CN104240253A - Oil spilling detection method based on polarization decomposition - Google Patents

Abstract

The invention relates to the oil field, in particular to an oil spilling detection method based on polarization decomposition. The spilling detection method based on the polarization decomposition comprises the steps that multi-polarization SAR data are acquired, wherein the multi-polarization SAR data comprise geographic coordinates, scattering entropy and CPD standard deviations, and the scattering entropy and the CPD standard deviations correspond to the geographic coordinates; according to the values of the scattering entropy corresponding to the geographic coordinates, the scattering entropy distribution range of oil spilling is obtained; afterwards, the geographic coordinates of oil spilling are determined. According to the spilling detection method based on the polarization decomposition, the multi-polarization SAR data are acquired, and the scattering entropy H and the CPD standard deviations are calculated according to the multi-polarization SAR data; thus, whether oil spilling exists or not is determined according to the two data, it effectively distinguishes between oil spilling and similar oil spilling, the deficiency of distinguishment between oil spilling and similar oil spilling through single data is made up for, and the defects in the prior art are overcome.

Description

A kind of oil spilling detection method of polarization decomposing

Technical field

The present invention relates to oil field, in particular to a kind of oil spilling detection method of polarization decomposing.

Background technology

Along with the fast development of China's economy, to the sharp increase of oil consumption, cause marine transportation oil amount to increase, increase the possibility that marine major oil spill accident occurs simultaneously.Oil spill accident often causes the oil pollution of large area sea, and not only make ocean, air physical environment, ecological resources suffer damage, cause halobiontic mass mortality, economy sustains a loss, and serious harm health.The fire that oil spill accident causes, also may cause marine and littoral facility, boats and ships etc. damage.Oil spilling, to the pollution of offshore sea waters, has caused the attention of national governments.Since 2006, China seas under its jurisdiction often have a lot of oil spill events to occur every year, and especially Talien New Port oil pipeline broke and caused marine oil overflow in July, 2010, caused huge harm to environment.In June, 2011 Bohai Sea PL19-3 oil platform generation oil spill accident, the time is long, and coverage is wide, and harm is large, and oil spilling, to the pollution of offshore sea waters, has caused the extensive attention in the whole nation.

Synthetic-aperture radar SAR (Synthetic Aperture Radar) have many weathers, little by weather effect, comparatively significantly feature is detected to oil spilling, be widely used in remote sensing monitoring operation system.Utilize satellite remote sensing technology in time, accurately to monitor to multiaspect marine oil spill to pollute; actively take removing and preventive measure remote sensing technology in time, accurately to monitor to multiaspect marine oil spill to pollute; taking precautions against natural calamities with ocean for protecting the marine environment and subtract, has been instant important process.Adopt multipolarization Radarsat-2 data, polarization decomposing research has been carried out for marine oil overflow, start with from the scattering matrix of oil spilling, research polarization SAR Electromagnetic Scattering, explore oil spilling scattering mechanism, analyze the polarization scattering characteristics of oil spilling, calculate oil spilling polarization parameter, the phase differential (CPD) carrying out oil spilling and seawater is analyzed, thus the scattering signatures of research oil spilling.Oil spilling information is extracted, for the remote sensing detection of oil spilling provides scientific basis based on polarization parameter research.

Use polarization SAR to carry out oil spilling and detect the importance having become remote sensing monitoring, existing oil spilling Detection Techniques are normally: based on SAR-C data, carry out the spilled oil monitoring of polarization SAR, extract polarization parameter entropy H, scattering angle α and anti-entropy A, use SVM supervised classification method, carry out oil spilling information extraction.

At present, based on Polarimetric SAR Image offshore spilled oil recognition methods mainly with remote sensing image processing and mode identification technology for means, extract as training sample by carrying out morphological feature to typical offshore spilled oil and doubtful oil film, offshore spilled oil detection and indentification is carried out by the means such as artificial neural network, fuzzy logic inference, but existing technical scheme is unsatisfactory for the removal effect of class oil spilling phenomenon, namely cannot determine whether as oil spilling.

Summary of the invention

The object of the present invention is to provide a kind of oil spilling detection method of polarization decomposing, to solve the above problems.

Provide a kind of oil spilling detection method of polarization decomposing in an embodiment of the present invention, comprising:

Obtain multipolarization SAR data, described multipolarization SAR data comprises geographic coordinate and the scattering entropy H corresponding to described geographic coordinate and CPD standard deviation;

Scattering entropy H corresponding to described geographic coordinate draws the scattering entropy H distribution range of oil spilling;

Determine the scattering entropy H distribution range meeting described oil spilling, and the geographic coordinate that there is the position of described CPD standard deviation is the geographic coordinate of oil spilling.

Preferably, described scattering entropy H corresponding to described geographic coordinate show that the scattering entropy H distribution range of oil spilling comprises:

Draw the H histogram of the scattering entropy H frequency of occurrences corresponding to described geographic coordinate;

According to the scattering entropy H distribution range of described H histogram determination oil spilling, the scattering entropy H distribution range of described oil spilling comprises the scattering entropy H of oil spilling and the scattering entropy H of class oil spilling.

Preferably, described H histogram comprises the function curve of at least two scattering entropy H;

The described scattering entropy H distribution range according to described H histogram determination oil spilling comprises:

Determine that function curve that the distribution range of H value meets default extent of oil spill is the scattering entropy H function curve of oil spilling;

Using the scattering entropy H distribution range of the scope corresponding to the H value corresponding to the scattering entropy H function curve of described oil spilling as oil spilling.

Preferably, describedly determine the scattering entropy H distribution range meeting described oil spilling, and the geographic coordinate that the geographic coordinate that there is the position of described CPD standard deviation is oil spilling comprises:

H-CPD coordinate X-Y scheme is generated according to described scattering entropy H and described CPD standard deviation;

The H-CPD coordinate points of coordinate points as oil spilling of the scattering entropy H distribution range of described oil spilling will be met in described H-CPD coordinate X-Y scheme;

Using the geographic coordinate of the geographic coordinate corresponding to described H-CPD coordinate points as oil spilling.

Preferably, the ordinate of described H-CPD coordinate X-Y scheme is scattering entropy H, and horizontal ordinate is CPD standard deviation.

Preferably, described acquisition multipolarization SAR data, comprising:

Obtain the T3 of described SAR data, or C3 matrix;

Determine described T3, or the coherence matrix proper vector of C3 matrix;

Described scattering entropy H is calculated according to described coherence matrix proper vector.

Preferably, calculate described scattering entropy H according to described coherence matrix proper vector to comprise:

Use and p _i=λ _i/ (λ ₁+ λ ₂+ λ ₃) calculate described scattering entropy H, wherein,

λ ₁, λ ₂and λ ₃for the proper vector of T3 or C3 matrix.

Preferably, described acquisition multipolarization SAR data also comprises:

Use formula calculate the value of CPD standard deviation, wherein L be the image of SAR data look number, S _hHt3 or C3 entry of a matrix element, S _vV* be T3, or element S in C3 matrix _vVconjugate number, ρ for representing the calculating of phase differential, without physical meaning.

Preferably, also comprise calculating correlation coefficient ρ, if when described ρ is greater than default numerical value, then determine to adopt according to described scattering entropy H and described CPD standard deviation determination oil spilling;

wherein S _vV* be T3, or element S in C3 matrix _vVconjugate number, S _hHand S _vVbe T3, or the element in C3 matrix.

Preferably, the scattering entropy H distribution range that described scattering entropy H corresponding to described geographic coordinate calculates oil spilling comprises:

Determine that scope that scattering entropy numerical value is less than 0.3 is the scattering entropy distribution range of oil spilling.

The oil spilling detection method of a kind of polarization decomposing that the embodiment of the present invention provides, with of the prior art by single to parameter entropy H, scattering angle α and anti-entropy A extracts, and then determine whether oil spilling, effectively cannot distinguish oil spilling compared with class oil spilling, it is by obtaining polarization SAR data, and calculate scattering entropy H and CPD standard deviation by these data, and then whether determined by two data be oil spilling, effectively oil spilling and class oil spilling are distinguished, compensate for the single data passed through and carry out the deficiency that oil spilling and class oil spilling carry out distinguishing, thus solve deficiency of the prior art.

Accompanying drawing explanation

Fig. 1 shows the oil spilling detection method basic flow sheet of a kind of polarization decomposing of the embodiment of the present invention;

Fig. 2 a shows the first reference figure of the correlation technique of the oil spilling detection method of a kind of polarization decomposing of the embodiment of the present invention;

Fig. 2 b shows the second of the correlation technique of the oil spilling detection method of a kind of polarization decomposing of the embodiment of the present invention with reference to figure;

Fig. 2 c shows the third reference figure of the correlation technique of the oil spilling detection method of a kind of polarization decomposing of the embodiment of the present invention;

The correlation technique that Fig. 3 a shows the oil spilling detection method of a kind of polarization decomposing of the embodiment of the present invention is a kind of with reference to figure;

The correlation technique that Fig. 3 b shows the oil spilling detection method of a kind of polarization decomposing of the embodiment of the present invention is another kind of with reference to figure;

Fig. 4 shows the H histogram of the oil spilling detection method of a kind of polarization decomposing of the embodiment of the present invention.

Embodiment

Also by reference to the accompanying drawings the present invention is described in further detail below by specific embodiment, as shown in Figure 1.

S101, obtains multipolarization SAR data, and multipolarization SAR data comprises scattering entropy H corresponding to geographic coordinate and geographic coordinate and CPD standard deviation;

S102, the scattering entropy H corresponding to geographic coordinate draw the scattering entropy H distribution range of oil spilling;

S103, determine the scattering entropy H distribution range meeting oil spilling, and the geographic coordinate that there is the position of CPD standard deviation is the geographic coordinate of oil spilling.

In step S101, multipolarization SAR data is got by synthetic-aperture radar SAR.Multipolarization SAR data is made up of the characteristic parameter that a group can be expressed sponge material physical characteristics, comprising having scattering angle α, scattering entropy H and the anti-entropy A etc. that polarizes.CPD standard deviation is by coherence matrix T3, or the parameter in C3 is calculated.

Step S102, draws the scattering entropy H distribution range of oil spilling, it should be noted that, the scattering entropy H distribution range of oil spilling can be empirical value, also can be the value range by calculating.As can be depicted the frequency of occurrences distribution plan of the scattering entropy H of certain each coordinate points of region according to scattering entropy H.Wherein, the H distribution range corresponding to coordinate set meeting a certain characteristic in figure is just the scattering entropy H distribution range of oil spilling.Certainly, this oil spilling distribution range may include the similar oil spilling of some, and this part similar oil spilling needs to be removed by CPD standard deviation.

In step S103, calculate by the scattering entropy H distribution range of oil spilling determined and the CPD standard deviation calculated the oil spilling determined.In step S102, after drawing the scattering entropy H distribution range of oil spilling, owing to having got rid of a certain amount of class oil spilling (value of class oil spilling data after CPD standard deviation calculates and the value of oil spilling there are differences) in CPD standard deviation, further, the scattering entropy H distribution range of oil spilling and CPD standard deviation are combined, to remove class oil spilling more accurately, and determine the H corresponding to oil spilling, and/or CPD standard deviation, finally, again according to the scattering entropy H of the oil spilling determined, and/or the CPD standard deviation of oil spilling, corresponding geographic coordinate is searched in original SAR data, the geographic coordinate found i.e. the position of oil spilling, thus the position determined residing for oil spilling, so that clear up in time oil spilling.

Further, in order to more accurately and intuitively determine the scattering entropy H distribution range of oil spilling, step S102 can also comprise following two steps:

Draw the H histogram of the scattering entropy H frequency of occurrences corresponding to geographic coordinate;

According to the scattering entropy H distribution range of H histogram determination oil spilling, the scattering entropy H distribution range of oil spilling comprises the scattering entropy H of oil spilling and the scattering entropy H of class oil spilling.

The scattering entropy H distribution range of the oil spilling determined refers to the numerical range of a H, comprise by the numerical value of the scattering entropy H of whole oil spillings within the scope of this, but in the scope corresponding to numerical value of the scattering entropy H of H oil spilling, incessantly comprise by oil spilling, also have certain similar oil spilling (but be not in H value in the scattering entropy H distribution range of oil spilling be then decided to be similar oil spilling).Namely, be in the scattering entropy H of the scattering entropy H distribution range of oil spilling, still have a certain amount of class oil spilling, but on certain proportion, oil spilling and class oil spilling are distinguished.

Further, H histogram comprises the function curve of at least two scattering entropy H;

Scattering entropy H distribution range according to H histogram determination oil spilling comprises:

Determine that function curve that the distribution range of H value meets default extent of oil spill is the scattering entropy H function curve of oil spilling; Using the scattering entropy H distribution range of the scope corresponding to the H value corresponding to the scattering entropy H function curve of oil spilling as oil spilling.

In H histogram, usually by two, or many curves composition, due to the characteristic of oil spilling and class oil spilling, the curve of curve corresponding for oil spilling and class oil spilling can be distinguished in the curve formed.As shown in Figure 4, horizontal ordinate is the numerical value of H, and ordinate is the frequency of occurrences.In figure, clearly define two curves, the curve that one of them peak value is higher, the curve being namely arranged in figure left is that (the H value corresponding to usual oil spilling is relatively little to the curve of oil spilling, and the frequency that H occurs is higher than the frequency of class oil spilling), the curve that another peak value is lower is then the curve of class oil spilling.By the curve of two in comparison diagram, clearly curve can be divided into the H curve of oil spilling and the H curve of class oil spilling.

After the curve determining oil spilling, the H value change that this Curves is corresponding can determine, and as shown in Figure 4, scope is between 0 to 0.5.Namely determine the scattering entropy H distribution range of oil spilling.

Further, in order to improve, oil spilling is judged more accurately, can by the range shorter of H value corresponding for oil spilling Curves.As in Fig. 4, can be (numerical value of scattering entropy H) 0-0.4 by the numerical value of horizontal ordinate, or the scope of 0-0.3 be as the scattering entropy H distribution range of oil spilling.Although the H value of the oil spilling of definite part can be got rid of by the scattering entropy H distribution range reducing oil spilling, can recognize this curve by observing curve, substantially in normal distribution, namely the value of the edge of curve is less.The value of curve illustrates the frequency of appearance, and so the frequency of occurrences at curved edge place is namely less, and the frequency of occurrences of class oil spilling is then higher.Thus, the H value of the class oil spilling participating in this scope can be reduced by the scattering entropy H distribution range reducing oil spilling, thus improve the correct probability assert oil spilling.

Step S103 can also be divided into following steps:

S1031, generates H-CPD coordinate X-Y scheme according to scattering entropy H and CPD standard deviation;

S1032, will meet the H-CPD coordinate points of coordinate points as oil spilling of the scattering entropy H distribution range of oil spilling in H-CPD coordinate X-Y scheme;

S1033, using the geographic coordinate of the geographic coordinate corresponding to H-CPD coordinate points as oil spilling.

In step 1031, the H-CPD coordinate X-Y scheme generated, include oil spilling and class oil spilling, by calculating CPD standard deviation, the oil spilling of a part can be removed, again by confirming that the scattering entropy H distribution range of oil spilling also can get rid of the scope at certain class oil spilling place, after namely two kinds of methods being merged, the position at oil spilling place just can be determined further.It should be noted that, first calculate CPD standard deviation, the scattering entropy H distribution range still first calculating oil spilling does not need to do requirement, and both can carry out simultaneously.And first can mark out corresponding coordinate points according to scattering entropy H and CPD standard deviation in H-CPD figure, then limit the scope at oil spilling place further according to the scattering entropy H distribution range of the oil spilling confirmed.Namely can be confirmed the coordinate of H-CPD by scattering entropy H and CPD standard deviation, oil spilling and class oil spilling all can show on figure, are normally divided into two, or multiple different region.Further, by the scattering entropy H distribution range of oil spilling obtained, confirm which region is only real oil spilling region further.

In step 1032, namely determine the coordinate of oil spilling in H-CPD figure.

In step 1033, namely determine the geographic coordinate of concrete oil spilling.

In the H-CPD figure generated, its transverse and longitudinal coordinate is respectively scattering entropy H and CPD standard deviation, but considers service condition, and the ordinate of H-CPD coordinate X-Y scheme is scattering entropy H, and horizontal ordinate is CPD standard deviation.

Concrete, step S101 comprises: following steps:

S1011, obtains the T3 of SAR data, or C3 matrix;

S1012, determines T3, or the coherence matrix proper vector of C3 matrix;

S1013, calculates scattering entropy H according to coherence matrix proper vector.

Wherein, step S1013, calculates scattering entropy H according to coherence matrix proper vector and comprises:

Use and p _i=λ _i/ (λ ₁+ λ ₂+ λ ₃) calculate scattering entropy H, wherein, λ ₁, λ ₂and λ ₃for the proper vector of T3 or C3 matrix.

Further, step S101, calculates CPD standard deviation according to SAR data and also comprises:

Use formula calculate the value of CPD standard deviation, wherein L be the image of SAR data look number, SHH is T3 or C3 entry of a matrix element, and SVV* is T3, or the conjugate number of element S VV in C3 matrix, ρ for representing the calculating of phase differential, without physical meaning.

Polarization SAR system is using scattering matrix S as master record unit, and multiple scattering matrix have expressed contacting between incident electromagnetic wave and scattering electromagnetic wave Jones vector:

${\underset{\‾}{E}}_s=\frac{e^{-\mathrm{jkr}}}{r}S{\underset{\‾}{E}}_i=\frac{e^{-\mathrm{jkr}}}{r}\left[\begin{array}{cc}{S}_{\mathrm{HH}}& S_{\mathrm{HV}}\\ S_{\mathrm{VH}}& S_{\mathrm{VV}}\end{array}\right]{\underset{\‾}{E}}_i$

$S=\left[\begin{array}{cc}{S}_{\mathrm{HH}}& S_{\mathrm{HV}}\\ S_{\mathrm{VH}}& S_{\mathrm{VV}}\end{array}\right]$

Complex scattering coefficients can be write as the form of amplitude and phase place:

$S_{\mathrm{pq}}=\left|S_{\mathrm{pq}}\right|e^{{\mathrm{j\φ}}_{\mathrm{pq}}}$

Wherein, | S _pq| be scattering amplitude, φ is phase place, and p, q=H, V represent launching and receiving electromagnetic polarization mode respectively.Thus:

$S=\left[\begin{array}{cc}\left|S_{\mathrm{HH}}\right|e^{{\mathrm{j\φ}}_{\mathrm{HH}}}& \left|S_{\mathrm{HV}}\right|e^{{\mathrm{j\φ}}_{\mathrm{HV}}}\\ \left|S_{\mathrm{VH}}\right|e^{{\mathrm{j\φ}}_{\mathrm{VH}}}& \left|S_{\mathrm{VV}}\right|e^{{\mathrm{j\φ}}_{\mathrm{VV}}}\end{array}\right]$

Using the relative phase of scattering matrix first element S HH as reference phase, then above formula is expressed as with relative phase form:

$S{=e}^{{\mathrm{j\φ}}_{\mathrm{HH}}}\left[\begin{array}{cc}\left|S_{\mathrm{HH}}\right|& \left|S_{\mathrm{HV}}\right|e^{{\mathrm{j\φ}}_x}\\ \left|S_{\mathrm{VH}}\right|e^{{\mathrm{j\φ}}_x}& \left|S_{\mathrm{VV}}\right|e^{{\mathrm{j\φ}}_c}\end{array}\right]$

Wherein, φ _x, φ _cbe called cross polarization phase differential and same polarization phase differential:

φ _x＝φ _HV-φ _HH＝φ _VH-φ _HH；

φ _c＝φ _VV-φ _HH；

Usually, each absolute phase φ _hH, φ _vV, φ _hVall meet being uniformly distributed between [-π, π], therefore do not carry any effective geography of Target scalar and the dielectric property information (φ in formula _cbe also called C13 parameter).But phase differential no longer may meet and is uniformly distributed, thus be associated with Target scalar character.In fact, cross polarization phase difference _xbeing uniformly distributed on [-π, π] is still met for most natural target thing, and same polarization phase difference _cthen be similar to and meet Gaussian distribution, and the average of its distribution function, variance and object and sensor characteristic have close association.Same polarization phase difference probability density profile broadening is anti-phase relevant to HH, VV POLARIZATION CHANNEL backscattering echo signal multiple correlation coefficient, and HH, VV POLARIZATION CHANNEL multiple correlation coefficient can be expressed as:

$\mathrm{\ρ}=\frac{S_{\mathrm{HH}}\·S_{\mathrm{VV}}^{*}}{\sqrt{{\left|S_{\mathrm{HH}}\right|}^2\·{\left|S_{\mathrm{VV}}\right|}^2}};$

When related coefficient is 0, φ _cbe similar to φ _x, distribution curve is evenly distributed feature, does not include effective information; When related coefficient is 1, distribution curve then shows as Dirac function.

CPD standard deviation is defined as:

$\mathrm{CPDstd}=\mathrm{std\ρ}\left[\frac{1}{L}{\mathrm{\Σ}}_{i=1}^L(S_{\mathrm{HH}}\×S_{\mathrm{VV}}^{*})\right];$

Std represents the calculating carrying out standard deviation, L be the image of SAR data look number, ρ ^ρfor representing the calculating of phase differential, without physical meaning, SHH is T3 or C3 entry of a matrix element, and SVV* is T3, or the conjugate number of element S VV in C3 matrix.

CPD standard deviation and passage related coefficient can be evaluated interchannel correlativity.In low correlation region, the change of phase place to related coefficient has stronger susceptibility, can well region class like oil spilling and oil spilling.

Because similar oil spilling is different to different parameters sensitivity from oil spilling, based on the estimation of C3 matrix of refining, then in H-CPDstd plane, careful differentiation is carried out to similar oil spilling and oil spilling, finally obtain sophisticated category result, to eliminate the impact of class oil spilling in oil spilling judges.

The oil spilling detection method of a kind of polarization decomposing that the application provides, also comprises calculating correlation coefficient ρ, if when ρ is greater than default numerical value, then determines to adopt according to scattering entropy H and CPD standard deviation determination oil spilling;

wherein SVV* is T3, or the conjugate number of element S VV in C3 matrix, SHH and SVV is T3, or the element in C3 matrix.

Wherein, C3 is polarization covariance matrix, and T3 is the matrix that C3 is formed through reciprocity, namely coherence matrix.CPD standard deviation and passage related coefficient can be evaluated interchannel correlativity.In low correlation region, the change of phase place to related coefficient has stronger susceptibility, can well region class like oil spilling and oil spilling, thus, when high relevant range (related coefficient is larger), the mode of scattering entropy H and the verification of CPD standard deviation compound (namely the application provide mode) is preferably adopted to distinguish oil spilling and class oil spilling.

On the other hand, step S102, the scattering entropy H corresponding to geographic coordinate show that the scattering entropy H distribution range of oil spilling also comprises the steps:

Determine that scope that scattering entropy numerical value is less than 0.3 is the scattering entropy distribution range of oil spilling.

Directly determine that the zone of reasonableness of H also can directly use based on experience value, if use this scope, then do not need to be confirmed by the histogrammic mode of H.

On the whole in correlation technique, use the mode that oil spilling is detected of polarization parameter.

Polarization SAR data can comprise several as follows,

1, scattering angle α,

The value of scattering angle α and the physical mechanism of scattering process connect each other, correspond to from odd scattering (surface scattering) (α=0 °) change to dipole scattering (volume scattering) (α=45 °) to even scattering (dihedral angle scattering) (α=90 °).

2, scattering entropy H,

Scattering entropy (Entropy) (0≤H≤1) illustrates scattering medium by the randomness of isotropic scatterning (H=0) to completely random scattering (H=1).

If H value very low (H<0.3), then the weak depolarization of system, advantage scattering mechanism is eigenvalue of maximum characteristic of correspondence vector; If H value is very high, then illustrate that target depolarisation effect is stronger; If H value is 1, then polarization information is 0, and target scattering is actually a random noise processes.Computing formula is:

wherein, p _i=λ _i/ (λ ₁+ λ ₂+ λ ₃), λ 1, λ 2, λ 3 are the proper vector of coherence matrix (T3, or C3).

3, polarize anti-entropy A,

Polarization scattering entropy provides the information of overall scattering mechanism in same resolution element, but, at low entropy or middle constant entropy (λ ₁> λ ₂, λ ₃) in situation, scattering entropy can not provide relevant two less eigenvalue λ ₂, λ ₃between relation, now need to further consider the anti-entropy A of polarization.Expression formula is as follows: anti-entropy A reflect Cloude decompose in magnitude relationship between the relatively weak scattering component of beyond advantage scattering mechanism two, be the supplementary parameter of polarization entropy.Only as H>0.7, A just can be used as the source identified further, otherwise A comprises too high random noise.H value becomes large, then polarization can decline by sub-category number.When H value increases, A is obvious as scattering category significance.

Use the product combination of scattering entropy, anti-entropy to form polarization characteristic spectrum, there are 4 parts, be divided into a kind of scattering parameter to be (1-H) (1-A); A kind of scattering mechanism is taken as the leading factor, and the relatively little parameter of another two kinds of scatterings is H (1-A); Two kinds of scattering phases are to being HA; Single, even, random scatter are relatively unanimously A (1-H).

According to scattering entropy H and scattering angle α plane relation, setting T3 eigenvalue λ 1, λ 2, λ 3, by descending order, Probability p 1 corresponding to it, p2, p3 also press descending order, in conjunction with the scattering angle of H-Alpha and the relation of scattering entropy, set up the scattered information of Eigenvalues Decomposition, disclose the scattering mechanism of oil spilling.

Application H-α feature space divides atural object scattering mechanism, obtains H-alpha taxonomy result figure.Seawater, oil spilling Bragg scattering have the feature of low scattering entropy, low scattering angle, and its scattering signatures value is positioned at H-α feature space lower left quarter.Because oil spilling is to the inhibiting effect of waves of seawater, thus make sea more tranquil.And accordingly, research oil spilling is in the characteristic of H-α feature space.

Itself can carry out from following three aspects with the contrast of correlation technique.

1, application single polarization SAR carries out oil spilling detection, mostly carry out from image procossing aspect, different Methods of Mathematical Physics is adopted to carry out oil spilling information extraction, but on image, show multiple types all easily cause error like the blackspot of oil spilling, as low wind, stream etc. are easily judged as oil spilling, precision is subject to certain impact.Adopt multipolarization SAR data can obtain more oil spilling scattered information, precision increases, and still there is some and is difficult to remove with the phenomenon that oil spilling has identical scattering properties, detection accuracy also has not enough.Thus, according to the entropy of polarization decomposing and the standard deviation construction feature space (H-CPD coordinate X-Y scheme) of same polarization phase difference between channels, carry out the research of oil spilling polarized spectrum, deeply excavate polarization characteristic spectrum parameter and information extraction ability thereof, express the information of oil spilling scattering process.The combination carrying out phase differential and entropy parameter can be extracted effectively to oil spilling information.

2, oil film and class oil film cover the existence decay to a certain degree compared with backscattering of ocean water coefficient of backscattering from ocean surfaces coefficient, thus form luminance contrast on image, and contrast level and electromagnetic wave bands, polarization mode etc. have certain relation.But under some specific condition, only the doubtful oil films such as biological membrane and offshore spilled oil oil film effectively cannot be distinguished by backscattering coefficient, adopt multipolarization SAR data can obtain the multiple scattered information of oil spilling, research oil spilling polarization characteristic, set up multiple oil spilling parameter, be mainly polarization entropy, anti-entropy, SPAN, Alpha angle, CPD standard deviation, co-pol related coefficient, for the uncertainty (as the oil spilling of CPD and the differentiation threshold value of class oil film) of eigenwert to oil spilling information extraction, the extracting method of research multiparameter.

The scattering main manifestations of marine oil overflow is single scattering, show comparatively obvious symmetrical scattering, HH and VV polarization phases is not had an impact, because oil spilling and class oil film exist difference in phase place, study the scattering properties of oil spilling in H-Alpha plane, set up parameter combinations, mainly carry out the plane research of entropy and polarization phases standard deviation, carry out the scrutiny of this plane, to study the small differences of oil spilling and class oil spilling, do to become more meticulous further to oil spilling and class oil film and extract.

3 polarization parameter entropys are combined the technology detecting oil spilling that becomes more meticulous and verify in the following areas with phase differential.

1) polarization parameter detection oil spilling compares

Polarization scattering parameter entropy, scattering angle, anti-entropy diagram, be shown in Fig. 2 a to Fig. 2 c.Fig. 2 a is the image obtained according to scattering entropy, and Fig. 2 b is the image obtained according to scattering angle, and Fig. 2 c is the image obtained according to the anti-entropy of plan, and in Bragg scattering, parameter entropy, scattering angle can obviously reflect oil spilling information; Anti-entropy is that of entropy supplements, as H>0.7, anti-entropy could as the parameter of carrying out identifying, mainly when entropy is lower, the impact that second and third eigenwert is subject to noise is very large, to such an extent as to anti-entropy also has a lot of noises, the entropy parameter value of Fig. 2 a is generally below 0.7, thus the noise of anti-entropy parameter seems comparatively large, and oil spilling information is not embodied, concrete as Fig. 2 c.

2) scattering entropy (0≤H≤1) represents that scattering matter is from isotropic scatterning (H=0) to the randomness of completely random scattering (H=1).Oil spilling occurs at sea, oil spilling and seawater mix, and also can produce the chemical processes such as emulsification, in this case, oil spilling scattering matrix is not unique oil spilling scattering, has the scatterings such as profit, be a random noise processes, be thus non-homogeneous value in the performance of entropy, scattering angle.ɑ presentation surface scatters to the AVERAGE MECHANISM of two area scatterings, and seawater shows as Bragg scattering, be thus dark-coloured at image, and oil spilling and water is potpourri, show the feature of two area scatterings, are light tone at image.Anti-entropy parameter is supplementing of entropy parameter, because entropy is lower, effectively can not extract oil spilling information from seawater.

By calculating T3 matrix, obtain phase differential (CPD), on this basis, calculate the standard deviation of CPD, obtain Fig. 3 a (according to the image that CPD standard deviation obtains, the position irised out in figure is class oil spilling, namely by part that method provided by the present invention is more accurately got rid of), in figure oil spilling and ship information obvious, but the side lobe effect of oil well does not react, noise decrease, at foundation entropy, the oil spilling information of the parameter extractions such as scattering angle, Fig. 3 b is (according to the image that scattering entropy H obtains, the position irised out in figure is oil spilling, namely by part that method provided by the present invention is more accurately got rid of) in oil spilling and ship show in the picture and there is no difference, and significantly distinguished by CPD result of calculation display ship and oil spilling information in Fig. 3 a, this part false-alarm is removed by phase difference calculating, oil spilling information extraction precision is improved.

The oil spilling detection method of a kind of polarization decomposing that the application provides, by obtaining polarization SAR data, and calculate scattering entropy H and CPD standard deviation by these data, and then the position of oil spilling is determined by two data, effectively oil spilling and class oil spilling are distinguished, compensate for the single data passed through and carry out the deficiency that oil spilling and class oil spilling carry out distinguishing, and, the scattering entropy H distribution range of oil spilling is determined by the histogram of scattering entropy H, and eliminate by CPD standard deviation the class oil spilling part (class oil spilling also can be described as false-alarm) that scattering entropy H cannot determine further, thus oil spilling information extraction precision is improved, better solve deficiency of the prior art.

The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, for a person skilled in the art, the present invention can have various modifications and variations.Within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (10)

1. an oil spilling detection method for polarization decomposing, is characterized in that, comprising:

Obtain multipolarization SAR data, described multipolarization SAR data comprises geographic coordinate and the scattering entropy corresponding to described geographic coordinate and CPD standard deviation;

Scattering entropy numerical value corresponding to described geographic coordinate draws the scattering entropy distribution range of oil spilling;

Determine the scattering entropy distribution range meeting described oil spilling, and the geographic coordinate that there is the position of described CPD standard deviation is the geographic coordinate of oil spilling.

2. the oil spilling detection method of a kind of polarization decomposing according to claim 1, is characterized in that, described scattering entropy numerical value corresponding to described geographic coordinate show that the scattering entropy distribution range of oil spilling comprises:

Draw the scattering entropy histogram of the scattering entropy frequency of occurrences corresponding to described geographic coordinate;

According to the scattering entropy distribution range of described scattering entropy histogram determination oil spilling, the scattering entropy distribution range of described oil spilling comprises the scattering entropy of oil spilling and the scattering entropy of class oil spilling.

3. the oil spilling detection method of a kind of polarization decomposing according to claim 2, is characterized in that, described scattering entropy histogram comprises the function curve of at least two scattering entropys;

The described scattering entropy distribution range according to described scattering entropy histogram determination oil spilling comprises:

Determine that function curve that the distribution range of scattering entropy meets default extent of oil spill is the scattering entropy function curve of oil spilling;

Using the scattering entropy distribution range of the scope corresponding to scattering entropy corresponding for the scattering entropy function Curves of described oil spilling as oil spilling.

4. the oil spilling detection method of a kind of polarization decomposing according to claim 1, is characterized in that, describedly determines the scattering entropy distribution range meeting described oil spilling, and the geographic coordinate that the geographic coordinate that there is the position of described CPD standard deviation is oil spilling comprises:

H-CPD coordinate X-Y scheme is generated according to described scattering entropy and described CPD standard deviation;

The H-CPD coordinate points of coordinate points as oil spilling of the scattering entropy distribution range of described oil spilling will be met in described H-CPD coordinate X-Y scheme;

Using the geographic coordinate of the geographic coordinate corresponding to described H-CPD coordinate points as oil spilling.

5. the oil spilling detection method of a kind of polarization decomposing according to claim 4, is characterized in that, the ordinate of described H-CPD coordinate X-Y scheme is scattering entropy, and horizontal ordinate is CPD standard deviation.

6. the oil spilling detection method of a kind of polarization decomposing according to claim 1, is characterized in that, described acquisition multipolarization SAR data, comprising:

Obtain the T3 of described SAR data, or C3 matrix;

Determine described T3, or the coherence matrix proper vector of C3 matrix;

Described scattering entropy is calculated according to described coherence matrix proper vector.

7. the oil spilling detection method of a kind of polarization decomposing according to claim 6, is characterized in that, calculates described scattering entropy comprise according to described coherence matrix proper vector:

Use and p _i=λ _i/ (λ ₁+ λ ₂+ λ ₃) calculate described scattering entropy, wherein, λ ₁, λ ₂and λ ₃for the proper vector of T3 or C3 matrix.

8. the oil spilling detection method of a kind of polarization decomposing according to claim 1, is characterized in that, described acquisition multipolarization SAR data also comprises:

Use formula calculate the value of CPD standard deviation, wherein L be the image of SAR data look number, S _hHt3 or C3 entry of a matrix element, S _vV* be T3, or element S in C3 matrix _vVconjugate number, ρ for representing the calculating of phase differential, without physical meaning.

9. the oil spilling detection method of a kind of polarization decomposing according to claim 6, is characterized in that, also comprises calculating correlation coefficient ρ, if when described ρ is greater than default numerical value, then determines to adopt according to described scattering entropy and described CPD standard deviation determination oil spilling;

wherein S _vV* be T3, or element S in C3 matrix _vVconjugate number, S _hHand S _vVbe T3, or the element in C3 matrix.

10. the oil spilling detection method of a kind of polarization decomposing according to claim 1, is characterized in that, described scattering entropy numerical value corresponding to described geographic coordinate show that the scattering entropy distribution range of oil spilling also comprises:

Determine that scope that scattering entropy numerical value is less than 0.3 is the scattering entropy distribution range of oil spilling.