CN111025291A - Ocean oil spill detection method based on new characteristics of fully-polarized SAR - Google Patents
Ocean oil spill detection method based on new characteristics of fully-polarized SAR Download PDFInfo
- Publication number
- CN111025291A CN111025291A CN201911076780.7A CN201911076780A CN111025291A CN 111025291 A CN111025291 A CN 111025291A CN 201911076780 A CN201911076780 A CN 201911076780A CN 111025291 A CN111025291 A CN 111025291A
- Authority
- CN
- China
- Prior art keywords
- scattering
- polarization
- matrix
- polarized
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
- G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
- G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
- G01S13/9021—SAR image post-processing techniques
- G01S13/9027—Pattern recognition for feature extraction
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
- G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
- G01S13/904—SAR modes
- G01S13/9076—Polarimetric features in SAR
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/024—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using polarisation effects
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S7/411—Identification of targets based on measurements of radar reflectivity
- G01S7/412—Identification of targets based on measurements of radar reflectivity based on a comparison between measured values and known or stored values
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
Landscapes
- Engineering & Computer Science (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Electromagnetism (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Astronomy & Astrophysics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a marine oil spill detection method based on new characteristics of a fully-polarized SAR, which comprises the following steps: extracting a polarization coherent matrix; eigenvalue and eigenvector decomposition; constructing new characteristics of the kini coefficient; dividing a threshold value; the method of the invention provides a new polarization characteristic based on the eigenvalue decomposition theory, which is called polarization damping coefficient; the characteristic can reflect the polarization state among different targets in the set, can describe the purity of different scattering types in statistical sense, and indicates that the scattering mechanism types in the corresponding set are consistent when the purity is 0; the new characteristics obtain different scattering mechanism types through a target decomposition theory, oil spill information is extracted through calculating the impurity degree of the dominant scattering mechanism of the pixel in the statistical sense, the characteristics are extracted through the method for the fully polarized SAR oil spill image, the problem that a biological oil film and a mineral film cannot be distinguished is solved, and the new characteristic extraction technology has the capability of inhibiting coherent speckle noise.
Description
Technical Field
The invention relates to the technical field of marine oil spill detection, in particular to a marine oil spill detection method based on new characteristics of a fully-polarized SAR.
Background
Offshore oil spill is one of the main factors causing pollution damage to the ocean environment, the ocean oil pollution caused by pipeline rupture, oil tanker collision and drilling platform explosion is a global concern, and since 1970, the oil pollution generated every year all over the world exceeds 10000 tons, so that how to effectively detect the offshore oil spill is very important. After the offshore oil spill occurs, the position, the oil spill amount and the diffusion trend of the occurrence are generally known, and in an already-put monitoring system, satellite remote sensing is one of the most important and effective means, and plays an increasingly important role in offshore oil spill monitoring and emergency response. At present, the most commonly used satellite remote sensing sensors include visible light, infrared, ultraviolet, laser fluorescence, microwave sensors and the like, wherein the sensors are widely applied by the characteristics that a Synthetic Aperture Radar (SAR) can penetrate through clouds and fog and is not influenced by day and night;
oil slick usually appears as dark spots in SAR images due to the main mechanism of SAR sea surface imaging. When an oil film exists on the sea surface, capillary waves and short gravity waves on the sea surface are continuously subjected to a damping effect, the roughness of the sea surface is reduced, and therefore radar backscatter echo signals are weakened. Generally, conventional single-polarized oil spill detection can be divided into the following three steps: (1) dark spot detection: for example, identification of a region of interest (ROI); (2) feature extraction: in order to further distinguish the composition of dark spots, it is necessary to extract various features of the image after dark spot detection, and to pick appropriate features for the subsequent classification task; (3) and (4) classification: the phenomenon of misclassification of an oil film and an oil-like film is reduced;
with the development of the polarization mode of the SAR from single polarization to dual polarization and full polarization, a new oil spill detection research based on polarized SAR data is carried out, compared with single polarization, a full polarization SAR image not only contains intensity information, but also contains phase information, so that the full polarization scattering characteristic of a sea surface target can be effectively obtained, and the geometric shape and the physical characteristic of the sea surface target can be more comprehensively reflected;
in the existing polarization characteristic extraction technology, oil films and seawater can be well distinguished, but for a full polarization SAR oil spill image, the image is not only oil films and seawater, but also comprises oil film phenomena such as a low wind speed area, a rainball, a biological oil film and the like, particularly, the distinction of the biological oil film and a mineral oil film is more and more emphasized in recent years, and the existing characteristic extraction technology cannot well distinguish the biological oil film and the mineral oil film and is easily influenced by sea coherent speckle noise, so that the invention provides a marine oil spill detection method based on a new characteristic of the full polarization SAR to overcome the defects in the prior art.
