CN106373159A - Simplified unmanned aerial vehicle multi-target location method - Google Patents

Abstract

The present invention discloses a simplified unmanned aerial vehicle multi-target location method. The simplified unmanned aerial vehicle multi-target location method can perform location in real time while realizing multiple targets, and is simple in hardware device. An area array CCD camera is employed to perform grounding imaging, and obtaining the pixel coordinates of each target; performing distortion correction, and obtaining the ideal imaging points of each target; constructing the sight vectors of a main target, a sub target and a lower machine point in the camera coordinate systems; employing the coordinate transformation to obtain the sight vectors of the main target, the sub target and the lower machine point in the aerial carrier geographic coordinate system; calculating the coordinates of each target in the aerial carrier geographic coordinate system according to the distance from the main target and the sub target to a photoelectricity platform and the sight vectors of the main target and the sub target in the aerial carrier geographic coordinate system; and finally, obtaining the geodetic coordinates through the coordinate transformation, namely locating the result.

Description

A kind of unmanned plane multi-target orientation method of simplification

Technical field

The present invention relates to photoelectronic imaging field of measuring technique is and in particular to a kind of unmanned plane Multi-target position side of simplification Method.

Background technology

General UAV electro-optical's platform measures the angle of itself relative aircraft by angular transducer, by range finder using laser Export itself targeted distance of relative image central cross silk, be combined with aircraft gps position data and inertial navigation attitude data, Target positioning is realized using coordinate transformation method, this localization method based on attitude measurement/laser ranging model is to multiple mesh When mark implements positioning, need the space frequently changing photoelectric platform to point to and carry out multiple bearing, elapsed time length is it is difficult to simultaneously right Multiple targets are implemented to position it is impossible to adapt to the situation that modern battlefield situation is changeable in real time, more than destination number in real time or quasi real time.

Existing multi-target orientation method is based primarily upon multiple sensor platforms, and such as cn201410078624.5 discloses one Plant the method that Multi-target position is carried out using multigroup video camera, it is fixed that it adopts optics Convergent measurement localization method to realize multiple target Position, but its hardware device is complicated, and real-time is poor.

Content of the invention

In view of this, the invention provides a kind of unmanned plane multi-target orientation method of simplification, it is capable of to multiple target Real-time positioning, and hardware device is simple.

The unmanned plane multi-target orientation method of the simplification of the present invention, comprises the steps:

Step 1, unmanned plane is imaged over the ground using face battle array ccd camera, obtains the pixel coordinate of each target, wherein, positioned at taking the photograph The target of camera field of view center is referred to as major heading, the target of other positions referred to as time target in visual field；

Step 2, the pixel coordinate using secondary target step 1 being obtained based on aberration rate method is modified, and obtains each time The pixel coordinate of the ideal image point of target；The pixel coordinate of major heading is the pixel coordinate of its ideal image point；According to master The pixel coordinate of ideal image point of target and time target and camera origin, construction major heading and time target are being taken the photograph respectively Sight line vector under camera coordinate system；The sight line vector under camera coordinate system is put, as camera initial point points under construction machine The vector of point under machine；

Step 3, according between camera coordinate system and carrier aircraft coordinate system, between carrier aircraft coordinate system and carrier aircraft geographic coordinate system Spin matrix, the sight line vector put under camera coordinate system under major heading, secondary target and machine is transformed into carrier aircraft is geographical to sit In mark system；

Step 4, the distance according to primary and secondary target and photoelectric platform and primary and secondary target are in carrier aircraft geographic coordinate system Sight line vector, calculates coordinate in carrier aircraft geographic coordinate system for each target；

Wherein, major heading and the distance of photoelectric platform are λ₁, obtained by range finder using laser measurement；Secondary target and photoelectric platform Distance be λ₂, by formulaCalculate and obtain, h=λ₁Cos α, α are sight line vector and the point under machine of major heading The pixel angle of sight between sight line vector,Pixel sight line between the sight line vector of point under sight line vector for secondary target and machine Angle；

Step 5, in conjunction with unmanned plane in the course angle beta of the position coordinateses of the earth's core rectangular coordinate system in space and aircraft, pitching Angle ε and roll angle γ, Coordinate Conversion in carrier aircraft geographic coordinate system for each target that step 4 is obtained is sat to the earth's core space right-angle In mark system；Then according to the transition matrix between the earth's core rectangular coordinate system in space and earth coordinates, each target is empty in the earth's core Between Coordinate Conversion in rectangular coordinate system in earth coordinates, obtain the geodetic coordinates of each target, the earth of described each target Coordinate is the positioning result of corresponding target.

Further, in described step 1, multiple targets are obtained using image segmentation, frame difference method or optical flow method simultaneously Pixel coordinate.

