CN102608604B - Sea oil spill detection method based on co-polarization characteristic power map - Google Patents

Abstract

The invention provides a method for detecting sea surface oil spillage on basis of a common polarimetric characteristic power diagram, which comprises the following steps of: obtaining the common polarimetric characteristic power diagram of the polarimetric synthetic aperture radar (SAR) data according to a scattering coherent matrix; calculating the polarimetric base frame height and the maximum polarimetric characteristic power of each pixel point in a polarimetric SAR observation area; computing the maximum polarimetric characteristic power attenuation ratios of the sea return data at different incidence angles; utilizing a parzen window function to respectively estimate the sea clutter statistical distributions of the polarimetric base frame height and the charactertistic power attenuation ratios; and presetting the detection false alarm rate, and respectively comparing the polarimetric base frame height and the characteristic attenuation ratios with the corresponding calculated detection threshold to acquire the oil spillage detection result. The method does not need additional calculated quantity, the correct detection probability of an oil slick can be improved, the sea surface oil spillage condition is efficiently detected, and the area of the oil slick can be correctly estimated by a maximum connected domain analytical method.

Description

Offshore spilled oil detection method based on co-polarization feature power diagram

Technical field

The present invention relates to polarimetric synthetic aperture radar ocean remote sensing applied technical field, particularly a kind of offshore spilled oil detection method based on co-polarization feature power diagram.

Background technology

The pollution that Oil spills causes is one of at present serious marine pollution phenomenon, and ecologic environment and people's life has all been caused to very large impact.Oil pollution on sea is generally caused by two kinds of reasons, and a kind of is to be caused by oil accident, and fuel oil discharge in violation of regulations another kind of in sea-freight process causes.The pollution that oil accident causes comprises the crude oil leakage of oil platform, the fuel leakage that shipwreck causes etc.; the leakage of oil that this class accident causes conventionally can be by media report out; greasy dirt area is large; leakage of oil approximate location is clear and definite; the leakage of oil of for example, the leakage of oil of Gulfian BP company oil platform in 2010, Peng Lai 19-3 oil gas field platform in 2011 and leakage of oil of New Zealand's " thunder is received " number cargo ship in 2011 etc.The discharge in violation of regulations of transportation ship in sea-freight process causes oil pollution, and this class phenomenon is unpredictable in advance, and the normally narrow and long strip-type of the shape of consequent slick.Because sea oil pollution is huge on the impact of human being's production, life, therefore slick phenomenon is carried out Real-Time Monitoring and cleared up and be very important in time.At present, the monitoring means that offshore spilled oil is polluted is mainly to utilize SAR(Synthetic Aperture Radar, synthetic-aperture radar) monitor, SAR due to have round-the-clock, round-the-clock, penetrable, the feature such as can observe on a large scale, become the technical way in ocean remote sensing application.Yet, when traditional single polarization SAR system is observed target, only can obtain a backscatter intensity information of target, quantity of information few and target is differentiated limited in one's ability, and complete polarization SAR system can obtain the back scattering information that target is complete, be therefore subject to people's extensive attention.

Slick is a kind of organism membranous surface activity material, there is toughness, very large with the viscoelasticity difference of seawater surface, thereby its impact on sea maximum is the gravity capillary wave of the ocean surface minimizing reflectivity of can decaying, and then show as dark areas on SAR image, it is practicable therefore utilizing SAR to monitor ocean surface oil spilling phenomenon.Its physical interpretation is: under the irradiation of radar wave, gravity capillary wave and the radar wave of ocean surface can produce resonance under certain conditions, and this resonance effect is called as Bragg scattering phenomenon.Oil film is due to this Bragg resonance that decayed, so its backscattering coefficient is low than surrounding environment, and mechanism that SAR detects offshore spilled oil that Here it is.Yet, due to marine environment more complicated, other oceanographic phenomenas also can produce similar dark areas on SAR image, such as low wind speed region, internal wave of ocean, unimolecule top layer biological membrane, grease glacier, rainfall, ocean current and upward flow region etc., these factors have affected the detection of SAR to offshore spilled oil greatly.The detection of oil film is distinguished slick and oil film analog exactly.Through people a large amount of research, show, under low sea situation (sea wind speed is 2.5m/s～12.5m/s), optimum utilizes SAR system to detect slick, now the contrast on oil film and sea is larger.

The major technique of utilizing SAR system to detect Algorithms for Oil Slick has experienced the development course from single polarization to complete polarization.When utilizing single polarization (VV polarization) data to detect oil film, the technological means of Main Basis is the attenuation ratio of backscattering coefficient.According to the Marangoni attenuation model of monomolecular surdace film, M.Gade utilized SIR-C/X-SAR data research in 1998 biological membrane and artificial rust rear to attenuation characteristic, has pointed out that the attenuation characteristic of surface film is subject to the impact of radar line of sight, wind direction and wind speed.But the attenuation ratio of the backscattering coefficient of different surfaces membrane substance slightly difference is not obvious.2010, rear to attenuation ratio under different wind speed of oil film under VV polarization mode that Kim has utilized TerraSAR-X data research, assay has also used the data of ERS-2, ENVISAT, RADARSAT-1 and ALOS-PALSAR to be verified simultaneously, result shows that TerraSAR-X data can test to oil film effectively, but the detection based on attenuation characteristic is sometimes difficult to distinguish oil film and biological membrane.Developed subsequently the offshore spilled oil detection method based on polarization SAR, polarization SAR data can be strengthened the detection performance to oil film.The technological means being applied to has the biological membrane sorting technique of decomposing based on polarization characteristic value, and the entropy that wherein polarizes, polarizing angle and anti-entropy can be for distinguishing different biological membranes.Also have the Oil spill detection method based on the filtering of Mueller matrix, and used SIR-C/X-SAR data and the ALOS-PALSAR data of looking to be verified more.Slick is detected and also also has the maximum likelihood classification algorithm distributing based on Wishart, but the method needs the prior imformation of oil film.

When existing method detects oil film, some method has only been applied the backward attenuation characteristic of oil film, and some method has only been applied the polarization characteristic of oil film.Such detection method is suitable for when thicker the lower SAR system of noise limit or oil film, but when oil film thickness less, inapplicable when around the side lobe effect of the metal target such as naval vessel is excessive.Oil film thickness is less, not obvious to the decay of backscattering coefficient, and the impact that attenuation characteristic is now subject to speckle noise is large; Around oil film, the side lobe effect of metal target is excessive, can affect the analysis to oil film region polarization characteristic.Therefore, said method is difficult to exactly oil film be detected and area estimation under sea environment or sea situation complicated situation.

Summary of the invention

Object of the present invention is intended at least solve one of above-mentioned technological deficiency.

For this reason, the present invention proposes a kind of offshore spilled oil detection method based on co-polarization feature power diagram, the method does not need extra calculated amount, can greatly improve the correct detection probability of oil film, and can keep the integrality to oil film Region Segmentation.

For achieving the above object, the embodiment of the invention discloses a kind of offshore spilled oil detection method based on co-polarization feature power diagram, comprise the steps:

According to the scattering coherence matrix of polarimetric synthetic aperture radar target, calculate the co-polarization feature power diagram of each pixel;

According to the feature power diagram of described polarimetric synthetic aperture radar data, retain the maximum feature performance number of described each pixel, the polarization pedestal height in calculating observation region;

According to the maximum feature power diagram of described observation area, calculate the feature power attenuation ratio of each pixel of described observation area under identical incident angle;

Utilize window function to estimate respectively the extra large clutter statistical distribution of described polarization pedestal height and the extra large clutter statistical distribution of described feature power attenuation ratio;

False alarm rate is set, according to the extra large clutter statistical distribution of the extra large clutter statistical distribution of described false alarm rate, described polarization pedestal height and described feature power attenuation ratio, calculate respectively the detection threshold of described polarization pedestal height and the detection threshold of described feature power attenuation ratio, by described polarization pedestal height and described feature power attenuation than comparing to obtain oil spilling testing result with corresponding detection threshold respectively.

According to the offshore spilled oil detection method of the embodiment of the present invention, reduced calculated amount, improved the contrast between oil film and sea, therefore can greatly improve the correct detection probability of oil film, and can keep the integrality to oil film Region Segmentation, can to offshore spilled oil situation, monitor timely and effectively.

In one embodiment of the invention, the feature power diagram of described polarimetric synthetic aperture radar data can calculate under co-polarization antenna receiving-sending mode.

In one embodiment of the invention, the polarization characteristic power representing in the feature power diagram of described polarimetric synthetic aperture radar data is:

$P(\mathrm{\ϵ},\mathrm{\τ})=\frac{2\mathrm{\π}}{k^2}h{(\mathrm{\ϵ},\mathrm{\τ})}^T\mathrm{Kg}(\mathrm{\ϵ},\mathrm{\τ})$

Wherein, k is electromagnetic wave number, and K is Kennaugh matrix, the in the situation that of co-polarization antenna receiving-sending,

H (ε, τ)=g (ε, τ)=[1 cos2 τ cos2 ε sin2 τ cos2 ε sin2 ε] ^t, h (ε, τ) ^tfor the transposition of h (ε, τ), ε ∈ [π/4, π/4], the ellipticity of expression polarization of ele ellipse, τ ∈ [pi/2, pi/2], the polarization tilt angle of expression polarization of ele ellipse.

In one embodiment of the invention, according to the feature power diagram of described polarimetric synthetic aperture radar data, show that described polarization pedestal height is:

$h_p=\frac{\min \left(P(\mathrm{\ϵ},\mathrm{\τ})\right)}{\max \left(P(\mathrm{\ϵ},\mathrm{\τ})\right)}$

Wherein, h _pfor described polarization pedestal height, max (P (ε, τ)) represents the maximal value of co-polarization feature power, and min (P (ε, τ)) represents the minimum value of co-polarization feature power.

In one embodiment of the invention, according to the feature power diagram of described polarimetric synthetic aperture radar data, the feature power attenuation ratio of described sea echo is:

$d_r=\frac{P_{\max }{(\mathrm{\ϵ},\mathrm{\τ})}_{\mathrm{sea}}}{P_{\max }(\mathrm{\ϵ},\mathrm{\τ})}$

Wherein, d _rfor the feature power attenuation ratio of described sea echo, P _max(ε, τ)=max (P (ε, τ)), max () represents to ask the maximal value of array, P _max(ε, τ) _searepresent the maximum polarization characteristic power in Wu Mo overlay area, sea.

According to the offshore spilled oil detection method of the embodiment of the present invention, attenuation characteristic and the polarization characteristic of having applied oil film detect simultaneously, and these two characteristic parameters can obtain simultaneously from the co-polarization feature power diagram of target, do not need extra calculated amount, utilize in addition the attenuation characteristic of maximum feature power calculation oil film, reduce the impact of SAR data speckle noises, improved the contrast between oil film and sea simultaneously.

In one embodiment of the invention, utilize parzen window function to estimate respectively the extra large clutter statistical distribution of described polarization pedestal height and the extra large clutter statistical distribution of described feature power attenuation ratio, and the kernel function of parzen window function adopt lognormal distribution.

Utilize parzen window function to carry out matching to extra large clutter statistical distribution, in the situation that sample number is abundant, the probability density that window method is estimated can converge to real probability distribution.

In one embodiment of the invention, according to described false alarm rate, utilize following formula to calculate the detection threshold of described polarization pedestal height,

$p_{\mathrm{fa}}={\∫}_{Z_{{\mathrm{Th}}_p}}^{\∞}{\hat{p}}_{h_p}\left(h_p\right)\mathrm{dx},$

Wherein, p _fafor described false alarm rate,

for the extra large clutter statistical distribution of described polarization pedestal height, detection threshold for described polarization pedestal height.

In one embodiment of the invention, according to described false alarm rate, utilize following formula to calculate the detection threshold of described feature power attenuation ratio,

$p_{\mathrm{fa}}={\∫}_{Z_{{\mathrm{Td}}_r}}^{\∞}{\hat{p}}_{d_r}\left(d_r\right)\mathrm{dx},$

Wherein, p _fafor described false alarm rate,

for the extra large clutter statistical distribution of described feature power attenuation ratio,

detection threshold for described feature power attenuation ratio.

In one embodiment of the invention, described polarization pedestal height and described feature power attenuation, than comparing to obtain oil spilling testing result with detection threshold and the feature power attenuation of corresponding polarization pedestal height than detection threshold respectively, are comprised the steps:

According to the extra large clutter statistical distribution of described false alarm rate and described polarization pedestal height, obtain the testing result of described polarization pedestal height;

According to the extra large clutter statistical distribution of described false alarm rate and described feature power attenuation ratio, obtain the testing result of described feature power attenuation ratio;

The testing result of the testing result of described polarization pedestal height and described feature power attenuation ratio is merged to obtain described oil spilling testing result.

According to the offshore spilled oil detection method of the embodiment of the present invention, will polarize pedestal height and feature power attenuation are than (h _pand d _r) testing result combine, greatly improved the correct detection probability of oil film.

In one embodiment of the invention, the offshore spilled oil detection method based on co-polarization feature power diagram also comprises the steps:

Described oil spilling testing result is carried out to largest connected regional analysis to obtain offshore spilled oil area, and wherein, described offshore spilled oil area is:

S=connected(res)×ρ _a×ρ _gr

Wherein, ρ _gr=ρ _r/ sin (θ _i), θ _ifor the incident angle of radar beam, ρ _afor the size of orientation to pixel cell, ρ _rfor the size of distance to pixel cell, connected (res) represents the largest connected domain analysis process to the described oil spilling testing result of bianry image.

Area based on largest connected domain analysis oil film can significantly reduce the impact that false-alarm is estimated film area, and can keep the integrality to oil film Region Segmentation.

The aspect that the present invention is additional and advantage in the following description part provide, and part will become obviously from the following description, or recognize by practice of the present invention.

Accompanying drawing explanation

Above-mentioned and/or the additional aspect of the present invention and advantage will become from the following description of the accompanying drawings of embodiments and obviously and easily understand, wherein:

Fig. 1 is the FB(flow block) of the offshore spilled oil detection method based on co-polarization feature power diagram according to one embodiment of the invention;

Fig. 2 is the process flow diagram of the offshore spilled oil detection method based on co-polarization feature power diagram in accordance with another embodiment of the present invention;

Fig. 3 a-Fig. 3 c is respectively according to three of the embodiment of the present invention groups of general power figure for the full polarimetric SAR data of testing;

Fig. 4 is the polarization pedestal height h according to different target in the embodiment of the present invention _pand d is compared in feature power attenuation _rtwo-dimensional distribution;

Fig. 5 a and Fig. 5 b are respectively the polarization pedestal height h according to the embodiment of the present invention _pand d is compared in feature power attenuation _rclutter statistical Butut;

Fig. 6 a-Fig. 6 c is that in Fig. 3 a-Fig. 3 c three groups are for offshore spilled oil testing result figure corresponding to the full polarimetric SAR data of testing; With

Fig. 7 a-Fig. 7 b is largest connected regional analysis result figure corresponding to the offshore spilled oil testing result in Fig. 6 a-Fig. 6 b.

Embodiment

Describe embodiments of the invention below in detail, the example of described embodiment is shown in the drawings, and wherein same or similar label represents same or similar element or has the element of identical or similar functions from start to finish.Below by the embodiment being described with reference to the drawings, be exemplary, only for explaining the present invention, and can not be interpreted as limitation of the present invention.

Disclosing below provides many different embodiment or example to be used for realizing different structure of the present invention.Of the present invention open in order to simplify, hereinafter the parts of specific examples and setting are described.Certainly, they are only example, and object does not lie in restriction the present invention.In addition, the present invention can be in different examples repeat reference numerals and/or letter.This repetition is in order to simplify and object clearly, itself do not indicate the relation between discussed various embodiment and/or setting.In addition, the various specific technique the invention provides and the example of material, but those of ordinary skills can recognize the property of can be applicable to of other techniques and/or the use of other materials.In addition, First Characteristic described below Second Characteristic it " on " structure can comprise that the first and second Characteristics creations are for the direct embodiment of contact, also can comprise the embodiment of other Characteristics creation between the first and second features, such the first and second features may not be direct contacts.

In description of the invention, it should be noted that, unless otherwise prescribed and limit, term " installation ", " being connected ", " connection " should be interpreted broadly, for example, can be mechanical connection or electrical connection, also can be the connection of two element internals, can be to be directly connected, and also can indirectly be connected by intermediary, for the ordinary skill in the art, can understand as the case may be the concrete meaning of above-mentioned term.

With reference to description and accompanying drawing below, these and other aspects of embodiments of the invention will be known.These describe and accompanying drawing in, specifically disclose some specific implementations in embodiments of the invention, represent to implement some modes of the principle of embodiments of the invention, still should be appreciated that the scope of embodiments of the invention is not limited.On the contrary, embodiments of the invention comprise spirit and all changes within the scope of intension, modification and the equivalent that falls into additional claims.

Below with reference to Fig. 1 to Fig. 7, describe according to the offshore spilled oil detection method based on co-polarization feature power diagram of the embodiment of the present invention.

As shown in Figure 1, the offshore spilled oil detection method based on co-polarization feature power diagram that the embodiment of the present invention proposes, comprises the following steps:

Step S101: according to the scattering coherence matrix of polarimetric synthetic aperture radar datum target; calculate the co-polarization feature power diagram of each pixel; that is to say; according on the basis of measured polarimetric synthetic aperture radar (SAR) data, calculate the co-polarization feature power diagram of institute observation area;

Step S102: according to the feature power diagram of polarimetric synthetic aperture radar data, the maximum feature performance number that retains each pixel, the polarization pedestal height in calculating observation region, that is to say, according to the feature power diagram of polarization SAR data, calculate the polarization pedestal height of each pixel in polarization SAR observation area, and retain maximum polarization characteristic power;

Step S103: according to the maximum feature power diagram of observation area, calculate the feature power attenuation ratio of each pixel of observation area under identical incident angle, that is to say, according to the maximum feature power diagram obtaining, the back scattering attenuation ratio in calculating observation marine site;

Step S104: utilize parzen window function, estimate respectively the extra large clutter statistical distribution of polarization pedestal height and feature power attenuation ratio;

Step S105: false alarm rate is set, according to false alarm rate, the extra large clutter statistical distribution of polarization pedestal height and the extra large clutter statistical distribution of feature power attenuation ratio, calculate respectively the detection threshold of polarization pedestal height and the detection threshold of feature power attenuation ratio, by polarization pedestal height and feature power attenuation than comparing to obtain oil spilling testing result with corresponding detection threshold respectively.

According to the offshore spilled oil detection method of the embodiment of the present invention, attenuation characteristic and the polarization characteristic of having applied oil film detect simultaneously, and these two characteristic parameters can obtain from the co-polarization feature power diagram of target simultaneously, do not need extra calculated amount; Utilize the attenuation characteristic of maximum feature power calculation oil film, reduced the impact of SAR data speckle noises, improved the contrast between oil film and sea simultaneously; Utilize parzen window function to carry out matching to extra large clutter statistical distribution, in the situation that sample number is abundant, the probability density that window method is estimated can converge to real probability distribution; By two test statistics (h _pand d _r) testing result combine, therefore can greatly improve the correct detection probability of oil film, can to offshore spilled oil situation, detect timely and effectively.

In one embodiment of the invention, the feature power diagram of polarization SAR data can calculate under co-polarization antenna receiving-sending mode and any mode of cross polarised antenna transmitting-receiving mode.These two kinds of transmitting-receiving modes act on identical when the received energy of evaluating objects Optimal polarization state, therefore no matter adopt which kind of transmitting-receiving mode to analyze clarification of objective power, and effect is the same.

In another embodiment of the present invention, as shown in Figure 2, describe particularly the offshore spilled oil detection method based on co-polarization feature power diagram, and utilize this offshore spilled oil detection method to verify test zone, oil film and oil film analog are distinguished.Selected test zone as shown in Fig. 3 a-Fig. 3 c, experiment oleyl alcohol (OLA) full polarimetric SAR data in the experiment oil film full polarimetric SAR data in three data are respectively 19-3 oil field region, Peng Lai, RADARSAT2C wave band Bohai Sea Gulf on the 19th August in 2011 full polarimetric SAR data, SIR-C/X-SAR C-band on October 8th, 1994 and SIR-C/X-SAR C-band on April 15th, 1994.Three groups of space pixel sizes corresponding to data are respectively 4.7m * 5.5m, 12.5m * 12.5m and 12.5m * 12.5m.Wherein, the dark areas in Fig. 3 a data and Fig. 3 b data is slick, and dark areas in Fig. 3 c data is oil film analog.

Step S201, calculates co-polarization feature power diagram according to polarization SAR data.

According to above-mentioned offshore spilled oil detection method, polarization SAR data represent with the form of the relevant matrix T of scattering conventionally:

$T=\frac{1}{N}{\stackrel{N}{\mathrm{\&Sigma;}}\stackrel{\&RightArrow;}{k}}_p\&times;{\stackrel{\&RightArrow;}{k}}_P^H=\left[\begin{array}{ccc}{2A}_0& C-\mathrm{iD}& H+\mathrm{iG}\\ C+\mathrm{<figure-callout\; id="0"\; label="iD\; B"\; filenames="HDA0000143526660000052.png"\; state="\{\{state\}\}">iD</figure-callout>}& B_{}{}_0+B& E+\mathrm{iF}\\ H-\mathrm{iG}& E-\mathrm{iF}& B_0-B\end{array}\right],---\left(1\right)$

${\stackrel{\&RightArrow;}{k}}_p={\left[\frac{1}{\sqrt{2}}(S_{\mathrm{HH}}+S_{\mathrm{VV}})\frac{1}{\sqrt{2}}(S_{\mathrm{HH}}-S_{\mathrm{VV}})\sqrt{2}{S}_{\mathrm{HV}}\right]}^T.$

Wherein, subscript T represents transposition, the Scattering of Vector that represents target, S _ijthe scattering coefficient of target while being illustrated in linear polarization mode, A ₀, B ₀, B, C, D, E, H, G, F be Huynen parameter, i represents horizontal polarization H, j represents vertical polarization V.Kennaugh matrix K is the another kind of form of expression of polarization SAR data, and itself and the scattering relation between matrix T that is concerned with can be expressed as,

$K=\left[\begin{array}{cccc}{K}_{11}& K_{12}& K_{13}& K_{14}\\ K_{12}& K_{22}& K_{23}& K_{24}\\ K_{13}& K_{23}& K_{33}& K_{34}\\ K_{14}& K_{24}& K_{34}& K_{44}\end{array}\right]=\left[\begin{array}{cccc}{A}_0+B_0& C& H& \mathrm{<figure-callout\; id="0"\; label="F\; C\; A"\; filenames="HDA0000143526660000052.png"\; state="\{\{state\}\}">F</figure-callout>}& \\ C& A_{}{}_0+B& E& G\\ H& E& A_0-B& D\\ F& G& D& B_0-{\mathrm{<figure-callout\; id="0"\; label="A"\; filenames="HDA0000143526660000052.png"\; state="\{\{state\}\}">A</figure-callout>}}_0\end{array}\right].---\left(2\right)$

Further, in one embodiment of the invention, when antenna receives target back scattering energy in co-polarization mode, the co-polarization feature power of target can be expressed as:

$P(\mathrm{\&epsiv;},\mathrm{\&tau;})=\frac{2\mathrm{\&pi;}}{k^2}h{(\mathrm{\&epsiv;},\mathrm{\&tau;})}^T\&times;K\&times;h(\mathrm{\&epsiv;},\mathrm{\&tau;})---\left(3\right)$

Wherein, k represents the wave number of emitting electromagnetic wave, and K is Kennaugh matrix, and h (ε, τ) represents the polarized state of dual-mode antenna, and

h(ε,τ)=[1 cos2τcos2ε sin2τcos2ε sin2ε] ^T （4）

Wherein, ε ∈ [π/4, π/4], the ellipticity of expression polarization of ele ellipse, τ ∈ [pi/2, pi/2], the polarization tilt angle of expression polarization of ele ellipse.In an example of the present invention, 0.1rad and 0.2rad are got respectively in the interval of ε and τ.

Step S202, calculates polarization pedestal height h _p, then perform step S204.

Step S203, d is compared in calculated characteristics power attenuation _r, then perform step S205.

It should be noted that, step S202 and step S203 can executed in parallel.

Below the implementation of step S202 and step S203 is described.

According to (3) formula, can obtain the co-polarization feature power diagram of each pixel, retain maximum feature power and minimal characteristic power in co-polarization feature power diagram, can calculate needed two oil spilling test statistics: h _pand d _r.

$h_p=\frac{\min \left(P(\mathrm{\&epsiv;},\mathrm{\&tau;})\right)}{\max \left(P(\mathrm{\&epsiv;},\mathrm{\&tau;})\right)};$ $d_r=\frac{P_{\max }{(\mathrm{\&epsiv;},\mathrm{\&tau;})}_{\mathrm{sea}}}{P_{\max }(\mathrm{\&epsiv;},\mathrm{\&tau;})}---\left(5\right)$

Wherein, h _pfor polarization pedestal height, max (P (ε, τ)) represents the maximal value of co-polarization feature power, and min (P (ε, τ)) represents the minimum value of co-polarization feature power; d _rfor the feature power attenuation ratio of sea echo, P _max(ε, τ) _seafor the maximum polarization characteristic power in Wu Mo overlay area, sea, P _max(ε, τ)=max (P (ε, τ)).。

Only the attenuation characteristic according to target detects and can not distinguish offshore spilled oil and oil film analog (for example oleyl alcohol) oil film, and only according to the polarization characteristic of target, offshore spilled oil is detected and can not correctly film area be extracted.In one embodiment of the invention, choose respectively the targets such as sea in Fig. 3 a-Fig. 3 c, naval vessel, oil film, oleyl alcohol, naval vessel secondary lobe, investigate them at h _pand d _ron two-dimensional distribution, as shown in Figure 4.As can be seen from Figure 4, utilize separately polarization pedestal height h _por attenuation ratio d _rbe difficult to by above-mentioned different target area separately, but it is passable to utilize these two features to carry out joint-detection.

Step S204, utilizes the statistical distribution of parzen window function matching sea clutter, obtains h _pprobability density distribution, then perform step S206.

Step S205, utilizes the statistical distribution of parzen window function matching sea clutter, obtains d _rprobability density distribution, then perform step S207.

It should be noted that, step S204 and step S205 can executed in parallel.

Below the implementation of step S204 and step S205 is described.

In step S202 and step S203, obtain above-mentioned two oil spilling detection statistic h _pand d _rafterwards, utilize the statistical distribution of parzen window function matching sea clutter.If the sample set of extra large clutter is X={x ₁, x ₂..., x _m, due to the test statistics h obtaining _pand d _rbe on the occasion of, so choose the kernel function K (u) of lognormal distribution, clutter is carried out to matching.

$K\left(u\right)=\frac{1}{\mathrm{<figure-callout\; id="2"\; label="u"\; filenames="HDA0000143526660000042.png,HDA0000143526660000052.png"\; state="\{\{state\}\}">u</figure-callout>}\sqrt{2\mathrm{\&pi;}}}{e}^{-\frac{{\left(1\mathrm{nu}\right)}^2}{2}}---\left(6\right)$

The probability density distribution of parzen window matching is:

Wherein, h is the width of kernel function, and M is number of samples.Choose the data instance of Fig. 3 b, estimate two test statistics h _pand d _rstatistical distribution, result is as shown in Fig. 5 a and Fig. 5 b.

The probability density distribution obtaining according to formula (7)

, and determine detection threshold Z according to given false alarm rate _t.If false alarm rate is P _fa, detection threshold is determined according to following formula.

$p_{\mathrm{fa}}={\&Integral;}_{Z_T}^{\&infin;}{\hat{p}}_X\left(x\right)\mathrm{dx}---\left(8\right)$

The polarize detection threshold of pedestal height,

$p_{\mathrm{fa}}={\&Integral;}_{Z_{{\mathrm{Th}}_p}}^{\&infin;}{\hat{p}}_{h_p}\left(h_p\right)\mathrm{dx}---\left(9\right)$

Wherein, p _fafor false alarm rate,

for the extra large clutter statistical distribution of polarization pedestal height, detection threshold for polarization pedestal height.

According to false alarm rate, utilize the detection threshold of following formula calculated characteristics power attenuation ratio,

$p_{\mathrm{fa}}={\&Integral;}_{Z_{{\mathrm{Td}}_r}}^{\&infin;}{\hat{p}}_{d_r}\left(d_r\right)\mathrm{dx}---\left(10\right)$

Wherein, p _fafor false alarm rate,

for the extra large clutter statistical distribution of feature power attenuation ratio, detection threshold for feature power attenuation ratio.

Because false alarm rate is generally lower, for example 10 ^-3..., 10 ^-5.In an example of the present invention, given false alarm rate is p _fa=10 ^-3.

Step S206, obtains polarization pedestal height PRELIMINARY RESULTS res ₁.

Step S207, obtains feature power attenuation than PRELIMINARY RESULTS res ₂.

That is to say, according to the extra large clutter statistical distribution of described false alarm rate and described polarization pedestal height, obtain the testing result of described polarization pedestal height; According to the extra large clutter statistical distribution of described false alarm rate and described feature power attenuation ratio, obtain the testing result of described feature power attenuation ratio.Say, pedestal height h will polarize respectively _pcompare d with feature power attenuation _rcompare with corresponding threshold value, obtain polarizing pedestal height and feature power attenuation, than corresponding testing result, are respectively: ${\mathrm{<figure-callout\; id="1"\; label="res"\; filenames="HDA0000143526660000052.png"\; state="\{\{state\}\}">res</figure-callout>}}_1=h_p>Z_{{\mathrm{Th}}_p},$ ${\mathrm{<figure-callout\; id="2"\; label="res"\; filenames="HDA0000143526660000042.png,HDA0000143526660000052.png"\; state="\{\{state\}\}">res</figure-callout>}}_2=d_r>Z_{{\mathrm{Td}}_r}.$

Step S208, carries out AND-operation by two testing results, and then can obtain final offshore spilled oil testing result.Speech, merges the testing result of the testing result of described polarization pedestal height and described feature power attenuation ratio with obtain described oil spilling testing result.

res=res ₁&res ₂ （11）

Three width polarization SAR data in Fig. 3 a-Fig. 3 c are carried out to slick detection, obtain final oil spilling testing result as shown in Fig. 6 a-Fig. 6 c, three result orders in figure are corresponding to the data in Fig. 3 a-Fig. 3 c.The testing result of Fig. 6 c does not almost have continuous oil film region, therefore it can be sentenced and removes from Oil spill detection.Therefore, from result, can find out, utilize the joint-detection of attenuation characteristic and polarization characteristic oil film and oil film analog can be distinguished.

In one embodiment of the invention, the offshore spilled oil detection method based on co-polarization feature power diagram also comprises the steps:

Step S209, to oil spilling testing result, res carries out largest connected regional analysis.

Step S210, calculates offshore spilled oil area according to largest connected regional analysis result.Wherein, connected domain refers to the set that the pixel that has same level in image and communicate with each other forms, and largest connected domain analysis refers to the region in all connected set with maximum pixel number.If the size of each pixel of SAR slant-range image is ρ _a* ρ _r, the oblique distance estimated result of offshore spilled oil area is,

S=connected(res)×ρ _a×ρ _r （12）

Wherein, connected (res) represents the largest connected domain analysis process to bianry image res.If desired estimate practically apart from the oil spill area of image, need to utilize following formula,

S=connected(res)×ρ _a×ρ _gr （13）

Wherein, ρ _gr=ρ _r/ sin (θ _i), θ _ifor the incident angle of radar beam, ρ _afor the size of orientation to pixel cell, ρ _rfor the size of distance to pixel cell, connected (res) represents the largest connected domain analysis process to the oil spilling testing result of bianry image.

In an example of the present invention, to select the regional connectivity of 8 neighborhoods, and think when the pixel number of a connected domain is greater than 16, this connected region is an oil film target.According to formula (12), can obtain the slant-range image estimated result of offshore spilled oil area, wherein, connected (res) refers to that pixel number is greater than the set of 16 connected domain.Here choosing 16 main cause is, if in orientation to distance to form target area to be selected by 4 pixels simultaneously, this region is likely the interference of oil film target rather than background clutter so.

For example, the offshore spilled oil testing result of three width polarization SAR data in Fig. 6 a-Fig. 6 c is done to largest connected domain analysis, the number that obtains oil film pixel in three width data is respectively 5947,19555 and 0.The analysis result in largest connected territory is as shown in Fig. 7 a-Fig. 7 b, because Fig. 6 c does not detect oil film region, so last piece image is saved herein.In conjunction with the space size of adopted SAR data pixels point, obtain the estimated value of corresponding slick area on SAR slant-range image and be respectively 0.15373km ², 3.0555km ²and 0km ².

Area based on largest connected domain analysis oil film can significantly reduce the impact that false-alarm is estimated film area, and can keep the integrality to oil film Region Segmentation, correctly slick area is estimated.

In the description of this instructions, the description of reference term " embodiment ", " some embodiment ", " example ", " concrete example " or " some examples " etc. means to be contained at least one embodiment of the present invention or example in conjunction with specific features, structure, material or the feature of this embodiment or example description.In this manual, the schematic statement of above-mentioned term is not necessarily referred to identical embodiment or example.And the specific features of description, structure, material or feature can be with suitable mode combinations in any one or more embodiment or example.

Although illustrated and described embodiments of the invention, for the ordinary skill in the art, be appreciated that without departing from the principles and spirit of the present invention and can carry out multiple variation, modification, replacement and modification to these embodiment, scope of the present invention is by claims and be equal to and limit.

Claims (11)

1. the offshore spilled oil detection method based on co-polarization feature power diagram, is characterized in that, comprises the steps:

According to the scattering coherence matrix of polarimetric synthetic aperture radar datum target, calculate the co-polarization feature power diagram of each pixel;

According to the feature power diagram of described polarimetric synthetic aperture radar data, retain the maximum feature performance number of described each pixel, the polarization pedestal height in calculating observation region;

According to the maximum feature power diagram of described observation area, calculate the feature power attenuation ratio of each pixel of described observation area under identical incident angle;

Utilize window function to estimate respectively the extra large clutter statistical distribution of described polarization pedestal height and the extra large clutter statistical distribution of described feature power attenuation ratio; And

False alarm rate is set, according to the extra large clutter statistical distribution of the extra large clutter statistical distribution of described false alarm rate, described polarization pedestal height and described feature power attenuation ratio, calculate respectively the detection threshold of described polarization pedestal height and the detection threshold of described feature power attenuation ratio, by described polarization pedestal height and described feature power attenuation than comparing to obtain oil spilling testing result with corresponding detection threshold respectively.

2. offshore spilled oil detection method as claimed in claim 1, is characterized in that, the feature power diagram of described polarimetric synthetic aperture radar data obtains under co-polarization antenna receiving-sending mode.

3. offshore spilled oil detection method as claimed in claim 1, is characterized in that, the co-polarization feature power representing in the feature power diagram of described polarimetric synthetic aperture radar data is:

$P(\mathrm{\&epsiv;},\mathrm{\&tau;})=\frac{2\mathrm{\&pi;}}{k^2}h{(\mathrm{\&epsiv;},\mathrm{\&tau;})}^T\mathrm{Kg}(\mathrm{\&epsiv;},\mathrm{\&tau;})$

Wherein, k is electromagnetic wave number, and K is Kennaugh matrix, the in the situation that of co-polarization antenna receiving-sending,

H (ε, τ)=g (ε, τ)=[1 cos2 τ cos2 ε sin2 τ cos2 ε sin2 ε] ^t, h (ε, τ) ^tfor the transposition of h (ε, τ), ε ∈ [π/4, π/4], the ellipticity of expression polarization of ele ellipse, τ ∈ [pi/2, pi/2], the polarization tilt angle of expression polarization of ele ellipse.

4. offshore spilled oil detection method as claimed in claim 1, is characterized in that, described polarization pedestal height is:

$h_p=\frac{\min \left(P(\mathrm{\&epsiv;},\mathrm{\&tau;})\right)}{\max \left(P(\mathrm{\&epsiv;},\mathrm{\&tau;})\right)}$

Wherein, h _pfor described polarization pedestal height, max (P (ε, τ)) represents the maximal value of co-polarization feature power, and min (P (ε, τ)) represents the minimum value of co-polarization feature power.

5. offshore spilled oil detection method as claimed in claim 1, is characterized in that, the feature power attenuation ratio of described sea echo is:

$d_r=\frac{P_{\max }{(\mathrm{\&epsiv;},\mathrm{\&tau;})}_{\mathrm{sea}}}{P_{\max }(\mathrm{\&epsiv;},\mathrm{\&tau;})}$

Wherein, d _rfor the feature power attenuation ratio of described sea echo, P _max(ε, τ)=max (P (ε, τ)), max () represents to ask the maximal value of array, P _max(ε, τ) _searepresent the maximum polarization characteristic power in Wu Mo overlay area, sea.

6. offshore spilled oil detection method as claimed in claim 1, is characterized in that, utilizes parzen window function to estimate respectively the extra large clutter statistical distribution of described polarization pedestal height and the extra large clutter statistical distribution of described feature power attenuation ratio.

7. offshore spilled oil detection method as claimed in claim 6, is characterized in that, the kernel function of described parzen window function adopts lognormal distribution.

8. offshore spilled oil detection method as claimed in claim 1, is characterized in that, according to described false alarm rate, utilizes following formula to calculate the detection threshold of described polarization pedestal height,

$p_{\mathrm{fa}}={\&Integral;}_{Z_{{\mathrm{Th}}_p}}^{\&infin;}{\hat{p}}_{h_p}\left(h_p\right){\mathrm{dh}}_p,$

Wherein, p _fafor described false alarm rate,

for the extra large clutter statistical distribution of described polarization pedestal height,

detection threshold for described polarization pedestal height.

9. offshore spilled oil detection method as claimed in claim 1, is characterized in that, according to described false alarm rate, utilizes following formula to calculate the detection threshold of described feature power attenuation ratio,

$p_{\mathrm{fa}}={\&Integral;}_{Z_{{\mathrm{Td}}_r}}^{\&infin;}{\hat{p}}_{d_r}\left(d_r\right){\mathrm{dd}}_r,$

Wherein, p _fafor described false alarm rate, for the extra large clutter statistical distribution of described feature power attenuation ratio,

detection threshold for described feature power attenuation ratio.

10. offshore spilled oil detection method as claimed in any one of claims 1-9 wherein, it is characterized in that, described described polarization pedestal height and described feature power attenuation are compared with corresponding detection threshold than respectively, obtain testing result separately, and described testing result is merged, obtain the net result that oil spilling detects, comprise the steps:, according to the detection threshold of described polarization pedestal height and polarization pedestal height, to obtain the testing result of described polarization pedestal height; According to the detection threshold of described feature power attenuation ratio and feature power attenuation ratio, obtain the testing result of described feature power attenuation ratio; And the testing result of the testing result of described polarization pedestal height and described feature power attenuation ratio is merged to obtain the net result that described oil spilling detects.

11. offshore spilled oil detection methods as claimed in claim 10, is characterized in that, also comprise the steps: described oil spilling testing result to carry out largest connected regional analysis to obtain offshore spilled oil area, and wherein, described offshore spilled oil area is:

S=connected(res)×ρ _a×ρ _gr

Wherein, ρ _gr=ρ _r/ sin (θ _i), θ _ifor the incident angle of radar beam, ρ _afor the size of orientation to pixel cell, ρ _rfor the size of distance to pixel cell, connected (res) represents the largest connected domain analysis process to the described oil spilling testing result of bianry image.

Images