Disclosure of Invention
Aiming at the problems, the invention provides a marine oil spill detection method based on a new fully-polarized SAR characteristic, a new polarization characteristic is provided based on a characteristic value decomposition theory, the polarization characteristic is called polarization damping coefficient, the characteristic can reflect the polarization state among different targets in a set and can describe the purity of different scattering types in a statistical sense, the new characteristic obtains different scattering mechanism types through a target decomposition theory, oil spill information is extracted by calculating the impurity degree of a dominant scattering mechanism of a pixel in the statistical sense, the problem that a biological oil film and a mineral film cannot be distinguished is solved, and a new characteristic extraction technology has the capability of inhibiting coherent speckle noise.
The invention provides a marine oil spill detection method based on new characteristics of a fully-polarized SAR (synthetic aperture radar), which comprises the following steps of:
the method comprises the following steps: extracting a polarized coherence matrix comprising an original scattering matrix S and a coherence matrix T3The polarized scattering matrix S can completely describe the electromagnetic scattering characteristics of the SAR target, and is defined asThe coherence matrix T is shown in equation (1)3Generated by the outer product of the target vector k and the conjugate transpose vector of the target vector k, and a coherent matrix T3The definition is shown in formula (2);
in the formula (2), T represents the conjugation transposition and conjugation respectively, and < > represents the overall average value;
step two: eigenvalue and eigenvector decomposition. Based on the target decomposition theory, 3 multiplied by 3 Hermite average coherent matrix T3The eigenvalues and eigenvectors obtained from the calculation produce a diagonalized version of the coherence matrix, coherence matrix T3Is defined as shown in formula (3):
where Σ is a 3 × 3 non-negative real diagonal matrix, eigenvalue (λ)1≥λ2≥λ3≥0);U3=[e1e2e3]Is a 3 x 3 special unitary matrix SU (3) where e1,e2,e3Are respectively a coherence matrix T3Of the feature vector of (1), then the coherence matrix T3Redefined, the new definition is shown in formula (4);
step three: and constructing new characteristics of the kini coefficient. And (3) defining a polarization-kini coefficient based on the eigenvalue and eigenvector decomposition in the step two as shown in a formula (5):
wherein,piCorresponding to the characteristic value lambdaiThe obtained pseudo probability satisfiesp1+p2+p3=1,λ1≥λ2≥λ3The characteristic value is rotation invariant, and the polarization-based damping coefficient also has rotation invariance;
step four: and (5) carrying out threshold segmentation. Different scattering mechanism types and corresponding scattering mechanisms are obtained through a target decomposition theory, and the impure degree and the ordered degree of the different scattering mechanisms in the statistical sense are judged according to the new characteristics of the Gini coefficient to extract the oil spill information.
The further improvement lies in that: s in the formula (1)HVRepresenting horizontally polarized transmission, vertically polarized reception; sVHRepresenting vertical polarization transmission, horizontal polarization reception; sHHRepresenting horizontally polarized transmission and reception; sVVRepresenting vertically polarized transmission and reception.
The further improvement lies in that: when the reciprocity theorem is satisfied in the formula (1), SHV=SVHAnd obtaining a target scattering vector k by using a Pauli base decomposition method for the polarized scattering matrix S, wherein the definition of the polarized scattering matrix S is shown as a formula (6):
wherein the factorFor ensuring that the norm of the target vector is equal to the total scattered power.
The further improvement lies in that: diagonal elements of the scattering matrix S in the first step represent the relationship between an incident field and a scattering field in the same polarization mode, and are represented by a 'same polarization' item; the off-diagonal elements represent the incident angle and scattered field relationship for the orthogonal polarization mode, represented by the term "cross polarization".
The further improvement lies in that: solving the feature vector e in the second step1,e2,e3Then, 3 mutually independent targets are obtained for constructing a statistical model, and a coherent matrix T is obtained3Expanding the mixture into the sum of 3 mutually independent targets, wherein each target corresponds to a determined scattering mechanism and then is represented by an equivalent simple scattering matrix, and the weight of a component i of the determined scattering mechanism in the whole scattering process is represented by a characteristic value lambdaiDescription, type of scattering mechanism and normalized feature vector eiAnd (4) correlating.
The further improvement lies in that: in the formula (5) in the third step, when the characteristic value has only one non-zero value (lambda)1≠0,λ2=λ30), then the coherence matrix T3The scattering situation of the point object defined by the polarization scattering matrix S is degraded.
