CN103487035A - Vehicle image based monocular positioning method and system thereof - Google Patents

Vehicle image based monocular positioning method and system thereof Download PDF

Info

Publication number
CN103487035A
CN103487035A CN201310472446.XA CN201310472446A CN103487035A CN 103487035 A CN103487035 A CN 103487035A CN 201310472446 A CN201310472446 A CN 201310472446A CN 103487035 A CN103487035 A CN 103487035A
Authority
CN
China
Prior art keywords
target
ground
picture
coordinate
che
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.)
Granted
Application number
CN201310472446.XA
Other languages
Chinese (zh)
Other versions
CN103487035B (en
Inventor
刘进
李德仁
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 D & W SPATIAL INFORMATION TECHNOLOGY Co Ltd
Original Assignee
SHENZHEN D & W SPATIAL INFORMATION 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 D & W SPATIAL INFORMATION TECHNOLOGY Co Ltd filed Critical SHENZHEN D & W SPATIAL INFORMATION TECHNOLOGY Co Ltd
Priority to CN201310472446.XA priority Critical patent/CN103487035B/en
Publication of CN103487035A publication Critical patent/CN103487035A/en
Application granted granted Critical
Publication of CN103487035B publication Critical patent/CN103487035B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C11/00—Photogrammetry or videogrammetry, e.g. stereogrammetry; Photographic surveying
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C11/00—Photogrammetry or videogrammetry, e.g. stereogrammetry; Photographic surveying
    • G01C11/04—Interpretation of pictures
    • G01C11/06—Interpretation of pictures by comparison of two or more pictures of the same area

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Navigation (AREA)
  • Length Measuring Devices By Optical Means (AREA)
  • Image Analysis (AREA)

Abstract

The invention relates to a vehicle image based monocular positioning method and a system thereof. The method comprises an off-line process and an online process, the off-line process comprises the steps of training a target master drawing and establishing a target classifier database, and the online process comprises the following steps: detecting an interested target in an input image based on the target classifier database to obtain a corresponding target rectangular block area, carrying out ground monocular positioning of the corresponding target rectangular block area against a natural object type target to obtain a three dimensional coordinate, and transiting to a target latitude and longitude, wherein one side edge of the natural object type target is grounded; and carrying out ground binocular positioning of the corresponding target rectangular block area against a non-natural object type target to obtain a three dimensional coordinate, and transiting to a target latitude and longitude, wherein each of the side edges of the non-natural object type target is not grounded.