Further, in described step 2, the computational methods of the pixel coordinate of ideal image point of each target are as follows:

Step 2.1, the pixel coordinate (u of the target being obtained according to step 1_d,v_d) calculate the physical coordinates (x of this target_d, y_d):

x_d=(u_d-u₁)d_x, y_d=(v_d-v₁)d_y

In formula, u₁And v₁It is the pixel coordinate of picture centre, d_xAnd d_yIt is the physical size in x and y direction for the single pixel；

Step 2.2, according to lens distortion rate d, calculates target ideal as the physical coordinates (x of imaging point_r,y_r):

$x_r=\frac{x_d}{1+d},y_r=\frac{y_d}{1+d}$

Step 2.3, by target ideal as the physical coordinates (x of imaging point_r,y_r) calculate target ideal as the pixel of imaging point Coordinate (u_r,v_r)

u_r=u₁+x_r/d_x, v_r=v₁+y_r/d_y

Further, in described step 3,It is respectively under major heading, secondary target and machine and put in carrier aircraft geography Sight line vector under coordinate system；

$\stackrel{\→}{s_v}=r_{vb}{r}_{bc}\stackrel{\→}{s_c}=\left[\begin{array}{ccc}{c}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}& -c_{\mathrm{\γ}}{s}_{\mathrm{\β}}+s_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}& s_{\mathrm{\γ}}{s}_{\mathrm{\β}}+c_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}\\ c_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}& c_{\mathrm{\γ}}{c}_{\mathrm{\β}}+s_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}& -s_{\mathrm{\γ}}{c}_{\mathrm{\β}}+c_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}\\ -s_{\mathrm{\ϵ}}& s_{\mathrm{\γ}}{c}_{\mathrm{\ϵ}}& c_{\mathrm{\γ}}{c}_{\mathrm{\ϵ}}\end{array}\right]\left[\begin{array}{ccc}{c}_{\mathrm{\θ}}{s}_{\mathrm{\ψ}}& -s_{\mathrm{\θ}}& c_{\mathrm{\θ}}{c}_{\mathrm{\ψ}}\\ s_{\mathrm{\ψ}}{s}_{\mathrm{\θ}}& c_{\mathrm{\θ}}& s_{\mathrm{\θ}}{c}_{\mathrm{\ψ}}\\ -c_{\mathrm{\ψ}}& 0& s_{\mathrm{\ψ}}\end{array}\right]\stackrel{\→}{s_c}$

$\stackrel{\→}{t_v}=r_{vb}{r}_{bc}\stackrel{\→}{t_c}=\left[\begin{array}{ccc}{c}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}& -c_{\mathrm{\γ}}{s}_{\mathrm{\β}}+s_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}& s_{\mathrm{\γ}}{s}_{\mathrm{\β}}+c_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}\\ c_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}& c_{\mathrm{\γ}}{c}_{\mathrm{\β}}+s_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}& -s_{\mathrm{\γ}}{c}_{\mathrm{\β}}+c_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}\\ -s_{\mathrm{\ϵ}}& s_{\mathrm{\γ}}{c}_{\mathrm{\ϵ}}& c_{\mathrm{\γ}}{c}_{\mathrm{\ϵ}}\end{array}\right]\left[\begin{array}{ccc}{c}_{\mathrm{\θ}}{s}_{\mathrm{\ψ}}& -s_{\mathrm{\θ}}& c_{\mathrm{\θ}}{c}_{\mathrm{\ψ}}\\ s_{\mathrm{\ψ}}{s}_{\mathrm{\θ}}& c_{\mathrm{\θ}}& s_{\mathrm{\θ}}{c}_{\mathrm{\ψ}}\\ -c_{\mathrm{\ψ}}& 0& s_{\mathrm{\ψ}}\end{array}\right]\stackrel{\→}{t_c}$

$\stackrel{\→}{j_v}={\left[\begin{array}{ccc}0& 0& 1\end{array}\right]}^t$

Wherein,It is respectively and put the sight line vector under camera coordinate system under major heading, secondary target and machine； r_vbRepresent the spin matrix being tied to carrier aircraft geographic coordinate system from carrier aircraft coordinate, r_bcRepresent and be tied to carrier aircraft coordinate from camera coordinates The spin matrix of system, c_*Represent cos (*), s_*Represent sin (*)；θ is airborne photoelectric platform azimuth；ψ is airborne photoelectric platform The angle of site；β is carrier aircraft course angle；ε is the carrier aircraft angle of pitch；γ is carrier aircraft roll angle.

Further, in described step 5, the positioning result of continuous multiple frames image is carried out at recurrence least square filtering Reason, filter result is final positioning result.

Further, when recurrence least square Filtering Processing is carried out to continuous multiple frames framing result, using boat position Projectional technique obtains the aircraft position coordinate of interframe.

Beneficial effect:

The present invention is used for unmanned plane multi-target orientation method, and available one side battle array ccd sensor enters to multiple targets simultaneously Row geo-location, improves reconnaissance efficiency and real-time, thus adapting to the feelings that modern battlefield situation is changeable in real time, more than destination number Condition；Multi-target orientation method is combined with target detection technique and can enter Mobile state locating and tracking to multiple moving targets, obtains The parameter such as the movement locus of multiple moving targets and movement velocity, to realize multiobject scout with strike integrated have important Meaning.Process the location data of multiple image, multiple static targets to ground with dead reckoning using the method that rls filtering is combined Quickly it is accurately positioned, the gps of available general precision and navigation attitude measuring apparatus obtain higher target location accuracy, thus Significantly reduce equipment cost.

Brief description

Fig. 1 is localization method schematic flow sheet of the present invention.

Fig. 2 is one side battle array ccd sensor Multi-target position model schematic of the present invention.Wherein,

Fig. 3 is definition and its mutual relation schematic diagram of each coordinate system of preferred embodiment of the present invention Multi-target position model.

Fig. 4 is the Coordinate Conversion schematic flow sheet of preferred embodiment of the present invention multi-target orientation method.

Fig. 5 is recurrence least square filtering algorithm schematic flow sheet.

Wherein, f_cFor camera coordinate system；f_bFor carrier aircraft coordinate system；f_vFor carrier aircraft geographic coordinate system；G projects for video camera Center；P is major heading, positioned at the target of picture centre；Q is time target, positioned at the target of image other positions；K is point under machine； J is point k picture point on the image plane under machine；F is major heading p picture point on the image plane；T is that time target q is put down in image Ideal image point on face；T ＇ is time target q actual image point on the image plane；λ₁For between major heading p and aircraft platform Distance；λ₂For the distance between secondary target q and aircraft platform；H is the relative altitude between each target and aircraft platform；C is to take the photograph Camera coordinate system；B is carrier aircraft coordinate system；V is carrier aircraft geographic coordinate system；E is the earth's core rectangular coordinate system in space；G is geodetic coordinates System；θ is photoelectric platform azimuth；ψ is the photoelectric platform angle of site；β is carrier aircraft course angle；ε is the carrier aircraft angle of pitch；γ is carrier aircraft Roll angle；x_c, y_c, z_cIt is respectively three coordinate axess of camera coordinate system；x_b, y_b, z_bIt is respectively three seats of carrier aircraft coordinate system Parameter；x_v, y_v, z_vIt is respectively three coordinate axess of carrier aircraft geographic coordinate system；x_e, y_e, z_eIt is respectively the earth's core rectangular coordinate system in space Three coordinate axess；U, v are respectively two coordinate axess of image pixel coordinates system, the row of u axle labelling image, v axle labelling image Row, unit be pixel；X, y are respectively two coordinate axess of image physical coordinates system, and x-axis is parallel with u axle, and y-axis is put down with v axle OK, this coordinate system is in units of m or mm；ψ₁For the complementary angle of photoelectric platform angle of site ψ, equal to 90 ° of-ψ；l₀The earth for carrier aircraft Longitude；m₀Geodetic latitude for carrier aircraft；h₀Geodetic height for carrier aircraft；D, e are respectively carrier aircraft geographic coordinate system and the earth's core space right-angle Interim coordinate system between coordinate system, d axle and z_eAxle is parallel, e axle and y_eAxle is parallel；x_kOriginal positioning number for kth frame image According to, k=1,2 ..., t, wherein t are the totalframes of input picture；x_kFor the location data obtaining after rls filtering.

Specific embodiment

Develop simultaneously embodiment below in conjunction with the accompanying drawings, describes the present invention.

The invention provides a kind of unmanned plane multi-target orientation method of simplification, by setting up tight positioning geometry mould Type, calculates the distance between target and UAV electro-optical's platform and angular relationship, realizes single-frame imagess by coordinate transform many Target real-time geographic positioning, with solve existing based on the multi-target orientation method of unmanned plane multisensor platform lead to hard The technical problem of part equipment complexity, real-time and poor reliability.

Present invention is generally directed to the photoelectric platform that small and medium size unmanned aerial vehicles are carried, it is imaged over the ground using one side battle array ccd sensor, Implement low-to-medium altitude reconnaissance flight, flying height relatively low (generally less than 3km), the angle of visual field is 30 °～40 °, and single image ground is covered Lid scope is 1～2km, does not in most cases have too big fluctuating it is believed that near flat, in single image each target with fly Relative altitude between machine is identical, sets up Multi-target position model according to the image-forming principle of one side battle array ccd sensor, using pixel Sight line vector method calculates the distance between each target and photoelectric platform and angular relationship, obtains single width by homogeneous coordinate transformation The geodetic coordinates of each target in image, realizes that single image is multiobject to be positioned in real time or quasi real time.In the present invention, take the photograph being located at The target of camera field of view center is referred to as major heading, and the target referred to as time target of other positions in visual field, this target positions Method can achieve the geo-location of optional position target in camera field of view, and flow chart is as shown in figure 1, specifically include following step Rapid:

Step 1, according to the image-forming principle of one side battle array ccd sensor, obtains the pixel of each target using object detection unit Coordinate, as shown in Figure 1 and Figure 2.

Wherein, object detection unit can be detected using image segmentation, frame difference method or optical flow method simultaneously multiple quiet Only or moving target pixel coordinate:

Image is divided into target area and background area according to gray threshold or marginal information by image segmentation, calculates mesh Mark regional center coordinate is as the pixel coordinate of target；

Frame difference method passes through the moving target of consecutive frame image pixel difference energy quick detection pixel characteristic change, realizes letter Single, requirement of real-time can be met；

Optical flow method estimates the sports ground of image by the changing features of sequential frame image respective pixel, by similar motion arrow Amount merges into moving target.

Step 2, is modified to the target picture point skew that camera lens distortion causes using the method based on aberration rate, Obtain its ideal image position so as to meet pin-hole imaging model.Then the pixel according to primary and secondary target ideal image point is sat Point under mark, camera origin, machine, respectively construction major heading, put under secondary target and machine sight line under camera coordinate system to Amount.

The optical lens of general designed, designed all can provide corresponding aberration rate, or is obtained in ground survey by optical device To the aberration rate of camera lens, then it is modified accordingly, aberration rate is defined as follows

$d=\frac{\mathrm{\η}-\mathrm{\ζ}}{\mathrm{\ζ}}\×100\%---\left(1\right)$

In formula, d is lens distortion rate, and η is actual imaging height, and ζ is ideal image height.

According to the aberration rate of optical lens, ideal image position ζ can be released as follows

$\mathrm{\ζ}=\frac{\mathrm{\η}}{1+d}---\left(2\right)$

The specifically comprising the following steps that of lens distortion calibration

Step 2.1, calculates its physical coordinates according to pixel coordinate in fault image for the target:

x_d=(u_d-u₁)d_x, y_d=(v_d-v₁)d_y(3)

U in formula₁, v₁It is the pixel coordinate of picture centre, d_xAnd d_yIt is the physical size in x and y direction for the single pixel.

Step 2.2, aberration rate d being given according to system, calculate the physical coordinates of ideal image point

$x_r=\frac{x_d}{1+d},y_r=\frac{y_d}{1+d}---\left(4\right)$

Step 2.3, calculates its pixel coordinate by the physical coordinates of ideal image point

u_r=u₁+x_r/d_x, v_r=v₁+y_r/d_y(5)

After corrective lens distortion, picture point on image for secondary target q moves to ideal position t, camera by distorted position t ＇ Projection centre g, secondary target q and its on image 3 points of corresponding ideal position t point-blank, meet pin-hole imaging mould Type, as shown in Fig. 2 construct the sight line vector of each target, wherein, major heading according to the pixel coordinate of the ideal image point of each target Pixel coordinate be the pixel coordinate of its ideal image point；If the sight line vector of point k is respectively under major heading p, secondary target q, machine ForThe sight line vector put under machine under camera coordinate system is under camera initial point sensing machine The vector of point.

Construct following coordinate system: camera coordinate system, image physical coordinates system, carrier aircraft coordinate system, carrier aircraft geographic coordinate system, The earth's core rectangular coordinate system in space, earth coordinates, as shown in Figure 3.

Wherein, camera coordinate system (c or f_c): initial point is video camera projection centre g, x_cAxle, y_cAxle respectively with image pixel Coordinate system u axle (row of labelling image, unit be pixel), v axle (row of labelling image) is parallel and direction is consistent.

Image physical coordinates system, it is principle point location that initial point is located at camera optical axis with the intersection point of the plane of delineation, and x-axis, y-axis are divided Not with u axle, v axle is parallel and direction is consistent, this coordinate system is in units of m or mm.

Carrier aircraft coordinate system (b or f_b): initial point is navigation attitude measuring system barycenter, and general navigation attitude measuring system is installed on light level On the level reference of platform, navigation attitude measuring system barycenter and video camera projection centre, apart from very little, can be approximately considered both and overlap, x_bAxle is 0 ° of direction of navigation attitude measuring system, y_bAxle is 90 ° of directions of navigation attitude measuring system, z_bIt is true that axle passes through right-hand screw rule Fixed, the photoelectric platform of the azimuth angle theta of photoelectric platform interior angle encoder output and angle of site ψ and range finder using laser output with Distance lambda between field of view center target₁It is this coordinate system relative.

Carrier aircraft geographic coordinate system (v or f_v): initial point be located at navigation attitude measuring system barycenter, be ned (north east down, East northeast ground) coordinate system, the carrier aircraft course angle beta of navigation attitude measuring system output, angle of pitch ε is this coordinate system relative with roll angle γ 's.

Wgs-84 the earth's core rectangular coordinate system in space (e): initial point is earth centroid, z_eAxle points to the association of bih1984.0 definition View earth polar (ctp) direction, x_eAxle points to the zero degree meridian plane of bih1984.0 and the intersection point in ctp equator, y_eAxle and z_e、x_eAxle is constituted Right-handed coordinate system.

Wgs-84 earth coordinates (g): zero and the sensing of three axles are identical with rectangular coordinate system in space, using the earth warp Degree (l), geodetic latitude (m) to describe locus, the carrier aircraft position (l of gps output with geodetic height (h)₀,m₀,h₀) it is relatively large Ground coordinate system.

Target position fixing process according to the position and attitude parameter of photoelectric platform and aircraft, by target from camera coordinate system Pixel coordinate is transformed into the geodetic coordinates under wgs-84 earth coordinates, thus obtaining the actual geographic coordinate of each target.

Coordinate Conversion flow process is as shown in Figure 4.

Step 3, calculates the pixel angle of sight of itself and picture centre major heading according to the sight line vector of each target.

Wherein, the sight line vector of major heading pSight line vector with secondary target qCoordinate in camera coordinate system is respectively ForWherein f is camera focus, and unit is pixel, (u₀,v₀) it is point f Pixel coordinate, after distortion correction major heading p on image corresponding picture point f be located at picture centre；(u, v) is the pixel of point t Coordinate.

Calculate the sight line vector of point k under major heading p, secondary target q and machineSeat in carrier aircraft geographic coordinate system It is designated as

$\stackrel{\→}{s_v}=r_{vc}\stackrel{\→}{s_c}=r_{vb}{r}_{bc}\stackrel{\→}{s_c}=\left[\begin{array}{ccc}{c}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}& -c_{\mathrm{\γ}}{s}_{\mathrm{\β}}+s_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}& s_{\mathrm{\γ}}{s}_{\mathrm{\β}}+c_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}\\ c_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}& c_{\mathrm{\γ}}{c}_{\mathrm{\β}}+s_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}& -s_{\mathrm{\γ}}{c}_{\mathrm{\β}}+c_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}\\ -s_{\mathrm{\ϵ}}& s_{\mathrm{\γ}}{c}_{\mathrm{\ϵ}}& c_{\mathrm{\γ}}{c}_{\mathrm{\ϵ}}\end{array}\right]\left[\begin{array}{ccc}{c}_{\mathrm{\θ}}{s}_{\mathrm{\ψ}}& -s_{\mathrm{\θ}}& c_{\mathrm{\θ}}{c}_{\mathrm{\ψ}}\\ s_{\mathrm{\ψ}}{s}_{\mathrm{\θ}}& c_{\mathrm{\θ}}& s_{\mathrm{\θ}}{c}_{\mathrm{\ψ}}\\ -c_{\mathrm{\ψ}}& 0& s_{\mathrm{\ψ}}\end{array}\right]\stackrel{\→}{s_c}---\left(6\right)$

$\stackrel{\→}{t_v}=r_{vc}\stackrel{\→}{t_c}=r_{vb}{r}_{bc}\stackrel{\→}{t_c}=\left[\begin{array}{ccc}{c}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}& -c_{\mathrm{\γ}}{s}_{\mathrm{\β}}+s_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}& s_{\mathrm{\γ}}{s}_{\mathrm{\β}}+c_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}\\ c_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}& c_{\mathrm{\γ}}{c}_{\mathrm{\β}}+s_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}& -s_{\mathrm{\γ}}{c}_{\mathrm{\β}}+c_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}\\ -s_{\mathrm{\ϵ}}& s_{\mathrm{\γ}}{c}_{\mathrm{\ϵ}}& c_{\mathrm{\γ}}{c}_{\mathrm{\ϵ}}\end{array}\right]\left[\begin{array}{ccc}{c}_{\mathrm{\θ}}{s}_{\mathrm{\ψ}}& -s_{\mathrm{\θ}}& c_{\mathrm{\θ}}{c}_{\mathrm{\ψ}}\\ s_{\mathrm{\ψ}}{s}_{\mathrm{\θ}}& c_{\mathrm{\θ}}& s_{\mathrm{\θ}}{c}_{\mathrm{\ψ}}\\ -c_{\mathrm{\ψ}}& 0& s_{\mathrm{\ψ}}\end{array}\right]\stackrel{\→}{t_c}---\left(7\right)$

$\stackrel{\→}{j_v}={\left[\begin{array}{ccc}0& 0& 1\end{array}\right]}^t---\left(8\right)$

In formula, r_vbRepresent the spin matrix being tied to carrier aircraft geographic coordinate system from carrier aircraft coordinate, r_bcRepresent from camera coordinates It is tied to the spin matrix of carrier aircraft coordinate system, r_vcRepresent the spin matrix being tied to carrier aircraft geographic coordinate system from camera coordinates, wherein, c_*=cos (*), s_*=sin (*).

CalculateWithBetween pixel angle of sight α,WithBetween the pixel angle of sightAs follows respectively

Assume that the corresponding ground region of single image is flat, the relative altitude between each target and photoelectric platform is identical, profit Obtain major heading p and the distance between aircraft platform λ with range finder using laser measurement₁, it is calculated as follows each target and photoelectric platform Relative altitude h and secondary target and photoelectric platform distance lambda₂.

Step 4, the range finder using laser measurement according within photoelectric platform obtains the distance between major heading and aircraft, according to The angular encoder measurement of platform interior obtains camera optical axis relative to the azimuth of aircraft platform and the angle of site, in conjunction with each target The pixel angle of sight and major heading between calculates angle and distance relation between each target and aircraft platform；

According to secondary target range value λ₂And its sight line vector in carrier aircraft geographic coordinate systemCalculate time target carrying Coordinate in machine geographic coordinate system is

${\left[\begin{array}{ccc}{x}_v& y_v& z_v\end{array}\right]}^t={\mathrm{\λ}}_2\frac{\stackrel{\→}{t_v}}{\left|\right|t_v\left|\right|}---\left(11\right)$

Step 5, in conjunction with aircraft position data (the longitude l of gps alignment system output₀, latitude m₀With geodetic height h₀), aviation The aspect data (course angle beta, angle of pitch ε and roll angle γ) that attitude measurement system (imu) exports, is become by homogeneous coordinates The method of changing calculates the geodetic coordinates of multiple targets in single image.

Specifically, by x_v, y_v, z_vValue substitute into formula (12) calculate seat in the rectangular coordinate system in space of the earth's core for the target It is designated as

$\left[\begin{array}{c}{x}_e\\ y_e\\ z_e\\ 1\end{array}\right]=\left[\begin{array}{cccc}-c_{l_0}{s}_{m_0}& -s_{l_0}& -c_{l_0}{c}_{m_0}& \[n+h_0\]c_{m_0}{c}_{l_0}\\ -s_{l_0}{s}_{m_0}& c_{l_0}& -c_{m_0}{s}_{l_0}& \[n+h_0\]c_{m_0}{s}_{l_0}\\ c_{m_0}& 0& -s_{m_0}& \[n(1-e^2)+h_0\]s_{m_0}\\ 0& 0& 0& 1\end{array}\right]\left[\begin{array}{c}{x}_v\\ y_v\\ z_v\\ 1\end{array}\right]---\left(12\right)$

Then the conversion formula according to the earth's core rectangular coordinate system in space to earth coordinates (13)～(16), obtain target Geodetic coordinates is as follows

$u=arctan\frac{{\mathrm{az}}_e}{b\sqrt{x_e^2+y_e^2}}---\left(13\right)$

$m=arctan\frac{z_e+{\mathrm{be}}^{\′2}{\sin }^{3}u}{\sqrt{x_e^2+y_e^2}-{\mathrm{ae}}^2{\cos }^{3}u}---\left(15\right)$

$h=\frac{\sqrt{x_e^2+y_e^2}}{\cos m}-n---\left(16\right)$

In formula (12)～(16), l, m and h are respectively geodetic longitude, geodetic latitude and the geodetic height of target, semimajor axis of ellipsoid A=6378137.0m, semiminor axis of ellipsoid b=6356752.0m, ellipsoid first eccentricityEllipsoid second is inclined Heart rateEllipsoid radius of curvature in prime vertical $n=a/\sqrt{1-e^2{\sin }^{2}m}.$

Further, using recursive least squares (rls, recursive least squares) to multiple image Positioning result be filtered process, reduce random error, improve target location accuracy.

If the original location data of t two field picture is x_k(k=1,2 ..., t), rls filtering algorithm flow process is as shown in Figure 5.

I in Fig. 5_1×1Unit matrix for 1 × 1, initial data x_kThe data obtaining after rls filtering is x_k(k=1, 2 ..., t), x here can be longitude l, latitude m and geodetic height h.

Further, for improving the convergence rate of rls filtering algorithm, manage to improve the image for positioning in the unit interval Frame number, takes the number of image frames being used for positioning in 1s consistent with video image frame frequency, when known to the speed of aircraft, the corresponding moment Gps data can be determined using dead reckoning.

Under the rectangular coordinate system in space of wgs-84 the earth's core, the projected coordinate of aircraft is:

$x_e=x_{e0}+{\&integral;}_0^n{v}_x\·dt---\left(17\right)$

$y_e=y_{e0}+{\&integral;}_0^n{v}_y\·dt---\left(18\right)$

X in formula_e0, y_e0It is the coordinate of initial time, v_x, v_yIt is the aircraft speed of a ship or plane in x, the component in y direction.

According to dead reckoning formula (formula (17) and formula (18)), the error that gps data updating rate causes can be compensated, make Rls algorithm rapidly converges to stationary value, realizes being quickly accurately positioned to the multiple static target in ground.

In sum, these are only presently preferred embodiments of the present invention, be not intended to limit protection scope of the present invention. All any modification, equivalent substitution and improvement within the spirit and principles in the present invention, made etc., should be included in the present invention's Within protection domain.

Claims (6)

1. a kind of unmanned plane multi-target orientation method of simplification is it is characterised in that comprise the steps:

Step 1, unmanned plane is imaged over the ground using face battle array ccd camera, obtains the pixel coordinate of each target, wherein, positioned at video camera The target of field of view center is referred to as major heading, the target of other positions referred to as time target in visual field；

Step 2, the pixel coordinate using secondary target step 1 being obtained based on aberration rate method is modified, and obtains each target Ideal image point pixel coordinate；The pixel coordinate of major heading is the pixel coordinate of its ideal image point；According to major heading With pixel coordinate and the camera origin of the ideal image point of secondary target, construction major heading and time target are in video camera respectively Sight line vector under coordinate system；The sight line vector under camera coordinate system is put, as under camera initial point sensing machine under construction machine The vector of point；

Step 3, according to the rotation between camera coordinate system and carrier aircraft coordinate system, between carrier aircraft coordinate system and carrier aircraft geographic coordinate system Torque battle array, the sight line vector put under camera coordinate system under major heading, secondary target and machine is transformed into carrier aircraft geographic coordinate system In；

Step 4, sight line in carrier aircraft geographic coordinate system for the distance and primary and secondary target according to primary and secondary target and photoelectric platform Vector, calculates coordinate in carrier aircraft geographic coordinate system for each target；

Wherein, major heading and the distance of photoelectric platform are λ₁, obtained by range finder using laser measurement；Secondary target and photoelectric platform away from From for λ₂, by formulaCalculate and obtain, h=λ₁Cos α, α are the sight line with point under machine for the sight line vector of major heading The pixel angle of sight between vector,The pixel angle of sight between the sight line vector of point under sight line vector for secondary target and machine；

Step 5, in conjunction with unmanned plane in the course angle beta of the position coordinateses of the earth's core rectangular coordinate system in space and aircraft, the angle of pitch_εWith Roll angle γ, Coordinate Conversion in carrier aircraft geographic coordinate system for each target that step 4 is obtained is to the earth's core rectangular coordinate system in space In；Then according to the transition matrix between the earth's core rectangular coordinate system in space and earth coordinates, each target is straight in the earth's core space Coordinate Conversion in angular coordinate system, in earth coordinates, obtains the geodetic coordinates of each target, the geodetic coordinates of described each target It is the positioning result of corresponding target.

2. the unmanned plane multi-target orientation method of simplification as claimed in claim 1 is it is characterised in that in described step 1, adopt Image segmentation, frame difference method or optical flow method obtain the pixel coordinate of multiple targets simultaneously.

3. the unmanned plane multi-target orientation method of simplification as claimed in claim 1 is it is characterised in that in described step 2, each time The computational methods of the pixel coordinate of ideal image point of target are as follows:

Step 2.1, the pixel coordinate (u of the target being obtained according to step 1_d,v_d) calculate the physical coordinates (x of this target_d,y_d):

x_d=(u_d-u₁)d_x, y_d=(v_d-v₁)d_y

In formula, u₁And v₁It is the pixel coordinate of picture centre, d_xAnd d_yIt is the physical size in x and y direction for the single pixel；

Step 2.2, according to lens distortion rate d, calculates target ideal as the physical coordinates (x of imaging point_r,y_r):

$x_r=\frac{x_d}{1+d},y_r=\frac{y_d}{1+d}$

Step 2.3, by target ideal as the physical coordinates (x of imaging point_r,y_r) calculate target ideal as the pixel coordinate of imaging point (u_r,v_r)

u_r=u₁+x_r/d_x, v_r=v₁+y_r/d_y.

4. the unmanned plane multi-target orientation method of simplification as claimed in claim 1 is it is characterised in that in described step 3,It is respectively and put the sight line vector under carrier aircraft geographic coordinate system under major heading, secondary target and machine；

$\stackrel{\→}{s_v}=r_{vb}{r}_{bc}\stackrel{\→}{s_c}=\left[\begin{array}{ccc}{c}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}& -c_{\mathrm{\γ}}{s}_{\mathrm{\β}}+s_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}& s_{\mathrm{\γ}}{s}_{\mathrm{\β}}+c_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}\\ c_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}& c_{\mathrm{\γ}}{c}_{\mathrm{\β}}+s_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}& -s_{\mathrm{\γ}}{c}_{\mathrm{\β}}+c_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}\\ -s_{\mathrm{\ϵ}}& s_{\mathrm{\γ}}{c}_{\mathrm{\ϵ}}& c_{\mathrm{\γ}}{c}_{\mathrm{\ϵ}}\end{array}\right]\left[\begin{array}{ccc}{c}_{\mathrm{\θ}}{s}_{\mathrm{\ψ}}& -s_{\mathrm{\θ}}& c_{\mathrm{\θ}}{c}_{\mathrm{\ψ}}\\ s_{\mathrm{\ψ}}{s}_{\mathrm{\θ}}& c_{\mathrm{\θ}}& s_{\mathrm{\θ}}{c}_{\mathrm{\ψ}}\\ -c_{\mathrm{\ψ}}& 0& s_{\mathrm{\ψ}}\end{array}\right]\stackrel{\→}{s_c}$

$\stackrel{\→}{t_v}=r_{vb}{r}_{bc}\stackrel{\→}{t_c}=\left[\begin{array}{ccc}{c}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}& -c_{\mathrm{\γ}}{s}_{\mathrm{\β}}+s_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}& s_{\mathrm{\γ}}{s}_{\mathrm{\β}}+c_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{c}_{\mathrm{\β}}\\ c_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}& c_{\mathrm{\γ}}{c}_{\mathrm{\β}}+s_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}& -s_{\mathrm{\γ}}{c}_{\mathrm{\β}}+c_{\mathrm{\γ}}{s}_{\mathrm{\ϵ}}{s}_{\mathrm{\β}}\\ -s_{\mathrm{\ϵ}}& s_{\mathrm{\γ}}{c}_{\mathrm{\ϵ}}& c_{\mathrm{\γ}}{c}_{\mathrm{\ϵ}}\end{array}\right]\left[\begin{array}{ccc}{c}_{\mathrm{\θ}}{s}_{\mathrm{\ψ}}& -s_{\mathrm{\θ}}& c_{\mathrm{\θ}}{c}_{\mathrm{\ψ}}\\ s_{\mathrm{\ψ}}{s}_{\mathrm{\θ}}& c_{\mathrm{\θ}}& s_{\mathrm{\θ}}{c}_{\mathrm{\ψ}}\\ -c_{\mathrm{\ψ}}& 0& s_{\mathrm{\ψ}}\end{array}\right]\stackrel{\→}{t_c}$

j_v=[0 0 1]^t

Wherein,It is respectively and put the sight line vector under camera coordinate system under major heading, secondary target and machine；r_vbTable Show the spin matrix being tied to carrier aircraft geographic coordinate system from carrier aircraft coordinate, r_bcRepresent the rotation being tied to carrier aircraft coordinate system from camera coordinates Torque battle array, c_*Represent cos (*), s_*Represent sin (*)；θ is airborne photoelectric platform azimuth；ψ is airborne photoelectric platform height Angle；β is carrier aircraft course angle；ε is the carrier aircraft angle of pitch；γ is carrier aircraft roll angle.

5. the unmanned plane multi-target orientation method of the simplification as described in Claims 1 to 4 any one is it is characterised in that described In step 5, recurrence least square Filtering Processing is carried out to the positioning result of continuous multiple frames image, filter result is final determining Position result.

6. the unmanned plane multi-target orientation method of simplification as claimed in claim 5 is it is characterised in that to continuous multiple frames image The aircraft position coordinate of interframe when positioning result carries out recurrence least square Filtering Processing, is obtained using dead reckoning method.