CN110163916A - The matching process of circulating type circular markers based on rolling target - Google Patents
The matching process of circulating type circular markers based on rolling target Download PDFInfo
- Publication number
- CN110163916A CN110163916A CN201910433534.6A CN201910433534A CN110163916A CN 110163916 A CN110163916 A CN 110163916A CN 201910433534 A CN201910433534 A CN 201910433534A CN 110163916 A CN110163916 A CN 110163916A
- Authority
- CN
- China
- Prior art keywords
- point
- head
- tail
- circular markers
- pattern
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of matching process of circulating type circular markers based on rolling target, including creation feature coding mark point, the feature coding mark point refer to the combination of one or more identical circular markers;It is created around formula coding circular markers, circulating type circular markers, which refer to, is arranged in the feature coding mark point combination that missile rolling model surface is divided into head and tail portion, head circular label point set is denoted as { Head }, and trailer label point set is denoted as { Tail };Extract circular markers, the matching of circular markers.The present invention solves reference point identifying of the store Combinations wind tunnel model when 6 freedom degree directions move and matching problem, without being routed to store Combinations wind tunnel model aperture, the aerodynamic configuration of store Combinations wind tunnel model will not be destroyed, its structural strength and rigidity are not changed, reduce the design difficulty and cost of store Combinations wind tunnel model, decoding efficiency is improved, there is rapidity.
Description
Technical field
The present invention relates to the wind-tunnel technique field of Halo vest, in particular to a kind of store Combinations wind tunnel model is surround
Formula circular markers matching process.
Background technique
Moving object pose measurement is the aeroperformances such as determining moving object flight path, mode of motion in high speed complex flowfield
Important means.As science and technology is constantly progressive, when characteristic of the testee with high-speed motion and test environment is in complexity
Under flow field, the requirement to pose measurement accuracy is continuously improved, high to missile pose measurement especially in drop test in wind tunnel tunnel
The requirement of precision.In recent years, monocular vision pose measurement has measurement motion range wide, designs experimental model and wants without special
It asks, does not change model aerodynamic configuration, can adapt to the testee of complex experiment environment and high-speed motion, the processing time is fast, surveys
The advantages that accuracy of measurement is high is increasingly becoming the research hotspot of missile pose measurement.
Monocular pose measurement core content first is that target's feature-extraction and matching.Groundwork includes setting for mark point
Meter, identification, centralized positioning and reference points matching of the same name.Common label dot shape has circle, cross, rectangular etc..Mark point
Matching be to mark point carry out coding and decoding obtain unique identity information, the type of coded markings point is divided into colouring information class
The forms such as mark point, circular ring marks point, circulating type mark point.For the reference points matching problem of the same name of rolling body class missile,
The intelligent wind-tunnel measurements system that the research center NASA Langley established monocular video measurement early in 2001 is used to survey
Model attitude, deformation and bending are measured, wherein 16-Ft transonic speed tunnel master pattern has rolling type, using circulating type reflecting sign
Note point cooperation external light source carries out Image Acquisition.London Institute of Technology Lin Zhang in 2017 et al. to ultrasonic probe into
Row real-time tracking designs chessboard vertex and circular mark point set solves pose fuzzy problem using cylinder target as measured body.
Liu et al. people, which devises, spirally carries out auxiliary fuel tank separation marking around circular markers and based on color coding reflective marker point
Point matching work.
There are the movements of 6 freedom degrees for the missile testee of rolling type, and the testee size of rolling type is small, increase
The identification of mark point and matched difficulty.In order to realize, mark point is accurate and quick on the small rolling type measured target of size
Matching, this paper presents a kind of circulating type matching process based on rolling target.
Summary of the invention
The technical problems to be solved by the present invention are: the identification of signature point positions on store Combinations in high speed complex flowfield
With match be determining store Combinations stalling characteristic, the important foundation data of flight path.For small-sized rolling body class missile
Signature point matching problem, devises a kind of circulating type mark point coding-decoding method.This method, which has, not to be destroyed outside pneumatic
Shape, designs the advantages such as simple, manufacturing cost is low at not aperture wiring.Solving small-sized rolling body class missile, there are 6 freedom degrees
The reference points matching problem of direction movement.
The technical solution adopted by the present invention is as follows:
The matching process of circulating type circular markers based on rolling target, comprising the following steps:
Step 1: creation feature coding mark point, the feature coding mark point refers to one or more identical round marks
Remember the combination of point;
Step 2: being created around formula coding circular markers, circulating type circular markers, which refer to, is arranged in missile rolling
Model surface is divided into the feature coding mark point combination of head and tail portion, and head circular label point set is denoted as { Head }, tail portion
Label point set is denoted as { Tail }.
Step 3: extracting circular markers, the specific steps are as follows:
S31: posting the testee of circulating type circular markers using camera shooting, by the image transmitting of acquisition to calculating
In machine;
S32: the video camera that this patent uses is slr camera, and the picture of imaging is color image, if using industrial phase
The original image of machine acquisition is gray level image, then can save color image gray processing step;Due to by site environment and image
Noise can be generated in transmission process, need to carry out gaussian filtering to gray level image;Filtered image carries out gray scale stretching, increases quilt
Object and background contrasts are surveyed, the limb recognition of mark point is conducive to;Then self-adaption binaryzation, conversion are carried out in gray level image
At binary image;Eight neighborhood frontier tracing is finally carried out on binary image, calculates the area S and perimeter C of profile.
S33: according to the size of mark point, characteristic, the profile that constraint criterion filters out circular markers is set.This patent is adopted
Circular markers are the bounding box major and minor axis of the constraint of area perimeter, objective contour according to round characteristic design constraint criterion
The area ratio of the ratio between (length and width), the area of profile and bounding box sets threshold values, can filter out the profile of circular markers, about
Beam criterion are as follows:
Wherein, s1,s2For the threshold values of area, c1,c2For the threshold values of perimeter, a, b are the major and minor axis of bounding box, pass through profile
Ranks maximin calculate, T1,T2For a, the threshold values of the ratio between b.
S34: centralized positioning is carried out to circular markers, consults pertinent literature, robustness is optimal in centralized positioning algorithm
It is grey scale centre of gravity positioning[3,4].The completely white or black of the circular mark point design of this paper, the mark point pixel region in binaryzation
Grey value profile it is uniform, so being positioned using grey scale centre of gravity, formula are as follows:
Wherein, (x0,y0) be the profile central pixel coordinate, x=1 ... m, y=1 ... n are the profiles of the circular markers
Pixel coordinate number in region, I (x, y) are the gray values at pixel coordinate (x, y).
Step 4: the matching of circular markers, the specific steps are as follows:
S41: the centre coordinate for remembering circular markers is set Q (q1,…qn), ask each centre coordinate and other centers in Q
The distance between coordinate is denoted as matrixWherein row vectorRepresent qiPicpointed coordinate and other
The distance between picture point is embodied as:
dijIt is expressed as qiPicture point and qjThe distance of picture point, formula are as follows:
Wherein, i=(q1,…,qn), j=(q1,…,qn|j≠i).Element minimum value, is denoted as Dmin_ in calculating matrix D
I, the row vector where recording Dmin_i are denoted as dmin_q, according to Dmin_i is certain in first principle in code Design principle
The distance of adjacent marker point in a B, C, D, E pattern, then dmin_qBe expert at serial number representative is the point on head, is put into set of headers
In { Head }.
S42: d is calculatedmin_qElement maximum value in row vector, is denoted as Dmax.Known according to coding Article 2 principle and is marked with head
The farthest distance of note point must be the mark point of tail portion, therefore element Dmax column serial number is the point of tail portion, is denoted as q_max and puts
Enter in set { Tail }.
S43: classifying to the circular markers of head-tail, is respectively put into set { Head }, in { Tail }:
S431: minimum value element in row vector of the picture point q_max where in D is calculated, which is tail portion
Point, is denoted as q_max_min, is put into set { Tail }.
S432: in set Q remaining point and 3 points of q_max, q_max_min it is whether conllinear, if collinearly be tail portion point,
It is put into set { Tail }, is otherwise the point on head, is put into set { Head }.Conllinear calculation formula are as follows:
S44: according to head mark point and tail portion vectorIt is whether vertical, to set of headers
{ Head } is carried out with group classification, decoding.
S441: { Head } is gathered and carries out every group of pattern classification.I-th picture point and other pictures in { Head } in { Head } set
Whether the vector of point composition, ask the vector vertical with tail portion vector, and it is then same that the angle between vector, which meets 85 ° of 95 ° of < θ <,
The circular markers of group pattern are put into set Code_i (i is that the i-th picture point organizes set together in { Head }), if the i-th picture point and other
Picture point forms vector and vector all out of plumb in tail portion are then A patterns.
S442: every group echo point number in statistics Code_i, then be the coding of this group of pattern.Code set is denoted as
Decode_i。
S443: the distance between picture point and q_max in i-th group of pattern of Code_i are calculated, after Code_i is according to distance-taxis
It is put into set sortCode_i.Farthest picture point is the starting circle of this group of pattern, and starting circle coordinates are put into set startCode_
i。
S444: judging the point for whether having fracture in every group of pattern, and fracture leads to the coding of the false judgment pattern.It calculates
Originated in sortCode_i it is round finally justify with this group of pattern at a distance from and Dmin_i multiple, if multiple greater than if Decode_i value
There are fractures, update serial number, recalculate pattern encoding D ecode_i.Such as D pattern, if intermediate 2nd circle of missing, judgement should
Pattern is C, and calculating multiple is 3 times, then the circle serial number 4, and redefining this group of pattern is D.
S445: judge whether each starting circle is conllinear in startCode_i, collinearly then illustrate not leak identification starting circle, such as
Then give up the photo in the presence of not conllinear starting circle.According to the row coordinate of startCode_i or column coordinate ascending sort
StartCode_i, Decode_i and startCode_i (seat that the difference wonderful works of ranks coordinate is sorted in startCode_i
Mark).Decode_i is coding segment after then sorting.
S446:Decode_i encode fragment order successively with header code segment in cryptoprinciple Article 3 with sequence or
Inverted order is successively searched, and is found out the corresponding coding serial number of Decode_i coding segment, is denoted as Decode_number_i.
S45: being decoded tail portion, and the coding segment and code sequence of tail portion { Tail } are successively found out with header code segment
Number.
It chooses in S451:startCode_i in addition to A pattern, any one group of pattern, is found out in tail total with this group of pattern
The mark point of line, the coding serial number of conllinear corresponding head mark point is the coding serial number of conllinear corresponding trailer label point, if not having
It finds out collinearly, then chooses other group of pattern (except A pattern) in startCode_i and find out conllinear mark point.Conllinear head mark
Note point is denoted as head_coll, and trailer label point is denoted as tail_coll.
S452: wherein head_coll and startCode_i arbitrarily originates the composition vector of circleJudge vector coding sequence
Number it is sequence or inverted order, calculates the distance between mark point in tail_coll and { Tail }, withDirectionality judge distance
It is positive and negative, be then positive in the same direction, otherwise be negative.
S453: according to apart from ascending sort { Tail }, it is denoted as sort_tail, is existed according to the number serial number of tail_coll
The position sort_tail successively (withJudgement is that sequence or inverted order read tail portion coding in code Design principle Article 3) it reads
The coding segment and coding serial number of other mark points of sort_tail.
Further, the circulating type circular mark point design has the feature that
A, it is divided into two part circular mark point of head and tail portion on testee surface, head circular label point set is denoted as
{ Head }, trailer label point set are denoted as { Tail }.Head mark point is that one group of volume is formed by one or more circular markers
Code labeling point, is evenly distributed on the surface of rolling type testee.5 kinds of as shown in Figure 2 patterns of assemble pattern be respectively by
1,2,3,4,5 circular markers compositions, are denoted as the coded markings point of A, B, C, D, E pattern respectively.3 groups of coding marks of arbitrary continuation
Note point is denoted as a coding segment and coding segment assemble pattern does not repeat.Such as A, B, D pattern are continuously a coding segment,
Its number is 125.Trailer label point is evenly distributed on the surface of testee only with A pattern.
Then, the cryptoprinciple of the circulating type circular markers has following principle:
A, rolling type testee diameter is set as Lmm, and in addition to circular cone, the length of intermediate cylindrical meets > 2Lmm, considers
It include head-tail at least 4 groups of coded markings points, and adjacent circular mark point in every group of pattern in the image acquired under any attitude
Spacing be less than the spacing (spacing that spacing refers to the round outer surface between circle) of upper the following group pattern.Therefore design circular mark
Spot diameter is 1/10Lmm, and the round spacing between circle is 1~2mm in every group of pattern, and avoiding imaging rear profile from being connected can not shape
At the closed outline of single circular markers.Every group of pattern spacing is 1/10Lmm.
B, it is conllinear to design every group of head-tail of the circular markers center of circle.Head-tail is greater than the straight of testee apart from spacing
Diameter (distance that the spacing arbitrarily justified in head is both less than head-tail).
C, any one coding segment (the coded markings point of continuous 3 groups of patterns) is unique code clerk, part in head
It is as shown in Figure 3 to encode segment.Wherein the code clerk of each circular markers of tail portion is collinear header code number, is assigned
Every group of circular markers of head-tail are unique encodings serial number.Assuming that there are 15 groups of coded markings points on head, header code segment is successively
It is 123124135145214, corresponding conllinear tail portion circular markers coding segment is 123124135145214.Then head is compiled
Code serial number 1-15, tail portion encode serial number 16-30.Such as header code segment is 3124 in certain photo imaging, then corresponds to
Coding serial number 3-6, tail portion 1241, encode serial number 19-22.
In conclusion by adopting the above-described technical solution, the beneficial effects of the present invention are:
The matching process of circulating type circular markers proposed by the present invention based on rolling target, not only solves store Combinations
Reference point identifying of the wind tunnel model when 6 freedom degree directions move and matching problem, and with existing embedded encoded luminescent device
Technology compare, without to store Combinations wind tunnel model aperture be routed, the aerodynamic configuration of store Combinations wind tunnel model will not be destroyed, do not changed
Become its structural strength and rigidity, reduces the design difficulty and cost of store Combinations wind tunnel model.Compared with circulating type coding method,
Decoding efficiency is improved, there is rapidity.
Detailed description of the invention
Fig. 1 is the feature coding circular markers schematic diagram of the embodiment of the present invention.
Fig. 2 is the circulating type circular markers of the embodiment of the present invention in missile model schematic.
Fig. 3 is the code segment fragmentary views of the embodiment of the present invention.
Fig. 4 is that the circular markers of the embodiment of the present invention identify positioning flow figure.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention more comprehensible, right hereinafter, referring to the drawings and the embodiments,
The present invention is described in further details.
The matching process of circulating type circular markers based on rolling target.Comprising the following specific steps
Step 1: design feature coded markings point, the feature coding mark point refer to one or more identical round marks
The combination for remembering point, as shown in Figure 1, having the feature that
A, assemble pattern 5 kinds of patterns as shown in Figure 1 are made of 1,2,3,4,5 circular markers respectively, are remembered respectively
For the coded markings point of A, B, C, D, E pattern.
B, conllinear for the coded markings point center of circle of B, C, D, E pattern, and the spacing in every group of pattern between the center of circle is
1.1dmm (diameter that d is circular markers), the disk wheel of single circular markers can not be formed by avoiding imaging rear profile from being connected
It is wide.
If c, rolling type testee diameter is Lmm, design circular mark spot diameter is at most 1/10Lmm.
Step 2: circulating type circular markers of the creation based on rolling target are distributed in the surface of the model of missile, have
The design of body is as shown in Figure 2.It is divided into two part circular mark point of head and tail portion, head circular mark point on testee surface
Set is denoted as { Head }, and trailer label point set is denoted as { Tail }.Wherein specific cryptoprinciple is as follows:
A, it defines header code mark point Head to be made of A, B, C, D, E pattern, tail portion Tail only has A pattern composition.
3 groups of coded markings points of arbitrary continuation are denoted as a coding segment in head mark point and coding segment assemble pattern does not repeat.Example
If A, B, D pattern are continuously a coding segment, number is 125.Trailer label point only with A pattern, be evenly distributed on by
It surveys on the surface of object.
If b, rolling type testee diameter is Lmm, in addition to circular cone, the length of intermediate cylindrical meets > 2L mm, considers
It include head-tail at least 4 groups of coded markings points, and adjacent circular mark point in every group of pattern in the image acquired under any attitude
Spacing be less than the spacing (spacing that spacing refers to the round outer surface between circle) of upper the following group pattern.Therefore design circular mark
Spot diameter is 1/10Lmm, and the round spacing between circle is 1~2mm in every group of pattern, and avoiding imaging rear profile from being connected can not shape
At the closed outline of single circular markers.Every group of pattern spacing is 1/10Lmm.
C, it is conllinear to design every group of head-tail of the circular markers center of circle.Head-tail is greater than the straight of testee apart from spacing
Diameter (distance that the spacing arbitrarily justified in head is both less than head-tail).
D, any one coding segment (the coded markings point of continuous 3 groups of patterns) is unique code clerk, part in head
It is as shown in Figure 3 to encode segment.Wherein the code clerk of each circular markers of tail portion is collinear header code number, is assigned
Every group of circular markers of head-tail are unique encodings serial number.Assuming that there are 15 groups of coded markings points on head, header code segment is successively
It is 123124135145214, corresponding conllinear tail portion circular markers coding segment is 123124135145214.Then head is compiled
Code serial number 1-15, tail portion encode serial number 16-30.Such as header code segment is 3124 in certain photo imaging, then corresponds to
Coding serial number 3-6, tail portion 1241, encode serial number 19-22.
Step 3: extracting circular markers, identification positioning is carried out, design cycle is as shown in Figure 4, the specific steps are as follows:
S31: posting the testee of circulating type circular markers using camera shooting, by the image transmitting of acquisition to calculating
In machine;
S32: the video camera that this patent uses is slr camera, and the picture of imaging is color image, if using industrial phase
The original image of machine acquisition is gray level image, then can save color image gray processing step;Due to by site environment and image
Noise can be generated in transmission process, need to carry out gaussian filtering to gray level image;Filtered image carries out gray scale stretching, increases quilt
Object and background contrasts are surveyed, the limb recognition of mark point is conducive to;Then self-adaption binaryzation, conversion are carried out in gray level image
At binary image;Eight neighborhood frontier tracing is finally carried out on binary image, calculates the area S and perimeter C of profile.
S33: according to the size of mark point, characteristic, the profile that constraint criterion filters out circular markers is set.This patent is adopted
Circular markers are the bounding box major and minor axis of the constraint of area perimeter, objective contour according to round characteristic design constraint criterion
The area ratio of the ratio between (length and width), the area of profile and bounding box sets threshold values, can filter out the profile of circular markers, about
Beam criterion is
Wherein, s1,s2For the threshold values of area, c1,c2For the threshold values of perimeter, a, b are the major and minor axis of bounding box, pass through profile
Ranks maximin calculate, T1,T2For a, the threshold values of the ratio between b.
S34: centralized positioning is carried out to circular markers, consults pertinent literature, robustness is optimal in centralized positioning algorithm
It is grey scale centre of gravity positioning[3,4].The completely white or black of the circular mark point design of this paper, the mark point pixel region in binaryzation
Grey value profile it is uniform, so being positioned using grey scale centre of gravity, formula is
Wherein, (x0,y0) be the profile central pixel coordinate, x=1 ... m, y=1 ... n are the profiles of the circular markers
Pixel coordinate number in region, I (x, y) are the gray values at pixel coordinate (x, y).
Step 4: the matching of circular markers, the specific steps are as follows:
S41: the centre coordinate for remembering circular markers is set Q (q1,…qn), ask each centre coordinate and other centers in Q
The distance between coordinate is denoted as matrixWherein row vectorRepresent qiPicpointed coordinate and other
The distance between picture point is embodied as
dijIt is expressed as qiPicture point and qjThe distance of picture point, formula are
Wherein, i=(q1,…,qn), j=(q1,…,qn|j≠i).Element minimum value, is denoted as Dmin_ in calculating matrix D
I, the row vector where recording Dmin_i are denoted as dmin_q, according to Dmin_i is certain in first principle in code Design principle
The distance of adjacent marker point in a B, C, D, E pattern, then dmin_qBe expert at serial number representative is the point on head, is put into set of headers
In { Head }.
S42: d is calculatedmin_qElement maximum value in row vector, is denoted as Dmax.Known according to coding Article 2 principle and is marked with head
The farthest distance of note point must be the mark point of tail portion, therefore element Dmax column serial number is the point of tail portion, is denoted as q_max and puts
Enter in set { Tail }.
S43: classifying to the circular markers of head-tail, is respectively put into set { Head }, in { Tail }:
S431: minimum value element in row vector of the picture point q_max where in D is calculated, which is tail portion
Point, is denoted as q_max_min, is put into set { Tail }.
S432: in set Q remaining point and 3 points of q_max, q_max_min it is whether conllinear, if collinearly be tail portion point,
It is put into set { Tail }, is otherwise the point on head, is put into set { Head }.Collinearly calculation formula is
S44: according to head mark point and tail portion vectorIt is whether vertical, to set of headers
{ Head } is carried out with group classification, decoding.
S441: { Head } is gathered and carries out every group of pattern classification.I-th picture point and other pictures in { Head } in { Head } set
Whether the vector of point composition, ask the vector vertical with tail portion vector, and it is then same that the angle between vector, which meets 85 ° of 95 ° of < θ <,
The circular markers of group pattern are put into set Code_i (i is that the i-th picture point organizes set together in { Head }), if the i-th picture point and other
Picture point forms vector and vector all out of plumb in tail portion are then A patterns.
S442: every group echo point number in statistics Code_i, then be the coding of this group of pattern.Code set is denoted as
Decode_i。
S443: the distance between picture point and q_max in i-th group of pattern of Code_i are calculated, after Code_i is according to distance-taxis
It is put into set sortCode_i.Farthest picture point is the starting circle of this group of pattern, and starting circle coordinates are put into set startCode_
i。
S444: judging the point for whether having fracture in every group of pattern, and fracture leads to the coding of the false judgment pattern.It calculates
Originated in sortCode_i it is round finally justify with this group of pattern at a distance from and Dmin_i multiple, if multiple greater than if Decode_i value
There are fractures, update serial number, recalculate pattern encoding D ecode_i.Such as D pattern, if intermediate 2nd circle of missing, judgement should
Pattern is C, and calculating multiple is 3 times, then the circle serial number 4, and redefining this group of pattern is D.
S445: judge whether each starting circle is conllinear in startCode_i, collinearly then illustrate not leak identification starting circle, such as
Then give up the photo in the presence of not conllinear starting circle.According to the row coordinate of startCode_i or column coordinate ascending sort
StartCode_i, Decode_i and startCode_i (seat that the difference wonderful works of ranks coordinate is sorted in startCode_i
Mark).Decode_i is coding segment after then sorting.
S446:Decode_i encode fragment order successively with header code segment in cryptoprinciple Article 3 with sequence or
Inverted order is successively searched, and is found out the corresponding coding serial number of Decode_i coding segment, is denoted as Decode_number_i.
S45: being decoded tail portion, and the coding segment and code sequence of tail portion { Tail } are successively found out with header code segment
Number.
It chooses in S451:startCode_i in addition to A pattern, any one group of pattern, is found out in tail total with this group of pattern
The mark point of line, the coding serial number of conllinear corresponding head mark point is the coding serial number of conllinear corresponding trailer label point, if not having
It finds out collinearly, then chooses other group of pattern (except A pattern) in startCode_i and find out conllinear mark point.Conllinear head mark
Note point is denoted as head_coll, and trailer label point is denoted as tail_coll.
S452: wherein head_coll and startCode_i arbitrarily originates the composition vector of circleJudge vector coding sequence
Number it is sequence or inverted order, calculates the distance between mark point in tail_coll and { Tail }, withDirectionality judge distance
It is positive and negative, be then positive in the same direction, otherwise be negative.
S453: according to apart from ascending sort { Tail }, it is denoted as sort_tail, is existed according to the number serial number of tail_coll
The position sort_tail successively (withJudgement is that sequence or inverted order read tail portion coding in code Design principle Article 3) it reads
The coding segment and coding serial number of other mark points of sort_tail.
Claims (6)
1. the matching process of the circulating type circular markers based on rolling target, which comprises the following steps:
Step 1: creation feature coding mark point, the feature coding mark point refer to one or more identical circular markers
Combination;
Step 2: being created around formula coding circular markers, circulating type circular markers, which refer to, is arranged in missile rolling model
Surface is divided into the feature coding mark point combination of head and tail portion, and head circular label point set is denoted as { Head }, trailer label
Point set is denoted as { Tail };
Step 3: circular markers are extracted,
Step 4: the matching of circular markers.
2. the matching process of the circulating type circular markers according to claim 1 based on rolling target, which is characterized in that
The step three extracts circular markers, the specific steps are as follows:
S31: the testee of circulating type circular markers is posted using camera shooting, by the image transmitting of acquisition to computer
In;
S32: if video camera is slr camera, the picture of imaging is that color image carries out gray processing, if using industrial camera
The original image of acquisition is gray level image, then can save color image gray processing step;
Gaussian filtering is carried out to gray level image;Filtered image carries out gray scale stretching, increases testee and background contrasts,
Be conducive to the limb recognition of mark point;Then self-adaption binaryzation is carried out in gray level image, is converted into binary image;Finally exist
Eight neighborhood frontier tracing is carried out on binary image, calculates the area S and perimeter C of profile;
S33: according to the size of mark point, characteristic, the profile that constraint criterion filters out circular markers is set;The round mark of use
Remember point, according to round characteristic design constraint criterion be the constraint of area perimeter, objective contour the ratio between bounding box major and minor axis (length and width),
The area of profile and the area ratio of bounding box set threshold values, can filter out the profile of circular markers, constraint criterion is
Wherein, s1,s2For the threshold values of area, c1,c2For the threshold values of perimeter, a, b are the major and minor axis of bounding box, pass through the ranks of profile
Maximin calculates, T1,T2For a, the threshold values of the ratio between b;
S34: carrying out centralized positioning to circular markers, consults pertinent literature, and what robustness was optimal in centralized positioning algorithm is ash
Spend center of gravity positioning;The completely white or black of circular mark point design, the grey value profile of mark point pixel region in binaryzation
Uniformly, it is positioned using grey scale centre of gravity, formula are as follows:
Wherein, x=1 ... m, y=1 ... n are pixel coordinates in the contour area of the circular markers, and I (x, y) is pixel coordinate
Gray value at (x, y).
3. the matching process of the circulating type circular markers according to claim 1 based on rolling target, which is characterized in that
The matching of step four, the circular markers, the specific steps are as follows:
S41: the centre coordinate for remembering circular markers is set Q (q1,…qn), ask each centre coordinate and other centre coordinates in Q
The distance between, it is denoted as matrixWherein row vectorRepresent qiPicpointed coordinate and other picture points
The distance between, it is embodied as:
dijIt is expressed as qiPicture point and qjThe distance of picture point, formula are as follows:
Wherein, i=(q1,…,qn), j=(q1,…,qn|j≠i);
Element minimum value in calculating matrix D is denoted as Dmin_i, and the row vector where recording Dmin_i is denoted as dmin_q, according to coding
Know that Dmin_i is the distance of adjacent marker point in some B, C, D, E pattern in design principle in first principle, then dmin_qInstitute
Serial number of being expert at representative is the point on head, is put into set of headers { Head };
S42: d is calculatedmin_qElement maximum value in row vector, is denoted as Dmax;Known and head mark point according to coding Article 2 principle
Farthest distance must be the mark point of tail portion, therefore element Dmax column serial number is the point of tail portion, is denoted as q_max and is put into collection
It closes in { Tail };
S43: classifying to the circular markers of head-tail, is respectively put into set { Head }, in { Tail }:
S44: according to head mark point and tail portion vectorIt is whether vertical, to set of headers { Head }
It carries out with group classification, decoding;
S45: being decoded tail portion, and the coding segment and coding serial number of tail portion { Tail } are successively found out with header code segment.
4. the matching process of the circulating type circular markers according to claim 3 based on rolling target, which is characterized in that
The S43 the following steps are included:
S431: calculating minimum value element in row vector of the picture point q_max where in D, which is the point of tail portion, note
For q_max_min, it is put into set { Tail };
S432: in set Q remaining point and 3 points of q_max, q_max_min it is whether conllinear, if collinearly be tail portion point, be put into
Gather in { Tail }, is otherwise the point on head, is put into set { Head };Conllinear calculation formula are as follows:
5. the matching process of the circulating type circular markers according to claim 4 based on rolling target, which is characterized in that
The S44 the following steps are included:
S441: { Head } is gathered and carries out every group of pattern classification;I-th picture point and other picture point groups in { Head } in { Head } set
At vector, ask the vector whether vertical with tail portion vector, angle between vector meets 85 ° of 95 ° of < θ <, then is with group sample
The circular markers of formula are put into set Code_i, and i is that the i-th picture point is with set is organized in { Head }, if the i-th picture point and other picture points
It forms vector and vector all out of plumb in tail portion is then A pattern;
S442: every group echo point number in statistics Code_i, then be the coding of this group of pattern;Code set is denoted as Decode_i;
S443: calculating the distance between picture point and q_max in i-th group of pattern of Code_i, and Code_i after distance-taxis according to being put into
Set sortCode_i;Farthest picture point is the starting circle of this group of pattern, and starting circle coordinates are put into set startCode_i;
S444: judging the point for whether having fracture in every group of pattern, and fracture leads to the coding of the false judgment pattern;It calculates
Originated in sortCode_i it is round finally justify with this group of pattern at a distance from and Dmin_i multiple, if multiple greater than if Decode_i value
There are fractures, update serial number, recalculate pattern encoding D ecode_i;
S445: judge whether each starting circle is conllinear in startCode_i, collinearly then illustrate not leak identification starting circle, such as exist
Not conllinear starting circle then gives up the photo;According to the row coordinate of startCode_i or column coordinate ascending sort
StartCode_i, Decode_i and startCode_i, then Decode_i is coding segment after sorting;
S446:Decode_i encodes fragment order successively with header code segment in cryptoprinciple Article 3 with sequence or inverted order
It successively searches, finds out the corresponding coding serial number of Decode_i coding segment, be denoted as Decode_number_i.
6. the matching process of the circulating type circular markers according to claim 5 based on rolling target, which is characterized in that
The S45 the following steps are included:
It is chosen in S451:startCode_i in addition to A pattern, any one group of pattern, is found out in tail conllinear with this group of pattern
Mark point, the coding serial number of conllinear corresponding head mark point is the coding serial number of conllinear corresponding trailer label point, if not finding out
Collinearly, then it chooses in startCode_i except other group of pattern of A pattern finds out conllinear mark point;Conllinear head mark point note
For head_coll, trailer label point is denoted as tail_coll;
S452: wherein head_coll and startCode_i arbitrarily originates the composition vector of circleJudge that vector coding serial number is suitable
Sequence or inverted order calculate the distance between mark point in tail_coll and { Tail }, withDirectionality judge distance just
It is negative, it is then positive, otherwise is negative in the same direction;
S453: according to apart from ascending sort { Tail }, it is denoted as sort_tail, according to the number serial number of tail_coll in sort_
The position tail is successively read the coding segment and coding serial number of other mark points of sort_tail.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910433534.6A CN110163916B (en) | 2019-05-23 | 2019-05-23 | Rolling target-based surrounding type circular mark point matching method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910433534.6A CN110163916B (en) | 2019-05-23 | 2019-05-23 | Rolling target-based surrounding type circular mark point matching method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110163916A true CN110163916A (en) | 2019-08-23 |
CN110163916B CN110163916B (en) | 2022-06-28 |
Family
ID=67632111
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910433534.6A Active CN110163916B (en) | 2019-05-23 | 2019-05-23 | Rolling target-based surrounding type circular mark point matching method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110163916B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112150546A (en) * | 2020-10-23 | 2020-12-29 | 西华大学 | Monocular vision pose estimation method based on auxiliary point geometric constraint |
CN114509089A (en) * | 2021-12-31 | 2022-05-17 | 成都弓网科技有限责任公司 | Non-contact rail transit train speed direction mileage detection method and system |
CN115439630A (en) * | 2022-08-04 | 2022-12-06 | 思看科技(杭州)股份有限公司 | Mark point splicing method, photogrammetry method, device and electronic device |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130108136A1 (en) * | 2009-11-25 | 2013-05-02 | Dental Imaging Technologies Corporation | Marker identification and processing in x-ray images |
CN103310215A (en) * | 2013-07-03 | 2013-09-18 | 天津工业大学 | Detecting and identifying method for annular coding mark point |
CN105444982A (en) * | 2015-11-24 | 2016-03-30 | 中国空气动力研究与发展中心高速空气动力研究所 | Monocular video measurement method for external store separating locus wind tunnel test |
CN106372702A (en) * | 2016-09-06 | 2017-02-01 | 深圳市欢创科技有限公司 | Positioning identification and positioning method thereof |
CN107452054A (en) * | 2017-08-08 | 2017-12-08 | 西华大学 | A kind of decoding method of store Combinations wind tunnel model circulating type coded markings point |
CN107578051A (en) * | 2017-09-14 | 2018-01-12 | 西华大学 | A kind of detection and recognition methods of ring-type coded markings point |
CN108510545A (en) * | 2018-03-30 | 2018-09-07 | 京东方科技集团股份有限公司 | Space-location method, space orientation equipment, space positioning system and computer readable storage medium |
CN108898635A (en) * | 2018-05-07 | 2018-11-27 | 深圳大学 | A kind of control method and system improving camera calibration precision |
CN109215016A (en) * | 2018-08-03 | 2019-01-15 | 湖南科技大学 | A kind of recognition positioning method of coding maker |
-
2019
- 2019-05-23 CN CN201910433534.6A patent/CN110163916B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130108136A1 (en) * | 2009-11-25 | 2013-05-02 | Dental Imaging Technologies Corporation | Marker identification and processing in x-ray images |
CN103310215A (en) * | 2013-07-03 | 2013-09-18 | 天津工业大学 | Detecting and identifying method for annular coding mark point |
CN105444982A (en) * | 2015-11-24 | 2016-03-30 | 中国空气动力研究与发展中心高速空气动力研究所 | Monocular video measurement method for external store separating locus wind tunnel test |
CN106372702A (en) * | 2016-09-06 | 2017-02-01 | 深圳市欢创科技有限公司 | Positioning identification and positioning method thereof |
CN107452054A (en) * | 2017-08-08 | 2017-12-08 | 西华大学 | A kind of decoding method of store Combinations wind tunnel model circulating type coded markings point |
CN107578051A (en) * | 2017-09-14 | 2018-01-12 | 西华大学 | A kind of detection and recognition methods of ring-type coded markings point |
CN108510545A (en) * | 2018-03-30 | 2018-09-07 | 京东方科技集团股份有限公司 | Space-location method, space orientation equipment, space positioning system and computer readable storage medium |
CN108898635A (en) * | 2018-05-07 | 2018-11-27 | 深圳大学 | A kind of control method and system improving camera calibration precision |
CN109215016A (en) * | 2018-08-03 | 2019-01-15 | 湖南科技大学 | A kind of recognition positioning method of coding maker |
Non-Patent Citations (4)
Title |
---|
周单: "《基于单目视觉的运动物位姿测量及优化研究》", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
孟祥丽 等: "《一种圆形编码标志点的设计及解码算法研究》", 《传感器与微系统》 * |
董辉 等: "《基于圆弧边缘特征的圆检测算法》", 《浙江工业大学学报》 * |
金滔 等: "《一种环状编码标记点的设计及解码算法》", 《计算机应用研究》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112150546A (en) * | 2020-10-23 | 2020-12-29 | 西华大学 | Monocular vision pose estimation method based on auxiliary point geometric constraint |
CN112150546B (en) * | 2020-10-23 | 2023-11-21 | 西华大学 | Monocular vision pose estimation method based on auxiliary point geometric constraint |
CN114509089A (en) * | 2021-12-31 | 2022-05-17 | 成都弓网科技有限责任公司 | Non-contact rail transit train speed direction mileage detection method and system |
CN115439630A (en) * | 2022-08-04 | 2022-12-06 | 思看科技(杭州)股份有限公司 | Mark point splicing method, photogrammetry method, device and electronic device |
CN115439630B (en) * | 2022-08-04 | 2024-04-19 | 思看科技(杭州)股份有限公司 | Mark point splicing method, photogrammetry method, device and electronic device |
Also Published As
Publication number | Publication date |
---|---|
CN110163916B (en) | 2022-06-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108549873B (en) | Three-dimensional face recognition method and three-dimensional face recognition system | |
Zou et al. | A robust license plate recognition model based on bi-lstm | |
CN106707296B (en) | It is a kind of based on the unmanned machine testing of Based on Dual-Aperture photo electric imaging system and recognition methods | |
CN104200495B (en) | A kind of multi-object tracking method in video monitoring | |
CN110163916A (en) | The matching process of circulating type circular markers based on rolling target | |
CN107403163B (en) | A kind of laser SLAM closed loop automatic testing method based on deep learning | |
CN106092086B (en) | A kind of quick, high robust robot indoor orientation method based on panoramic vision | |
CN106683137B (en) | Artificial mark based monocular and multiobjective identification and positioning method | |
CN109144072A (en) | A kind of intelligent robot barrier-avoiding method based on three-dimensional laser | |
CN110555475A (en) | few-sample target detection method based on semantic information fusion | |
CN105469418A (en) | Photogrammetry-based wide-field binocular vision calibration device and calibration method | |
CN105091782A (en) | Multilane laser light plane calibration method based on binocular vision | |
CN102773862A (en) | Quick and accurate locating system used for indoor mobile robot and working method thereof | |
CN108154102A (en) | A kind of traffic sign recognition method | |
CN103136525B (en) | A kind of special-shaped Extended target high-precision locating method utilizing Generalized Hough Transform | |
CN108334881A (en) | A kind of licence plate recognition method based on deep learning | |
CN106384079A (en) | RGB-D information based real-time pedestrian tracking method | |
CN106595702B (en) | A kind of multisensor spatial registration method based on astronomy calibration | |
CN102788572B (en) | Method, device and system for measuring attitude of lifting hook of engineering machinery | |
CN109035329A (en) | Camera Attitude estimation optimization method based on depth characteristic | |
CN112966571B (en) | Standing long jump flight height measurement method based on machine vision | |
CN110929710A (en) | Method and system for automatically identifying meter pointer reading based on vision | |
CN110296707A (en) | Indoor service robot based on VLC technology | |
CN104376328B (en) | Coordinate-based distributed coding mark identification method and system | |
CN109271868B (en) | Dense connection convolution network hypersphere embedding-based target re-identification method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |