CN106845506A - A kind of target surface location of pixels coding method - Google Patents
A kind of target surface location of pixels coding method Download PDFInfo
- Publication number
- CN106845506A CN106845506A CN201611195876.1A CN201611195876A CN106845506A CN 106845506 A CN106845506 A CN 106845506A CN 201611195876 A CN201611195876 A CN 201611195876A CN 106845506 A CN106845506 A CN 106845506A
- Authority
- CN
- China
- Prior art keywords
- target surface
- pixel
- target
- loop wire
- point
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of target surface location of pixels coding method, the method is used for chest silhouette target, comprises the following steps：The initial target surface a of chest silhouette target is obtained, and target surface d is extracted from initial target surface ai；To target surface diIn loop wire be fitted；In target surface diIn, the loop wire according to fitting carries out position encoded to target surface pixel.The present invention is in chest silhouette target positioning, using the target position identification technology based on color characteristic, target surface is easily big by complex background, illumination variation interference, artificial interference in solving the problems, such as chest silhouette target positioning, and chest silhouette target positioning precision is high, manual intervention is few, and is not disturbed by illumination condition.
Description
Technical field
Automatic target-indicating field the present invention relates to be based on image procossing, more particularly to a kind of target surface location of pixels coding
With the automatic target-indicating method encoded based on target surface location of pixels.
Background technology
The acquisition of target surface initial information is shell hole identification and the premise and the basis that judge number of rings, decides automatic scoring round target system
The quality of performance.In traditional automatic scoring round target system based on image procossing, the work that target surface initial information fetching portion is done
Only image filtering and image distortion correction, are obtained in that more regular target sheet image, but do not examine in theoretical research
Consider the present situation of target range ambient interferences and paste target, apply has significant limitation in actually shooting indication of shots；In addition, army shoots
Indication of shots when number of rings judges use " even if touching on only lightly " even if rule, i.e., using shell hole most inner side point as calculating this shooting grade
Foundation point.If using traditional mode that achievement shell hole achievement is calculated according to distance between two points, will enter after shooting every time
The substantial amounts of calculating of row, it is impossible to ensure ageing.Therefore, research target surface information acquisition method and a kind of new target mode of sentencing turn into base
In the key that image procossing automatic target-indicating is popularized.
At present, image procossing theoretical research is just towards the trend development of multidisciplinary multidirectional mixing together, substantial amounts of engineering reality
Trample result to also indicate that, merging multidisciplinary multi-field target surface image procossing has turned into the research heat for solving automatic target-indicating practical problem
Point.
The content of the invention
Do not consider that interference, the number of rings decision procedures such as paste target do not meet for target surface acquisition of information in above-mentioned existing hit telling system
The actual problem of indication of shots, it is an object of the invention to provide a kind of target surface location of pixels coding method.
A kind of target surface location of pixels coding method, the method is used for chest silhouette target, comprises the following steps：
Step 1, obtains the initial target surface a of chest silhouette target, and target surface d is extracted from initial target surface ai；
Including：
Step 11, extracts and the pixel of condition A is met in initial target surface a forms target surface b, the condition A and be：The pixel
R color component values more than the pixel G color component values, and the pixel R color component values be more than the pixel
B color component values；
Step 12, HSI color spaces are transformed to by target surface b from RGB color, and target surface b is in HIS color spaces for extraction
Meet condition B pixel formed target surface c, the condition B be：The colourity of the pixel is more than 60 and less than 180, and the picture
The saturation degree of vegetarian refreshments is more than 0.3；
Step 13, carries out all pixels point of target surface c binaryzation and negates, obtains target surface d；
Step 14, carries out Xiao Bo projections to target surface d in a coordinate system, obtains projection fs of the target surface d in X-axisxX () takes
The projection f of value space and target surface d in Y-axisyY the interval of (), f is intercepted from target surface dx(x) and fyThe interval of (y)
As target surface di；The coordinate origin is in tetra- drift angles of target surface d, and X-axis is horizontal direction, and Y-axis is vertical direction；
Wherein, hxRepresent the quantity of the pixel that the grey scale pixel value projected in X-axis in target surface d is 0, hyRepresent target surface d
In the grey scale pixel value that projects in Y-axis be 0 pixel quantity；A-1, C-1 are respectively the wide and height of target surface d, gray (x,
Y) it is the gray value at pixel (x, y) place.
Step 2, to target surface diIn loop wire be fitted；
Including：
Step 21, to target surface diRim detection is carried out, edge chest silhouette target d is obtainedj, setting target surface diOn scoring ring region be
The area of K, K is SK；
K is divided into n-1 radius for Rn, area be SnConcentric circles, that is, form n region Qn；
Wherein, Q1=S1,Q2=S2-S1,…,Qn-1=Sn-1-Sn-2,Qn=SK-Sn-1, n=1,2 ...；
Step 22, performs step 221~step 224 to each region respectively, obtains target surface df；
Step 221, takes region QjMiddle any pixel point is traveled through 8 centered on the current pixel point as current pixel point
All pixels point in neighborhood, all pixels point in the neighborhood of current pixel point 8 is merged into the current pixel point same
Individual loop wire Cq；Wherein, CqZhou Changwei LCq, q=1,2 ...；
Step 222, repeat step 221, until region QjIn all of pixel all by as current pixel point or
Person is traversed, and performs step 223；
Step 223, removes LCq, q=1,2 ... preceding 20% loop wire in descending arrangement；
Step 224, if region QjMiddle Ring Line CqRadius be DqkIf,Then remove loop wire
Cq, q=1,2 ...；
Wherein, RjIt is region QjRadius；
Step 23, in region Qj, j=1 ... chooses target surface d successively in nfMiddle any pixel point is used as central pixel point, j
=1 ..., n；
If to each M in horizontal both sides of the central pixel pointmaxPixel is searched in/2, and the pixel for searching is with
The distance of imago vegetarian refreshments is more than or equal to Mmin/ 2, then retain the pixel for searching, the pixel for searching otherwise is removed, retain
Pixel constitute horizontal direction loop wire primitive Bl；Using the direction of X-axis as horizontal direction；
If to each M in vertical both sides of the central pixel pointmaxPixel is searched in/2, and the pixel for searching is with
The distance of imago vegetarian refreshments is more than or equal to Mmin/ 2, then retain the pixel for searching, the pixel for searching otherwise is removed, retain
Pixel constitute vertical direction loop wire primitive Br；Using the direction of Y-axis as vertical direction；
Wherein, M is target surface dfAll loop wires in adjacent ring linear distance average value, Mmin=(1-sec22.5 °) M, Mmax
=(1+sec22.5 °) M；
Loop wire primitive Bl, Br are constituted into loop wire primitive, and removes length in loop wire primitive and be less than LCminLoop wire primitive, obtain
To the parallel line primitives B of loop wire；Wherein, LCminIt is target surface dfIn most becate line girth；
Step 25, in region QjIn, j=1 ..., n, by region QjParallel line primitives B in length primitive most long make
It is initial seed point, is then scanned for according to adaptive optimal control degree function, connect parallel line primitives B, finally gives the ring of fitting
Line chart Bf。
Further, also include：
Step 3, in target surface diIn, the loop wire combined standard target surface according to fitting carries out position encoded to target surface pixel；
Including：
Step 31, to target surface diBinarization operation is carried out, and the circle for selecting radius to be 2 is corroded, after being corroded
Target surface e；
Rim detection is carried out to target surface e and obtains e0, and by e0Middle gray value is attached into J connection for the pixel of " 1 "
Composition, if any coordinator ejArea be Sej, j=1,2 ..., J, J are the natural number more than or equal to 2, retain Sej, j=
The coordinator of the first two, obtains target surface e in the arrangement of 1,2 ..., J descending1；
To target surface e1Hough loop truss is carried out, target surface e is obtained1Center of circle o0With radius R0；
Retain target surface e again1The maximum coordinator of area, obtains target surface e in middle coordinatorf；
efBoundary pixel point coordinates g (x, y) of middle effective coverage, the effective coverage is target surface efThe maximum company of middle area
It is connected into point；
In target surface efSpace coordinate conversion relation is set up and standard target surface between, coordinate transformation parameter a is obtainedmn,bmn：
Wherein, amn、bmnIt is coordinate transformation parameter, m is the integer more than or equal to 0 and less than or equal to 2, and n is more than or equal to 0
And the integer less than or equal to 2, x ', y ' are respectively target surface efThe coordinate of middle pixel, u, v are respectively pixel in standard target surface
Coordinate；
Step 32, appoints and takes target surface efIn pixel an O, the pixel O coordinate be (x0, y0If)
The then number of rings value OV=0 of pixel O；
If x0∈[gxmin,gxmax] andThe number of rings value O of note pixel OV=0；
Wherein, gxminIt is target surface efMinimum value of the boundary pixel point of middle effective coverage in x-axis, gxmaxIt is target surface efIn
Maximum of the boundary pixel point of effective coverage in x-axis；gyminIt is target surface efThe boundary pixel point of middle effective coverage is on the y axis
Minimum value, gymaxIt is target surface efThe boundary pixel point of middle effective coverage maximum on the y axis；
Otherwise, O points and o are connected0Line segment L, calculate the number of times T of pixel grey scale Spline smoothing on line segment L, pixel O's
Number of rings value OV=10-T；
Step 33, by chest silhouette target efThe coordinate (x, y) of middle all pixels point, gray value gray, place number of rings value OVWith array
Form storage, obtain target surface location of pixels coding F.
Further, coordinate origin described in step 14 is the upper left corner of target surface d.
Compared with prior art, the present invention has following technique effect：
(1) present invention, using the target position identification technology based on color characteristic, solves chest silhouette target and determines in chest silhouette target positioning
, easily by the problem that complex background, illumination variation interference, artificial interference are big, chest silhouette target positioning precision is high, manual intervention for target surface in position
Less, and not by illumination condition disturbed；
(2) present invention is fitted skill based on parallel lines when loop wire is fitted to the loop wire for extracting and genetic algorithm is attached
Art, solve paste target image loop wire fitting Middle Ring Line discontinuously, there are problems that abnormal loop wire, loop wire fitting precision is high, effect
It is really good；
(3) present situation that the present invention is big for target range ambient interferences in shooting, target surface pastes target, by the positional information of target surface pixel
It is combined with number of rings information, and this method is processed in units of pixel, and the process is completed in preshot, it is not take up penetrating
Hit the time, and shell hole achievement judged while shell hole is recognized, when being easy to meet the actual indication of shots of army with shell hole most on the inside of point
Place number of rings sentences target rule as shooting grade, takes into account indication of shots accuracy rate and ageing.
Brief description of the drawings
Fig. 1 (a) is initial target surface a；B () is through just extracting the target surface b for obtaining；C () is through extracting the target surface c for obtaining again；
D () is to carry out binaryzation to target surface c to negate the target surface d for obtaining；
Fig. 2 is target surface di；
Fig. 3 (a) is edge chest silhouette target dj；B () is target surface df；
Fig. 4 (a) is the loop wire primitive Bl of horizontal direction；B () is the loop wire primitive Br of vertical direction；C () is parallel lines base
First B；D () is the loop wire figure B of fittingf；
Fig. 5 (a) is the target surface e after corrosion；B () carries out rim detection and obtains e to target surface e0；C () is target surface e1；D () is target
Face e2；
Fig. 6 is target surface ef；
Fig. 7 is encoded for target surface location of pixels.
Specific embodiment
Below by embodiment and accompanying drawing, the present invention will be further described.
Embodiment
A kind of target surface location of pixels coding method is present embodiments provided, the method is used for chest silhouette target, comprises the following steps：
Step 1, such as shown in Fig. 1 (a), obtains the initial target surface a of chest silhouette target, and target surface d is extracted from initial target surface ai, such as
Fig. 2；
Detailed process is：
Step 11, because chest silhouette target paper color has " green constancy ", it is possible to select RGB color to extract chest
Target paper green characteristic, just splits chest silhouette target structural form.In automatic scoring round target system, for chest silhouette target target sheet region segmentation and calmly
For position, the profile parameters in chest silhouette target region are important, such as chest silhouette target contour edge, the coordinate position of each angle point.As for
The detailed information such as loop wire, numeral in chest silhouette target region can be ignored in the process.Therefore, the green characteristic according to target sheet is thick
The elementary contour structure that lines ground extracts chest silhouette target is the important content of chest silhouette target segmentation.
Extract satisfaction (g in initial target surface a (Fig. 1 (a))value＞ bvalue)∩(gvalue＞ rvalue) all pixels point P,
By pixel p-shaped into target surface b, such as Fig. 1 (b)；
Wherein, gvalue、bvalue、rvalueR, G, B color component value of pixel in respectively initial target surface a；
Step 12, due to the influence of illumination, the Chest bitmap extracted based on chest silhouette target green characteristic in RGB color
As in, a small amount of qualified background area is contained, in order to reduce influence of the illumination variation to recognizing, eliminate a small amount of background and do
Disturb, in the first Chest bitmap picture for extracting transformed into HSI (tone, saturation degree and brightness) color space, and based on tone
(Hue, H) component and saturation degree (Saturation, S) further extract chest silhouette target region.
Target surface b is transformed into HSI color spaces from RGB color, then during pixel P corresponds to HSI color spaces
Pixel P ', if pixel P ' meets 60 ＜ H ＜ 180 and S ＞ 0.3, extracts pixel p ', and pixel p ' is formed into target surface c,
Such as Fig. 1 (c)；
Wherein, H is the colourity of pixel p ', and s is pixel p ' saturation degrees, and I is pixel p ' brightness；
After successfully chest silhouette target target and background is separated, in order to the scope of image procossing is locked in into target area, also
It needs to be determined that position of the chest silhouette target in entire image, i.e., will obtain the coordinate of chest silhouette target profile.The present invention by extracting again
Chest bitmap picture carries out binaryzation, obtains corresponding binary map；Then projective transformation is carried out to chest silhouette target binary map, determines chest ring
Position of the target in figure.
Detailed process is：
Step 13, the present invention is negated by the way that all pixels point of target surface c is carried out into binaryzation, obtains target surface d, such as Fig. 1
(d)；
Step 14, carries out Xiao Bo projections to target surface d in a coordinate system, obtains projection fs of the target surface d in X-axisxX () takes
The projection f of value space and target surface d in Y-axisyY the interval of (), f is intercepted from target surface dx(x) and fyThe interval of (y)
As target surface di；The coordinate origin is the upper left corner of target surface d, and X-axis is horizontal direction, and Y-axis is vertical direction；
Wherein, hxRepresent the quantity of the pixel that the grey scale pixel value projected in X-axis in target surface d is 0, hyRepresent target surface d
In the grey scale pixel value that projects in Y-axis be 0 pixel quantity；A-1, C-1 are respectively the wide and height of target surface d, gray (x,
Y) it is the gray value at pixel (x, y) place.In the present embodiment, the wide and high of target surface d is 150 pixels.
In order to save target sheet in army's gunnery training, the frequency for changing target sheet is reduced, generally by the way of paste target manually
Instead of changing target, paste target is essentially to fill up shell hole with the paper block of same color, but due to time restriction and paper is block-shaped, area
Influence, unavoidably occur that white loop wire, white occur in the block-shaped excessive covering part loop wire of paper or green ring value region
There is green paper block situation in color region, the severe jamming continuity of scoring ring loop wire, integrality and authenticity.In addition, though preceding
Phase is corrected to distortion target sheet, but still there is partial deformation, in the Chest bitmap picture for finally giving, loop wire is discontinuous,
There is substantial amounts of funiclar curve noise jamming around loop wire and loop wire is one group of concentric, asymmetric class ellipse.
The present invention is to target surface diIn loop wire be fitted；
Step 21, to target surface diRim detection is carried out, edge chest silhouette target d is obtainedj(Fig. 3 (a)), setting target surface diOn target
Ring region is K, and the area of K is SK, and K has n bar scoring ring loop wires；
K is divided into n-1 radius for Rn, area be SnConcentric circles, that is, form n region Qn；In the present embodiment, such as
Shown in Fig. 3, n takes 5, under 256 × 256 pixel resolutions, R1Take 22 pixels, Rn-Rn-1It is 15 pixels；
Wherein, Q1=S1,Q2=S2-S1,…,Qn-1=Sn-1-Sn-2,Qn=SK-Sn-1, n=1,2 ...；
Step 22, appoints and takes region QnIn any one region Qj, j≤n；
Step 221, takes region QjMiddle any pixel point is traveled through 8 centered on the current pixel point as current pixel point
All pixels point (x, y) in neighborhood, all pixels point in the neighborhood of current pixel point 8 is merged into the current pixel point
Same loop wire Cq；Wherein, CqLength be LCq, q=1,2 ...；
Step 222, repeat step 221, until region QjIn all of pixel all by as current pixel point or
Untill person is traversed；
Step 223, removes LCq, q=1,2 ..., descending arrangement in preceding 20% loop wire；
Step 224, if region QjMiddle Ring Line CqWith concentric circles KnThe distance in the center of circle, i.e. loop wire CqRadius be DqkIf,Then remove loop wire Cq；
Wherein, RjIt is region QjRadius；
Step 23, to all region QnStep 22 is carried out, target edge d is obtainedf(Fig. 3 (b)), now target surface dfIn it is most short
Loop wire length is LCmin, LC in the present embodimentminTake 30 π pixels.
Step 24, in region QjIn choose target surface d successivelyfMiddle any pixel point as central pixel point, j=1 ..., n；
If to each side M of the central pixel pointmaxPixel is searched in/2, and the pixel for searching is with
The distance of imago vegetarian refreshments is more than or equal to Mmin/ 2, then retain the pixel for searching, the pixel for searching otherwise is removed, obtain
Loop wire the primitive Bl, such as Fig. 4 (a) of horizontal direction；
If to each M in both sides up and down of the central pixel pointmaxPixel is searched in/2, and the pixel for searching is with
The distance of imago vegetarian refreshments is more than or equal to Mmin/ 2, then retain the pixel for searching, the pixel for searching otherwise is removed, obtain
Loop wire the primitive Br, such as Fig. 4 (b) of vertical direction；
Wherein, M is the average value of distance between adjacent two loop wires in target surface df, Mmin=(1-sec22.5 °) M, Mmax=
(1+sec22.5°)M；In the present embodiment, M is 15 pixels.
Loop wire primitive Bl, Br are constituted into loop wire primitive, and removes length in loop wire primitive and be less than LCminLoop wire primitive, obtain
To the parallel line primitives B, such as Fig. 4 (c) of loop wire.
Step 25, for chest silhouette target concentric target rings loop wire complex situations, to overcome Hough transform and least square method computing
The shortcoming that amount is big, the fitting degree of accuracy is low.
In region QjIn, by region QjParallel line primitives B in during length primitive most long is input into as genetic algorithm
Initial seed point, then scans for according to adaptive optimal control degree function, connects parallel line primitives B, finally gives the loop wire of fitting
Figure Bf, such as Fig. 4 (d).
Table 1, table 2 sets forth the fitting loop wire circularity and run time calculated using 3 kinds of algorithms of different.
Table 1 is fitted the value of loop wire circularity
The algorithm operation time of table 2
From the results, it was seen that either fitting loop wire is still fitted loop wire and reality with the size of actual loop wire registration
On the height of loop wire circularity similitude, the degree of accuracy of scoring ring loop wire that the fitting of this paper algorithms is obtained be superior to Hough transform and
The loop wire that least square fitting is obtained, for the accurate indication of shots of automatic scoring round target system next step provides authentic data；The algorithm of table 2
The time of operation shows that the Algorithms T-cbmplexity is better than Hough transform more than least square method, substantially meets system requirements.
In target surface diIn, the loop wire according to fitting is encoded to target surface location of pixels.
From the point of view of the result of target position identification, the Chest bitmap picture of extraction also includes partial white background area.According to shooting
The rule that achievement judges, the white portion that bullet falls beyond chest silhouette target profile is considered as and misses the target, and achievement is designated as 0 ring, therefore, must
The effective coverage of chest silhouette target, the i.e. edge contour of chest silhouette target must first be determined.
In order to eliminate the loop wire in chest silhouette target and numerical portion.
Step 31, to target surface diBinarization operation is carried out, and the circle for selecting radius to be 2 is corroded, after being corroded
Target surface e, such as Fig. 5 (a)；
Rim detection is carried out to target surface e and obtains e0(Fig. 5 (b)), and gray value is more for the pixel of " 1 " is attached into
Individual closed area, coordinator is labeled as by each closed area, if the area of coordinator is Sei, retain SeiDescending is arranged
The coordinator of middle the first two, obtains target surface e1, such as Fig. 5 (c)；The loop wire and numerical portion in chest silhouette target are eliminated, but is schemed
In the interference of segment word and background is still present.
To target surface e1Hough loop truss is carried out, center of circle o is obtained0With radius R0, obtain target surface e2, such as Fig. 5 (d)；
To target surface e2Zone marker is carried out again, retains the coordinator of area maximum in coordinator, obtain target surface ef,
Such as Fig. 6；
Processed more than, complete extraction eliminates the chest silhouette target edge contour of loop wire and numeral.Chest silhouette target will be located at
The coordinate of pixel is stored on edge contour, is subsequently to judge whether point of impact misses the target there is provided data foundation.
Storage efBoundary pixel point coordinates g (x, y) of middle effective coverage；
In target surface efSpace coordinate conversion relation is set up and standard target surface between, coordinate transformation parameter a is obtainedmn,bmn：
Wherein, amn、bmnIt is coordinate transformation parameter, m is 0~2 integer, and n is 0~2 integer, and x, y are respectively target surface ef
The coordinate of middle pixel, u, v are respectively the coordinate of pixel in standard target surface；
Step 32, appoints and takes target surface efIn a pixelIfThen pixelNumber of rings value OV
=0；
IfAndNote OV=0；
Wherein, gxminIt is target surface efMinimum value of the boundary pixel point of middle effective coverage in x-axis, gxmaxIt is target surface efIn
Maximum of the boundary pixel point of effective coverage in x-axis；gyminIt is target surface efThe boundary pixel point of middle effective coverage is on the y axis
Minimum value, gymaxIt is target surface efThe boundary pixel point of middle effective coverage maximum on the y axis；
Otherwise, O points and o are connected0Line segment L, calculate the number of times T, O of pixel grey scale Spline smoothing on line segment LV=10-T.
Due to loop wire figure BfThe gray value of the pixel on Middle Ring Line is 1, and the gray value of the pixel on non-loop wire is 0,
So calculating the number of times T of pixel grey scale Spline smoothing on line segment L, O points and o are can obtain0Between by several loop wires；
Target surface e is traveled through successivelyfMiddle all pixels point, calculates the number of rings value of pixel position.
Step 33, by chest silhouette target efCoordinate (x, y), gray value gray, the place number of rings value v of middle all pixels point are with array
Form storage, obtain target surface location of pixels coding F.
Claims (3)
1. a kind of target surface location of pixels coding method, the method is used for chest silhouette target, it is characterised in that comprise the following steps：
Step 1, obtains the initial target surface a of chest silhouette target, and target surface d is extracted from initial target surface ai；
Including：
Step 11, extracts and the pixel of condition A is met in initial target surface a forms target surface b, the condition A and be：The R of the pixel
Color component value more than the pixel G color component values, and the pixel R color component values more than the pixel B
Color component value；
Step 12, HSI color spaces are transformed to by target surface b from RGB color, are extracted target surface b and are met in HIS color spaces
The pixel of condition B forms target surface c, the condition B：The colourity of the pixel is more than 60 and less than 180, and the pixel
Saturation degree be more than 0.3；
Step 13, carries out all pixels point of target surface c binaryzation and negates, obtains target surface d；
Step 14, carries out Xiao Bo projections to target surface d in a coordinate system, obtains projection fs of the target surface d in X-axisxThe valued space of (x)
Projection f with target surface d in Y-axisyY the interval of (), f is intercepted from target surface dx(x) and fyY the interval of () is used as target
Face di；The coordinate origin is in tetra- drift angles of target surface d, and X-axis is horizontal direction, and Y-axis is vertical direction；
Wherein, hxRepresent the quantity of the pixel that the grey scale pixel value projected in X-axis in target surface d is 0, hyRepresent throwing in target surface d
Grey scale pixel value on shadow to Y-axis is the quantity of 0 pixel；A-1, C-1 are respectively the wide and height of target surface d, and gray (x, y) is
The gray value at pixel (x, y) place.
Step 2, to target surface diIn loop wire be fitted；
Including：
Step 21, to target surface diRim detection is carried out, edge chest silhouette target d is obtainedj, setting target surface diOn scoring ring region be K, K's
Area is SK；
K is divided into n-1 radius for Rn, area be SnConcentric circles, that is, form n region Qn；
Wherein, Q1=S1,Q2=S2-S1,…,Qn-1=Sn-1-Sn-2,Qn=SK-Sn-1, n=1,2 ...；
Step 22, performs step 221~step 224 to each region respectively, obtains target surface df；
Step 221, takes region Qj, used as current pixel point, traversal is centered on the current pixel point in 8 neighborhoods for middle any pixel point
All pixels point, all pixels point in the neighborhood of current pixel point 8 and the current pixel point are merged into same loop wire
Cq；Wherein, CqZhou Changwei LCq, q=1,2 ...；
Step 222, repeat step 221, until region QjIn all of pixel all by as current pixel point or by time
Go through, perform step 223；
Step 223, removes LCq, q=1,2 ... preceding 20% loop wire in descending arrangement；
Step 224, if region QjMiddle Ring Line CqRadius be DqkIf,Then remove loop wire Cq, q=
1,2,...；
Wherein, RjIt is region QjRadius；
Step 23, in region Qj, j=1 ... chooses target surface d successively in nfMiddle any pixel point is used as central pixel point, j=
1,…,n；
If to each M in horizontal both sides of the central pixel pointmaxPixel, and the pixel for searching and middle imago are searched in/2
The distance of vegetarian refreshments is more than or equal to Mmin/ 2, then retain the pixel for searching, otherwise remove the pixel for searching, the picture of reservation
Vegetarian refreshments constitutes the loop wire primitive Bl of horizontal direction；Using the direction of X-axis as horizontal direction；
If to each M in vertical both sides of the central pixel pointmaxPixel, and the pixel for searching and middle imago are searched in/2
The distance of vegetarian refreshments is more than or equal to Mmin/ 2, then retain the pixel for searching, otherwise remove the pixel for searching, the picture of reservation
Vegetarian refreshments constitutes the loop wire primitive Br of vertical direction；Using the direction of Y-axis as vertical direction；
Wherein, M is target surface dfAll loop wires in adjacent ring linear distance average value, Mmin=(1-sec22.5 °) M, Mmax=(1+
sec22.5°)M；
Loop wire primitive Bl, Br are constituted into loop wire primitive, and removes length in loop wire primitive and be less than LCminLoop wire primitive, obtain ring
The parallel line primitives B of line；Wherein, LCminIt is target surface dfIn most becate line girth；
Step 25, in region QjIn, j=1 ..., n, by region QjParallel line primitives B in length primitive most long as first
Beginning seed point, then scans for according to adaptive optimal control degree function, connects parallel line primitives B, finally gives the loop wire figure of fitting
Bf。
2. the method for obtaining target surface loop wire as claimed in claim 1, it is characterised in that also include：
Step 3, in target surface diIn, the loop wire combined standard target surface according to fitting carries out position encoded to target surface pixel；
Including：
Step 31, to target surface diBinarization operation is carried out, and the circle for selecting radius to be 2 is corroded, the target surface e after being corroded；
Rim detection is carried out to target surface e and obtains e0, and by e0Middle gray value is connected into for the pixel of " 1 " is attached into J
Point, if any coordinator ejArea be Sej, j=1,2 ..., J, J are the natural number more than or equal to 2, retain Sej, j=1,
The coordinator of the first two, obtains target surface e in the arrangement of 2 ..., J descending1；
To target surface e1Hough loop truss is carried out, target surface e is obtained1Center of circle o0With radius R0；
Retain target surface e again1The maximum coordinator of area, obtains target surface e in middle coordinatorf；
efBoundary pixel point coordinates g (x, y) of middle effective coverage, the effective coverage is target surface efMiddle area is maximum to be connected into
Point；
In target surface efSpace coordinate conversion relation is set up and standard target surface between, coordinate transformation parameter a is obtainedmn,bmn：
Wherein, amn、bmnIt is coordinate transformation parameter, m is the integer more than or equal to 0 and less than or equal to 2, and n is more than or equal to 0 and is less than
Integer equal to 2, x ', y ' are respectively target surface efThe coordinate of middle pixel, u, v are respectively the coordinate of pixel in standard target surface；
Step 32, appoints and takes target surface efIn pixel an O, the pixel O coordinate be (x0, y0If)
The then number of rings value OV=0 of pixel O；
If x0∈[gxmin,gxmax] andThe number of rings value O of note pixel OV=0；
Wherein, gxminIt is target surface efMinimum value of the boundary pixel point of middle effective coverage in x-axis, gxmaxIt is target surface efMiddle effective district
Maximum of the boundary pixel point in domain in x-axis；gyminIt is target surface efThe boundary pixel point of middle effective coverage minimum on the y axis
Value, gymaxIt is target surface efThe boundary pixel point of middle effective coverage maximum on the y axis；
Otherwise, O points and o are connected0Line segment L, calculate the number of times T of pixel grey scale Spline smoothing on line segment L, the number of rings value of pixel O
OV=10-T；
Step 33, by chest silhouette target efThe coordinate (x, y) of middle all pixels point, gray value gray, place number of rings value OVWith the shape of array
Formula is stored, and obtains target surface location of pixels coding F.
3. target surface location of pixels coding method as claimed in claim 1, it is characterised in that coordinate system described in step 14 is former
Point is the upper left corner of target surface d.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611195876.1A CN106845506B (en) | 2016-12-22 | 2016-12-22 | A kind of target surface location of pixels coding method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611195876.1A CN106845506B (en) | 2016-12-22 | 2016-12-22 | A kind of target surface location of pixels coding method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106845506A true CN106845506A (en) | 2017-06-13 |
| CN106845506B CN106845506B (en) | 2019-09-20 |
Family
ID=59136947
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611195876.1A Active CN106845506B (en) | 2016-12-22 | 2016-12-22 | A kind of target surface location of pixels coding method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106845506B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109635679A (en) * | 2018-11-26 | 2019-04-16 | 深圳深知未来智能有限公司 | A kind of real-time target sheet positioning and loop wire recognition methods |
| CN109948630A (en) * | 2019-03-19 | 2019-06-28 | 深圳初影科技有限公司 | Recognition methods, device, system and the storage medium of target sheet image |
| CN113750498A (en) * | 2021-09-15 | 2021-12-07 | 深圳市一岭高尔夫科技有限公司 | Golf scoring system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH1145335A (en) * | 1997-07-28 | 1999-02-16 | Ishikawajima Harima Heavy Ind Co Ltd | Method and device for detecting closed area |
| CN1546937A (en) * | 2003-12-04 | 2004-11-17 | 华南理工大学 | Ball firing system and method for identifying slow fire shot hole thereof |
| CN1546938A (en) * | 2003-12-04 | 2004-11-17 | 华南理工大学 | Ball firing system and method for identifying rapid fire shot hole thereof |
| CN101937509A (en) * | 2010-08-17 | 2011-01-05 | 西安理工大学 | Automatic target identifying method based on image processing technology |
| CN203464855U (en) * | 2013-07-15 | 2014-03-05 | 中国人民武装警察部队工程大学 | Training target convenient for shooting by two persons |
| CN105976366A (en) * | 2016-04-29 | 2016-09-28 | 浙江大华技术股份有限公司 | Method and device for target loop positioning |
-
2016
- 2016-12-22 CN CN201611195876.1A patent/CN106845506B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH1145335A (en) * | 1997-07-28 | 1999-02-16 | Ishikawajima Harima Heavy Ind Co Ltd | Method and device for detecting closed area |
| CN1546937A (en) * | 2003-12-04 | 2004-11-17 | 华南理工大学 | Ball firing system and method for identifying slow fire shot hole thereof |
| CN1546938A (en) * | 2003-12-04 | 2004-11-17 | 华南理工大学 | Ball firing system and method for identifying rapid fire shot hole thereof |
| CN101937509A (en) * | 2010-08-17 | 2011-01-05 | 西安理工大学 | Automatic target identifying method based on image processing technology |
| CN203464855U (en) * | 2013-07-15 | 2014-03-05 | 中国人民武装警察部队工程大学 | Training target convenient for shooting by two persons |
| CN105976366A (en) * | 2016-04-29 | 2016-09-28 | 浙江大华技术股份有限公司 | Method and device for target loop positioning |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109635679A (en) * | 2018-11-26 | 2019-04-16 | 深圳深知未来智能有限公司 | A kind of real-time target sheet positioning and loop wire recognition methods |
| CN109635679B (en) * | 2018-11-26 | 2023-01-24 | 深圳深知未来智能有限公司 | Real-time target paper positioning and loop line identification method |
| CN109948630A (en) * | 2019-03-19 | 2019-06-28 | 深圳初影科技有限公司 | Recognition methods, device, system and the storage medium of target sheet image |
| CN109948630B (en) * | 2019-03-19 | 2020-03-31 | 深圳初影科技有限公司 | Target paper image identification method, device and system and storage medium |
| CN113750498A (en) * | 2021-09-15 | 2021-12-07 | 深圳市一岭高尔夫科技有限公司 | Golf scoring system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106845506B (en) | 2019-09-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107239748A (en) | Robot target identification and localization method based on gridiron pattern calibration technique | |
| CN104751465A (en) | ORB (oriented brief) image feature registration method based on LK (Lucas-Kanade) optical flow constraint | |
| CN108447022A (en) | Moving target joining method based on single fixing camera image sequence | |
| CN108197642A (en) | A kind of seal discrimination method and device | |
| JP2009086926A (en) | Image recognition method and device | |
| CN107452028B (en) | Method and device for determining position information of target image | |
| CN110400278A (en) | A kind of full-automatic bearing calibration, device and the equipment of color of image and geometric distortion | |
| CN106845506B (en) | A kind of target surface location of pixels coding method | |
| CN109271937A (en) | Athletic ground Marker Identity method and system based on image procossing | |
| WO2021098163A1 (en) | Corner-based aerial target detection method | |
| CN114418899B (en) | Self-adaptive repairing method and system for self-color printing and readable storage medium | |
| CN108133460B (en) | Color type state indicator identification method suitable for electric power robot | |
| CN106097354A (en) | A kind of combining adaptive Gauss Face Detection and the hand images dividing method of region growing | |
| CN105488475B (en) | Method for detecting human face in mobile phone | |
| CN108830800A (en) | The luminance raising Enhancement Method of image under a kind of half-light scene | |
| CN107945221A (en) | A kind of three-dimensional scenic feature representation based on RGB D images and high-precision matching process | |
| CN103049731A (en) | Decoding method for point-distributed color coding marks | |
| CN110111387A (en) | A kind of pointer gauge positioning and reading algorithm based on dial plate feature | |
| CN114782953A (en) | Seal intelligent identification method | |
| CN110264397A (en) | A kind of method and apparatus of effective coverage that extracting fish eye images | |
| CN107992856A (en) | High score remote sensing building effects detection method under City scenarios | |
| CN110909772B (en) | High-precision real-time multi-scale dial pointer detection method and system | |
| CN109948461A (en) | A kind of sign language image partition method based on center coordination and range conversion | |
| CN112634262A (en) | Writing quality evaluation method based on Internet | |
| JPH05181411A (en) | Map information collation and update system |
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 |