Description

Monocular localization method and system based on installed video
Technical field
The present invention relates to the traverse measurement field, relate in particular to a kind of monocular localization method and system based on installed video.
Background technology
The vehicle-mounted mobile measuring system can be taken live-action image static city parts are positioned and measure, be applied to the city management parts and collect management, yet ubiquity efficiency is low at present, automaticity is not high, acquisition cost is high, upgrades the problems such as movement properties difficult, that can't obtain object.Classic method utilizes manual observation of manpower to select attention object, select the point on two images to carry out the binocular location, all need 2 or more images to position atural object at every turn, this has not only greatly reduced target and has collected efficiency, also improve acquisition cost, and be difficult to the movement velocity direction of Quick Measurement object.Therefore in the urgent need to a kind of vehicle-mounted automatic collection, locate the dynamic live-action image method for measurement of targets of interest.
Summary of the invention
The object of the invention is to overcome above-mentioned technical disadvantages, a kind of monocular location technology scheme based on installed video is provided.
Technical scheme of the present invention is a kind of monocular localization method based on installed video, comprises off-line procedure and at line process,
Off-line procedure, comprise the training objective master drawing, sets up target classification device storehouse;
At line process, comprise the interesting target in based target sorter storehouse detection input picture, obtain respective objects rectangular block zone; For the target of type of ground objects, carry out monocular location, ground according to respective objects rectangular block zone and obtain the coordinate in three-dimensional scenic, finally convert the longitude and latitude of target to; The target of described type of ground objects, a certain lateral edges ground connection of feeling the pulse with the finger-tip target.
And ground monocular location implementation is as follows,
Set up road surface intersection point coordinate system, from roof GPS, do a vertical line perpendicular to ground, the ground intersection point O of usining sets up intersection point O coordinate system X-Y-Z as initial point, this coordinate system X-Z axle is close to ground, X-axis is with respect to car towards the right side, and Z axis points to the dead ahead of car, and Y-axis is pointed to underground perpendicular to ground; h gpsfor the height of roof GPS with respect to ground;
The lower edge in described target rectangle piece zone is in ground, and the upper left corner coordinate in target rectangle piece zone is u, v, and the frame in target rectangle piece zone is wide=w, and frame is high=h;
Attitude transition matrix R for the ground intersection point picture-Owith the displacement transition matrix T of camera with respect to the ground intersection point picture-O, decompose the matrix H that obtains 3 * 3 o,
H o3x3=sK[r picture-Che 1r picture-Che 3-R picture-Chet picture-O]=s[r picture-Che 1r picture-Che 3-R picture-Che(T picture-Che+ T car-O)]
Wherein, s is Arbitrary Coefficient, and K is camera internal reference matrix, R picture-Chethe attitude transition matrix of camera with respect to car, r picture-Che 1and r picture-Che 3be respectively matrix R picture-Chethe 1st row and the 3rd be listed as, T picture-Chethe displacement transition matrix of camera with respect to car, T car-Othe displacement transition matrix of car with respect to the ground intersection point;
If the object space three-dimensional coordinate in intersection point O coordinate system X-Y-Z is (X, Y, Z), by the ground intersection point, as coordinate u+w/2, v+h/2 calculates X, and Z is as follows,
[X?Z?1] T=zH O -1[u+w/2?v+h?1] T
Wherein, z is unknowm coefficient, by 3 equations of above-mentioned equation row, solves;
Gained X, Z is [X-h with respect to the car gps coordinate gpsz], [X-h gpsz] be converted to again and measure the city coordinate system, finally conversion obtains the ground latitude and longitude coordinates to terrestrial coordinate system.
And, for the target of non-type of ground objects, carry out binocular location, ground according to respective objects rectangular block zone and obtain the coordinate in three-dimensional scenic, finally convert the longitude and latitude of target to; The target of described non-type of ground objects, the feeling the pulse with the finger-tip mark does not have either side edge ground connection;
The upper left corner coordinate of supposing the target rectangle piece zone that detects on image 1 is u 1, v 1, frame is wide=w, and frame is high=h, obtains the coordinate x of center on image 1 in target rectangle piece zone 1=u 1+ w/2, y 1=v 1+ h/2 carries out match search on another image 2, obtains the coordinate x of the respective objects rectangular block regional center on image 2 2=u 2+ w/2, y 2=v 2+ h/2,
Separate the object space three-dimensional coordinate X that following equation obtains the object center, Y, Z
M 111 - x 1 M 131 M 112 - x 1 M 132 M 113 - x 1 M 133 M 121 - y 1 M 131 M 122 - y 1 M 132 M 123 - y 1 M 133 M 211 - x 2 M 231 M 212 - x 2 M 232 M 213 - x 2 M 233 M 221 - y 2 M 231 M 222 - y 2 M 232 M 223 - y 2 M 233 X Y Z = x 1 m 13 - m 11 y 1 m 13 - m 12 x 2 m 23 - m 21 y 2 m 23 - m 22
[M wherein im i] built the projection matrix of i camera
M i=K iR i?m i=-K iR iT i?i=1,2
M ithe matrix of 3 row 3 row, m ithe matrix of 3 row 1 row, M ircmean i camera M ithe capable c train value of the r of matrix, m irmean i camera m icapable 1 train value of the r of matrix; K iit is the inner parameter matrix of i camera;
Gained [X Y Z] is converted to measures the city coordinate system, and finally conversion obtains the ground latitude and longitude coordinates to terrestrial coordinate system.
The present invention is also corresponding provides a kind of monocular positioning system based on installed video, comprises off-line part and online part,
The off-line part, comprise the training module for the training objective master drawing, for setting up the classifier modules in target classification device storehouse;
Online part, comprise with lower module,
Detection module, the interesting target for based target sorter storehouse detection input picture, obtain respective objects rectangular block zone; The monocular locating module, for the target for type of ground objects, carry out monocular location, ground according to respective objects rectangular block zone and obtain the coordinate in three-dimensional scenic, finally converts the longitude and latitude of target to; The target of described type of ground objects, a certain lateral edges ground connection of feeling the pulse with the finger-tip target.
And online part comprises the binocular locating module, for the target for non-type of ground objects, carry out binocular location, ground according to respective objects rectangular block zone and obtain the coordinate in three-dimensional scenic, finally convert the longitude and latitude of target to; The target of described non-type of ground objects, the feeling the pulse with the finger-tip mark does not have either side edge ground connection.
Monocular location technology scheme effect based on installed video provided by the invention is:
1. can locate in real time the movement of the street and static attention object with measuring car, for ground object target, can directly carry out list as the objective location, ground object target does not need binocular to locate; Realize simply, the corresponding system device can be arranged on to be measured on car.
2. versatility is good, and the work off-line of setting up object library completes, can support the batch processing of various interesting targets to collect;
3. the binocular location, ground to non-ground object target be can expand, ground object target and non-ground object target two class location supported simultaneously, efficient and convenient.
The accompanying drawing explanation
Fig. 1 is the schematic flow sheet of the embodiment of the present invention.
Fig. 2 is the schematic diagram of the center of the going to the bottom corresponding ground intersection point of ground object target frame on the road ground image of the embodiment of the present invention.
Embodiment
The monocular location that the present invention introduces ground object target combines with the binocular location of non-atural object, and adopts the mode establishing target sorter storehouse of sorter training, in order to support automatic collection and the three-dimensional fix of target, has greatly improved efficiency.According to the position characteristics of target, target is divided into to atural object and the large class of non-atural object two.
Ground object target refers to a certain lateral edges ground connection of target, comprises the pedestrian, vehicle, and well lid, wire pole tower, the building construction of power transformer, Stall ground connection, the shop front, the stall with goods spread out on the ground for sale retailer who goes up in the street waits the ground connection target.
Non-ground object target feeling the pulse with the finger-tip mark, without any a lateral edges ground connection, comprises traffic mark board, the high-rise architecture against regulations, traffic lights, various billboards, bank's signboard, government department's signboard, people's face.
Describe technical solution of the present invention in detail below in conjunction with drawings and Examples.
The present invention can adopt software engineering to realize operation automatically, and defiber and online two parts process are carried out, as Fig. 1:
Off-line procedure, comprise the training objective master drawing, set up target classification device storehouse: the off-line part of usually carrying out in advance can utilize the Haar features training in OpenCV to obtain the Cascade sorter, extract the Haar feature of sample image, adopt cascade sorter training algorithm to be trained great amount of samples, obtain sorter, specific implementation can be with reference to prior art, and it will not go into details in the present invention.Can adopt different classifiers combination constituent class devices storehouse for identifying various interesting target.In embodiment, at first collect the image of a large amount of targets of interest, trained, each group target image sample obtains a sorter, organizes the different target training more and obtains sorter set C 1... C nrepresentative is the sorter of n target altogether, forms target classification device storehouse, wherein sorter C ican be used for judging in image, whether a rectangle frame target belongs to i class target, the value of i is 1 ... n.During concrete enforcement, training can adopt the software modularity mode, in the mode of training module, provides, and sample image input training module, obtain sorter.
At line process, comprise the interesting target in based target sorter storehouse detection input picture, obtain respective objects rectangular block zone.The present invention proposes the monocular location for the type of ground objects target, while specifically implementing, for improving range of application of the present invention, can integrated existing binocular location technology.
For the target of type of ground objects, carry out monocular location, ground according to respective objects rectangular block zone and obtain the coordinate in three-dimensional scenic, finally convert the longitude and latitude of target to;
For the target of non-type of ground objects, carry out binocular location, ground according to respective objects rectangular block zone and obtain the coordinate in three-dimensional scenic, finally convert the longitude and latitude of target to.
The realization of the online part of embodiment is described as follows:
(1) at first carry out target detection, based on sorter i scanning full figure, search obtains target frame set rect i1, rect i2, rect i3Rect ijmean that the i classification is marked on j the target found on image.The rectangle frame that the minimum area-encasing rectangle that the target frame is target forms.
(2) whether target localization, comprise the attribute according to sorter i, be ground object target, these rectangle frames entered respectively to (2.1) (2.2) and processed.
(2.1) the installed video ground object list picture of embodiment location realizes being described as follows:
Positioning result obtains the ground latitude and longitude coordinates according to image ground object target coordinate (generally adopt upper left corner coordinate u, v is as image origin) exactly.To ground object, target detection as a result gained target frame lower limb just in time on the ground, as pedestrian's pin stands in ground, rectangle frame just comprises pedestrian's pin.As Fig. 2, in the middle of highway sideline, it is pavement of road, the target frame that target rectangle piece zone forms is positioned on road, the upper left corner coordinate in target rectangle piece zone is u, v, and the frame in target rectangle piece zone is wide=w, frame is high=h, in Fig. 2, the go to the bottom cross at center of rectangle frame means intersection point, and the center point coordinate of going to the bottom in target rectangle piece zone is u+w/2, v+h.
Set up road surface intersection point coordinate system, from roof GPS, do a vertical line perpendicular to ground.The ground intersection point O of usining sets up intersection point O coordinate system X-Y-Z as initial point.This coordinate system X-Z axle is close to ground, and X-axis is with respect to car towards the right side, and Z axis points to the dead ahead of car, and Y-axis is pointed to underground perpendicular to ground.
Suppose that GPS device on car is fixed as h with respect to the height on ground gps(be on vertical line O to the distance h of GPS gps).
Under ideal state, it is certain that the posture position relation on onboard system and ground keeps, and camera is with respect to the T of car picture-Che, R picture-Chealso fixing, so
T picture-O=T picture-Che+ T car-O
R picture-O=R picture-Cher car-O=R picture-Che
T wherein car-O=[0 0-h gps] R car-O=I
T picture-Chethe displacement transition matrix of camera with respect to car, T picture-Othe displacement transition matrix of camera with respect to the ground intersection point, T car -Othe displacement transition matrix of car with respect to the ground intersection point;
R picture-Chethe attitude transition matrix of camera with respect to car, R picture-Othe attitude transition matrix of camera with respect to the ground intersection point, R car -Othe attitude transition matrix of car with respect to the ground intersection point.
So by R picture-Oand T picture-Ocan decompose the H that obtains 3 * 3 omatrix,
H o3x3=sK[r picture-Che 1r picture-Che 3-R picture-Chet picture-O]=s[r picture-Che 1r picture-Che 3-R picture-Che(T picture-Che+ T car-O)]
S is Arbitrary Coefficient, and K is camera internal reference matrix,
K = f u 0 c u 0 f v c v 0 0 1
C u, c vthe principal point of camera, f u, f vthe focal length pixel that is camera is long;
R picture-Che 1and r picture-Che 3be respectively matrix R picture-Chethe 1st row and the 3rd be listed as
T picture-Chethe fixed amount on earth of calibration, therefore no matter which road surface, plane garage sails on, H oall fix.H o3x3reflected that ground any point in the intersection point coordinate system is to the mapping relations on image.H conversely o -1reflected that any point on the road surface on the image is to the topocentric mapping relations in the intersection point coordinate system.
The object space three-dimensional coordinate of setting a trap in the road surface coordinate system X-Y-Z of section is (X, Y, Z), and by the ground intersection point, as coordinate u+w/2, v+h calculates X, and Z is as follows,
[X?Z?1] T=zH O -1[u+w/2?v+h?1] T
Z is a unknowm coefficient, can solve by 3 equations of top equation row.(two of left and right of above formula equation 3x1 vector equates, equation of every a line equation row, three equations just in time solve three unknown number X Z z)
Note X, Z is that the coordinate be converted to respect to the car gps coordinate is [X-h with respect to local road surface intersection point coordinate system gpsz].[X-h gpsz] be converted to again and measure the city coordinate system, finally conversion can obtain the ground latitude and longitude coordinates to terrestrial coordinate system.
(2.2) for the sake of ease of implementation, introduce the non-ground object fast locating algorithm of installed video as follows:
Positioning result be exactly according to the non-ground object target rectangle frame of image as coordinate (adopt upper left corner coordinate u, v) reach the wide w of frame, the high h of frame obtains the ground latitude and longitude coordinates.
If the 1st camera taken 1, the 2 camera of image and taken image 2, the upper left corner coordinate of hypothetical target frame on image i is u i, v i, frame is wide=w i, frame is high=h i, the value of i is 1,2.Two camera target image size approximately equals of taking side by side, make w 1=w 2=w, h 1=h 2=w.
The upper left corner coordinate of supposing to detect the target frame on image 1 is u 1, v 1, frame is wide=w, and frame is high=h, now also the unnecessary rectangle frame lower limb of asking is positioned at ground.The coordinate of target Kuang center on image 1 is x 1=u 1+ w/2, y 1=v 1+ h/2.
On another image 2 along the core line pixel distance D that fluctuates, D arranges (such as D=0.01 figure image height) in proportion according to the image size, carry out the rectangular block search of the high h pixel of wide w in this belt-like zone, the rectangular block center of searching for and the pixel distance≤D of core line, can adopt prior art to calculate the rectangular block of search and the coupling facies relationship numerical value of the target frame on image 1, if maximum coupling facies relationship numerical value has surpassed a certain predetermined threshold value, maximum is mated to the match bit of rectangle frame position on image 2 that facies relationship numerical value is corresponding, otherwise illustrate and do not search match bit.So obtain the target frame centre coordinate x on image 2 2=u 2+ w/2, y 2=v 2+ h/2.
The following equation of binocular stereo imaging solution obtains the object space three-dimensional coordinate X at object center, Y, Z
M 111 - x 1 M 131 M 112 - x 1 M 132 M 113 - x 1 M 133 M 121 - y 1 M 131 M 122 - y 1 M 132 M 123 - y 1 M 133 M 211 - x 2 M 231 M 212 - x 2 M 232 M 213 - x 2 M 233 M 221 - y 2 M 231 M 222 - y 2 M 232 M 223 - y 2 M 233 X Y Z = x 1 m 13 - m 11 y 1 m 13 - m 12 x 2 m 23 - m 21 y 2 m 23 - m 22
[M wherein im i] built the projection matrix of i camera
M i=K iR i?m i=-K iR iT i?i=1,2
M ithe matrix of 3 row 3 row, m ithe matrix of 3 row 1 row, M ircmean i camera M ithe capable c train value of the r of matrix, m irmean i camera m icapable 1 train value of the r of matrix.
K ithe inner parameter matrix of i camera,
K i = f ui 0 c ui 0 f vi c ui 0 0 1
C ui, c vithe principal point of camera i, f ui, f vithe focal length pixel that is camera i is long;
R iit is the attitude matrix of i camera;
T ithe location matrix of i camera, i.e. camera photocentre position.
Note X, Y, Z is the coordinate with respect to the car gps coordinate.[X Y Z] is converted to and measures the city coordinate system, and finally conversion can obtain the ground latitude and longitude coordinates to terrestrial coordinate system.
During concrete enforcement, can adopt the software solidification mode that the corresponding monocular positioning system based on installed video is provided, mainly be divided into off-line and online two large divisions, off-line part training objective master drawing, set up target classification device storehouse; Online part detects interesting target in real time according to the sorter storehouse.Off-line partly comprises training module and the classifier modules of sample image, is mainly to set up the sorter storehouse of interesting target by machine learning according to training sample image.Online part inclusion test module, monocular locating module also can arrange the binocular locating module:
It is guidance that detection module be take target classification device storehouse, and the scanning full figure, detect the interesting target in input picture automatically, obtains target rectangle piece zone.
The monocular locating module is for ground object target, and the target rectangle piece zone obtained according to detection module is carried out monocular location, ground and obtained the coordinate in three-dimensional scenic, finally converts the longitude and latitude of target to.
The binocular locating module is for non-ground object target, and the target rectangle piece zone obtained according to detection module is carried out binocular location, ground and obtained the coordinate in three-dimensional scenic, finally converts the longitude and latitude of target to.
Each module specific implementation is corresponding with method, and it will not go into details in the present invention.
Specific embodiment described herein is only to the explanation for example of the present invention's spirit.Those skilled in the art can make various modifications or supplement or adopt similar mode to substitute described specific embodiment, but can't depart from spirit of the present invention or surmount the defined scope of appended claims.

Claims (5)

1. the monocular localization method based on installed video is characterized in that: comprise off-line procedure and at line process,
Off-line procedure, comprise the training objective master drawing, sets up target classification device storehouse;
At line process, comprise the interesting target in based target sorter storehouse detection input picture, obtain respective objects rectangular block zone; For the target of type of ground objects, carry out monocular location, ground according to respective objects rectangular block zone and obtain the coordinate in three-dimensional scenic, finally convert the longitude and latitude of target to; The target of described type of ground objects, a certain lateral edges ground connection of feeling the pulse with the finger-tip target.
2. the monocular localization method based on installed video according to claim 1, it is characterized in that: ground monocular location implementation is as follows,
Set up road surface intersection point coordinate system, from roof GPS, do a vertical line perpendicular to ground, the ground intersection point O of usining sets up intersection point O coordinate system X-Y-Z as initial point, this coordinate system X-Z axle is close to ground, X-axis is with respect to car towards the right side, and Z axis points to the dead ahead of car, and Y-axis is pointed to underground perpendicular to ground; h gpsfor the height of roof GPS with respect to ground;
The lower edge in described target rectangle piece zone is in ground, and the upper left corner coordinate in target rectangle piece zone is u, v, and the frame in target rectangle piece zone is wide=w, and frame is high=h;
Attitude transition matrix R for the ground intersection point picture-Owith the displacement transition matrix T of camera with respect to the ground intersection point picture-O, decompose the matrix H that obtains 3 * 3 o,
H o3x3=sK[r picture-Che 1r picture-Che 3-R picture-Chet picture-O]=s[r picture-Che 1r picture-Che 3-R picture-Che(T picture-Che+ T car-O)]
Wherein, s is Arbitrary Coefficient, and K is camera internal reference matrix, R picture-Chethe attitude transition matrix of camera with respect to car, r picture-Che 1and r picture-Che 3be respectively matrix R picture-Chethe 1st row and the 3rd be listed as, T picture-Chethe displacement transition matrix of camera with respect to car, T car-Othe displacement transition matrix of car with respect to the ground intersection point;
If the object space three-dimensional coordinate in intersection point O coordinate system X-Y-Z is (X, Y, Z), by the ground intersection point, as coordinate u+w/2, v+h calculates X, and Z is as follows,
[X?Z?1] T=zH O -1[u+w/2?v+h?1] T
Wherein, z is unknowm coefficient, by 3 equations of above-mentioned equation row, solves;
Gained X, Z is [X-h with respect to the car gps coordinate gpsz], [X-h gpsz] be converted to again and measure the city coordinate system, finally conversion obtains the ground latitude and longitude coordinates to terrestrial coordinate system.
3. the monocular localization method based on installed video according to claim 2, it is characterized in that: for the target of non-type of ground objects, carry out binocular location, ground according to respective objects rectangular block zone and obtain the coordinate in three-dimensional scenic, finally convert the longitude and latitude of target to; The target of described non-type of ground objects, the feeling the pulse with the finger-tip mark does not have either side edge ground connection;
Binocular location, ground implementation is as follows,
The upper left corner coordinate of supposing the target rectangle piece zone that detects on image 1 is u 1, v 1, frame is wide=w, and frame is high=h, obtains the coordinate x of center on image 1 in target rectangle piece zone 1=u 1+ w/2, y 1=v 1+ h/2 carries out match search on another image 2, obtains the coordinate x of the respective objects rectangular block regional center on image 2 2=u 2+ w/2, y 2=v 2+ h/2,
Separate the object space three-dimensional coordinate X that following equation obtains the object center, Y, Z
M 111 - x 1 M 131 M 112 - x 1 M 132 M 113 - x 1 M 133 M 121 - y 1 M 131 M 122 - y 1 M 132 M 123 - y 1 M 133 M 211 - x 2 M 231 M 212 - x 2 M 232 M 213 - x 2 M 233 M 221 - y 2 M 231 M 222 - y 2 M 232 M 223 - y 2 M 233 X Y Z = x 1 m 13 - m 11 y 1 m 13 - m 12 x 2 m 23 - m 21 y 2 m 23 - m 22
[M wherein im i] built the projection matrix of i camera
M i=K iR i?m i=-K iR iT i?i=1,2
M ithe matrix of 3 row 3 row, m ithe matrix of 3 row 1 row, M ircmean i camera M ithe capable c train value of the r of matrix, m irmean i camera m icapable 1 train value of the r of matrix; K iit is the inner parameter matrix of i camera;
Gained [X Y Z] is converted to measures the city coordinate system, and finally conversion obtains the ground latitude and longitude coordinates to terrestrial coordinate system.
4. the monocular positioning system based on installed video is characterized in that: comprise off-line part and online part,
The off-line part, comprise the training module for the training objective master drawing, for setting up the classifier modules in target classification device storehouse;
Online part, comprise with lower module,
Detection module, the interesting target for based target sorter storehouse detection input picture, obtain respective objects rectangular block zone; The monocular locating module, for the target for type of ground objects, carry out monocular location, ground according to respective objects rectangular block zone and obtain the coordinate in three-dimensional scenic, finally converts the longitude and latitude of target to; The target of described type of ground objects, a certain lateral edges ground connection of feeling the pulse with the finger-tip target.
5. the monocular positioning system based on installed video according to claim 4, it is characterized in that: online part comprises the binocular locating module, for the target for non-type of ground objects, carry out binocular location, ground according to respective objects rectangular block zone and obtain the coordinate in three-dimensional scenic, finally convert the longitude and latitude of target to; The target of described non-type of ground objects, the feeling the pulse with the finger-tip mark does not have either side edge ground connection.
CN201310472446.XA 2013-10-11 2013-10-11 Based on monocular localization method and the system of installed video Active CN103487035B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201310472446.XA CN103487035B (en) 2013-10-11 2013-10-11 Based on monocular localization method and the system of installed video

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201310472446.XA CN103487035B (en) 2013-10-11 2013-10-11 Based on monocular localization method and the system of installed video

Publications (2)

Publication Number Publication Date
CN103487035A true CN103487035A (en) 2014-01-01
CN103487035B CN103487035B (en) 2015-10-21

Family

ID=49827450

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201310472446.XA Active CN103487035B (en) 2013-10-11 2013-10-11 Based on monocular localization method and the system of installed video

Country Status (1)

Country Link
CN (1) CN103487035B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105512646A (en) * 2016-01-19 2016-04-20 腾讯科技（深圳）有限公司 Data processing method, data processing device and terminal
CN110243339A (en) * 2019-06-25 2019-09-17 重庆紫光华山智安科技有限公司 A kind of monocular cam localization method, device, readable storage medium storing program for executing and electric terminal

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006010610A (en) * 2004-06-29 2006-01-12 Tokyo Univ Of Science Three-dimensional positioning device and method by monocular vision
CN101441769A (en) * 2008-12-11 2009-05-27 上海交通大学 Real time vision positioning method of monocular camera
US20110299730A1 (en) * 2010-03-16 2011-12-08 Elinas Pantelis Vehicle localization in open-pit mining using gps and monocular camera

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006010610A (en) * 2004-06-29 2006-01-12 Tokyo Univ Of Science Three-dimensional positioning device and method by monocular vision
CN101441769A (en) * 2008-12-11 2009-05-27 上海交通大学 Real time vision positioning method of monocular camera
US20110299730A1 (en) * 2010-03-16 2011-12-08 Elinas Pantelis Vehicle localization in open-pit mining using gps and monocular camera

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
赵天云等: "基于单目视觉的空间定位算法", 《西北工业大学学报》, vol. 27, no. 1, 28 February 2009 (2009-02-28), pages 47 - 51 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105512646A (en) * 2016-01-19 2016-04-20 腾讯科技（深圳）有限公司 Data processing method, data processing device and terminal
CN110243339A (en) * 2019-06-25 2019-09-17 重庆紫光华山智安科技有限公司 A kind of monocular cam localization method, device, readable storage medium storing program for executing and electric terminal

Also Published As

Publication number Publication date
CN103487035B (en) 2015-10-21

Similar Documents

Publication Publication Date Title
US20200041284A1 (en) Map road marking and road quality collecting apparatus and method based on adas system
Huang et al. The apolloscape dataset for autonomous driving
CN110174093B (en) Positioning method, device, equipment and computer readable storage medium
CN103559791B (en) A kind of vehicle checking method merging radar and ccd video camera signal
CN104280036B (en) A kind of detection of transport information and localization method, device and electronic equipment
CN105930819A (en) System for real-time identifying urban traffic lights based on single eye vision and GPS integrated navigation system
EP3650814A1 (en) Vision augmented navigation
JP5388082B2 (en) Stationary object map generator
CN106127153A (en) The traffic sign recognition methods of Vehicle-borne Laser Scanning cloud data
CN103625649B (en) A kind of aircraft independent landing area judging method
CN111275960A (en) Traffic road condition analysis method, system and camera
CN111179300A (en) Method, apparatus, system, device and storage medium for obstacle detection
CN110119698A (en) For determining the method, apparatus, equipment and storage medium of Obj State
CN102208013A (en) Scene matching reference data generation system and position measurement system
CN102435174A (en) Method and device for detecting barrier based on hybrid binocular vision
CN114359181B (en) Intelligent traffic target fusion detection method and system based on image and point cloud
JP2012208525A (en) Stationary object map generation device
KR20210081983A (en) System and method of Automatically Generating High Definition Map Based on Camera Images
JP6298221B2 (en) Path detection system based on solar blind ultraviolet light signal
Vu et al. Traffic sign detection, state estimation, and identification using onboard sensors
CN110210384A (en) A kind of road global information extract real-time and indicate system
KR101510745B1 (en) Autonomous vehicle system
CN114639085A (en) Traffic signal lamp identification method and device, computer equipment and storage medium
CN111506069A (en) All-weather all-ground crane obstacle identification system and method
CN103093214B (en) A kind of pedestrian detection method based on vehicle mounted infrared camera

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant