CN101937083A - Method for inhibiting mountain shadow by combining airborne interference SAR with geographic coding - Google Patents

Abstract

The invention discloses a method for inhibiting a mountain shadow by combining an airborne interference SAR (Synthetic Aperture Radar) with geographic coding. All image data of a mountain region are obtained by observing the mountain region in a plurality of directions, the inhibition of the mountain shadow is achieved by selectively combining with a geographic coding technique in a geographic coding stage to generate a DEM (Digital Elevation Model). The method comprises the following steps of: a) firstly, extracting a shadow region by using relevant coefficient combined amplitude information as a shadow extraction data source and using optimal threshold value segmentation; b) calculating the interpolation weight coefficient of each data pixel point; and c) processing the data pixel points one by one, carrying out geographic coding on a non-shadow region and discarding the pixel point for the shadow region.

Description

A kind of airborne interference SAR joint geography coding suppresses the massif shadow method

Technical field

The information of the invention belongs to is obtained and processing technology field, relates to a kind of airborne interference SAR joint geography coding particularly and suppresses the massif shadow method.

Technical background

Be subjected to the restriction of side-looking radar working system, unavoidably have shade in the data that airborne InSAR system obtains, this phenomenon is very obvious in the massif district.The shadow region does not have echoed signal, and the coherence is poor, and phase noise is serious, and the existence of shade phenomenon has reduced precision and the quality that the InSAR system obtains DEM.

Existing InSAR system Shadows Processing method comprises utilizes various filtering algorithms to suppress shadow region interferometric phase noise, and final shadow region DEM is carried out interpolation etc.These methods can't be obtained the shadow region real information, and filtering method can lose the resolution of normal region again.

Summary of the invention

The object of the present invention is to provide a kind of airborne interference SAR joint geography coding to suppress the massif shadow method.

For achieving the above object, airborne interference SAR joint geography coding provided by the invention suppresses the massif shadow method, by the massif district is carried out multi-direction observation, obtain the whole image datas in massif district, realize the inhibition of massif shade by selectivity joint geography coding techniques in the geocoding stage, generate DEM (Digital Elevation Model, digital elevation model), its key step is:

A) at first utilize related coefficient to extract data source as shade, utilize optimal threshold to cut apart the extraction of realization the shadow region in conjunction with amplitude information;

B) calculate the interpolation weights coefficient of each data pixels point, final interpolation weights coefficient is ρ '=ρ γ; ρ is the interpolation weights coefficient of each data pixels point in the formula, and γ is a related coefficient;

C) handle by pixel, handle if nonshaded area then carries out geocoding, if throw aside to this point in the shadow region.

Wherein, the multi-direction data that are observed reverse direction and the same side observation.

Wherein, related coefficient is to calculate by following formula:

$\mathrm{\γ}=\frac{\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\underset{j=1}{\overset{N}{\mathrm{\Σ}}}\left|S_1(i,j){S_2}^{*}(i,j)\right|}{\sqrt{\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\left|S_1(i,j)\right|}^2}\sqrt{\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\left|S_2(i,j)\right|}^2}}$

In the formula, M, N represent window size, S ₁, S ₂Represent master and slave complex pattern respectively, * number expression complex conjugate.

Wherein, the interpolation weights coefficient of each data pixels point is to calculate by following formula:

$\mathrm{\ρ}=\frac{\sin \left(\mathrm{\π\Δx}\right)}{\mathrm{\Δx}}\frac{\sin \left(\mathrm{\π\Δy}\right)}{\mathrm{\Δy}}$

Δ x, Δ y are illustrated respectively in the interpolation window ranges interior pixel point position of setting and the reference point location distance in x, y direction in the formula.

Description of drawings

Fig. 1 is the magnitude image of the whole image datas in massif district;

Fig. 2 is the gray-scale statistical histogram that shade extracts data source;

Fig. 3 is a computational data piece pixel interpolation weights coefficient process flow diagram;

Fig. 4 is the DEM that generates of the independent geocoding of each a data block and joint geography encoding process comparison diagram as a result;

Fig. 5 is that the joint geography coding suppresses massif shade process flow diagram.

Embodiment

The present invention by the massif district is carried out multi-direction observation, obtains the whole image datas in massif district in order to obtain the true elevation information in shadow region, and the inhibition in the geocoding stage by selectivity joint geography coding techniques realization massif shade generates high-quality DEM.Method of the present invention specifically describes as follows:

Figure 1 shows that two width of cloth images of areal, there is tangible massif landform in the image, ridge blocks and forms tangible shadow region, Fig. 1 (a) is the image data of flying from West to East, Fig. 1 (b) is from east orientation west flight image data, radar all is that mode of operation is looked on the right side, and two width of cloth images have comprised whole images in massif district, and its shadow region is the both sides of corresponding massif respectively.

Figure 2 shows that the gray-scale statistical histogram that extracts data source at Fig. 1 (a) lower right corner area shading, the basic thought that threshold value is asked for is that the histogram of image is divided into two groups with a certain gray-scale value, when the variance between two groups that are divided into is maximum, just as the optimal threshold of image segmentation, it is 48 that adopting said method gets optimal threshold as calculated to this gray-scale value.The data source of wherein extracting shade is multiplied each other by magnitude image and related coefficient and obtains.The data source image that extracts shade simultaneously stretches to increase the dynamic range of image through gray level.Therefore consider the difference of target scattering characteristics, entire image is used same threshold value can produce than large deviation, use semi-automatic artificial delineation target area, the variance between algorithm be divided into the target area image histogram two groups is to the maximum according to calculating optimal threshold.

The computing formula of related coefficient is:

$\mathrm{\γ}=\frac{\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\underset{j=1}{\overset{N}{\mathrm{\Σ}}}\left|S_1(i,j){S_2}^{*}(i,j)\right|}{\sqrt{\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\left|S_1(i,j)\right|}^2}\sqrt{\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\left|S_2(i,j)\right|}^2}}$

Wherein, M, N represent window size, S ₁, S ₂Represent master and slave complex pattern respectively, * number expression complex conjugate.

According to following formula

$\mathrm{\ρ}=\frac{\sin \left(\mathrm{\π\Δx}\right)}{\mathrm{\Δx}}\frac{\sin \left(\mathrm{\π\Δy}\right)}{\mathrm{\Δy}}$

Calculate the interpolation weights coefficient of each data pixels point; Wherein Δ x, Δ y are illustrated respectively in the interpolation window ranges interior pixel point position of setting and the reference point location distance in x, y direction.

Final interpolation weights coefficient is: ρ '=ρ γ.

See also Fig. 3, shown the flow process of the interpolation weights coefficient of computational data piece pixel of the present invention, step is:

1) two-dimensional coordinate (x, the y) extreme value of statistics all data blocks pixel is provided with 2 meters * 2 meters in geocoding DEM product distance space.

2) the DEM parameter that is provided with according to step 1 is asked for each reference point coordinate figure (x ₀, y ₀), its computing formula is as follows;

x ₀＝x _min+2i

y ₀＝y _min+2j

x _Min, y _MinBe respectively the x that step 1 obtains, the minimum value of y coordinate, i, j are the row, column sequence number of reference point, and 2 is the distance space that step 1 is provided with.

3) by the two-dimensional coordinate value of data block pixel (x, y), reference point coordinate figure (x ₀, y ₀) and related coefficient calculate the interpolation weights coefficient of each pixel.

It is 3 * 3 that the interpolation window is set, and promptly only to influence with this point be reference point in 6 meters * 6 meters scopes at center to each data block pixel.Can calculate the final interpolation weights coefficient of each data block pixel by coordinate figure, reference point coordinate figure and the related coefficient of data block pixel

$\mathrm{\ρ}=\frac{\sin \left(\mathrm{\π}\right|x-x_0\left|\right)}{|x-x_0|}\frac{\sin \left(\mathrm{\π}\right|y-y_0\left|\right)}{|y-y_0|}\mathrm{\γ}.$

Fig. 4 (a) and (b) are depicted as the two data blocks DEM that geocoding generates that respectively hangs oneself, and the shadow region noise phenomenon is very obvious.Fig. 4 (c) suppresses standard DEM behind the shade for joint geography coding.

Fig. 5 suppresses massif shadow method process flow diagram for the joint geography coding, and its step is as follows:

1) obtains whole image datas in massif district

2) utilize related coefficient and amplitude information to extract data source, utilize optimal threshold to cut apart the extraction of realization the shadow region as shade;

3), calculate the interpolation weights coefficient of each data pixels point according to geocoding DEM product parameters;

4) handle by pixel, handle if nonshaded area then carries out geocoding, if throw aside to this point in the shadow region.

Claims (4)

1. an airborne interference SAR joint geography coding suppresses the massif shadow method, by the massif district is carried out multi-direction observation, obtain the whole image datas in massif district, realize the inhibition of massif shade by selectivity joint geography coding techniques in the geocoding stage, generate digital elevation model, its key step is:

A) at first utilize related coefficient to extract data source as shade, utilize optimal threshold to cut apart the extraction of realization the shadow region in conjunction with amplitude information;

B) calculate the interpolation weights coefficient of each data pixels point, final interpolation weights coefficient is ρ '=ρ γ; ρ is the interpolation weights coefficient of each data pixels point in the formula, and γ is a related coefficient;

C) handle by pixel, handle if nonshaded area then carries out geocoding, if throw aside to this point in the shadow region.

2. airborne interference SAR joint geography coding as claimed in claim 1 suppresses the massif shadow method, wherein, and the multi-direction data that are observed reverse direction and the same side observation.

3. airborne interference SAR joint geography coding as claimed in claim 1 suppresses the massif shadow method, and wherein, related coefficient is to calculate by following formula:

$\mathrm{\γ}=\frac{\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\underset{j=1}{\overset{N}{\mathrm{\Σ}}}\left|S_1(i,j){S_2}^{*}(i,j)\right|}{\sqrt{\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\left|S_1(i,j)\right|}^2}\sqrt{\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\left|S_2(i,j)\right|}^2}}$

In the formula, M, N represent window size, S ₁, S ₂Represent master and slave complex pattern respectively, * number expression complex conjugate.

4. airborne interference SAR joint geography coding as claimed in claim 1 suppresses the massif shadow method, and wherein, the interpolation weights coefficient of each data pixels point is to calculate by following formula:

$\mathrm{\ρ}=\frac{\sin \left(\mathrm{\π\Δx}\right)}{\mathrm{\Δx}}\frac{\sin \left(\mathrm{\π\Δy}\right)}{\mathrm{\Δy}}$

Δ x, Δ y are illustrated respectively in the interpolation window ranges interior pixel point position of setting and the reference point location distance in x, y direction in the formula.

Images