The further improvement lies in that: in the formula (5) in the third step, when all the characteristic values are not zero and are equal (lambda)1=λ2=λ3Not equal to 0), the coherence matrix T is obtained3It represents a completely depolarized random scattering state, and the object between these two extremes is partially polarized, when the coherence matrix T is present3Are greater than 1 and are not completely equal to each other.
The further improvement lies in that: in the fourth step, when the impurity degree is lower, the order degree of a corresponding scattering mechanism is higher, and the polarization coefficient is smaller; when the impurity degree is higher, the order degree of a corresponding scattering mechanism is lower, and the polarization coefficient is higher; when the impurity degree reaches the minimum value, the scattering mechanisms in the corresponding set are consistent in type; when the impurity reaches a maximum, it indicates that the scattering of the target in the set exhibits a random noise state.
The further improvement lies in that: from the perspective of the polarization state in the fourth step, when the value of the polarization kini coefficient is lower, the system can be regarded as weakly depolarized, the dominant scattering mechanism is regarded as a target scattering mechanism of a specific equivalent point, the feature vector corresponding to the maximum feature value is selected according to the mechanism, and other feature vectors are ignored; when the value of the polarization kiney coefficient is higher, the average scatterer in the set is in a depolarization state, and no single scattering target exists any more, so that the mixing proportion of all possible point target scattering types from the whole characteristic value distribution spectrum needs to be considered; when the polarization coefficient is further increased, the number of scattering mechanisms identified from the polarization measurement data will also gradually decrease, and when the polarization coefficient reaches a maximum value, the polarization information will be zero, and then the target scattering is completely a random noise process.
The invention has the beneficial effects that: the method provides a new polarization characteristic based on a characteristic value decomposition theory, wherein the polarization characteristic is called polarization prime coefficient, the characteristic can reflect polarization states among different targets in a set and can describe the impure degree of different scattering types in a statistical sense, when the impure degree is 0, the scattering mechanism types in the corresponding set are consistent, the new characteristic obtains the different scattering mechanism types through the target decomposition theory, and the oil spill information is extracted through calculating the impure degree of a dominant scattering mechanism of the pixel in the statistical sense; the method is used for extracting the characteristics of the fully polarized SAR oil spill image, the problem that a traditional method cannot effectively distinguish a biological oil film from a mineral film is well solved, and the new characteristic extraction technology has the capability of inhibiting speckle noise of the SAR image.
Drawings
FIG. 1 is a schematic diagram of a new polarization feature extraction process based on eigenvalue decomposition according to the present invention;
FIG. 2 is a diagram illustrating the extraction of new polarization features based on eigenvalue decomposition according to the present invention;
FIG. 3 is a schematic diagram of a decomposed feature image of a coherence matrix according to the present invention;
FIG. 4 is a schematic diagram of the polar Keyny coefficient of the new feature map extracted by the method of the present invention;
FIG. 5 is a diagram illustrating the threshold segmentation of the new characteristic polarization Boyni coefficient of the present invention;
FIG. 6 shows the polarization Boyny coefficient and span,H. P, A, CPD, Rho feature extraction results comparison diagram;
FIG. 7 shows an embodiment of the present inventionMedium polarization coefficient and span,H. The evaluation results of the probability density curves of P, A, CPD and Rho are compared with a schematic diagram;
FIG. 8 shows the polarization Boyny coefficient and span,H. The evaluation results of the probability density curves of P, A, CPD and Rho are compared with a schematic diagram;
FIG. 9 shows the polarization Boyny coefficient and span,H. The evaluation results of the probability density curves of P, A, CPD and Rho are compared with a schematic diagram;
Detailed Description
In order to further understand the present invention, the following detailed description will be made with reference to the following examples, which are only used for explaining the present invention and are not to be construed as limiting the scope of the present invention.
According to fig. 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10, the present embodiment provides a method for detecting marine oil spill based on a new characteristic of a fully-polarized SAR, including the following steps:
the method comprises the following steps: extracting a polarized coherence matrix comprising an original scattering matrix S and a coherence matrix T3The polarized scattering matrix S can completely describe the electromagnetic scattering characteristics of the SAR target and is defined as formula (1), and the coherence matrix T3Generated by the outer product of the target vector k and the conjugate transpose vector of the target vector k, and a coherent matrix T3The definition is shown in formula (2);
in the formula (1), SHVRepresenting horizontally polarized transmission, vertically polarized reception; sVHRepresenting vertical polarization transmission, horizontal polarization reception; sHHRepresenting horizontally polarized transmission and reception; sVVRepresenting vertically polarized transmission and reception;
in the formula (2), T represents the conjugation transposition and conjugation respectively, and < > represents the overall average value;
when the reciprocity theorem is satisfied, SHV=SVHAnd obtaining a target scattering vector k by using a Pauli base decomposition method for the polarized scattering matrix S, wherein the definition of the polarized scattering matrix S is shown as a formula (6):
wherein the factorThe system is used for ensuring that the norm of the target vector is equal to the total scattering power;
diagonal elements of the scattering matrix S represent the relationship between an incident field and a scattering field in the same polarization mode and are represented by a 'homopolarization' item; the off-diagonal element represents the relationship between the incident angle and the scattered field in an orthogonal polarization mode and is represented by a cross polarization term;
step two: decomposing the eigenvalue and eigenvector, and based on the target decomposition theory, dividing the 3X 3 Hermite flat phase coherence matrix T3The eigenvalues and eigenvectors obtained from the calculation produce a diagonalized version of the coherence matrix, coherence matrix T3As shown in equation (3):
where Σ is 3 × 3 negationNegative real diagonal array, U3Is a 3 x 3 special unitary matrix SU (3) if λ1,λ2,λ3Respectively representing a coherence matrix T3Is not a negative characteristic value, and λ1≥λ2≥λ3≥0;e1,e2,e3Respectively representing a coherence matrix T3The eigenvector of (1), the coherence matrix T3Further development is possible as shown in equation (4):
solving for the feature vector e1,e2,e3Then, 3 mutually independent targets are obtained for constructing a statistical model, and a coherent matrix T is obtained3Expanding the mixture into the sum of 3 mutually independent targets, wherein each target corresponds to a determined scattering mechanism and then is represented by an equivalent simple scattering matrix, and the weight of a component i of the determined scattering mechanism in the whole scattering process is represented by a characteristic value lambdaiDescription, type of scattering mechanism and normalized feature vector eiIn this embodiment, the decomposed picture is shown in fig. 3;
step three: and constructing new characteristics of the kini coefficient. And (3) defining a polarization-kini coefficient based on the eigenvalue and eigenvector decomposition in the step two as shown in a formula (5):
wherein p isiCorresponding to the characteristic value lambdaiThe obtained pseudo probability satisfiesAnd p is1+p2+p3=1,λ1≥λ2≥λ3The characteristic value is rotation invariant, and the polarization-based damping coefficient also has rotation invariance;
when the characteristic value has only one non-zero value (lambda)1≠0,λ2=λ30), then the coherence matrixT3Degenerates to the scattering behavior of the point object defined by the polarization scattering matrix S; when all the characteristic values are not zero and are equal (lambda)1=λ2=λ3Not equal to 0), the coherence matrix T is obtained3It represents a completely depolarized random scattering state, and the object between these two extremes is partially polarized, when the coherence matrix T is present3Are greater than 1 and are not completely equal to each other;
step four: and (4) threshold segmentation, obtaining different scattering mechanism types and corresponding scattering mechanisms through a target decomposition theory, and judging the impure degree and the ordered degree of the different scattering mechanisms in the statistical sense according to the new characteristics of the Gini coefficient to extract the oil spill information.
When the impurity degree is lower, the order degree of a corresponding scattering mechanism is higher, and the polarization coefficient is smaller; when the impurity degree is higher, the order degree of a corresponding scattering mechanism is lower, and the polarization coefficient is higher; when the impurity degree reaches the minimum value, the scattering mechanisms in the corresponding set are consistent in type; when the impurity degree reaches the maximum value, indicating that the scattering of the targets in the set presents a random noise state;
from the perspective of the polarization state, when the value of the polarization kini coefficient is low, the system can be regarded as weakly depolarized, the dominant scattering mechanism is regarded as a target scattering mechanism of a specific equivalent point, and the eigenvector corresponding to the maximum eigenvalue is selected according to the mechanism, while other eigenvectors are ignored; when the value of the polarization kiney coefficient is higher, the average scatterer in the set is in a depolarization state, and no single scattering target exists any more, so that the mixing proportion of all possible point target scattering types from the whole characteristic value distribution spectrum needs to be considered; when the polarization coefficient is further increased, the number of scattering mechanisms identified from the polarization measurement data will also gradually decrease, and when the polarization coefficient reaches a maximum value, the polarization information will be zero, and then the target scattering is completely a random noise process.
