CN103234552A - Optical navigation target satellite analog simulation image generating method - Google Patents

Abstract

The invention provides an optical navigation target satellite analog simulation image generating method. The method allows a navigation target satellite analog simulation image library having different characteristics to be generated through setting the observation orientation angle, the illumination intensity and the celestial body structure; the method can effectively analogue the optical images of a target celestial body having different characteristics in the stages of flyover, fly-around, close-range orbital transfer, celestial body rendezvous and the like of a deep space detector, the generated analog simulation images can really display real images, and the analog precision is high; and the generated reliable image library can be directly used for designing, verifying and assessing autonomous optical navigation related algorithms, so the validity and the reliability of the designing of the algorithms contained in the autonomous optical navigation of the deep space detector and comprising image processing, characteristic information extraction, navigation filtering and the like are guaranteed.

Description

Optical guidance target star analog simulation image generating method

Technical field

The invention belongs to the survey of deep space technical field, especially relate to optical guidance target star analog simulation image generating method.

Background technology

The success of any survey of deep space task all is to finish on the basis that is based upon the effective navigation of deep space probe and control.The spacecraft airmanship is the basic problem of Spacecraft Control, also is the basis of spacecraft rail control system.Along with the development in robot space flight and manned field, more and more higher to the requirement of spacecraft independent navigation ability.Simple to rely on land station's data to postpone for the remote transmission signal of survey of deep space too big, and real-time is too poor, so deep space probe possesses good independent navigation ability and becomes inevitable requirement.The autonomous optical research of current survey of deep space mainly concentrates on image and handles and information extraction navigation Study of filtering algorithm afterwards, comprised the dynamics of orbits modeling, navigation measures the Observability Analysis of model and navigation scheme, association areas such as navigation filtering technique, then less relatively to the research of image processing algorithm.Owing to lack the perfect survey of deep space navigation celestial body database of a cover, just can't be in advance the algorithm reliability of autonomous navigation system be checked, so can't ensure the success ratio of the algorithm in actual task.Therefore, developing and a kind of survey of deep space optical guidance target star analog simulation image generation technique, to understanding the characteristic of deep space objective optics image, handle design and checking with information extraction algorithm so that better finish image, all is very necessary.

Summary of the invention

Technical matters to be solved by this invention is to overcome the deficiencies in the prior art, and design and checking for the optical guidance algorithm of deep space probe provide optical guidance target star analog simulation image generating method.Described method can be that survey of deep space autonomous optical navigation system sets up one group of simulation celestial body database according to user's request, and the celestial image type that may exist in a large number in the simulation survey of deep space task is used for the assessment to the autonomous optical navigation algorithm of survey of deep space.

For solving the problems of the technologies described above, the technical solution adopted in the present invention is: optical guidance target star analog simulation image generating method comprises the steps:

Steps A generates the simplified model of simulating celestial image, calculates all pixels of celestial body surface and projects to coordinate corresponding on the camera focal plane; During calculating, the respective coordinates domain transformation is the observable reflection region in celestial body surface;

Under the inertial coordinates system celestial body surface take up an official post 1 P in the observable reflection region of meaning (x, _y, z), its coordinate P'(x under the camera imaging coordinate system, y z) is

P'(x,y,z)=T _IC·P(x,y,z)

At focal plane O _XyOn corresponding projection coordinate be P'(x, y), T wherein _ICBe tied to the transition matrix of imaging coordinate system for inertial coordinate;

Step B generates the light reflectance model of simulating celestial image, adopts bidirectional reflectance theoretical modeling celestial body to the reflection model of sunshine, calculates solar irradiation to the radiation intensity of celestial body surface back reflection light, the i.e. light intensity of imaging surface;

Catoptrical radiation intensity is expressed as:

I=r(i,θ,g)·S ₀

Wherein

$r(i,\mathrm{\θ},g)=\frac{\mathrm{\ω}}{4\mathrm{\π}}\·\frac{\cos i}{\cos i+\cos \mathrm{\θ}}\·\left\{\right[1+B\left(g\right)]P\left(g\right)+H\left(\cos i\right)H\left(\cos \mathrm{\θ}\right)-1\}$

$H\left(x\right)=\frac{1+2x}{1+2\mathrm{\γx}},$ γ=(1-ω) ^1/2

R represents bidirectional reflectance, and I represents to simulate the catoptrical radiancy in celestial body surface, S ₀Be the irradiance of simulation celestial body surface incident light, i is incident angle, and θ is reflection angle, and g is solar azimuth, and ω is average single scattering albedo, and P (g) is the average phase angle function, and B (g) is the backscattering function, H () and the γ intermediate variable for calculating;

Step C generates the texture model of simulating celestial image, covers the celestial body surface texture mapping that comprises atmosphere texture and rugged face of land texture, calculates the light reflectance model of texture model;

Step D generates the target celestial body analog simulation image under different observed ranges, the different observation angle, constitutes the analog image storehouse.

Beneficial effect of the present invention: the present invention proposes optical guidance target star analog simulation image generating method, described method can generate the optical guidance target star analog simulation image library of multiple different characteristic by the setting to observed bearing angle, intensity of illumination, celestial body texture; This method can effectively be simulated comprising the optical imagery that deep space probe leapt, is diversion, closely became the target celestial body different characteristic in stages such as rail, intersection celestial body, the analog simulation image that generates can be true to nature the situation of reflection true picture, the simulation precision height; The a large amount of image libraries reliably that generate thus can be directly used in design, checking and the assessment to the autonomous optical navigation related algorithm, thereby ensure that image related in the deep space probe autonomous optical navigation is handled, characteristic information extracts, validity and the reliability of the design of navigation filtering scheduling algorithm.

Description of drawings

Fig. 1 is that the simulation celestial body is simplified geometric model.

Fig. 2 celestial body surface bidirectional reflectance geometric model

Mercury image that Fig. 3 actual detection task is captured

The simulation celestial image of the smooth face of land of Fig. 4 texture

The simulation celestial image of Fig. 5 atmosphere texture

The simulation celestial image of the rugged face of land of Fig. 6 shell case texture

Fig. 7 histogram distribution comparison diagram

Embodiment

Below in conjunction with accompanying drawing, the optical guidance target star analog simulation image generating method that the present invention is proposed is elaborated:

Optical guidance target star analog simulation image generating method of the present invention, its implementation process is as follows:

Steps A, simulation celestial body are simplified geometric model and are generated

Only consider the celestial body geometric shape, ignore and comprise texture, light reflection strength, calculate target celestial body from the projection of three dimensions navigation camera imaging focal plane, wherein the respective coordinates domain transformation only need be considered the observable reflection region in celestial body surface.

To celestial body surface under the inertial system take up an official post 1 P in the observable reflection region of meaning (x, y, z), its coordinate P'(x under the camera imaging coordinate system, y z) is

P'(x,y,z)=T _IC·P(x,y,z)

Corresponding focal plane O _XyProjection coordinate is P'(x, y)

T wherein _ICBe tied to the transition matrix of imaging coordinate system for inertia.

Step B, simulation celestial light reflection model generates

Set the solar source radiation intensity, finish resolving of reflection model correlation parameter variable, to ask for respective coordinates position light intensity value.Employing bidirectional reflectance theory is come the celestial body reflection model of simulated solar irradiation and is introduced the Lommel-Seeliger reflection model and simplify.

Intensity of reflected light is expressed as

I=r(i,θ,g)·S ₀

Wherein

$r(i,\mathrm{\θ},g)=\frac{\mathrm{\ω}}{4\mathrm{\π}}\·\frac{\cos i}{\cos i+\cos \mathrm{\θ}}\·\left\{\right[1+B\left(g\right)]P\left(g\right)+H\left(\cos i\right)H\left(\cos \mathrm{\θ}\right)-1\}$

$H\left(x\right)=\frac{1+2x}{1+2\mathrm{\γx}},$ γ=(1-ω) ^1/2

R represents bidirectional reflectance, the catoptrical radiancy of I presentation surface, S ₀Be the irradiance of this surface incident light, i is incident angle, and θ is reflection angle, and g is solar azimuth, and ω is average single scattering albedo, and P (g) is the average phase angle function, and B (g) is the backscattering function, H () and the γ intermediate variable for calculating.

Step C, simulation day solid texture model generates

Selecting texture, texture type at user's request is the texture type feature of celestial bodies such as known terrestrial planet, giant planet, asteroid.Choose smooth texture mapping, the overlay area is the view field of target celestial body on the camera focal plane that calculates in the steps A, B carries out the calculating of light reflection strength set by step, generates the many groups target celestial body analog simulation image under different observed ranges, the different observation angle, constitutes the analog image storehouse.

Other content that is not described in detail in the instructions of the present invention belongs to this area professional and technical personnel's known prior art.

Embodiment:

Example background: resolve the navigation observed quantity because deep space probe need obtain the target celestial body image when closely the celestial body intersection stage is adopted autonomous optical navigation, but need the reliability of the designed algorithm of test and evaluation in advance in the design phase, therefore complete, the general celestial body analog image storehouse of a cover need be arranged.Be example with " MESSENGER " detection mission intersection Mercury, utilize the method for the invention to finish the generation of terrestrial planet image library.

Computing platform: MATLAB (2008a)

Implementation step:

Steps A, simulation celestial body are simplified geometric model and are generated

Terrestrial planet is generally the elliposoidal celestial body, and celestial body surface 1 P of meaning that takes up an official post satisfies among Fig. 1

${\mathrm{<figure-callout\; id="1"\; label="P\; T"\; filenames="HDA00002984380000011.png,HDA00002984380000025.png"\; state="\{\{state\}\}">P</figure-callout>}}^T\left[\begin{array}{ccc}1/a^2& 0& 0\\ 0& 1/b^2& 0\\ 0& 0& 1/c^2\end{array}\right]P=P^T\mathrm{AP}$

With its particularization, regard it as spherical regular celestial body, make O ₁A=O ₁B=O ₁C=O ₁P=R; Every some pixel respective distances of simulation celestial body profile offset d:

$\mathrm{\&Delta;d}=\sqrt{O_1{O_2}^2+R^2}(1+\sin \mathrm{\&theta;})$

Thereby try to achieve 1 P (x in any observable reflection region, y, z) projection coordinate on the camera imaging focal plane, wherein a, b, c are ellipsoid three semiaxis length, A is the ellipsoidal parameter matrix, θ represents is the angle of vector between sight line vector and celestial body and fixed star between detector and celestial body, or is called solar azimuth, and R is the celestial body radius.

Step B, simulation celestial light reflection model generates

Adopt the celestial body reflection model of improved Lommel-Seeliger bidirectional reflectance theoretical modeling sunshine, as shown in Figure 2, bidirectional reflectance r is expressed as

r(i,θ,g)=I/S ₀

What wherein I represented is the radiancy of the surface reflection on the assigned direction, S ₀Be the irradiance of this surface incident light,

(1) finds the solution bidirectional reflectance r

$r(i,\mathrm{\&theta;},g)=\frac{\mathrm{\&omega;}}{4\mathrm{\&pi;}}\&CenterDot;\frac{\cos i}{\cos i+\cos \mathrm{\&theta;}}\&CenterDot;\left\{\right[1+B\left(g\right)]P\left(g\right)+H\left(\cos i\right)H\left(\cos \mathrm{\&theta;}\right)-1\}$

$H\left(x\right)=\frac{1+2x}{1+2{(1-\mathrm{\&omega;})}^{1/2}x}$

$i=a\cos \left[e_n^T{e}_{\mathrm{sun}}\right]$

$\mathrm{\&theta;}=a\cos [-e_n^T{\left(e_i\right)}_I]$

$g=a\cos [-e_{\mathrm{sun}}^T{\left(e_i\right)}_I]$

I is incident angle, and θ is reflection angle, and g is solar azimuth, and ω is average single scattering albedo, e _nBe celestial body surface curve unit normal vector, e _SunBe the relative sun unit vector of celestial body, e _iThe relative detector unit vector of celestial body, () ^TBe matrix transpose, () _IExpression is transformed under the inertial coordinates system.

Calculate average phase angle function P (g) and backscattering function B (g)

$P\left(g\right)=\frac{4\mathrm{\&pi;}}{5}[\frac{\sin g+(\mathrm{\&pi;}-g)\cos g}{\mathrm{\&pi;}}+\frac{{(1-\cos g)}^2}{10}]$

$B\left(g\right)=\left\{\begin{array}{cc}{B}_0[1-\frac{1}{2\mathrm{\&kappa;}}(3-e^{-\mathrm{\&kappa;}})(1-e^{-\mathrm{\&kappa;}})]& \left|g\right|<\mathrm{\&pi;}/2\\ 0& \left|g\right|<\mathrm{\&pi;}/2\end{array}\right.$

κ=h/tan|g|

B ₀=exp(-ω ²/2)

Wherein h is Planck's constant, and k is Boltzmann constant, || expression takes absolute value.

(2) find the solution the irradiance S of incident light ₀

${\mathrm{<figure-callout\; id="0"\; label="S"\; filenames="HDA00002984380000012.png,HDA00002984380000025.png"\; state="\{\{state\}\}">S</figure-callout>}}_0=S_{\mathrm{ref}}{\left(\frac{r_{\mathrm{ref}}}{\left|\right|r_{\mathrm{sc}}\left|\right|}\right)}^2$

Wherein, represent with the light quantum form

$S_{\mathrm{ref},\mathrm{\&lambda;}}=\frac{2\mathrm{\&pi;}{\mathrm{hc}}^2}{{\mathrm{\&lambda;}}^2}{[\exp \left(\frac{\mathrm{hc}}{\mathrm{\&lambda;kT}}\right)-1]}^{-1}{\left(\frac{r_{\mathrm{Sun}}}{r_{\mathrm{ref}}}\right)}^2$

r _SunBe solar radius, c is the light velocity, λ _Min～λ _MaxThe spectral wavelength scope, λ _MinBe the minimum value of spectral wavelength, λ _MaxBe the maximal value of spectral wavelength, r _ScBe the position vector of the relative sun of detector, S _RefBe apart from sun r _RefThe reference irradiance at place, T is temperature, S _{Ref, λ}Expression λ wavelength correspondence is with reference to irradiance.

Calculate the catoptrical radiation intensity I in celestial body surface

$I=r\left[\left({\&Integral;}_{{\mathrm{\&lambda;}}_{\min }}^{{\mathrm{\&lambda;}}_{\max }}\frac{\mathrm{\&lambda;}}{\mathrm{hc}}{S}_{\mathrm{ref},\mathrm{\&lambda;}}\mathrm{d\&lambda;}\right){\left(\frac{r_{\mathrm{ref}}}{\left|\right|r_{\mathrm{sc}}\left|\right|}\right)}^2\right]={\mathrm{rS}}_{\mathrm{ref}}{\left(\frac{r_{\mathrm{ref}}}{\left|\right|r_{\mathrm{sc}}\left|\right|}\right)}^2$

|| || matrix norm is got in expression.

Step C, simulation day solid texture model generates

Choose smooth texture mapping, atmosphere texture mapping and crater texture mapping, spherical celestial body can directly be done the sphere processing to texture, calculate texture region simultaneously to the radiation intensity of solar reflection optical, generate the celestial image under corresponding specific range and the observation angle, the generation that finish flat surface simulation celestial image, possesses the simulation celestial image of atmosphere and possess the simulation celestial image of rugged face of land shell case texture.

Simulating, verifying:

The starting condition of simulation calculation is set at

Table 1 simulation celestial body basic parameter

Simulation parameter	Value	Unit
			The CCD camera focus	200	mm
The pixel ratio	83.8	Pixel/mm
			Focal plane, celestial body center coordinate	[250,250]	Pixel
The target celestial body diameter	6000	km

Image resolution ratio	500×500	Pixel
			Sun phase angle	π/3	deg

What the present invention adopted is the histogram analysis method, chooses the leading indicator that average, variance and entropy are estimated as the one dimension histogram, and carries out the reference comparison with the captured Mercury image of " MESSENGER " detection mission that Fig. 3 provides.

Generation gray level image f (x, y), probability density function p (s _k)

p(s _k)=n _k/n(k=0,1,2,...,L-1)

S wherein _kBe image f (x, k level gray-scale value y), n _kBe that (x has gray-scale value s in y) to f _kNumber of pixels, n is total number of image pixels.

Computed image average μ, variances sigma ²With entropy H

$\mathrm{\&mu;}=\underset{k=0}{\overset{L-1}{\mathrm{\&Sigma;}}}{s}_{k}P\left(s_k\right)$

${\mathrm{\&sigma;}}^2=\underset{k=0}{\overset{L-1}{\mathrm{\&Sigma;}}}{(s_k-\mathrm{\&mu;})}^{2}P\left(s_k\right)$

$H=-\underset{k=0}{\overset{L-1}{\mathrm{\&Sigma;}}}P\left(s_k\right)\log \left(P\left(s_k\right)\right)$

Table 2 true picture and the contrast of analog image statistics

Image	True picture	Analog image	1	Analog image 2	Analog image 3
						Average	0.0987	0.1741	0.0997	0.1146
Standard deviation	0.1509	0.2932	0.1616	0.1862
					Entropy	2.3838	3.7015	3.6935	3.7662

What average illustrated is the average light and shade of image, what variance illustrated is the overall contrast of image, what information entropy illustrated is information content of image, and in the image average light and shade reflection be the radiation power of scene, what contrast reflected is radiation contrast in the scene, one true to nature, believable emulating image at first must be that ir signature is accurate, being embodied in the gray level image is exactly the image average, contrast and quantity of information will meet truth, Fig. 4, what Fig. 5 and Fig. 6 provided is respectively the simulation celestial image of smooth face of land texture, consider the simulation celestial image of atmosphere texture and the simulation celestial image of rugged face of land shell case texture.Observe histogram distribution situation shown in Figure 7, the image average that associative list 2 is listed, standard deviation and entropy data, get rid of black background and celestial body shady face, only consider effective characteristics of image in solar irradiation zone, three groups of analog images have all been inherited the characteristic statistics data of true picture to a great extent, be enough to satisfy complicacy and the diversity requirement of autonomous optical navigation image in the deep space environment, the related algorithm design phase carrying out test and evaluation for navigation provides reliable material, for the smooth implementation of survey of deep space autonomous optical navigation task provides necessary guarantee.

Claims (1)

1. optical guidance target star analog simulation image generating method is characterized in that, comprises the steps:

Steps A generates the simplified model of simulating celestial image, calculates all pixels of celestial body surface and projects to coordinate corresponding on the camera focal plane; During calculating, the respective coordinates domain transformation is the observable reflection region in celestial body surface;

Under the inertial coordinates system celestial body surface take up an official post 1 P in the observable reflection region of meaning (x, y, z), its coordinate P'(x under the camera imaging coordinate system, y z) is

P'(x,y,z)=T _IC·P(x,y,z)

At focal plane O _XyOn corresponding projection coordinate be P'(x, y);

Wherein, T _ICBe tied to the transition matrix of imaging coordinate system for inertial coordinate;

Step B generates the light reflectance model of simulating celestial image, adopts bidirectional reflectance theoretical modeling celestial body to the reflection model of sunshine, calculates solar irradiation to the radiation intensity of celestial body surface back reflection light, the i.e. light intensity of imaging surface;

Catoptrical radiation intensity is expressed as:

I=r(i,θ,g)·S ₀

Wherein

$r(i,\mathrm{\&theta;},g)=\frac{\mathrm{\&omega;}}{4\mathrm{\&pi;}}\&CenterDot;\frac{\cos i}{\cos i+\cos \mathrm{\&theta;}}\&CenterDot;\left\{\right[1+B\left(g\right)]P\left(g\right)+H\left(\cos i\right)H\left(\cos \mathrm{\&theta;}\right)-1\}$

$H\left(x\right)=\frac{1+2x}{1+2\mathrm{\&gamma;x}},$ γ=(1-ω) ^1/2

R represents bidirectional reflectance, and I represents to simulate the catoptrical radiancy in celestial body surface, S ₀Be the irradiance of simulation celestial body surface incident light, i is incident angle, and θ is reflection angle, and g is solar azimuth, and ω is average single scattering albedo, and P (g) is the average phase angle function, and B (g) is the backscattering function, H () and the γ intermediate variable for calculating;

Step C generates the texture model of simulating celestial image, covers the celestial body surface texture mapping that comprises atmosphere texture and rugged face of land texture, calculates the light reflectance model of texture model;

Step D generates the target celestial body analog simulation image under different observed ranges, the different observation angle, constitutes the analog image storehouse.

Images