CN102831619A - Expansion target stable trace point extraction method - Google Patents

Abstract

The invention discloses an expansion target stable trace point extraction method which comprises the following steps of: firstly, recovering the expansion target successfully identified by early association into a binary image; then, performing contour tracing to obtain the contour point array arranged in order; smoothing the target contour by use of a method based on chord-arc ratio filtering, removing the sharp arc section on the target contour, and linearly connecting two endpoints of the removed arc section to form a new closed contour after filtering; calculating the head direction of the target by combining the target motion direction and the main shaft direction, and then obtaining the forefront point in the head direction of the target; and calculating the centroid in the local round area as a trace point by taking the forefront point as the center. Through the method, the finally calculated target trace point is stable and reliable, severe shake is avoided, the tracing stability is improved, and the risk of tracing loss is reduced.

Description

A kind of expansion target tenacious tracking point extracting method

Technical field

The invention belongs to image processing field, relate to a kind of target following point extracting method, particularly a kind of based on the method for distilling of string arc than the expansion target tenacious tracking point of contour smoothing.

Background technology

Along with the development of optical imagery track demand,, especially more and more urgent to the tenacious tracking demand of expansion target to the tenacious tracking of target.When target and imaging tracking equipment close together, target is rendered as the characteristic of expansion target, occupied the visual field 1/3 in addition be full of the visual field.With regard to the extracting position of target following point, originally no longer be suitable for expanding the tracking of target based on the tracking mode of the centre of form, barycenter, very easily cause cutting apart and lost efficacy and lose objects.Also exist this moment because the trace point drifting problem that the accumulation of tracking error causes based on the mode of correlation tracking, and finally cause tracking to be lost.

At present; In order to solve the problem that trace point is chosen; Have document to adopt the method for based target shape analysis, through the contour feature of evaluating objects, the rapid wear point that searches out target is as trace point; But these methods all exist the restriction that receives targeted attitude and potentially unstable, and the contour feature that perhaps causes extracting owing to the reason of cutting apart is unstable and finally influence the extraction of trace point.Other class methods are to adopt the method for extracting the target angle point, obtain trace point through the coupling of angle point, but these class methods receive noise effect easily; The angle point that extracts is unstable; Smoother when objective contour, even angle point causes easily to follow the tracks of and loses when not existing.

Summary of the invention

The present invention provides a kind of and stablizes method for distilling based on the string arc than contour smoothing and target head tracking point in order to overcome the deficiency of prior art, has solved in original method when following the tracks of the expansion target trace point and has slided easily or acutely rock and cause following the tracks of the problem of losing.

In order to achieve the above object, the extraction step of a kind of method for distilling of expanding target tenacious tracking point provided by the invention is following:

Step S1: for the expansion target in the visual field of passing through binaryzation and success association, identifying, the score segment table after the usage flag should be expanded target and revert to the two-value target image, made and had only the expansion target in the binary image;

Step S2: utilize contour tracing method that the profile of two-value target image is followed the tracks of, obtain the point array of arranging in order of two-value target, the point sum is designated as lenArcContour;

Step S3: utilize tactic point array; Direction according to appointment travels through the profile array clockwise or counterclockwise; Write down smallest chord arc in the corresponding regional area of each point than the information of point, obtain level and smooth segmental arc, and according to the threshold decision sharp-pointed segmental arc of needs filtering whether; If not filtering segmental arc does not then connect the end points of segmental arc; If the filtering segmental arc then connects the end points of segmental arc, constitute new sealing objective contour;

Step S4: combine the direction of motion and the major axes orientation of two-value target, confirm the cephalad direction of two-value target;

Step S5: utilize target cephalad direction and new sealing objective contour; Seek the forefront point of new sealing objective contour upper edge target cephalad direction; Calculating is the centre of form of the regional area at center with the forefront point, is final trace point with this centre of form that calculates, after the present frame processing finishes; Return S1, carry out next frame and handle.

The present invention chooses to compare with the conventional target trace point has following advantage:

(1) trace point of choosing is positioned at the target head, makes target and background in image, occupy certain proportion all the time separately, helps cutting apart of target, can avoid target to be full of the failure of cutting apart that the visual field causes.

(2) adopt and to delete on the profile unstablely based on the mode of contour smoothing, sharp-pointed part helps extracting trace point at the stable position of segmentation result, can avoid extracting the problem of the violent saltus step of point owing to the tracking that reason causes such as cutting apart.

(3) trace point that adopts the present invention to propose is expanded target following, has reduced the tracking risk of missing, makes and follows the tracks of more robust.

Trace point and the difference between the actual value that the inventive method is extracted are littler, and rock forr a short time, and it is littler than former method to extract the influence that point changes by the target body, and the stability that trace point extracts obviously is superior to original method.

Description of drawings

Fig. 1 expands the process flow diagram that target tenacious tracking point extracts for the present invention;

Fig. 2 a is the binary map that expansion score segment table recovers among the embodiment; Fig. 2 b is a profile tracking results of the present invention;

Fig. 3 is the synoptic diagram of contour smoothing of the present invention and connection;

Fig. 4 is the contour smoothing process flow diagram based on string arc ratio of the present invention;

Fig. 5 is to Fig. 2 b contour smoothing and connects the back result;

Fig. 6 a, Fig. 6 b are for calculating the synoptic diagram of target head tracking point;

Fig. 7 is for calculating the process flow diagram of target cephalad direction;

Fig. 8 is the true reference value of directions X trace point and original trace point extraction result who directly gets forefront point methods and the inventive method without profile filtering;

Fig. 9 directly gets forefront point methods and the trace point result of the inventive method extraction and the difference between the reference value for directions X is original without the profile filtering method;

Figure 10 is the true reference value of Y direction trace point and original trace point extraction result who directly gets forefront point methods and the inventive method without profile filtering;

Figure 11 directly gets forefront point methods and the trace point result of the inventive method extraction and the difference between the reference value for the Y direction is original without the profile filtering method;

Embodiment

For making the object of the invention, technical scheme and advantage clearer, below in conjunction with specific embodiment, and with reference to accompanying drawing, to further explain of the present invention.

The extraction of expansion target tenacious tracking point of the present invention can realize by hardware mode, also can realize by software mode.For example on personal computer, industrial computer and server, install and carry out, also can method of the present invention make embedded chip and embody with the form of hardware with the form of software.Specific embodiments of the invention describes with reference to the accompanying drawings.

The present invention is applicable to that imaging tracking equipment follows the tracks of the expansion target in the visual field, and especially when the target range imaging device was nearer, the target body formed images comparatively obvious in the visual field, and target has strong wake flame or tail cigarette simultaneously.Concrete, the flow process that the inventive method expansion target tenacious tracking point extracts is as shown in Figure 1, comprises that the expansion target reverts to bianry image, is used to carry out profile and follows the tracks of; Profile is followed the tracks of, and is used to obtain tactic profile array; Based on the contour smoothing of string arc ratio, be used to delete the string arc than sharp-pointed segmental arc less than the setting threshold thresholding; The target cephalad direction is confirmed, through extracting target travel direction and major axes orientation, finally confirms the direction of target head, is used for finally extracting the target following point in head zone; Trace point extracts, and through in the centre of form of head forward position point for calculating impact point in the regional area at center, obtains final tracking and extracts point;

Concrete, the present invention includes following steps:

1. the expansion target reverts to bianry image: target is being passed through pre-service with dividing mark and after carrying out related identification; Obtained the correct target of current tracking; Follow-up for ease profile Tracking Realization through the successful score segment table of association, reverts to bianry image with target.Remaining false target has all obtained rejecting in this image.

2. profile is followed the tracks of: to the above-mentioned two-value target figure that recovers, utilize contour tracing method to obtain the point array of arranging in order of this target, add up the point sum simultaneously, be designated as lenArcContour.The method that profile is followed the tracks of is referring to < Visual C++ Digital Image Processing >, He Bin, Ma Tianyu, Wang Yunjian etc., People's Telecon Publishing House, 2001,435-444.

3. contour smoothing: utilize the order point array that obtains, the segmental arc that filtering is comparatively sharp-pointed also connects with straight line the segmental arc after the filtering, constitutes new closed outline, and specifically implementation method is following:

3a: by point array clockwise or that the counter clockwise direction traversal is arranged in order, first point that note is worked as pre-treatment is P _Cur, if the first point P _CurBe not marked as and handle, then carry out 3b, if the first point P _CurBe marked as and handle, then carry out 3e;

3b: by with 3a in each drops on the first point P in the same direction search profile array that travels through _CurFor the center of circle, R the second point P in the part-circular zone of radius _iCalculate along the same direction first point P _CurWith the second point P _iArc length lenArc, if lenArc≤lenArcContour/2, wherein lenArcContour is a point sum, then calculates the first point P _CurWith the second point P _iBetween chord length lenChord, and calculate string arc between these 2 than RatioCAcur [i], RatioCAcur [i]=lenChord/lenArc calculates the unidirectional second point P in part-circular _iSearch finishes, if otherwise lenArc＞lenArcContour/2 explains that then the arc of the point-to-point transmission of current statistics is a major arc, is left intact, continue to calculate the string arc ratio between the next point in first point and the part-circular zone.Wherein: i=1 ... N; N is in the part-circular zone and the first point P _CurSatisfy the number of the point of lenArc≤lenArcContour/2;

3c: calculate the first point P in the part-circular zone _CurWith all the other second point P _iThe minimum value of string arc ratio is expressed as Ratio _Min=min{RatioCAcur [i] }, the Ratio that record is obtained _MinWith the first point P _CurThe pairing second point P _iWherein, min is for getting minimum value function;

Chord length lenChord between first point and second point _ThValue is 7～10, lenChord _ThValue rule be the chord length of two end points needing the filtering segmental arc.

Judge whether segmental arc is the threshold value Ratio of sharp-pointed segmental arc _ThValue is 0.5～0.7, Ratio _ThValue rule be the upper bound of the string arc ratio that needs the filtering segmental arc.

3d: if Ratio _Min＜Ratio _ThAnd corresponding chord length lenChord satisfies lenChord＞lenCho _Th, then with the current first point P _CurBe labeled as and should be connected, and with the first point P _CurBe labeled as treated, with the second point P _iBe made as current process points, P _Cur← P _i, change 3a, on the contrary if Ratio _Min>=Ratio _ThThen explain and do not have point and current point P in the part-circular zone _CurThe segmental arc that constituent ratio is more sharp-pointed is only to the first point P _CurBe labeled as treatedly, change step 3a; Ratio wherein _ThExpression judges that segmental arc is the threshold value of sharp-pointed segmental arc, is set at 0.5 or 0.64 or 0.7 during enforcement; LenChord _ThRepresent the chord length between first point and second point, value is 7 or 9 or 10 during enforcement;

3e: the 2nd traversal profile array, if the current traversal first point P _CurBe marked as effectively also and should be connected, then from the first point P _CurBeginning is until the second point P that first point need connect _iBetween point be labeled as invalid, the first point P wherein _CurWith the second point P _iIt is invalid not to be labeled as, and promptly deletes the first point P _CurWith the second point P _iBetween segmental arc, and adopt the method for interpolation to connect the first point P _CurWith the second point P _i, the point of inserting is joined in the profile array, then P _Cur← P _i, finish up to the point traversal.

4. the target cephalad direction is extracted: through calculating the direction of motion and the major axes orientation of two-value target, obtain the direction of target head, direction of motion and the major axes orientation concrete steps of asking for the two-value target are following:

A) direction of motion of calculating target, mode is following:

V _Ai＝A _i-A _i-1 (1)

V _Ei＝E _i-E _i-1 (2)

V wherein _Ai, V _EiBe respectively Azimuth, Speed, Altitude and the angular altitude speed of i moment target with respect to survey station; The aggregate velocity of Azimuth, Speed, Altitude and angular altitude speed is the current movement velocity of target; Movement velocity is transformed into the picture plane of image, the angle theta of the aggregate velocity that obtains projection and X axle forward on the picture plane _v, A wherein _i, E _iBe respectively the composite value of i moment target, A with respect to survey station position angle and angular altitude _I-1, E _I-1Be the corresponding constantly composite value of i-1.

B) with the moment of inertia m of number of contours set of calculated target ₁₁, m ₂₀, m ₀₂

Wherein: $m_{i,j}=\underset{(x,y)}{\mathrm{\&Sigma;}}\underset{\&Element;R}{\mathrm{\&Sigma;}}{(x-\stackrel{\&OverBar;}{x})}^i{(x-\stackrel{\&OverBar;}{y})}^j---\left(3\right)$

θ ₁＝tan ^-1(2m ₁₁/(m ₂₀-m ₀₂))/2 (4)

θ ₂＝tan ^-1(2m ₁₁/(m ₂₀-m ₀₂))/2+π/2 (5)

I″(θ)＝[2(m ₂₀-m ₀₂)cos2θ+4m ₁₁sin2θ] (6)

m _{I, j}Expression i * j rank central moment, wherein i, j are the exponent number of central moment, θ ₁, θ ₂Be respectively the moment of inertia I (θ that asks _i) orientation angle, I " (θ _i) be I (θ _i) second derivative,

With If X coordinate and the Y coordinate of representing the target barycenter respectively are with θ ₁With θ ₂Bring (6) formula respectively into, as I " (θ _i)＞0, i=1,2 o'clock, this θ _iBe the minimum principal axis of inertia direction of being asked, this direction is the bearing of trend of target, establishes angle maximum deviation Δ θ=10 °, gets θ _iWith θ _vBetween angle be that the cephalad direction of target is designated as θ less than the direction of Δ θ _Axi

5. the extraction implementation of target following point is following:

A) objective contour is pressed cephalad direction and be rotated, calculate forefront point P along cephalad direction _HeadBe:

xvalue[i]＝P _ix×cosθ _axi+P _iy×sinθ _axi (7)

P wherein _iX and P _iY is respectively the horizontal ordinate and the ordinate value of i point, cos θ _AxiWith as sin θ _AxiBe respectively angle θ _AxiCosine value and sine value, xvalue [i] is that the profile coordinate points is according to θ _AxiAngle around the postrotational X coordinate figure of X axle, the point when getting maximal value then is the forefront point P of objective contour _Head

B) with the profile coordinate points according to θ _AxiAngle obtain peaked some P around the postrotational X coordinate figure of X axle xvalue [i] _HeadBe the regional final trace point P of centre of form conduct of the part-circular that belongs to impact point in the center calculation binary map _Track

The present invention has chosen an expansion target image sequence, and it is followed the tracks of, and wherein Fig. 1 is a process flow diagram of the present invention;

Fig. 2 a is the binary map that expansion score segment table recovers among the embodiment, and Fig. 2 b is a profile tracking results of the present invention, and the processing of contour smoothing filtering is based on that this tracking results carries out;

Fig. 3 is the synoptic diagram of contour smoothing of the present invention and connection, and solid stain representative is as the first point P of pre-treatment in the synoptic diagram _Cur, second point that soft dot representative is handled, circle is for being the center of circle with first point in the processing procedure, R is the process range of radius;

Fig. 4 is the contour smoothing process flow diagram based on string arc ratio of the present invention, and concrete performing step is following:

Step 4a: profile is followed the tracks of, formation order profile array;

Step 4b: dispose if having a few in the profile, change step 4h, do not dispose else if, change step 4c;

Step 4c: get next point P by assigned direction (clockwise or counterclockwise) _Cur

Step 4d: with the point P that gets among the step 4c _CurBe the center of circle, R is that the border circular areas of radius is interior by same assigned direction, gets the second point P successively _i, calculating the arc length lenArc of the minor arc that first point and second point purchase, judgment formula: lenArc≤lenArcContour/2 is adopted in the judgement of minor arc, and wherein lenArcContour is the point sum.If lenArc≤lenArcContour/2 then calculates the first point P _CurWith the second point P _iBetween chord length lenChord, and calculate string arc between these 2 than RatioCAcur [i], RatioCAcur [i]=lenChord/lenArc.If otherwise lenArc＞lenArcContour/2 explains that then the arc of the point-to-point transmission of current statistics is a major arc, be left intact, continue to calculate the string arc ratio between the next point in first point and the part-circular zone.Wherein: i=1 ... N; N is in the part-circular zone and P _CurSatisfy the number of the point of lenArc≤lenArcContour/2.If second point in the border circular areas has traveled through end, then change step 4e;

Step 4e:Ratio _Min=min{RatioCAcur [i] }, the Ratio that record is obtained _MinWith the first point P _CurThe pairing second point P _iWherein, min is for getting minimum value function, Ratio _MinP in the expression border circular areas _CurWith all the other P _iThe minimum value of string arc ratio is changeed step 4f;

Step 4f: judge whether smallest chord arc ratio compares threshold value less than predefined string arc in the border circular areas; Be set at 0.64 among the embodiment, and if corresponding chord length satisfy greater than setting thresholding, the embodiment value is 7; If; Change step 4g, otherwise do not have sharp-pointed segmental arc in the explanation border circular areas, change step 4b;

Step 4g: first point and the segmental arc between second point that setting obtains smallest chord arc ratio are invalid, promptly delete all point on this segmental arc.Mark first point should be connected, and writes down second point that first point need connect.Change step 4b;

Step 4h: travel through all point again,, then keep if current processing point effectively and not is marked as connection; Current else if processing point is effective and be labeled as connection, then adopts the mode of interpolation to insert new point to second point that with current point be connected of current processing point with record.Traversal finishes the back and forms new closed outline;

Fig. 5 is to Fig. 2 b contour smoothing and connects the back result, wherein figure below elliptical section be divided into carry out level and smooth and connect after new closed outline, the comparison that oblique line part segmental arc is disallowable on the figure sharply and between successive frame, show unsettled target body part;

Fig. 6 a is for new profile is the regional synoptic diagram of the selected calculating centre of form in the center of circle with the forefront point that the target cephalad direction calculates after the filtering to utilize, and Fig. 6 b is the part that is used to calculate the centre of form that selects;

Fig. 7 is for calculating the schematic flow sheet of target head tracking point, and step is following:

Step 7a: calculate the speed on target azimuth and the height direction,

V _Ai＝A _i-A _i-1 (8)

V _Ei＝E _i-E _i-1 (9)

V wherein _Ai, V _EiTarget is with respect to the Azimuth, Speed, Altitude and the angular altitude speed of survey station constantly to be respectively i, and its aggregate velocity is the current movement velocity of target, wherein A _i, E _iBe respectively the composite value of i moment target, A with respect to survey station position angle and angular altitude _I-1, E _I-1Be the corresponding constantly composite value of i-1.

Step 7b: target azimuth and the aggregate velocity of height on the direction convert the velocity reversal θ that the motion vector angle of planar pixel obtains into _v(with the angle of X axle forward);

Step 7c: the central moment m that calculates target ₁₁, m ₂₀, m ₀₂

Wherein: $m_{i,j}=\underset{(x,y)}{\mathrm{\&Sigma;}}\underset{\&Element;R}{\mathrm{\&Sigma;}}{(x-\stackrel{\&OverBar;}{x})}^i{(x-\stackrel{\&OverBar;}{y})}^j---\left(10\right)$

m _{I, j}Expression i * j rank central moment, wherein i, j are the exponent number of central moment, θ ₁, θ ₂Be respectively the moment of inertia I (θ that asks _i) orientation angle, I " (θ _i) be I (θ _i) second derivative,

With

X coordinate and the Y coordinate of representing the target barycenter respectively;

Step 7d: the direction of calculating target minimum rotation inertia

θ ₁＝tan ^-1(2m ₁₁/(m ₂₀-m ₀₂))/2 (11)

θ ₂＝tan ^-1(2m ₁₁/(m ₂₀-m ₀₂))/2+π/2 (12)

I″(θ)＝[2(m ₂₀-m ₀₂)cos2θ+4m ₁₁sin2θ] (13)

If with θ ₁With θ ₂Bring (13) formula respectively into, as I " (θ _i)＞0, i=1,2 o'clock, this θ _iBe the minimum principal axis of inertia direction of being asked, this direction is the bearing of trend of target.

Step 7e: calculate the target cephalad direction, obtain the close direction of minimum rotation inertia direction and direction of motion, establish angle maximum deviation Δ θ=10 °, get θ as the target cephalad direction _iWith θ _vBetween angle be the cephalad direction of target less than the direction of Δ θ, be designated as θ _Axi

Fig. 8 is the directions X miss distance, the true reference value of trace point and original trace point extraction result who directly gets forefront point methods and the inventive method without profile filtering; Can find out that the result that the inventive method is extracted approaches and actual value more.

Fig. 9 directly gets forefront point methods and the trace point result of the inventive method extraction and the difference between the reference value for directions X is original without the profile filtering method; Can find out that the result that the inventive method directions X extracts and the error of true value are much smaller than former method.

Figure 10 is a Y direction miss distance, the true reference value of trace point and original trace point extraction result who directly gets forefront point methods and the inventive method without profile filtering; Can find out that the result that the inventive method is extracted approaches and actual value more.

Figure 11 directly gets forefront point methods and the trace point result of the inventive method extraction and the difference between the reference value for the Y direction is original without the profile filtering method; Can find out that the inventive method Y direction is extracted the error of result and true value much smaller than former method.

As can be seen from the above results, trace point and the difference between the actual value that the inventive method is extracted are littler, and rock forr a short time, and it is littler than former method to extract the influence that point changes by the target body, and the stability that trace point extracts obviously is better than original method.

The above; Be merely the embodiment among the present invention, but protection scope of the present invention is not limited thereto, anyly is familiar with this technological people in the technical scope that the present invention disclosed; Conversion or the replacement expected can be understood, all of the present invention comprising within the scope should be encompassed in.

Claims (7)

1. a method for distilling of expanding target tenacious tracking point is characterized in that, it is following that this method comprises step:

Step S1: for the expansion target in the visual field of passing through binaryzation and success association, identifying, the score segment table after the usage flag should be expanded target and revert to the two-value target image, made and had only the expansion target in the binary image;

Step S2: utilize contour tracing method that the profile of two-value target image is followed the tracks of, obtain the point array of arranging in order of two-value target, the point sum is designated as lenArcContour;

Step S3: utilize tactic point array; Direction according to appointment travels through the profile array clockwise or counterclockwise; Write down the information of the smallest chord arc ratio point in the corresponding regional area of each point; And according to the threshold decision sharp-pointed segmental arc of needs filtering whether, if not filtering segmental arc does not then connect the end points of segmental arc; If the filtering segmental arc then connects the end points of segmental arc,, constitute new sealing objective contour to realize contour smoothing;

Step S4: combine the direction of motion and the major axes orientation of two-value target, confirm the cephalad direction of two-value target;

Step S5: utilize target cephalad direction and new sealing objective contour; Seek the forefront point of new sealing objective contour upper edge target cephalad direction; Calculating is the centre of form of the regional area at center with the forefront point, is final trace point with this centre of form that calculates, after the present frame processing finishes; Return S1, carry out next frame and handle.

2. expand the method for distilling of target following point according to claim 1, it is characterized in that: the step of said contour smoothing is following:

Step a: by point array clockwise or that the counter clockwise direction traversal is arranged in order, first point that note is worked as pre-treatment is P _Cur, if the first point P _CurBe not marked as and handle, if execution in step b then is the first point P _CurBe marked as and handle, then execution in step e;

Step b: by with step a in each drops on the first point P in the same direction search profile array that travels through _CurFor the center of circle, R the second point P in the part-circular zone of radius _iCalculate along the same direction first point P _CurWith the second point P _iArc length lenArc, if lenArc≤lenArcContour/2, wherein lenArcContour is a point sum, then calculates the first point P _CurWith the second point P _iBetween chord length lenChord, and calculate the first point P _CurWith the second point P _iBetween the string arc than RatioCAcur [i], RatioCAcur [i]=lenChord/lenArc calculates the unidirectional second point P in part-circular _iSearch finishes;, be left intact if otherwise lenArc＞lenArcContour/2 explains that then the arc of the point-to-point transmission of current statistics is a major arc; Continue to calculate the string arc ratio between the next point in first point and the part-circular zone, wherein: i=1 ... N; N is in the part-circular zone and the first point P _CurSatisfy the number of the point of lenArc≤lenArcContour/2;

Step c: calculate the first point P in the part-circular zone _CurWith all the other second point P _iThe minimum value of string arc ratio is expressed as Ratio _Min=min{RatioCAcur [i] }, the Ratio that record is obtained _MinWith the first point P _CurThe pairing second point P _i, wherein, min is for getting minimum value function;

Steps d: if Ratio _Min＜Ratio _ThAnd corresponding chord length lenChord satisfies lenChord＞lenChord _Th, then with the current first point P _CurBe labeled as and should be connected, and with the first point P _CurBe labeled as treated, with the second point P _iBe made as current process points, P _Cur← P _i, change step a, on the contrary if Ratio _Min>=Ratio _ThThen explain and do not have point and the current first point P in the part-circular zone _CurThe segmental arc that constituent ratio is more sharp-pointed is only to the first point P _CurBe labeled as treatedly, change step a; Ratio wherein _ThExpression judges whether segmental arc is the threshold value of sharp-pointed segmental arc; LenChord _ThRepresent the chord length between first point and second point;

Step e: the 2nd traversal profile array, if the current traversal first point P _CurBe marked as effectively also and should be connected, then from the first point P _CurBeginning is until the second point P that first point need connect _iBetween point be labeled as invalid, the first point P wherein _CurWith the second point P _iIt is invalid not to be labeled as, and promptly deletes the first point P _CurWith the second point P _iBetween segmental arc, and adopt the method for interpolation to connect the first point P _CurWith the second point P _i, the point of inserting is joined in the profile array, then P _Cur← P _i, finish up to the point traversal.

3. like the method for distilling of the said expansion target following point of claim 2, it is characterized in that: the chord length thresholding lenChord between first point and second point _ThValue is 7～10, lenChord _ThValue be the chord length that needs two end points of filtering segmental arc.

4. like the method for distilling of the said expansion target following point of claim 2, it is characterized in that: judge whether segmental arc is the threshold value Ratio of sharp-pointed segmental arc _ThValue is 0.5～0.7, Ratio _ThValue rule be the upper bound of the string arc ratio that needs the filtering segmental arc.

5. expand the method for distilling of target following point according to claim 1, it is characterized in that, direction of motion and the major axes orientation step of asking for the two-value target are following:

Step 51: calculate the direction of motion of target, mode is following:

V _Ai＝A _i-A _i-1 (1)

V _Ei＝E _i-E _i-1 (2)

V wherein _Ai, V _EiBe respectively Azimuth, Speed, Altitude and the angular altitude speed of i moment target with respect to survey station; The aggregate velocity of Azimuth, Speed, Altitude and angular altitude speed is the current movement velocity of target; Movement velocity is transformed into the picture plane of image, obtains the angle theta of aggregate velocity projection and X axle forward on the picture plane _v, A wherein _i, E _iBe respectively the composite value of i moment target, A with respect to survey station position angle and angular altitude _I-1, E _I-1Be the corresponding constantly composite value of i-1;

Step 52: with the moment of inertia m of number of contours set of calculated target ₁₁, m ₂₀, m ₀₂, further calculate

Wherein: $m_{i,j}=\underset{(x,y)}{\mathrm{\&Sigma;}}\underset{\&Element;R}{\mathrm{\&Sigma;}}{(x-\stackrel{\&OverBar;}{x})}^i{(x-\stackrel{\&OverBar;}{y})}^j---\left(3\right)$

θ ₁＝tan ^-1(2m ₁₁/(m ₂₀-m ₀₂))/2 (4)

θ ₂＝tan ^-1(2m ₁₁/(m ₂₀-m ₀₂))/2+π/2 (5)

I″(θ)＝[2(m ₂₀-m ₀₂)cos20+4m ₁₁sin2θ] (6)

m _{I, j}Expression i * j rank central moment, wherein i, j are the exponent number of central moment, θ ₁, θ ₂Be respectively the moment of inertia I (θ that asks _i) orientation angle, I " (θ _i) be I (θ _i) second derivative,

With

If X coordinate and the Y coordinate of representing the target barycenter respectively are with θ ₁With θ ₂Bring (6) formula respectively into, as I " (θ _i)＞0, i=1,2 o'clock, this θ _iBe the minimum principal axis of inertia direction of being asked, this direction is the bearing of trend of target, establishes angle maximum deviation Δ θ=10 °, gets θ _iWith θ _vBetween angle be that the cephalad direction of target is designated as θ less than the direction of Δ θ _Axi

6. expand the method for distilling of target following point according to claim 1, it is characterized in that: said forefront point P _HeadRepresent as follows: xvalue [i]=P _iX * cos θ _Axi+ P _iY * sin θ _Axi(7) P wherein _iX and P _iY is respectively the horizontal ordinate and the ordinate value of i point, cos θ _AxiWith as sin θ _AxiBe respectively angle θ _AxiCosine value and sine value, xvalue [i] is that the profile coordinate points is according to θ _AxiAngle around the postrotational X coordinate figure of X axle, the point when getting maximal value then is the forefront point P of objective contour _Head

7. like the method for distilling of the said expansion target following point of claim 6, it is characterized in that: with the profile coordinate points according to θ _AxiAngle obtain peaked some P around the postrotational X coordinate figure of X axle xvalue [i] _HeadBe the regional final trace point P of centre of form conduct of the part-circular that belongs to impact point in the center calculation bianry image _Track

Images