In the image of this embodiment, bragg scattering is dominant in clean seawater, and the scattering mechanism is single, and at this time, the coherent matrix T is3Is one greaterSo that the polarization-damping coefficient is small; the scattering mechanisms of mineral oils are complex and the number of scattering mechanisms that can be identified from the measurement data is small, when the coherence matrix T is present3Three approximately equal characteristic values exist, so that the polarization coefficient of the floating oil is larger than that of the clean seawater; besides the dominant scattering, the bio-oil film also comprises other complex target scattering mechanisms, such as mirror scattering and the like, and Bragg scattering is not dominant, so that the coherence matrix T is3There are two to three characteristic values that are not equal to each other, so the polarization coefficient should be between that of mineral oil and seawater. In general, the order of the big to the small of the Giny coefficient is mineral oil>Biological oil film>Seawater, the polarization damping coefficient characteristic extracted by the method of the invention is shown in fig. 4, mineral oil, a biological oil film and seawater can be clearly seen to show completely different distribution intervals on the polarization damping coefficient characteristic image, the polarization damping coefficient of the mineral oil is between 0.45 and 0.68, the polarization damping coefficient of the biological oil film is between 0.23 and 0.32, the polarization damping coefficient of the seawater is near 0.1, the polarization damping coefficient characteristic image is subjected to threshold segmentation by using an OTSU algorithm, the segmented image is shown in fig. 5, the mineral oil coverage area is well segmented, and the biological oil film and the seawater are combined into one class.
Using the process of the present invention and H, A,Comparing the polarization feature extraction results obtained by 7 other classical methods such as span, P, CPD and Rho, and qualitatively and quantitatively evaluating the feature extraction results by using a probability density curve, image contrast and local standard deviation, wherein FIGS. 7, 8 and 9 are qualitative evaluation results of the probability density curve, Table 1 is quantitative evaluation results of the image contrast, and Table 2 is quantitative evaluation results of the local standard deviation.
TABLE 1 evaluation of image contrast of crude oil and vegetable oil under different feature extraction methods
TABLE 2 local standard deviation evaluation of crude and vegetable oils under different feature extraction methods
The probability density curve can show the capability of each feature for distinguishing a bio-oil film (simulated by vegetable oil in the embodiment) from a mineral oil film, and in the probability density curves with different features, the longer the distance between the peaks of the curves of the bio-oil film and the mineral film is, the better the capability of the feature for distinguishing the bio-oil film from the mineral film is; the image contrast can measure the capability of the polarization characteristic to distinguish the information of the bio-oil film and the mineral oil, and if the contrast between the bio-oil film and the mineral film of a certain characteristic is higher, the characteristic is more beneficial to extracting the mineral oil film on the sea surface, namely the characteristic is more suitable for oil spill detection; the local standard deviation is used for measuring the speckle noise suppression capability of the extraction method of different polarization characteristics on the complete polarization SAR image, if the local standard deviation value is smaller, the noise of the image is smaller, and the characteristic is proved to have stronger noise suppression performance.
From qualitative angle evaluation, it can be visually seen in fig. 6 that the feature extraction result of the method is visually superior to most existing methods, after the polarization kiney coefficient is extracted, the bio-oil film and the mineral oil film have obvious difference, and from the probability density curves of fig. 7, 8, 9 and 10, it can be seen that the polarization kiney coefficients, H, span and P all show better separability;
from the quantitative evaluation, it can be seen from tables 1 and 2 that the image contrast and the local standard deviation of the polarization-damping coefficient exceed all the other features, which indicates that the polarization-damping coefficient can more effectively extract the oil spill information for other features, and therefore, the method is better than other polarization-feature extraction methods in oil spill and seawater separability and image noise suppression.
The method provides a new polarization characteristic based on a characteristic value decomposition theory, wherein the polarization characteristic is called polarization prime coefficient, the characteristic can reflect polarization states among different targets in a set and can describe the impure degree of different scattering types in a statistical sense, when the impure degree is 0, the scattering mechanism types in the corresponding set are consistent, the new characteristic obtains the different scattering mechanism types through the target decomposition theory, and the oil spill information is extracted through calculating the impure degree of a dominant scattering mechanism of the pixel in the statistical sense; the method is used for extracting the characteristics of the fully polarized SAR oil spill image, the problem that a traditional method cannot effectively distinguish a biological oil film from a mineral film is well solved, and the new characteristic extraction technology has the capability of inhibiting speckle noise of the SAR image.
The foregoing illustrates and describes the principles, general features, and advantages of the present invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (9)
1. A marine oil spill detection method based on new characteristics of a fully-polarized SAR is characterized by comprising the following steps:
the method comprises the following steps: extracting a polarized coherence matrix comprising an original scattering matrix S and a coherence matrix T3The extraction of (1). The polarized scattering matrix S can completely describe the electromagnetic scattering characteristics of the SAR target and is defined as the formula (1), and the coherence matrix T3Generated by the outer product of the target vector k and the conjugate transpose vector of the target vector k, and a coherent matrix T3The definition is shown in formula (2);
in the formula (2), T represents the conjugation transposition and conjugation respectively, and < > represents the overall average value;
step two: eigenvalue and eigenvector decomposition. Based on the target decomposition theory, 3 multiplied by 3 Hermite average coherent matrix T3The eigenvalues and eigenvectors obtained from the calculation produce a diagonalized version of the coherence matrix, coherence matrix T3Is defined as shown in formula (3):
where Σ is a 3 × 3 non-negative real diagonal matrix, U3Is a 3 x 3 special unitary matrix SU (3) if λ1,λ2,λ3Respectively representing a coherence matrix T3Is not a negative characteristic value, and λ1≥λ2≥λ3≥0;e1,e2,e3Respectively representing a coherence matrix T3The eigenvector of (1), the coherence matrix T3Further development is possible as shown in equation (4):
step three: and constructing new characteristics of the kini coefficient. And (3) defining a polarization-kini coefficient based on the eigenvalue and eigenvector decomposition in the step two as shown in a formula (5):
wherein p isiCorresponding to the characteristic value lambdaiThe obtained pseudo probability satisfiesp1+p2+p3=1,λ1≥λ2≥λ3The characteristic value is rotation invariant, and the polarization-based damping coefficient also has rotation invariance;
step four: and (5) carrying out threshold segmentation. Different scattering mechanism types and corresponding scattering mechanisms are obtained through a target decomposition theory, and the impure degree and the ordered degree of the different scattering mechanisms in the statistical sense are judged according to the new characteristics of the Gini coefficient to extract the oil spill information.
2. The marine oil spill detection method based on the new fully-polarized SAR characteristics according to claim 1, characterized in that: s in the formula (1)HVRepresenting horizontally polarized transmission, vertically polarized reception; sVHRepresenting vertical polarization transmission, horizontal polarization reception; sHHRepresenting horizontally polarized transmission and reception; sVVRepresenting vertically polarized transmission and reception.
3. The marine oil spill detection method based on the new fully-polarized SAR characteristics as claimed in claim 2, characterized in that: when the reciprocity theorem is satisfied in the formula (1), SHV=SVHAnd obtaining a target scattering vector k by using a Pauli base decomposition method for the polarized scattering matrix S, wherein the definition of the polarized scattering matrix S is shown as a formula (6):
4. The marine oil spill detection method based on the new fully-polarized SAR characteristics according to claim 1, characterized in that: diagonal elements of the scattering matrix S in the first step represent the relationship between an incident field and a scattering field in the same polarization mode, and are represented by a 'same polarization' item; the off-diagonal elements represent the incident angle and scattered field relationship for the orthogonal polarization mode, represented by the term "cross polarization".
5. The marine oil spill detection method based on the new fully-polarized SAR characteristics according to claim 1, characterized in that: solving the features in the second stepVector e1,e2,e3Then, 3 mutually independent targets are obtained for constructing a statistical model, and a coherent matrix T is obtained3Expanding the mixture into the sum of 3 mutually independent targets, wherein each target corresponds to a determined scattering mechanism and then is represented by an equivalent simple scattering matrix, and the weight of a component i of the determined scattering mechanism in the whole scattering process is represented by a characteristic value lambdaiDescription, type of scattering mechanism and normalized feature vector eiAnd (4) correlating.
6. The marine oil spill detection method based on the new fully-polarized SAR characteristics according to claim 1, characterized in that: in the formula (5) in the third step, when the characteristic value has only one non-zero value (lambda)1≠0,λ2=λ30), then the coherence matrix T3The scattering situation of the point object defined by the polarization scattering matrix S is degraded.
7. The marine oil spill detection method based on the new fully-polarized SAR characteristics according to claim 1, characterized in that: in the formula (5) in the third step, when all the characteristic values are not zero and are equal (lambda)1=λ2=λ3Not equal to 0), the coherence matrix T is obtained3It represents a completely depolarized random scattering state, and the object between these two extremes is partially polarized, when the coherence matrix T is present3Are greater than 1 and are not completely equal to each other.
8. The marine oil spill detection method based on the new fully-polarized SAR characteristics according to claim 1, characterized in that: in the fourth step, when the impurity degree is lower, the order degree of a corresponding scattering mechanism is higher, and the polarization coefficient is smaller; when the impurity degree is higher, the order degree of a corresponding scattering mechanism is lower, and the polarization coefficient is higher; when the impurity degree reaches the minimum value, the scattering mechanisms in the corresponding set are consistent in type; when the impurity reaches a maximum, it indicates that the scattering of the target in the set exhibits a random noise state.
9. The marine oil spill detection method based on the new fully-polarized SAR characteristics according to claim 1, characterized in that: from the perspective of the polarization state in the fourth step, when the value of the polarization kini coefficient is lower, the system can be regarded as weakly depolarized, the dominant scattering mechanism is regarded as a target scattering mechanism of a specific equivalent point, the feature vector corresponding to the maximum feature value is selected according to the mechanism, and other feature vectors are ignored; when the value of the polarization kiney coefficient is higher, the average scatterer in the set is in a depolarization state, and no single scattering target exists any more, so that the mixing proportion of all possible point target scattering types from the whole characteristic value distribution spectrum needs to be considered; when the polarization coefficient is further increased, the number of scattering mechanisms identified from the polarization measurement data will also gradually decrease, and when the polarization coefficient reaches a maximum value, the polarization information will be zero, and then the target scattering is completely a random noise process.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911076780.7A CN111025291B (en) | 2019-11-06 | 2019-11-06 | Ocean oil spill detection method based on new characteristics of fully-polarized SAR |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911076780.7A CN111025291B (en) | 2019-11-06 | 2019-11-06 | Ocean oil spill detection method based on new characteristics of fully-polarized SAR |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111025291A true CN111025291A (en) | 2020-04-17 |
CN111025291B CN111025291B (en) | 2021-07-16 |
Family
ID=70204944
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911076780.7A Active CN111025291B (en) | 2019-11-06 | 2019-11-06 | Ocean oil spill detection method based on new characteristics of fully-polarized SAR |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111025291B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111380809A (en) * | 2020-05-06 | 2020-07-07 | 长春理工大学 | Method for testing oil film type based on polarization characteristic |
CN112597434A (en) * | 2020-11-05 | 2021-04-02 | 中国科学院国家空间科学中心 | Rapid implementation method and system for polarized KHT decomposition |
CN116206203A (en) * | 2023-03-08 | 2023-06-02 | 中国石油大学(华东) | Oil spill detection method based on SAR and Dual-EndNet |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102096070A (en) * | 2010-11-16 | 2011-06-15 | 苏州科技学院 | Sea surface oil spilling detection method of complete polarization synthetic aperture radar |
WO2013095159A1 (en) * | 2011-12-19 | 2013-06-27 | Ispas As | Method of detecting oil spill at the sea by means of an oil spill radar, and such an oil spill radar |
US8570207B1 (en) * | 2010-06-09 | 2013-10-29 | Arrowhead Center, Inc. | Method, technique, and system for detecting Brillouin precursors at microwave frequencies for enhanced performance in various applications |
CN103971370A (en) * | 2014-05-15 | 2014-08-06 | 中国科学院遥感与数字地球研究所 | Intelligent ocean oil spill detection method for remote sensing large image |
US20140347213A1 (en) * | 2012-03-09 | 2014-11-27 | U.S. Army Research Laboratory Attn: Rdrl-Loc-I | Method and System for Estimation and Extraction of Interference Noise from Signals |
CN105866775A (en) * | 2016-03-28 | 2016-08-17 | 南京信息工程大学 | Method for detecting oil spill at sea based on full-polarized synthetic aperture radar image |
CN107609577A (en) * | 2017-08-23 | 2018-01-19 | 中国地质大学(武汉) | A kind of polarization SAR slick extracting method using random forest |
CN109490886A (en) * | 2018-11-28 | 2019-03-19 | 北京工业大学 | A kind of accurate extracting method in polarimetric synthetic aperture radar remote sensing offshore spilled oil region |
CN110070549A (en) * | 2019-04-25 | 2019-07-30 | 中国石油大学(华东) | A kind of soft dividing method of marine oil overflow SAR image based on optimal scale neighborhood information |
CN110412573A (en) * | 2019-08-08 | 2019-11-05 | 内蒙古工业大学 | Polarimetric SAR image decomposition method and storage medium |
-
2019
- 2019-11-06 CN CN201911076780.7A patent/CN111025291B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8570207B1 (en) * | 2010-06-09 | 2013-10-29 | Arrowhead Center, Inc. | Method, technique, and system for detecting Brillouin precursors at microwave frequencies for enhanced performance in various applications |
CN102096070A (en) * | 2010-11-16 | 2011-06-15 | 苏州科技学院 | Sea surface oil spilling detection method of complete polarization synthetic aperture radar |
WO2013095159A1 (en) * | 2011-12-19 | 2013-06-27 | Ispas As | Method of detecting oil spill at the sea by means of an oil spill radar, and such an oil spill radar |
US20140347213A1 (en) * | 2012-03-09 | 2014-11-27 | U.S. Army Research Laboratory Attn: Rdrl-Loc-I | Method and System for Estimation and Extraction of Interference Noise from Signals |
CN103971370A (en) * | 2014-05-15 | 2014-08-06 | 中国科学院遥感与数字地球研究所 | Intelligent ocean oil spill detection method for remote sensing large image |
CN105866775A (en) * | 2016-03-28 | 2016-08-17 | 南京信息工程大学 | Method for detecting oil spill at sea based on full-polarized synthetic aperture radar image |
CN107609577A (en) * | 2017-08-23 | 2018-01-19 | 中国地质大学(武汉) | A kind of polarization SAR slick extracting method using random forest |
CN109490886A (en) * | 2018-11-28 | 2019-03-19 | 北京工业大学 | A kind of accurate extracting method in polarimetric synthetic aperture radar remote sensing offshore spilled oil region |
CN110070549A (en) * | 2019-04-25 | 2019-07-30 | 中国石油大学(华东) | A kind of soft dividing method of marine oil overflow SAR image based on optimal scale neighborhood information |
CN110412573A (en) * | 2019-08-08 | 2019-11-05 | 内蒙古工业大学 | Polarimetric SAR image decomposition method and storage medium |
Non-Patent Citations (2)
Title |
---|
FERDINANDO NUNZIATA: "《On the Mueller Scattering Matrix for SAR Sea Oil Slick Observation》", 《IEEE GEOSCIENCE AND REMOTE SENSING LETTERS》 * |
谢广奇: "《基于简缩极化SAR的海洋溢油检测》", 《中国优秀硕士学位论文全文数据库 工程科技I辑》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111380809A (en) * | 2020-05-06 | 2020-07-07 | 长春理工大学 | Method for testing oil film type based on polarization characteristic |
CN112597434A (en) * | 2020-11-05 | 2021-04-02 | 中国科学院国家空间科学中心 | Rapid implementation method and system for polarized KHT decomposition |
CN112597434B (en) * | 2020-11-05 | 2024-03-22 | 中国科学院国家空间科学中心 | Method and system for rapidly realizing polarization KHT decomposition |
CN116206203A (en) * | 2023-03-08 | 2023-06-02 | 中国石油大学(华东) | Oil spill detection method based on SAR and Dual-EndNet |
CN116206203B (en) * | 2023-03-08 | 2023-08-18 | 中国石油大学(华东) | Oil spill detection method based on SAR and Dual-EndNet |
Also Published As
Publication number | Publication date |
---|---|
CN111025291B (en) | 2021-07-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111025291B (en) | Ocean oil spill detection method based on new characteristics of fully-polarized SAR | |
Salberg et al. | Oil spill detection in hybrid-polarimetric SAR images | |
Yeremy et al. | Ocean surveillance with polarimetric SAR | |
Minchew et al. | Polarimetric analysis of backscatter from the Deepwater Horizon oil spill using L-band synthetic aperture radar | |
Watanabe et al. | Analysis of urban areas affected by the 2011 off the Pacific Coast of Tohoku earthquake and tsunami with L-band SAR full-polarimetric mode | |
Mercier et al. | Partially supervised oil-slick detection by SAR imagery using kernel expansion | |
EP2283380B1 (en) | Target detection in a sar-imaged sea area | |
KR20190133442A (en) | Method for detecting oil spills on satellite sar images using artificial neural network | |
CN102109597B (en) | Method for identifying type of ship according to high-resolution SAR (synthetic aperture radar) image of ship | |
Quan et al. | Derivation of the orientation parameters in built-up areas: With application to model-based decomposition | |
Chaudhary et al. | Marine oil slicks detection using spaceborne and airborne SAR data | |
Vachon et al. | Analysis of Sentinel-1 marine applications potential | |
Hassani et al. | Oil spill four-class classification using UAVSAR polarimetric data | |
CN111257875B (en) | Sea surface oil spill monitoring method and device based on SAR | |
Sankaran | Protecting oceans from illicit oil spills: environment control and remote sensing using spaceborne imaging radars | |
Nakmuenwai et al. | Multi-temporal correlation method for damage assessment of buildings from high-resolution SAR images of the 2013 typhoon Haiyan | |
Zheng et al. | The polarimetric features of oil spills in full polarimetric synthetic aperture radar images | |
Sun et al. | SAR image classification of ice in the Gulf of Bothnia | |
Sankaran | Spaceborne radar remote sensing of ocean surfaces: electromagnetic modelling and applications | |
Salberg et al. | Model based oil spill detection using polarimetric SAR | |
Thomas | Spectral similarity algorithm-based image classification for oil spill mapping of hyperspectral datasets | |
Guo et al. | A novel strategy of nonnegative-matrix-factorization-based polarimetric ship detection | |
Kumar et al. | Dark spot detection for characterization of marine surface slicks using PolSAR remote sensing | |
Prajapati et al. | Feature combination of Pauli and H/A/alpha decomposition for improved oil spill detection using SAR | |
Singha et al. | Offshore oil spill detection using synthetic aperture radar |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |