CN101458328A - Remote sensing inversion method for moon surface reflective index - Google Patents

Abstract

The invention relates to a remote sensing inversion method for lunar surface reflectivity. In the method, a point-by-point inversion of the lunar surface reflectivity is carried out based on DEM data and remote sensing images of the lunar surface. Firstly, shadows of the remote sensing images are judged by utilizing the DEM and solar direction parameters, which match the remote sensing images, including solar altitude angles, so that pixels of the remote sensing images of the lunar surface are divided into two categories, including non-shadow pixels and shadow pixels and different methods are adopted for the reflectivity inversion on the two categories. As to the non-shadow pixels, the reflectivity is inverted by computing illumination intensity of direct sunlight and illumination intensity of reflected radiation of a neighboring slope by repeated iteration; as to the shadow pixels, the reflectivity is inverted by the illumination intensity of reflected radiation of the neighboring slope. The computation method of the inversion is original and is of great significance for remote sensing exploration of celestial bodies (including the moon) which human beings can not actually measure point by point with instruments. The method has important value in theory and technology of lunar exploration and methodology and epistemology of lunar study.

Description

Remote sensing inversion method for moon surface reflective index

Technical field:

The present invention relates to a kind of remote sensing inversion method for moon surface reflective index, belong to remote sensing technology and mapping science field.

Background technology:

The research of object wave spectral property (comprising reflectivity) is necessity foundation and the basis of work such as remote sensor optimization of working parameters, remote sensing test planning, target signature information extraction and remote sensing image processing analysis.Moon surface reflective index also is to carry out the radiation of moonscape neighboring slope reflection to calculate necessary parameter.Utilize moon remotely-sensed data to carry out moon surface reflective index pointwise Inversion Calculation, because the difference of the moon and earth environment, it is significantly different that this technology and conventional ground surface reflectance remote-sensing inversion have, and it is significant to use instrument to carry out celestial body (the comprising the moon) remote sensing of pointwise actual measurement for the mankind.

This invention is based on digital elevation model (DEM, digital elevation model) data and moonscape remote sensing images, carries out moon surface reflective index pointwise inverting.The moon surface reflective index remote-sensing inversion research of this paper belongs to original research, and relevant document does not have relevant report.

Summary of the invention:

The objective of the invention is to fill up the blank that prior art exists, remote sensing inversion method for moon surface reflective index is provided.

To achieve the above object of the invention, design of the present invention is:

The present invention is based on dem data and moonscape remote sensing images, carry out the pointwise inverting of moon surface reflective index: at first utilize the shade that carries out (comprising sun altitude and position angle) remote sensing images with the DEM and the solar azimuth parameter of remote sensing images coupling to judge, the pixel of moonscape remote sensing images is divided into two classes, non-shade pixel and shade pixel, and adopt diverse ways to carry out the reflectivity inverting to it respectively.For non-shade pixel, come the inverting reflectivity by iterate calculating direct sunlight illumination and neighboring slope reflection radiant illumination, the shade pixel utilizes the neighboring slope reflection radiant illumination to come the inverting reflectivity.

According to above-mentioned inventive concept, the technical solution used in the present invention is as follows:

A kind of remote sensing inversion method for moon surface reflective index is characterized in that the concrete operations step is as follows:

1) moon remote sensing digital image and digital elevation model dem data obtains and mate;

2) shade of moon remote sensing image is judged;

3) the reflectivity inverting of non-shade pixel;

4) the reflectivity inverting of shade pixel.

Above-mentioned steps 3) step of the reflectivity inverting of non-shade pixel is as follows:

In the moonscape remote sensing images, the remote sensing value of non-shade pixel is made up of two parts---direct sunlight radiance remote sensing component and neighboring slope reflection radiance remote sensing component, i.e. DN _Ii=DN _Sij+ DN _RijThe reflectivity of non-shade pixel adopts iterative algorithm to come Inversion Calculation, and progressively approaches its true value:

(1) the non-shade pixel of supposing the menology remote sensing images only obtains direct light DN _Ij=DN _Sij: because direct sunlight illumination is higher than the neighboring slope reflection radiant illumination greatly on the non-shade pixel, only obtain direct sunlight illumination, do not consider the neighboring slope reflection radiation so at first can suppose non-shade pixel.

(2) calculate direct light topographic correction coefficient F _Ij, ask the remote sensing value on the surface level ${\mathrm{DN}}_{\mathrm{sij}}^{\′}={\mathrm{DN}}_{\mathrm{sij}}/F_{\mathrm{ij}};$

Direct light topographic correction coefficient F _IjAccording to formula F _Ij=1-tg α _IjCtg θ _IjCos ω _IjTry to achieve, wherein α _IjBe the pixel slope angle; θ _IjBe sun altitude; ω _IjBe pixel aspect angle A _IjWith solar azimuth AL _IjThe absolute value of difference, α _IjAnd A _IjCan from corresponding D EM, obtain; θ _IjAnd AL _IjCan obtain according to the information calculations of substar in the remote sensing images;

(3) according to the remote sensing value formula on the surface level ${\mathrm{DN}}_{\mathrm{sij}}^{\′}={K\·R}_{\mathrm{ij}}\·E_{\mathrm{sij}}^{\′}/\mathrm{\π},$

Be respectively direct sunlight remote sensing component on the surface level and the direct sunlight illuminance component on the surface level, try to achieve initial reflectance R _Ij ⁰

(4) calculate pixel adjacent slope illumination E _Rij, try to achieve neighboring slope reflection radiance remote sensing component DN _Rij=KR _Ij ^mE _Rij/ π, R _Ij ^mBe the m time reflectivity after the iteration;

Pixel adjacent slope illumination E _RijWith neighboring slope reflection point P _KlReflectivity R _KlBetween relational expression be $E_{\mathrm{rij}}=\frac{1}{K}\underset{k=1}{\overset{M}{\mathrm{\Σ}}}\underset{l=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{DN}}_{\mathrm{kl}}\·{S\cos \mathrm{\θ}}_{\mathrm{kl}}\cos {\mathrm{\β}}_{\mathrm{kl}}/R_{\mathrm{kl}}\·{d_{\mathrm{kl}}}^2,$ DN wherein _KlBe neighboring slope reflection point P _KlRemote sensing value, S is the pixel area, θ _KlBe incident ray and reflection pixel P _KlThe angle of normal vector, β _KlBe neighboring slope reflection point P _KlIncident ray and the domatic pixel P of target _IjThe angle of normal vector, d _KlBe neighboring slope reflection point P _KlTo domatic pixel P _IjDistance, M, N are the line number and the columns of pixel;

(5) ask direct sunlight radiance remote sensing component DN _Sij=DN _Ij-DN _Rij

(6) calculate direct light topographic correction coefficient F once more _Ij, ask the remote sensing value on the surface level ${\mathrm{DN}}_{\mathrm{sij}}^{\′}={\mathrm{DN}}_{\mathrm{sij}}/F_{\mathrm{ij}};$

(7) according to formula ${\mathrm{DN}}_{\mathrm{sij}}^{\′}={K\·R}_{\mathrm{ij}}\·E_{\mathrm{sij}}^{\′}/\mathrm{\π},$ Try to achieve the reflectivity R of iteration m+1 time _Ij ^M+1

(8) judge whether to satisfy | R _Ij ^m+ 1-R _Ij ^m|＜ε, ε are an infinitesimal number, if satisfy then output reflection rate R _Ij ^M+1, otherwise with m=m+1, iteration execution in step (4) is to step (8), up to (8) the output reflection rate that satisfies condition.

Above-mentioned steps 4) step of the reflectivity inverting of shade pixel is as follows:

The shade pixel of moonscape remote sensing images does not have solar radiation, so the remote sensing value of pixel only is neighboring slope reflection radiance remote sensing component, i.e. DN _Ij=DN _RijThe reflectivity R ' of shade pixel _IjCan be by remote sensing formula R ' _Ij=π DN _Ij/ KE _RijObtain, wherein, E _RijBe pixel point P on the remote sensing images _IjThe neighboring slope reflection radiant illumination, DN _IjBe P _IjThe remote sensing value of point.

The present invention has following conspicuous outstanding substantive distinguishing features and remarkable advantage compared with prior art:

The present invention is based on dem data and moonscape remote sensing images, carries out the pointwise inverting of moon surface reflective index, and the computing method of this inverting belong to original, have significantly different with conventional ground surface reflectance remote-sensing inversion.It is significant that this method can't use instrument to carry out celestial body (the comprising the moon) remote sensing of pointwise actual measurement for the mankind.The pixel reflectivity of inverting moonscape remote sensing images of the present invention has been given prominence to the spectral characteristic of lunar surface material in moon sensor information, be the lunar surface important information that nature can't directly obtain.The present invention moon exploration in theory with technical, all there is important value the methodology of lunar studies and theory of knowledge aspect.

Description of drawings:

Fig. 1 is the process flow diagram of moon surface reflective index remote-sensing inversion of the present invention.

Fig. 2 is the process flow diagram of the non-shade pixel of moonscape emissivity inverting.

Embodiment:

A preferred embodiment of the present invention accompanying drawings is as follows: compare with the earth, moonscape does not have atmosphere, so atmospheric effect can be ignored.Therefore the pixel optical radiation illumination of moonscape remote sensing images mainly is made up of two parts: direct sunlight radiant illumination and neighboring slope reflection radiant light illumination.Be E _Ij=E _Sij+ E _RijWherein, E _Ij, E _SijAnd E _RijBe respectively the optical radiation illumination on moonscape remote sensing images pixel ground, direct sunlight illuminance component and neighboring slope reflection illuminance component.The remote sensing value of pixel also has same relational expression: DN _Ij=DN _Sij+ DN _Rij=KR _IjE _Ij/ π=KR _Ij(E _Sij+ E _Rij)/π.Wherein, DN _Ij, DN _SijDN _RijBe respectively moonscape remote sensing images element remote sensing value, direct sunlight radiance remote sensing component and neighboring slope reflection radiance remote sensing component.R _IjBe the reflectivity of target pixel, K is the sensor gain coefficient.

We utilize the shade that carries out remote sensing images with the DEM and the sun position parameter of remote sensing images coupling to judge, the pixel of moonscape remote sensing images is divided into two classes, non-shade pixel and shade pixel, and adopt diverse ways to carry out the reflectivity inverting to it respectively.For non-shade pixel, come the inverting reflectivity by iterate calculating direct sunlight illumination and neighboring slope reflection radiant illumination, the shade pixel utilizes the neighboring slope reflection radiant illumination to come the inverting reflectivity.The moon surface reflective index remote-sensing inversion comprises the reflectivity inverting of non-shade pixel and the reflectivity inverting of shade pixel.Its concrete operations step following (referring to Fig. 1):

1, moon remote sensing digital image and digital elevation model dem data obtains and mate;

2, the shade of moon remote sensing image is judged;

3, the reflectivity inverting of non-shade pixel;

4, the reflectivity inverting of shade pixel.

Obtain and the coupling of moon remote sensing digital image in the step 1 and digital elevation model dem data are the remote sensing routine operations.

It is not the main protection content of this patent that the shade of the moon remote sensing image in the step 2 is judged, so the method that shade is judged is omitted.

The step of the reflectivity inverting of the non-shade pixel in the step 3 following (referring to Fig. 2):

In the moonscape remote sensing images, the remote sensing value of non-shade pixel is made up of two parts, direct sunlight radiance remote sensing component and neighboring slope reflection radiance remote sensing component, i.e. DN _Ij=DN _Sij+ DN _RijThe reflectivity of non-shade pixel adopts iterative algorithm to come Inversion Calculation, and progressively approaches its true value.

(1) the non-shade pixel of supposing the menology remote sensing images only obtains direct light DN _Ij=DN _Sij: because direct sunlight illumination is higher than the neighboring slope reflection radiant illumination greatly on the non-shade pixel, only obtain direct sunlight illumination, do not consider the neighboring slope reflection radiation so at first can suppose non-shade pixel.

(2) calculate direct light topographic correction coefficient F _Ij, ask the remote sensing value on the surface level ${\mathrm{DN}}_{\mathrm{sij}}^{\′}={\mathrm{DN}}_{\mathrm{sij}}/F_{\mathrm{ij}}.$

Direct light topographic correction coefficient F _IjAccording to formula F _Ij=1-tg α _IjCtg θ _IjCos ω _IjTry to achieve, wherein α _IjBe the pixel slope angle; θ _IjBe sun altitude; ω _IjBe pixel aspect angle A _IjWith solar azimuth AL _IjThe absolute value of difference.α _IjAnd A _IjCan from corresponding D EM, obtain; θ _IjAnd AL _IjCan obtain according to the information calculations of substar in the remote sensing images.

(3) according to the remote sensing value formula on the surface level ${\mathrm{DN}}_{\mathrm{sij}}^{\′}={K\·R}_{\mathrm{ij}}\·E_{\mathrm{sij}}^{\′}/\mathrm{\π}$

Be respectively direct sunlight remote sensing component on the surface level and the direct sunlight illuminance component on the surface level), try to achieve initial reflectance R _Ij ⁰

(4) calculate pixel adjacent slope illumination E _Rij, try to achieve neighboring slope reflection radiance remote sensing component DN _Rij=KR _Ij ^mE _Rij/ π.(R _Ij ^mBe the m time reflectivity after the iteration).

Pixel adjacent slope illumination E _RijWith neighboring slope reflection point P _KlThe reflectivity R of (non-shade pixel) _KlBetween relational expression be $E_{\mathrm{rij}}=\frac{1}{K}\underset{k=1}{\overset{M}{\mathrm{\Σ}}}\underset{l=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{DN}}_{\mathrm{kl}}\·{S\cos \mathrm{\θ}}_{\mathrm{kl}}\cos {\mathrm{\β}}_{\mathrm{kl}}/R_{\mathrm{kl}}\·{d_{\mathrm{kl}}}^2.$ DN wherein _KlBe neighboring slope reflection point P _KlRemote sensing value, S is the pixel area, θ _KlBe incident ray and reflection pixel P _KlThe angle of normal vector, β _KlBe neighboring slope reflection point P _KlIncident ray and the domatic pixel P of target _IjThe angle of normal vector, d _KlBe neighboring slope reflection point P _KlTo domatic pixel P _IjDistance, M, N are the line number and the columns of pixel.

(5) ask direct sunlight radiance remote sensing component DN _Sij=DN _Ij-DN _Rij

(6) calculate direct light topographic correction coefficient F once more _Ij, ask the remote sensing value on the surface level ${\mathrm{DN}}_{\mathrm{sij}}^{\′}={\mathrm{DN}}_{\mathrm{sij}}/F_{\mathrm{ij}}.$

(7) according to formula ${\mathrm{DN}}_{\mathrm{sij}}^{\′}={K\·R}_{\mathrm{ij}}\·E_{\mathrm{sij}}^{\′}/\mathrm{\π},$ Try to achieve the reflectivity R of iteration m+1 time _Ij ^M+1

(8) judge whether to satisfy | R _Ij ^M+1-R _Ij ^m|＜ε (ε is an infinitesimal number), if satisfy then output reflection rate R _Ij ^M+1, otherwise with m=m+1, iteration execution in step (4) is to step (8), up to (8) the output reflection rate that satisfies condition.

The step of the reflectivity inverting of the shade pixel in the step 4 is as follows:

Shade pixel for the moonscape remote sensing images does not have solar radiation, so the remote sensing value of pixel only is neighboring slope reflection radiance remote sensing component, i.e. DN _Ij=DN _RijThe reflectivity R ' of shade pixel _IjCan be by remote sensing formula R ' _Ij=π DN _Ij/ KE _RijObtain.Wherein, E _RijBe pixel point P on the remote sensing images _IjThe neighboring slope reflection radiant illumination, DN _IjBe P _IjThe remote sensing value of point.

Claims (3)

1. remote sensing inversion method for moon surface reflective index, it is characterized in that based on dem data and moonscape remote sensing images, carry out the pointwise inverting of moon surface reflective index: at first utilize DEM and solar azimuth parameter with the remote sensing images coupling, comprise sun altitude and position angle, carrying out the shade of remote sensing images judges, the pixel of moonscape remote sensing images is divided into two classes, promptly non-shade pixel and shade pixel, and adopt diverse ways to carry out the reflectivity inverting to it respectively; For non-shade pixel, come the inverting reflectivity by iterate calculating direct sunlight illumination and neighboring slope reflection radiant illumination, utilize the neighboring slope reflection radiant illumination to come the inverting reflectivity for the shade pixel; The concrete operations step is as follows:

1) moon remote sensing digital image and digital elevation model dem data obtains and mate;

2) shade of moon remote sensing image is judged;

3) the reflectivity inverting of non-shade pixel;

4) the reflectivity inverting of shade pixel.

2. remote sensing inversion method for moon surface reflective index according to claim 1, the step of reflectivity inverting that it is characterized in that the non-shade pixel of described step 3) is as follows: in the moonscape remote sensing images, the remote sensing value of non-shade pixel is made up of two parts---direct sunlight radiance remote sensing component and neighboring slope reflection radiance remote sensing component, i.e. DN _Ij=DN _Sij+ DN _RijThe reflectivity of non-shade pixel adopts iterative algorithm to come Inversion Calculation, and progressively approaches its true value:

(1) the non-shade pixel of supposing the menology remote sensing images only obtains direct light DN _Ij=DN _Sij: because direct sunlight illumination is higher than the neighboring slope reflection radiant illumination greatly on the non-shade pixel, only obtain direct sunlight illumination, do not consider the neighboring slope reflection radiation so at first can suppose non-shade pixel.

(2) calculate direct light topographic correction coefficient F _Ij, ask the remote sensing value on the surface level ${\mathrm{DN}}_{\mathrm{sij}}^{\′}={\mathrm{DN}}_{\mathrm{sij}}/F_{\mathrm{ij}};$

Direct light topographic correction coefficient F _IjAccording to formula F _Ij=1-tg α _IjCtg θ _IjCos ω _IjTry to achieve, wherein α _IjBe the pixel slope angle; θ _IjBe sun altitude; ω _IjBe pixel aspect angle A _IjWith solar azimuth AL _IjThe absolute value of difference, α _IjAnd A _IjCan from corresponding D EM, obtain; θ _IjAnd AL _IjCan obtain according to the information calculations of substar in the remote sensing images;

(3) according to the remote sensing value formula on the surface level ${\mathrm{DN}}_{\mathrm{sij}}^{\′}=K\·R_{\mathrm{ij}}\·E_{\mathrm{sij}}^{\′}/\mathrm{\π},$

Be respectively direct sunlight remote sensing component on the surface level and the direct sunlight illuminance component on the surface level, try to achieve initial reflectance R _Ij ⁰

(4) calculate pixel adjacent slope illumination E _Rij, try to achieve neighboring slope reflection radiance remote sensing component DN _Rij=KR _Ij ^mE _Rij/ π, R _Ij ^mBe the m time reflectivity after the iteration;

Pixel adjacent slope illumination E _RijWith neighboring slope reflection point P _KlReflectivity R _KlBetween relational expression be $E_{\mathrm{rij}}=\frac{1}{K}\underset{k=1}{\overset{M}{\mathrm{\Σ}}}\underset{l=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{DN}}_{\mathrm{kl}}\·S\cos {\mathrm{\θ}}_{\mathrm{kl}}\cos {\mathrm{\β}}_{\mathrm{kl}}/R_{\mathrm{kl}}\·{d_{\mathrm{kl}}}^2,$ DN wherein _KlBe neighboring slope reflection point P _KlRemote sensing value, S is the pixel area, θ _KlBe incident ray and reflection pixel P _KlThe angle of normal vector, β _KlBe neighboring slope reflection point P _KlIncident ray and the domatic pixel P of target _IjThe angle of normal vector, d _KlBe neighboring slope reflection point P _KlTo domatic pixel P _IjDistance, M, N are the line number and the columns of pixel;

(5) ask direct sunlight radiance remote sensing component DN _Sij=DN _Ij-DN _Rij

(6) calculate direct light topographic correction coefficient F once more _Ij, ask the remote sensing value on the surface level ${\mathrm{DN}}_{\mathrm{sij}}^{\′}={\mathrm{DN}}_{\mathrm{sij}}/F_{\mathrm{ij}};$

(7) according to formula ${\mathrm{DN}}_{\mathrm{sij}}^{\′}=K\·R_{\mathrm{ij}}\·E_{\mathrm{sij}}^{\′}/\mathrm{\π},$ Try to achieve the reflectivity R of iteration m+1 time _Ij ^M+1

(8) judge whether to satisfy | R _Ij ^M+1-R _Ij ^m|＜ε, ε are an infinitesimal number, if satisfy then output reflection rate R _Ij ^M+1, otherwise with m=m+1, iteration execution in step (4) is to step (8), up to (8) the output reflection rate that satisfies condition.

3. remote sensing inversion method for moon surface reflective index according to claim 1, the step of reflectivity inverting that it is characterized in that described step 4) shade pixel is as follows: the shade pixel of moonscape remote sensing images, there is not solar radiation, therefore the remote sensing value of pixel only is neighboring slope reflection radiance remote sensing component, i.e. DN _Ij=DN _RijThe reflectivity R ' of shade pixel _IjCan be by remote sensing formula R ' _Ij=π DN _Ij/ KE _RijObtain, wherein, E _RijBe pixel point P on the remote sensing images _IjThe neighboring slope reflection radiant illumination, DN _IjBe P _IjThe remote sensing value of point.

Images