CN105014675B - A kind of narrow space intelligent mobile robot vision navigation system and method - Google Patents
A kind of narrow space intelligent mobile robot vision navigation system and method Download PDFInfo
- Publication number
- CN105014675B CN105014675B CN201410281773.1A CN201410281773A CN105014675B CN 105014675 B CN105014675 B CN 105014675B CN 201410281773 A CN201410281773 A CN 201410281773A CN 105014675 B CN105014675 B CN 105014675B
- Authority
- CN
- China
- Prior art keywords
- image
- cross
- location
- point
- analyzing
- 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
Links
Abstract
The invention provides a kind of narrow space intelligent mobile robot vision navigation system and method.System includes image acquisition end and analyzing and processing end, image acquisition end includes cross hairs laser projecting apparatus, the first microcam and the second microcam, control unit, and method comprises the steps: that cross hairs laser projecting apparatus projects cross hairs optical strip image；First microcam and the second microcam gather cross optical strip image respectively, and export cross optical strip image to analyzing and processing end；Image is analyzed by analyzing and processing end, draws the three-dimensional coordinate set of Rhizoma Dioscoreae (peeled)；Analyzing and processing end, according to the three-dimensional coordinate set of Rhizoma Dioscoreae (peeled), analyzes the regional location of the barrier that front occurs；Regional location is back to the control unit of image acquisition end by analyzing and processing end, completes to hide barrier controlling robot.The automatic obstacle avoiding navigation of robot in narrow working environment complicated, unknown, can be realized exactly according to the present invention.
Description
Technical field
The present invention relates to technical field of visual navigation, particularly one narrow space intelligent mobile robot vision guided navigation system
System and method.
Background technology
Intelligent mobile robot is that a class obtains ambient condition information and oneself state by self-contained sensor, complete
Become detection and the identification of target of barrier in environment, thus realize from original position autonomic movement to target location and right
Target object carries out the robot system of operation task.Itself there is the modules such as complete perception, analysis, decision-making and execution,
Production activity can be engaged in the most in the environment as autonomous mankind.The vision of intelligent robot under circumstances not known
Airmanship relates to artificial intelligence and divides with pattern recognition theory, Theory of Automatic Control, sensor technology, computer technology and image
Analysis and treatment technology etc., have become as artificial intelligence and the Frontier problem of sophisticated machine people's research field, be current intelligence
One of the focus of robot development's research and difficulties.
In known or structurized environment, the research of mobile robot autonomous navigation control theory and method achieved with
A lot of research and application achievement.In the actual application of current and future, mobile robot needs often to be operated in the unknown
Environment, and at present the robot autonomous navigation problem in the unknown or unstructured moving grids is the most well solved
Certainly.In mobile robot autonomous navigation technology, conventional sensor includes ultrasonic sensor, infrared sensor, Laser Measuring at present
Away from radar, chemical sensor and vision sensor.Compared with other sensor, vision sensor has higher intellectuality and place
The advantage of reason speed, and the scarcity of priori in circumstances not known and the uncertainty of environment etc. factor can be overcome, because of
The application in mobile robot of this vision sensor is increasingly subject to people's attention.
But, existing vision navigation system cannot realize identifying that motion front goes out accurately in narrow and small complex environment
Existing barrier and target object.
Accordingly, it would be desirable to a kind of narrow space intelligent mobile robot vision navigation system and method, to adapt to narrow and small work
Make environment, realize exactly unknown barrier and target object are detected and identified.
Summary of the invention
It is an object of the invention to provide a kind of narrow space intelligent mobile robot vision navigation system and method.
According to an aspect of the invention, it is provided a kind of narrow space intelligent mobile robot vision navigation system, bag
Include: image acquisition end, be used for gathering cross optical strip image, and by described cross optical strip image to analyzing and processing end output, described
Image acquisition end includes: cross hairs laser projecting apparatus, is used for projecting described cross optical strip image；First microcam and
Two microcams, project the described cross optical strip image on barrier for shooting respectively and obtain the first location image and the
Two location images, and described cross optical strip image is transmitted to described analyzing and processing end；Control unit, for according to from described point
Analysis processes the data message representing Obstacle Position that end receives, and controls the traveling of described image acquisition end；Analyzing and processing end, uses
According to the described cross optical strip image inputted respectively from the first microcam and described second microcam, complete ten
Word Rhizoma Dioscoreae (peeled) coupling on the first location image and the second location image, with the decussation point of described cross Rhizoma Dioscoreae (peeled) as boundary,
To lay respectively at described first location image and described second location image described in decussation point phase the same side Rhizoma Dioscoreae (peeled) in
Heart coordinate set, and, lay respectively at decussation point phase described in described first location image and described second location image
With the optical losses coordinate set of other side, analyze the three-dimensional coordinate set obtaining described cross optical strip image, and judge
The region of existing barrier, is finally back to described control unit by described region, to control the traveling of described image acquisition end.
Preferably, described analyzing and processing end, with the decussation point of described cross Rhizoma Dioscoreae (peeled) as boundary, obtain laying respectively at described
The optical losses coordinate set of decussation point phase the same side described in first location image and described second location image, with
And, lay respectively at the light of described first location image and the decussation identical other side of point described in described second location image
Bar centre coordinate set, algorithm be: a) described first position image and described second location image extract respectively described
Cross optical strip image, and choose gray threshold by described cross optical strip image binaryzation；B) image and institute are positioned described first
State in the second location image, scan described cross optical strip image, with described decussation point as boundary, obtain laying respectively at described
All Rhizoma Dioscoreae (peeled) collection of pixels of decussation point phase the same side in certain bit image and described second location image, and, respectively
It is positioned at described first location image and all Rhizoma Dioscoreae (peeled) pixels of the decussation identical other side of point in described second location image
Set；And c) obtain Rhizoma Dioscoreae (peeled) pixel center coordinate, thus extract and lay respectively at described first location image and described second calmly
The optical losses coordinate set of the some phase of decussation described in bit image the same side, and, lay respectively at described first location figure
Picture positions the optical losses coordinate set of the decussation identical other side of point described in image with described second.
Preferably, described analyzing and processing end determine according to formula (1) described cross optical strip image any point three
Dimension coordinate (x, y, z):
Wherein, αx,αyIt is respectively camera x, the effective focal length in y-axis, (u0,v0) it is the principal point image coordinate of camera, b is
Parallax range, and i ∈ [1,2], 1,2 represent respectively: lay respectively in described first location image and described second location image
The sequence number of the optical losses coordinate set of described decussation point phase the same side, and, lay respectively at described first location image
Sequence number with the optical losses coordinate set of the decussation identical other side of point described in described second location image.
Preferably, described analyzing and processing end analysis occurs that the algorithm in region of barrier is: first direction of advance divided
For upper left, upper right, bottom right, and region, four, lower-left, add up described four regions the most respectively at zPOwning in the range of≤H
The number of coordinate set is { NA,NB,NC,ND, wherein H is distance alarm threshold value, works as NiDuring >=TH, i ∈ (A, B, C, D), then recognize
Being to occur in that barrier in the i-th region, wherein, TH is the threshold value of coordinate number.
Preferably, the size of described image acquisition end is less than 50mm3。
Preferably, described first microcam and the second microcam are fixed on same horizontal plane.
Preferably, described first microcam and the second microcam are provided with LED light source.
According to a further aspect in the invention, it is provided that a kind of narrow space intelligent mobile robot vision navigation method, root
Including image acquisition end and analyzing and processing end according to the system of described method, described image acquisition end includes cross hairs laser projections
Device, the first microcam and the second microcam, control unit, described method comprises the steps: a) described cross
Line laser projector projection cross hairs optical strip image；B) described first microcam and described second microcam are respectively
Gather described cross optical strip image, and by the output of described cross optical strip image to described analyzing and processing end；C) described analyzing and processing
Image is analyzed by end, draws the three-dimensional coordinate set { x of Rhizoma Dioscoreae (peeled)P,yP,zP}；D) described analyzing and processing end is according to the three of Rhizoma Dioscoreae (peeled)
Dimension coordinate set { xP,yP,zP, analyze the regional location of the barrier that front occurs；And e) described analyzing and processing end by described
Regional location is back to the described control unit of described image acquisition end, completes to hide barrier controlling robot.
Preferably, in step c, image is analyzed by described analyzing and processing end, draws the three of described cross optical strip image
Dimension coordinate set { xP,yP,zPMethod comprise the steps: c1) described first position image and described second location image
Middle extract described cross optical strip image respectively, and choose gray threshold by described cross optical strip image binaryzation；C2) described
In certain bit image and described second location image, scan described cross optical strip image, with described decussation point as boundary, obtain
Lay respectively at all Rhizoma Dioscoreae (peeled) pixels of decussation point phase the same side in described first location image and described second location image
Set, and, lay respectively at described first location image and the decussation identical other side of point in described second location image
All Rhizoma Dioscoreae (peeled) collection of pixels；C3) obtain Rhizoma Dioscoreae (peeled) pixel center coordinate, thus extract lay respectively at described first location image and
The optical losses coordinate set of decussation point phase the same side described in described second location image, and, lay respectively at described
First location image positions the optical losses coordinate set of the decussation identical other side of point described in image with described second；
And, c4) determine that according to formula (1) (x, y z), thus obtain for the three-dimensional coordinate of any point of described cross optical strip image
Three-dimensional coordinate set { the x of described cross optical strip imageP,yP,zP}:
Wherein, αx,αyIt is respectively camera x, the effective focal length in y-axis, (u0,v0) it is the principal point image coordinate of camera, b is
Parallax range, and i ∈ [1,2], 1,2 represent respectively: lay respectively in described first location image and described second location image
The sequence number of the optical losses coordinate set of described decussation point phase the same side, and, lay respectively at described first location image
Sequence number with the optical losses coordinate set of the decussation identical other side of point described in described second location image.
Preferably, in step d, analyzing and processing end analysis occurs that the algorithm in the region of barrier is: first by direction of advance
Be divided into top left region A | xP∈[0,b/2],yP>=0}, right regions B | xP∈(b/2,b],yP>=0}, lower right area C | xP
∈(b/2,b],yP< 0}, and lower left region D | xP∈[0,b/2],yP< this four districts are added up in tetra-regions of 0} the most respectively
Territory is at zPThe number of all coordinate sets in the range of≤H is { NA,NB,NC,ND, wherein H is distance alarm threshold value, works as Ni≥TH
Time, it is believed that occurring in that barrier in the i-th region, wherein, i ∈ (A, B, C, D), TH are the threshold value of coordinate number.
Narrow space intelligent mobile robot vision navigation system according to the present invention and method, it is possible to narrow for the unknown
Assembly space moves robot self-navigation and the particularity of avoidance and required precision, in narrow working environment complicated, unknown,
Realize exactly unknown barrier and target object are detected and identified, the Autonomous Control of robot, and independently keep away
Barrier navigation.
Accompanying drawing explanation
With reference to the accompanying drawing enclosed, as follows by by embodiment of the present invention of the more purpose of the present invention, function and advantage
Description is illustrated, wherein:
Fig. 1 diagrammatically illustrates narrow space intelligent mobile robot vision navigation system and the structure of method of the present invention
Figure.
Fig. 2 diagrammatically illustrates the axonometric chart of image acquisition end.
Fig. 3 diagrammatically illustrates the method flow diagram that analyzing and processing end is analyzed processing.
Fig. 4 diagrammatically illustrates the first microcam and the schematic diagram of the second microcam coordinate system.
Fig. 5 diagrammatically illustrates the schematic diagram of Rhizoma Dioscoreae (peeled) region pre-stored.
Fig. 6 diagrammatically illustrates the schematic diagram of a coordinate system representing barrier region.
Fig. 7 diagrammatically illustrates the flow process of the narrow space intelligent mobile robot vision navigation method according to the present invention
Figure.
Detailed description of the invention
By with reference to one exemplary embodiment, the purpose of the present invention and function and for realizing the side of these purposes and function
Method will be illustrated.But, the present invention is not limited to one exemplary embodiment disclosed below；Can be come by multi-form
It is realized.The essence of description is only the detail helping the various equivalent modifications Integrated Understanding present invention.
Hereinafter, embodiments of the invention will be described with reference to the drawings.In the accompanying drawings, identical reference represents identical
Or similar parts, or same or similar step.
Fig. 1 diagrammatically illustrates the narrow space intelligent mobile robot vision navigation system block diagram of the present invention.Such as Fig. 1 institute
Showing, the narrow space intelligent mobile robot vision navigation system 100 of the present invention includes image acquisition end 110 and analyzing and processing
End 120.Wherein, image acquisition end 110 can be installed in mobile robot, and described image acquisition end 110 at least includes first
Microcam 111a, the second microcam 111b, cross wire laser projecting apparatus 112, and control unit 113.
Image acquisition end 110 projects the optical strip image of imaging on object for gathering by laser projecting apparatus 112, and
View data is exported to analyzing and processing end 120 by the way of wired or wireless.Analog video signal BNC is such as utilized to connect
The wired mode of head, video connecting line and video frequency collection card carries out data transmission, the most such as by the skill such as bluetooth, WiFi, NFC
Art carries out wireless data transmission.Especially, image acquisition end 110 is from power supply.Preferably, the size of image acquisition end 110
Less than 50mm3。
Image acquisition end 110 specifically includes:
First microcam 111a and the second microcam 111b, projects the cross light on object for shooting
Bar image.This cross optical strip image is sent by cross hairs laser projecting apparatus 112.For the cross optical strip image of the present invention, except
Beyond positive cross, it is also possible to be other cross, the most non-perpendicular intersection optical strip image.
Preferably, two video cameras are fixed on same horizontal plane.The optical axis keeping parallelism of two video cameras is placed, and two
Parallax range between person is between 30-50mm, preferably 40mm.It is highly preferred that the overall dimensions of two video cameras is 12
×12×12mm3。
Cross hairs laser projecting apparatus 112 is used for projecting cross optical strip image.Preferably, cross Rhizoma Dioscoreae (peeled) laser projecting apparatus is installed
In the middle of the first microcam 111a and the second microcam 111b, especially, if cross hairs laser projecting apparatus 112
It is positioned on the straight line being perpendicular to this line at two microcam 111a and place, 111b line midpoint.Further, its
Overall dimensions is preferably 10 × 20 × 10mm3.Each Rhizoma Dioscoreae (peeled) that cross hairs laser projecting apparatus 112 is launched is excellent with the angle of level
Elect 45 degree as.Overall overall dimensions is preferably more than 50 × 25 × 30mm3.Fig. 2 diagrammatically illustrates the vertical of image acquisition end
Body figure.As in figure 2 it is shown, further preferably install in the outer rim of the first microcam 111a and the second microcam 111b
There is illuminating LED light source 114, it is ensured that robot can work in dark environment, such as can be two shootings
The LED light source 114 of a ring shape it is respectively mounted in machine outer rim.
Control unit 113, for according to the data message representing Obstacle Position received from analyzing and processing end 120, control
The traveling of imaged collection terminal 110 is with avoiding obstacles.
The image acquisition end 110 of the present invention, compared with existing vision navigation system, has volume little, lightweight etc. excellent
Gesture, can be easily installed in narrow space and move robot front end.And image acquisition end 110 can be by analyzing and processing described later
Three-dimensional information in end 120 acquisition direction of advance, in order to robot avoids obstacle under complicated unknown on-plane surface environment smoothly
Thing arrives at.
Analyzing and processing end 120, it is according to the cross Rhizoma Dioscoreae (peeled) inputted from the first microcam and the second microcam
Image, completes cross Rhizoma Dioscoreae (peeled) coupling on left images, obtains the seat of cross Rhizoma Dioscoreae (peeled) pixel in the camera coordinates system of left and right
Mark set { uL,vL},{uR,vR, then analyze the three-dimensional coordinate set { x of described cross optical strip imageP,yP,zP, and judge
The region of the barrier occurred, is finally back to described control unit by the regional location of barrier, to control image acquisition end
Traveling.
Preferably, analyzing and processing end 120 includes the display for showing image analysis processing result.
Hereinafter describing for convenience, the first microcam 111a is also referred to as left camera, the cross optical strip image of its shooting
It is referred to as left image；Second microcam 111b is also referred to as right camera, and the cross optical strip image of its shooting is referred to as right image.
Fig. 3 schematically shows the method flow diagram that analyzing and processing end 120 is analyzed processing, as shown in Figure 3:
Step 301, completes cross Rhizoma Dioscoreae (peeled) coupling on left images, with decussation point w (with reference to Fig. 5) as boundary,
To the optical losses coordinate set (hereinafter referred to as 1# essence set) laying respectively at decussation point phase the same side in left and right image
(uL1,vL1) and (uR1,vR1), and, the optical losses laying respectively at the identical other side of decussation point in left and right image is sat
Mark set (hereinafter referred to as 2# essence set) (uL2,vL2) and (uR2,vR2)。
Specifically below to the 1# essence set obtaining coordinate set: (uL1,vL1) and (uR1,vR1) and 2# essence set:
(uL2,vL2) and (uR2,vR2) algorithm illustrate.Fig. 4 diagrammatically illustrates left camera and the signal of right camera coordinates system
Figure, as shown in Figure 4: OL-XLYLZLIt is the coordinate system of left camera, OR-XRYRZRIt it is the coordinate system of right camera.Left camera in this model
And right camera to be idealized as inner parameter identical, x-axis is identical, the perfect condition that y, z-axis are parallel.P is that cross hairs projector is thrown
Shadow is respectively P to any point on testee, coordinate corresponding on left images after collected by cameraL,PR。
First, left image and right image extract cross optical strip image respectively, and chooses suitable gray threshold by ten
Word optical strip image binaryzation.
Secondly, in left image or right image, progressively scan Rhizoma Dioscoreae (peeled) pixel, with decussation point W (with reference to Fig. 5) as boundary,
Obtain laying respectively at all Rhizoma Dioscoreae (peeled) collection of pixels (hereinafter referred to as 1# rough set of decussation point phase the same side in left and right image
Close), and, lay respectively in left and right image all Rhizoma Dioscoreae (peeled) collection of pixels of the identical other side of decussation point (hereinafter referred to as
Close for 2# rough set).Fig. 5 diagrammatically illustrates the schematic diagram of Rhizoma Dioscoreae (peeled) region pre-stored.As it is shown in figure 5, at same a line image pixel
On from left to right search for the region of Rhizoma Dioscoreae (peeled), with decussation point W as boundary, the Rhizoma Dioscoreae (peeled) region origin coordinates value on the limit that keeps left first run into
(x_b1, y) with end coordinate value (x_e1, y) it is stored in the conjunction of 1# rough set, the Rhizoma Dioscoreae (peeled) region origin coordinates value (x_ on remaining limit of keeping right
b2, y) with end coordinate value (x_e2, y) it is stored in the conjunction of 2# rough set.
Especially, although depositing ten with decussation point W as boundary respectively in 1# rough set closes and 2# rough set closes here
Left and right two parts of word optical strip image, but it is not limited to this, it is also possible to deposit upper and lower two parts of cross optical strip image, i.e. press
According to from top to bottom, or way of search by column from bottom to top scans for scanning to image.Such as searching by column from top to bottom
During rope scanning, the Rhizoma Dioscoreae (peeled) region origin coordinates value on the top limit first run into and end coordinate value can be stored in 1# rough set and close, remaining
The origin coordinates value in Rhizoma Dioscoreae (peeled) region on limit on the lower and end coordinate value be stored in 2# rough set and close.
Again, determine parameter a3 with formula 1, thus obtain Rhizoma Dioscoreae (peeled) pixel center coordinate (a3, yi), thus extract 1# essence collection
Close and 2# essence set.From (x_b in 1# gathers1,yi) begin stepping through to (x_e1,yi), wherein i ∈ [1, picture altitude H],
Gray value corresponding for their x-axis coordinate figure is recorded as I, then carries out gaussian curve approximation by following equation (1),
Wherein, a1, a2, a3And a4It is the parameter of curve.Result a of matching3It is the center of Gaussian curve, namely swashs
Centre coordinate (a of light Rhizoma Dioscoreae (peeled)3, yi).So can be obtained by all Rhizoma Dioscoreae (peeled) regions in the conjunction of 1# rough set and the conjunction of 2# rough set
Heart coordinate set, i.e. 1# essence set and 2# essence set, be designated as (uL1,vL1) and (uR1,vR1) and (uL2,vL2) and (uR2,
vR2).Wherein, (uL1,vL1) represent the 1# essence set in left image, (uR1,vR1) represent the 1# essence set in right image, (uL2,
vL2) represent the 2# essence set in left image, (uR2,vR2) represent the 2# essence set in right image.Fig. 5 diagrammatically illustrates Rhizoma Dioscoreae (peeled)
The schematic diagram of region pre-stored, Fig. 5 does not limit left image or right image here, tentatively sets Fig. 5 as left image here.As
Shown in Fig. 5, G (L1) and G (L2) represents the set of 1# essence and 2# essence set in left image respectively, represents cross Rhizoma Dioscoreae (peeled) the most respectively
Centre coordinate point set G (L2) on the left of the W of cross point refers to the centre coordinate point set on the right side of cross Rhizoma Dioscoreae (peeled) cross point W.
Step 302, obtain Rhizoma Dioscoreae (peeled) is three-dimensional coordinate set.Specifically, Rhizoma Dioscoreae (peeled) is analyzed according to the inside and outside parameter of camera
In space any point three-dimensional coordinate (x, y, z) be:
Wherein, αx,αyIt is respectively camera x, the effective focal length in y-axis, (u0,v0) it is the principal point image coordinate of camera, b is
Parallax range, and the sequence number that i ∈ [1,2] is 1# or 2# coordinate set.Travel through all of Rhizoma Dioscoreae (peeled) coordinate set { uL,vL},{uR,
vR, and
{uL,vL}={ uL1,vL1}∪{uL2,vL2}
{uR,vR}={ uR1,vR1}∪{uR2,vR2}
Additionally, U represents that abscissa, V represent vertical coordinate.
The three-dimensional coordinate set that i.e. can get Rhizoma Dioscoreae (peeled) is designated as { xP,yP,zP}。
Step 303, according to the three-dimensional coordinate set { x of Rhizoma Dioscoreae (peeled)P,yP,zP, it is judged that the region of the barrier that front occurs.Tool
Body ground, Fig. 6 is the O-x of a 3 d space coordinate systempypFloor map, zpIt is outside that axle is perpendicular to paper.Real thick line is two
Orthogonal laser Rhizoma Dioscoreae (peeled), the three-dimensional coordinate of the two intersection point is (b/2,0,0).The direction of advance of visual system can be divided into four
Individual region: A represent top left region A | xP∈[0,b/2],yP>=0}, B represent right regions B | xP∈(b/2,b],yP>=0}, C
Expression lower right area C | xP∈(b/2,b],yP< 0} and D represent lower left region D | xP∈[0,b/2],yP< 0}, as shown in the figure.
Additionally, surface represents A Yu B region, underface represents C Yu D region, and left side represents A Yu D region, and right side represents B Yu C region.
Add up these four regions respectively at zPThe number of all coordinate sets in the range of≤H (H is distance alarm threshold value) is { NA,NB,
NC,ND}.Work as NiDuring >=TH (i ∈ (A, B, C, D), TH are the threshold value of coordinate number), it is believed that the i-th region occurs in that barrier.
Different NiConstitute 16 kinds of different coding situation, it is assumed for example that wherein { NA,NB,NC,NDBeing all higher than TH, then encoded radio is
(1111), show that barrier occurs in front, without transitable direction.Remaining 15 kinds clog-free directions of can passing through accordingly are as follows
Shown in form:
Table 1 preceding object object location distribution situation signal table
After analyzing and processing end 120 completes the judgement to preceding object object location, positional information is back to image acquisition end
The control unit 113 of 110, smoothly completes and hides barrier controlling the robot at image acquisition end 110 place.About dividing
Analysis processes the mode that this positional information is back to the control unit 113 of image acquisition end 110 by end 120, it is possible to use simulation regards
Frequently the wired mode of signal BNC connector, video connecting line and video frequency collection card carries out data transmission, the most such as by bluetooth,
The technology such as WiFi, NFC carry out wireless data transmission.
Fig. 7 diagrammatically illustrates the flow process of the narrow space intelligent mobile robot vision navigation method according to the present invention
Figure, as shown in Figure 7:
Step 710, projects cross hairs optical strip image.The cross hairs laser projecting apparatus 112 of video acquisition end 110 launches cross
Optical strip image.Preferably, cross Rhizoma Dioscoreae (peeled) laser projecting apparatus is installed in the middle of left camera and right camera.Further, its overall dimensions
It is preferably 10 × 20 × 10mm3.Each Rhizoma Dioscoreae (peeled) that cross hairs laser projecting apparatus 112 is launched is preferably 45 with the angle of level
Degree.Overall overall dimensions is preferably more than 50 × 25 × 30mm3.Fig. 2 diagrammatically illustrates the axonometric chart of image acquisition end.As
Shown in Fig. 2, left camera and right camera each outer rim are further preferably installed an illuminating annular LED light source of circle, it is ensured that
Robot can work in dark environment.
Step 720, left camera and right camera gather cross optical strip image respectively, and are exported extremely by this cross optical strip image
Analyzing and processing end 120.That is, left camera and right camera projects on objects in front by cross hairs laser projecting apparatus 112 respectively
Cross optical strip image.Preferably, two video cameras are fixed on same horizontal plane, the optical axis keeping parallelism of two video cameras
Placing, parallax range therebetween is 40mm.It is highly preferred that the overall dimensions of video camera is 12 × 12 × 12mm3。
Step 730, image is analyzed by analyzing and processing end 120, draws the three-dimensional coordinate set of Rhizoma Dioscoreae (peeled).Specifically, bag
Include following steps:
Step a, extracts left image and right image (the alternatively referred to as first cross hairs laser strip image and the second cross hairs Rhizoma Dioscoreae (peeled)
Image), and choose suitable gray threshold by image binaryzation.
Step b, in left image or right image, progressively scans Rhizoma Dioscoreae (peeled) pixel, with decussation point W (with reference to Fig. 5) as boundary,
Obtain laying respectively at all Rhizoma Dioscoreae (peeled) collection of pixels (hereinafter referred to as 1# rough set of decussation point phase the same side in left and right image
Close), and, lay respectively in left and right image all Rhizoma Dioscoreae (peeled) collection of pixels of the identical other side of decussation point (hereinafter referred to as
Close for 2# rough set).Fig. 5 diagrammatically illustrates the schematic diagram of Rhizoma Dioscoreae (peeled) region pre-stored.As it is shown in figure 5, at same a line image pixel
On from left to right search for the region of Rhizoma Dioscoreae (peeled), the Rhizoma Dioscoreae (peeled) region origin coordinates value (x_b on the limit that keeps left first run into1, y) and end coordinate
Value (x_e1, y) it is stored in the conjunction of 1# rough set, the Rhizoma Dioscoreae (peeled) region origin coordinates value (x_b on remaining limit of keeping right2, y) with end coordinate value (x_
e2, y) it is stored in the conjunction of 2# rough set.
Step c, determines Rhizoma Dioscoreae (peeled) pixel center coordinate with formula 1, thus extracts 1# essence set and 2# essence set.At 1# collection
From (x_b in conjunction1,yi) begin stepping through to (x_e1,yi), wherein i ∈ [1, picture altitude H], their x-axis coordinate figure is corresponding
Gray value be recorded as I, then carry out gaussian curve approximation by following equation (2),
Wherein, a1, a2, a3And a4It is the parameter of curve.Result a of matching3It is the center of Gaussian curve, namely swashs
Centre coordinate (a of light Rhizoma Dioscoreae (peeled)3, yi).So can be obtained by all Rhizoma Dioscoreae (peeled) regions in the conjunction of 1# rough set and the conjunction of 2# rough set
Heart coordinate set, i.e. 1# essence set and 2# essence set, be designated as (uL1,vL1) and (uR1,vR1) and (uL2,vL2) and (uR2,
vR2).Wherein, (uL1,vL1) represent the 1# essence set in left image, (uR1,vR1) represent the 1# essence set in right image, (uL2,
vL2) represent the 2# essence set in left image, (uR2,vR2) represent the 2# essence set in right image.Fig. 5 diagrammatically illustrates Rhizoma Dioscoreae (peeled)
The schematic diagram of region pre-stored, (and Fig. 5 does not limit left image or right image here, tentatively sets Fig. 5 as left figure here
Picture) as it is shown in figure 5, G (L1) and G (L2) represents the set of 1# essence and 2# essence set in left image respectively, represent ten the most respectively
Centre coordinate point set G (L2) on the left of word Rhizoma Dioscoreae (peeled) cross point W refers to the centre coordinate point set on the right side of cross Rhizoma Dioscoreae (peeled) cross point W.
Step d, obtain Rhizoma Dioscoreae (peeled) is three-dimensional coordinate set.Specifically, analyze Rhizoma Dioscoreae (peeled) according to the inside and outside parameter of camera to exist
In space any point three-dimensional coordinate (x, y, z) be:
Wherein, αx,αyIt is respectively camera x, the effective focal length in y-axis, (u0,v0) it is the principal point image coordinate of camera, b is
Parallax range, and the sequence number that i ∈ [1,2] is 1# or 2# coordinate set.Travel through all of Rhizoma Dioscoreae (peeled) coordinate set { uL,vL},{uR,
vR, and
{uL,vL}={ uL1,vL1}∪{uL2,vL2}
{uR,vR}={ uR1,vR1}∪{uR2,vR2}
The three-dimensional coordinate set that i.e. can get Rhizoma Dioscoreae (peeled) is designated as { xP,yP,zP}。
Step 740, analyzing and processing end 120 is according to the three-dimensional coordinate set { x of Rhizoma Dioscoreae (peeled)P,yP,zP, it is judged that the barrier that front occurs
Hinder the regional location of thing.Analyzing and processing end analysis occurs that the algorithm in the region of barrier is: first direction of advance is divided into a left side
Upper region A | xP∈[0,b/2],yP>=0}, right regions B | xP∈(b/2,b],yP>=0}, lower right area C | xP∈(b/2,
b],yP< 0}, and lower left region D | xP∈[0,b/2],yP< these four regions are added up the most respectively at z in tetra-regions of 0}P≤
The number of all coordinate sets in the range of H is { NA,NB,NC,ND, wherein H is distance alarm threshold value, works as NiDuring >=TH, it is believed that
Occurring in that barrier in i-th region, wherein, i ∈ (A, B, C, D), TH are the threshold value of coordinate number.
Step 750, this position data is back to the control unit 113 of image acquisition end 110 by analyzing and processing end 120, with
The robot at control image acquisition end place smoothly completes to be hidden barrier, it is achieved independent navigation.
Narrow space intelligent mobile robot vision navigation system according to the present invention and method, it is possible to narrow for the unknown
Assembly space moves robot self-navigation and the particularity of avoidance and required precision, in narrow working environment complicated, unknown,
Realize exactly unknown barrier and target object are detected and identified, the Autonomous Control of robot, and independently keep away
Barrier navigation.
In conjunction with explanation and the practice of the present invention disclosed here, other embodiments of the present invention are for those skilled in the art
All it is easy to expect and understand.Illustrating and embodiment is to be considered only as exemplary, true scope and the purport of the present invention are equal
It is defined in the claims.
Claims (8)
1. a narrow space intelligent mobile robot vision navigation system, including:
Image acquisition end, is used for gathering cross optical strip image, and by described cross optical strip image to analyzing and processing end output, described
Image acquisition end includes:
Cross hairs laser projecting apparatus, is used for projecting described cross optical strip image；
First microcam and the second microcam, project the described cross Rhizoma Dioscoreae (peeled) on barrier for shooting respectively
Image obtains the first location image and the second location image, and is transmitted to described analyzing and processing end by described cross optical strip image；
Control unit, for according to the data message representing Obstacle Position received from described analyzing and processing end, controls described
The traveling of image acquisition end；
Analyzing and processing end, for according to described ten inputted respectively from the first microcam and described second microcam
Word optical strip image, completes cross Rhizoma Dioscoreae (peeled) coupling on the first location image and the second location image, with described cross Rhizoma Dioscoreae (peeled)
Decussation point is boundary, obtains laying respectively at decussation point described in described first location image and described second location image
The optical losses coordinate set of phase the same side, and, lay respectively in described first location image and described second location image
The optical losses coordinate set of the described decussation identical other side of point, analyzes the three-dimensional seat obtaining described cross optical strip image
Mark set, and judge the region of the barrier occurred, finally described region is back to described control unit, to control described figure
Traveling as collection terminal；
Wherein, described analyzing and processing end, with the decussation point of described cross Rhizoma Dioscoreae (peeled) as boundary, obtain laying respectively at described first fixed
The optical losses coordinate set of decussation point phase the same side described in bit image and described second location image, and, respectively
It is positioned at the optical losses of described first location image and the decussation identical other side of point described in described second location image
Coordinate set, algorithm be:
A) in described first location image and described second location image, extract described cross optical strip image respectively, and choose ash
Degree threshold value is by described cross optical strip image binaryzation；
B) position in image and described second location image described first, scan described cross optical strip image, with described cross
Cross point is boundary, obtains laying respectively at decussation point phase the same side in described first location image and described second location image
All Rhizoma Dioscoreae (peeled) collection of pixels, and, lay respectively at described first location image and described second location image in decussation
All Rhizoma Dioscoreae (peeled) collection of pixels of the identical other side of point；And
C) obtain Rhizoma Dioscoreae (peeled) pixel center coordinate, thus extract and lay respectively at described first location image and described second location image
Described in the optical losses coordinate set of decussation point phase the same side, and, lay respectively at described first location image and institute
State the optical losses coordinate set of the identical other side of decussation point described in the second location image.
System the most according to claim 1, it is characterised in that described analyzing and processing end determines described according to formula (1)
The three-dimensional coordinate of any point of cross optical strip image (x, y, z):
Wherein, αx,αyIt is respectively camera x, the effective focal length in y-axis, (u0,v0) it is the principal point image coordinate of camera, b is baseline distance
From, and i ∈ [1,2], 1,2 represent respectively: lay respectively at described in described first location image and described second location image ten
The sequence number of the optical losses coordinate set of phase the same side, word cross point, and, lay respectively at described first location image and described
The sequence number of the optical losses coordinate set of the decussation identical other side of point described in second location image.
System the most according to claim 1, it is characterised in that the region of barrier occurs in described analyzing and processing end analysis
Algorithm is: first direction of advance is divided into upper left, upper right, bottom right, and region, four, lower-left, adds up described four the most respectively
Individual region is at zPThe number of all coordinate sets in the range of≤H is { NA,NB,NC,ND, wherein H is distance alarm threshold value, works as Ni
During >=TH, i ∈ (A, B, C, D), then it is assumed that occur in that barrier in the i-th region, wherein, TH is the threshold value of coordinate number.
System the most according to claim 1, it is characterised in that the size of described image acquisition end is less than 50mm3。
System the most according to claim 1, it is characterised in that described first microcam and the second microcam
It is fixed on same horizontal plane.
System the most according to claim 1, it is characterised in that described first microcam and the second microcam
On LED light source is installed.
7. a narrow space intelligent mobile robot vision navigation method, includes image acquisition end according to the system of described method
And analyzing and processing end, described image acquisition end includes that cross hairs laser projecting apparatus, the first microcam and second are miniature
Video camera, control unit, described method comprises the steps:
A) described cross hairs laser projecting apparatus projection cross hairs optical strip image；
B) described first microcam and described second microcam gather described cross optical strip image respectively, and by institute
State the output of cross optical strip image to described analyzing and processing end；
C) image is analyzed by described analyzing and processing end, draws the three-dimensional coordinate set { x of Rhizoma Dioscoreae (peeled)P,yP,zP}；
Wherein, image is analyzed by described analyzing and processing end, draws the three-dimensional coordinate set { x of described cross optical strip imageP,
yP,zPMethod comprise the steps:
C1) in described first location image and described second location image, extract described cross optical strip image respectively, and choose
Gray threshold is by described cross optical strip image binaryzation；
C2) position in image and described second location image described first, scan described cross optical strip image, with described cross
Cross point is boundary, obtains laying respectively at decussation point phase the same side in described first location image and described second location image
All Rhizoma Dioscoreae (peeled) collection of pixels, and, lay respectively at described first location image and described second location image in decussation
All Rhizoma Dioscoreae (peeled) collection of pixels of the identical other side of point；
C3) obtain Rhizoma Dioscoreae (peeled) pixel center coordinate, thus extract and lay respectively at described first location image and described second location figure
The optical losses coordinate set of the some phase of decussation described in Xiang the same side, and, lay respectively at described first location image and
The optical losses coordinate set of the decussation identical other side of point described in described second location image；And
C4) determine according to formula (1) that (x, y z), thus obtain for the three-dimensional coordinate of any point of described cross optical strip image
Three-dimensional coordinate set { the x of described cross optical strip imageP,yP,zP}:
Wherein, αx,αyIt is respectively camera x, the effective focal length in y-axis, (u0,v0) it is the principal point image coordinate of camera, b is baseline distance
From, and i ∈ [1,2], 1,2 represent respectively: lay respectively at described in described first location image and described second location image ten
The sequence number of the optical losses coordinate set of phase the same side, word cross point, and, lay respectively at described first location image and described
The sequence number of the optical losses coordinate set of the decussation identical other side of point described in second location image；
D) described analyzing and processing end is according to the three-dimensional coordinate set { x of Rhizoma Dioscoreae (peeled)P,yP,zP, analyze the district of the barrier that front occurs
Position, territory；And
E) described regional location is back to the described control unit of described image acquisition end by described analyzing and processing end, to control machine
Device people completes to hide barrier.
Method the most according to claim 7, it is characterised in that in step d, there is the district of barrier in analyzing and processing end analysis
The algorithm in territory is: first direction of advance is divided into top left region A | xP∈[0,b/2],yP>=0}, right regions B | xP∈
(b/2,b],yP>=0}, lower right area C | xP∈(b/2,b],yP< 0}, and lower left region D | xP∈[0,b/2],yP< 0} tetra-
Individual region, adds up these four regions the most respectively at zPThe number of all coordinate sets in the range of≤H is { NA,NB,NC,ND,
Wherein H is distance alarm threshold value, works as NiDuring >=TH, it is believed that occur in that barrier in the i-th region, wherein, i ∈ (A, B, C, D), TH
Threshold value for coordinate number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410281773.1A CN105014675B (en) | 2014-06-20 | 2014-06-20 | A kind of narrow space intelligent mobile robot vision navigation system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410281773.1A CN105014675B (en) | 2014-06-20 | 2014-06-20 | A kind of narrow space intelligent mobile robot vision navigation system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105014675A CN105014675A (en) | 2015-11-04 |
| CN105014675B true CN105014675B (en) | 2016-08-17 |
Family
ID=54405157
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410281773.1A Active CN105014675B (en) | 2014-06-20 | 2014-06-20 | A kind of narrow space intelligent mobile robot vision navigation system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105014675B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105856227A (en) * | 2016-04-18 | 2016-08-17 | 呼洪强 | Robot vision navigation technology based on feature recognition |
| CN107305380A (en) * | 2016-04-20 | 2017-10-31 | 上海慧流云计算科技有限公司 | A kind of automatic obstacle-avoiding method and apparatus |
| CN106383518A (en) * | 2016-09-29 | 2017-02-08 | 国网重庆市电力公司电力科学研究院 | Multi-sensor tunnel robot obstacle avoidance control system and method |
| CN107020640A (en) * | 2017-04-28 | 2017-08-08 | 成都科力夫科技有限公司 | Robot interactive formula games system |
| CN107806857A (en) * | 2017-11-08 | 2018-03-16 | 沈阳上博智像科技有限公司 | Unpiloted movable equipment |
| CN109872558A (en) * | 2017-12-04 | 2019-06-11 | 广州市捷众智能科技有限公司 | A kind of indoor parking stall state-detection method based on Cross hair laser projection |
| CN113639748B (en) * | 2020-04-26 | 2024-04-05 | 苏州北美国际高级中学 | Pipeline trolley navigation method based on cross-shaped laser and monocular vision system |
| CN111765849B (en) * | 2020-07-31 | 2021-08-27 | 南京航空航天大学 | Device and method for measuring assembly quality of airplane in narrow space |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6141863A (en) * | 1996-10-24 | 2000-11-07 | Fanuc Ltd. | Force-controlled robot system with visual sensor for performing fitting operation |
| CN1161600C (en) * | 2001-04-30 | 2004-08-11 | 北京航空航天大学 | Structure-light 3D double-visual calibrating point generating method nad device |
| JP2004030445A (en) * | 2002-06-27 | 2004-01-29 | National Institute Of Advanced Industrial & Technology | Method, system, and program for estimating self-position of moving robot |
| CN102183216A (en) * | 2011-03-14 | 2011-09-14 | 沈阳飞机工业（集团）有限公司 | Three-dimensional measurement method and device based on linear structured light |
| CN102313536B (en) * | 2011-07-21 | 2014-02-19 | 清华大学 | Method for barrier perception based on airborne binocular vision |
-
2014
- 2014-06-20 CN CN201410281773.1A patent/CN105014675B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105014675A (en) | 2015-11-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105014675B (en) | A kind of narrow space intelligent mobile robot vision navigation system and method | |
| CA2950791C (en) | Binocular visual navigation system and method based on power robot | |
| Bai et al. | Smart guiding glasses for visually impaired people in indoor environment | |
| WO2020093436A1 (en) | Three-dimensional reconstruction method for inner wall of pipe | |
| CN109901590B (en) | Recharging control method of desktop robot | |
| CN104842358A (en) | Autonomous mobile multifunctional robot | |
| CN106407857B (en) | System and method for automation equipment pairing | |
| CN207115193U (en) | A kind of mobile electronic device for being used to handle the task of mission area | |
| CN207488823U (en) | A kind of mobile electronic device | |
| CN104268933B (en) | Scanning imaging method for three-dimensional environment in vehicle-mounted two-dimensional laser movement | |
| CN102435174A (en) | Method and device for detecting barrier based on hybrid binocular vision | |
| WO2019001237A1 (en) | Mobile electronic device, and method in mobile electronic device | |
| CN102622732A (en) | Front-scan sonar image splicing method | |
| CN103869824A (en) | Biological antenna model-based multi-robot underwater target searching method and device | |
| CN106162144A (en) | A kind of visual pattern processing equipment, system and intelligent machine for overnight sight | |
| CN207067803U (en) | A kind of mobile electronic device for being used to handle the task of mission area | |
| CN115597659B (en) | Intelligent safety management and control method for transformer substation | |
| KR20200020465A (en) | Apparatus and method for acquiring conversion information of coordinate system | |
| CN106370160A (en) | Robot indoor positioning system and method | |
| CN108931982A (en) | Vision navigation system and method for robot moving equipment | |
| CN105043351A (en) | Biological robot-based miniature wireless active omni-directional vision sensor | |
| Shen et al. | A multi-view camera-projector system for object detection and robot-human feedback | |
| CN107607939B (en) | Optical target tracking and positioning radar device based on real map and image | |
| CN206833252U (en) | A kind of mobile electronic device | |
| Mathews et al. | Supervised morphogenesis: morphology control of ground-based self-assembling robots by aerial robots. |
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 |