P,yp),(xQ,yQ)),
Wherein (x)P,yp)，(xQ,yQ) Respectively, the coordinates of the upper end point P and the lower end point Q.
In the present embodiment, preferably, when the straight line is detected in the plan view in step S104 and whether or not the horizon is correct is determined,
if the included angles between any two straight lines and the x-axis are parallel when the following conditions are met,
|α1-α2|≤ε
epsilon is a threshold value of the minimum angle difference for determining whether two straight lines are parallel, where epsilon is 5 deg..
Preferably, in the present embodiment, the angle between the x-axis and the x-axis can be obtained by the coordinates of the end points of the straight line:
Figure BDA0001030851380000071
wherein x isP,ypAnd xQ,yQRespectively, the coordinates of the upper end point P and the lower end point Q.
As a preferable example in the present embodiment, the determination of whether the horizon line is correct in step S105 includes the following conditions:
if the horizon of at least 3 frames in 10 consecutive frames meets the parallel condition, the horizon is considered correct.
The invention principle is as follows:
firstly, an inverse perspective transformation matrix T corresponding to the current horizon is generated by utilizing the functional relation between the inverse perspective transformation of the top view and the coordinates of the horizon-1. Within the next 10 consecutive frames of images, using this T-1Generating a top view and performing straight line detection. If the detected straight lines in more than 3 frames in 10 frames are in parallel relation, the search is quitted. If the exit condition can not be met, the horizon line is replaced, and the steps are repeated to continue verification until the search exits. And after the search is successful, performing perspective transformation on the parallel straight lines in the top view, and taking the intersection points of the parallel straight lines in the perspective view as vanishing points.
Specifically, the method introduces the search concept into vanishing point calibration and searches the horizon position circularly.
Fig. 2 is a schematic diagram of the simulation of the model for establishing the horizon position in fig. 1. The lines marked in the figure cover the position of the flat lines in the image for most camera installations. The ordinate corresponding to each horizon line is marked as yn. From y0To start, for each stripAnd (4) verifying continuous 10 frames of the horizon, and if the exit condition is not met, jumping to the next horizon to continue verification. If y is searched10If the search is not exited, then the search continues from y0And the process is restarted.
Since the images captured by the cameras change in real time in practical application scenarios, the amount of information contained in each frame of image is different. Moreover, the calibration of the vanishing point is not required to be completed immediately, and only needs to be completed as accurately as possible within a certain time. Therefore, the invention does not traverse all search targets in the same frame of image, but adopts a circular search method, distributes the search pressure to the time dimension, verifies a horizon every 10 frames, and does not generate a large amount of operations in one frame.
1) Top view generation
The first step of the horizon line verification process is to generate a top view, which is to convert the lane line region of the perspective view image into a top view through an inverse perspective transformation.
Computing an inverse perspective transformation matrix T-1The method of (2) is well established and there are corresponding functions in OPENCV, and the technical details are not described here. The method for calculating the inverse perspective transformation matrix according to the center point of the horizon line is mature, and the invention is not described again.
Here only T needs to be known-1Is a function of the horizon center point P0, the ordinate of P0 is the horizon coordinate yn, and the abscissa is half the width of the fluoroscopic image, which is a fixed value. Therefore, Tn -1Only yn.
Tn -1＝f-1(yn)
Fig. 3(a) -3 (c) are schematic top views generated in fig. 1, wherein fig. 3(a) -3 (c) are top views corresponding to y4, y5, and y6 horizons, respectively.
2) Line detection
The straight line detection is performed in a plan view, and the dimension of the plan view is much smaller than that of the original image, so that the speed of the straight line detection is high. Methods for line detection are very common, such as hough transform and enhanced projection, and the technical details are not described herein.
The detected straight line L is represented by the coordinates of the upper end point P and the lower end point Q:
L＝((xP,yP),(xQ,yQ))
3) verification of parallel lines
From the top view above, it can be seen that if the current horizon is correct, the straight lines that converge to the vanishing point (e.g., lane line) in the perspective view are parallel in the top view. By using this a priori knowledge, it can be determined whether the horizon is correct.
As shown in fig. 4, it is a schematic diagram of verifying whether the horizon is a parallel line in fig. 1, and if the included angle between any two straight lines and the x-axis satisfies the following condition, the two straight lines are parallel.
|α1-α2|≤ε
After a large number of experiments, epsilon is 5 deg..
The angle between the straight line and the x-axis can be obtained by the coordinates of the end points of the straight line:
Figure BDA0001030851380000091
if 3 frames or more than 3 frames in the continuous 10 frames meet the parallel condition, the horizon is considered to be correct, and the searching process is exited.
Finally, the parallel straight lines in the top view are transformed to the perspective view through perspective. The perspective transformation is the inverse of the inverse perspective transformation, and there is a corresponding function in OPENCV, which will not be described in detail herein.
Fig. 5 is a schematic diagram of the vanishing point finding method in fig. 1, and inverse transformation of inverse perspective transformation is performed to obtain perspective transformation:
(P'x,y,Q'x,y)＝f(Px,y,Qx,y)
taking the intersection point of the horizon lines in the perspective view as a vanishing point, and the vanishing point coordinate calculation method can be as follows:
Figure BDA0001030851380000101
Figure BDA0001030851380000102
fig. 6 is a schematic structural diagram of an automatic vanishing point calibration system based on horizon line search according to the present invention.
The automatic vanishing point calibration system 10 based on horizon line search in this embodiment includes:
the template unit 1 is used for establishing a horizon position template;
a top view generation unit 2 for generating an inverse perspective transformation matrix corresponding to the horizon according to the current position of the horizon; generating a top view corresponding to the horizon through the inverse perspective transformation matrix;
a line detection unit 3 for performing line detection in a plan view;
a parallel line verification unit 4 for determining whether the horizon is correct; if the straight lines are correct, the straight lines detected in the top view are converted to the perspective view through inverse perspective, and the intersection points of the straight lines in the perspective view are taken as vanishing points; and if not, taking the next horizon as the current horizon according to the horizon position template, and continuing to verify.
Preferably, the automatic vanishing point calibrating system 10 further includes a searching unit 5, configured to, when the parallel line verification unit determines that the horizon is incorrect, continue the verification by taking the next horizon as the current horizon according to the horizon position template. Until the verification is passed. If all the horizons fail to verify, the verification is repeated again starting from the first horizon.
In the automatic vanishing point calibration system in this embodiment, the search pressure is distributed to the time dimension by using a cyclic search method in the search unit 5, and the consumption of computational resources is low.
Those of ordinary skill in the art will understand that: the present invention is not limited to the above embodiments, and any modifications, equivalent substitutions, improvements, etc. within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (8)

1. An automatic vanishing point calibration method based on horizon search is characterized in that the method does not depend on prior conditions and comprises the following steps:
establishing a horizon position template;
generating an inverse perspective transformation matrix corresponding to the current horizon according to the current horizon position;
generating a top view corresponding to the current horizon through the inverse perspective transformation matrix;
performing straight line detection in a plan view, and judging whether the current horizon is correct or not;
the result of the line detection in the top view is represented by the two end points of the line:
detected straight line L ═ ((x)P,yP),(xQ,yQ)),
Wherein (x)P,yP)，(xQ,yQ) Coordinates of an upper end point P and a lower end point Q are respectively;
when straight line detection is carried out in the top view and whether the current horizon is correct or not is judged,
if the included angles between any two straight lines and the x-axis are parallel when the following conditions are met,
|α1-α2|≤ε
epsilon is a threshold value of the minimum angle difference for judging whether the two straight lines are parallel or not;
if the straight lines are correct, the straight lines detected in the top view are converted to the perspective view through inverse perspective, and the intersection points of the straight lines in the perspective view are taken as vanishing points;
and if not, taking the next horizon as the current horizon according to the horizon position template, continuously verifying and carrying out circular search.
2. The automatic vanishing point calibration method as claimed in claim 1, wherein 11 alternative horizon positions are selected, and the following horizon position templates are established:
{y0、y1、y2、y3、y4、y5、y6、y7、y8、y9、y10}。
3. the automatic vanishing point calibrating method according to claim 1, wherein the method for generating the inverse perspective transformation matrix corresponding to the current horizon is as follows:
computing an inverse perspective transformation matrix T-1With the center point P of the horizon line0A function of (a);
Tn -1＝f-1(yn)
wherein, P0The ordinate of (a) is the horizon coordinate ynThe abscissa is a fixed value which is half the width of the fluoroscopic image.
4. The automatic vanishing point calibrating method according to claim 1, wherein e is 5 °.
5. The automatic vanishing point calibrating method as claimed in claim 1, wherein the included angle with the x-axis is obtained by the coordinates of the end points of the straight line:
Figure FDA0002366523300000021
6. the automatic vanishing point calibration method as claimed in claim 1, wherein the determining whether the current horizon is correct comprises the following conditions:
if the horizon of at least 3 frames in 10 consecutive frames meets the parallel condition, the horizon is considered correct.
7. An automatic vanishing point calibration system based on horizon search is characterized by comprising:
the template unit is used for establishing a horizon position template;
the top view generating unit generates an inverse perspective transformation matrix corresponding to the current horizon according to the current horizon position; generating a top view corresponding to the current horizon through the inverse perspective transformation matrix;
a line detection unit configured to perform line detection in a top view, wherein a result of the line detection in the top view is represented by two end points of a line:
detected straight line L ═ ((x)P,yP),(xQ,yQ))，
Wherein (x)P,yP)，(xQ,yQ) Coordinates of an upper end point P and a lower end point Q are respectively;
when straight line detection is carried out in the top view and whether the current horizon is correct or not is judged,
if the included angles between any two straight lines and the x-axis are parallel when the following conditions are met,
|α1-α2|≤ε
epsilon is a threshold value of the minimum angle difference for judging whether the two straight lines are parallel or not;
verifying the parallel line unit to judge whether the current horizon is correct; if the straight lines are correct, the straight lines detected in the top view are converted to the perspective view through inverse perspective, and the intersection points of the straight lines in the perspective view are taken as vanishing points; and if not, taking the next horizon as the current horizon according to the horizon position template, continuously verifying and carrying out circular search.
8. The automatic vanishing point calibrating system of claim 7, further comprising a searching unit, when the parallel line verifying unit determines that the current horizon is incorrect, and based on the horizon position template, taking the next horizon as the current horizon, and continuing to verify until the verification is passed; if all the horizons fail to verify, the verification is repeated again starting from the first horizon.
CN201610492617.9A 2016-06-27 2016-06-27 Automatic vanishing point calibration method and system based on horizon line search Active CN106228531B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610492617.9A CN106228531B (en) 2016-06-27 2016-06-27 Automatic vanishing point calibration method and system based on horizon line search

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610492617.9A CN106228531B (en) 2016-06-27 2016-06-27 Automatic vanishing point calibration method and system based on horizon line search

Publications (2)

Publication Number Publication Date
CN106228531A CN106228531A (en) 2016-12-14
CN106228531B true CN106228531B (en) 2020-03-20

Family

ID=57519965

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610492617.9A Active CN106228531B (en) 2016-06-27 2016-06-27 Automatic vanishing point calibration method and system based on horizon line search

Country Status (1)

Country Link
CN (1) CN106228531B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107316331B (en) * 2017-08-02 2020-04-14 浙江工商大学 Vanishing point automatic calibration method for road image
CN111077893B (en) * 2019-12-31 2023-06-23 驭势科技（浙江）有限公司 Navigation method based on multiple vanishing points, electronic equipment and storage medium
CN111210411B (en) * 2019-12-31 2024-04-05 驭势科技（浙江）有限公司 Method for detecting vanishing points in image, method for training detection model and electronic equipment
CN111174796B (en) * 2019-12-31 2022-04-29 驭势科技（浙江）有限公司 Navigation method based on single vanishing point, electronic equipment and storage medium
CN113409235B (en) * 2020-03-17 2023-08-22 杭州海康威视数字技术股份有限公司 Vanishing point estimation method and apparatus

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102103747A (en) * 2009-12-16 2011-06-22 中国科学院电子学研究所 Method for calibrating external parameters of monitoring camera by adopting reference height
CN102324017A (en) * 2011-06-09 2012-01-18 中国人民解放军国防科学技术大学 FPGA (Field Programmable Gate Array)-based lane line detection method
CN105644785A (en) * 2015-12-31 2016-06-08 哈尔滨工业大学 Unmanned aerial vehicle landing method based on optical flow method and horizon line detection

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102103747A (en) * 2009-12-16 2011-06-22 中国科学院电子学研究所 Method for calibrating external parameters of monitoring camera by adopting reference height
CN102324017A (en) * 2011-06-09 2012-01-18 中国人民解放军国防科学技术大学 FPGA (Field Programmable Gate Array)-based lane line detection method
CN105644785A (en) * 2015-12-31 2016-06-08 哈尔滨工业大学 Unmanned aerial vehicle landing method based on optical flow method and horizon line detection

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
Detection and Classification of Painted Road Objects for Intersection Assistance Applications;Radu Danescu 等;《2010 13th International IEEE Annual Conference on Intelligent Transportation Systems》;20101231;433-438 *
基于单目视觉的结构化道路检测算法研究;焦欣欣 等;《现代计算机》;20110131;正文第2.1节第3段 *
基于线性逼近的车道线弯道识别方法;王宝锋 等;《北京理工大学学报》;20160531;第36卷(第5期);470-474 *
路况PTZ 摄像机自动标定方法;李勃 等;《北京邮电大学学报》;20090430;第32卷;正文第2节第2-3段 *

Also Published As

Publication number Publication date
CN106228531A (en) 2016-12-14

Similar Documents

Publication Publication Date Title
CN106228531B (en) Automatic vanishing point calibration method and system based on horizon line search
US10015394B2 (en) Camera-based speed estimation and system calibration therefor
WO2018196391A1 (en) Method and device for calibrating external parameters of vehicle-mounted camera
CN109583280A (en) Lane detection method, apparatus, equipment and storage medium
Nieto et al. Stabilization of inverse perspective mapping images based on robust vanishing point estimation
CN107305632B (en) Monocular computer vision technology-based target object distance measuring method and system
JP5075672B2 (en) Object detection apparatus and method
Kang et al. Multi-lane detection based on accurate geometric lane estimation in highway scenarios
JP4793324B2 (en) Vehicle monitoring apparatus and vehicle monitoring method
CN105335955A (en) Object detection method and object detection apparatus
US20180293447A1 (en) Road parameter calculator
KR20090088210A (en) Method and apparatus for detecting target parking location by using two reference point and parking assist system using same
CN111443704B (en) Obstacle positioning method and device for automatic driving system
JP6552448B2 (en) Vehicle position detection device, vehicle position detection method, and computer program for vehicle position detection
US20150178922A1 (en) Calibration device, method for implementing calibration, and camera for movable body and storage medium with calibration function
JP4797846B2 (en) Lane marking device and lane detection device
JP5310027B2 (en) Lane recognition device and lane recognition method
JP2020067698A (en) Partition line detector and partition line detection method
CN113393524B (en) Target pose estimation method combining deep learning and contour point cloud reconstruction
JP2012159469A (en) Vehicle image recognition device
JP7145770B2 (en) Inter-Vehicle Distance Measuring Device, Error Model Generating Device, Learning Model Generating Device, Methods and Programs Therefor
JP2012159470A (en) Vehicle image recognition device
CN111862146B (en) Target object positioning method and device
WO2014054124A1 (en) Road surface markings detection device and road surface markings detection method
KR102448944B1 (en) Method and Device for Measuring the Velocity of Vehicle by Using Perspective Transformation

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant