A method and system for identifying marine oil spills based on a synthetic aperture radar feature time spectrum library

Abstract

This invention discloses a method and system for identifying marine oil spills based on a synthetic aperture radar (SAR) feature time spectrum library, relating to the field of oil spill identification technology. The key technical points of this invention include: establishing a SAR feature time spectrum library for oil spill targets; wherein the library includes a SAR feature time spectrum of oil spill targets and feature parameters obtained from the SAR feature time spectrum; extracting suspected target feature parameters corresponding to multi-time-series SAR data of suspected oil spill targets; calculating the similarity between the suspected target feature parameters and multiple feature parameters in the SAR feature time spectrum library of oil spill targets, and averaging the multiple similarities; when the average exceeds a preset threshold, the suspected oil spill target is identified as an oil spill. This invention expands the response characteristics of oil spill targets from a time dimension, ensuring the quality and reliability of the SAR time spectrum library for marine oil spill targets, and improving the identification capability of oil spill targets.

Description

Technical Field

[0001] This invention relates to the field of oil spill identification technology, specifically to a marine oil spill identification method and system based on a synthetic aperture radar feature time spectrum library. Background Technology

[0002] Oil spills at sea originate from diverse sources, and remote sensing technology is widely used for oil spill identification and monitoring. This includes optical remote sensing and radar remote sensing (primarily referring to synthetic aperture radar remote sensing technology, hereinafter referred to as SAR remote sensing technology, or SAR). Because oil spills often occur in harsh environments, SAR is an important tool for oil spill identification and monitoring at sea.

[0003] However, in SAR images, in addition to oil slicks, some meteorological and oceanographic phenomena also present dark spots similar to oil slicks, called "oil slick-like" or "suspected oil slicks". This poses a challenge to SAR, which has relatively poor single-band imaging and spectral response characteristics, and affects SAR's ability to identify and monitor oil spill targets. Summary of the Invention

[0004] To address this, the present invention proposes a marine oil spill identification method and system based on a synthetic aperture radar feature time spectrum library, in an attempt to solve or at least alleviate at least one of the problems mentioned above.

[0005] According to one aspect of the present invention, a method for identifying marine oil spills based on a synthetic aperture radar feature time spectral library is proposed, the method comprising the following steps:

[0006] Step 1: Establish a synthetic aperture radar (SAR) feature time spectrum library for oil spill targets; wherein, the SAR feature time spectrum library for oil spill targets includes the SAR feature time spectrum of oil spill targets and feature parameters obtained from the SAR feature time spectrum;

[0007] Step 2: Acquire multi-temporal synthetic aperture radar data of suspected oil spill targets;

[0008] Step 3: Extract the suspected target feature parameters corresponding to the multi-temporal synthetic aperture radar data of suspected oil spill targets;

[0009] Step 4: Calculate the similarity between the suspected target feature parameters and multiple feature parameters in the synthetic aperture radar feature time spectrum library of oil spill targets, and take the average of the multiple similarities. When the average exceeds a preset threshold, the suspected oil spill target is identified as an oil spill.

[0010] Furthermore, the process of acquiring the synthetic aperture radar characteristic time spectrum of the oil spill target includes:

[0011] Acquire multi-temporal synthetic aperture radar data of oil spill targets;

[0012] The multi-time-series synthetic aperture radar data is preprocessed;

[0013] Feature extraction is performed on the preprocessed multi-temporal synthetic aperture radar data;

[0014] Principal component analysis is used to select extracted features to obtain multiple features corresponding to the synthetic aperture radar data at each time step.

[0015] The information entropy of the oil spill target is calculated based on the principal components obtained by feature selection, and the pure pixel of the oil spill target with the minimum information entropy is obtained.

[0016] The clean pixels of the oil spill target are mapped to multiple features obtained by feature selection, thereby obtaining the feature time series curves corresponding to each feature of the oil spill target. The multiple feature time series curves form the synthetic aperture radar feature time spectrum of the oil spill target.

[0017] Furthermore, the process of obtaining the feature parameters or suspected target feature parameters includes: extracting the feature parameters corresponding to the troughs on the feature time-series curve, wherein the feature parameters corresponding to the troughs include the trough position, trough width, trough symmetry, and trough depth; the trough position is the number of observation days corresponding to the minimum feature value; the trough width is the curve width at half the trough depth; and the trough symmetry refers to the degree of symmetry between the left and right parts of the vertical line passing through the trough position.

[0018] Furthermore, the synthetic aperture radar feature time spectrum library for oil spill targets also includes metadata, which includes observation target information, observation equipment information, observation location information, observation time information, and observation environment information.

[0019] Furthermore, the synthetic aperture radar data includes radar data acquired by the radar in single-polarization, dual-polarization, or full-polarization operating modes; the preprocessing includes radiometric calibration and filtering; and the features extracted during feature extraction include texture features and polarization features.

[0020] Furthermore, the process of calculating the information entropy of the oil spill target based on the principal components obtained from feature selection, and obtaining the clean pixel of the oil spill target with the minimum information entropy, includes:

[0021] The probability density function of a pixel is calculated using the principal components obtained from feature selection according to the following formula:

[0022] ;

[0023] in, Let be the probability density function of a pixel; Let be the gray value of the pixel in the i-th principal component; and The mean and variance of the pixels across multiple principal components;

[0024] The information entropy of a pixel is calculated using the following formula based on the pixel's probability density function. :

[0025] ;

[0026] in, Indicates the number of principal components;

[0027] Sort all oil spill pixels by information entropy and select the clean pixels with the lowest information entropy.

[0028] Furthermore, in step four, the similarity is calculated using the Pearson correlation coefficient.

[0029] According to another aspect of the present invention, a marine oil spill identification system based on a synthetic aperture radar feature time spectral library is proposed, the system comprising:

[0030] The feature time spectrum library establishment module is configured to establish a synthetic aperture radar feature time spectrum library for oil spill targets; wherein, the synthetic aperture radar feature time spectrum library for oil spill targets includes synthetic aperture radar feature time spectrum of oil spill targets and feature parameters obtained from synthetic aperture radar feature time spectrum;

[0031] The suspected oil spill target data acquisition module is configured to acquire multi-temporal synthetic aperture radar data of suspected oil spill targets;

[0032] The suspected oil spill target feature extraction module is configured to extract the suspected target feature parameters corresponding to the multi-temporal synthetic aperture radar data of suspected oil spill targets.

[0033] The oil spill identification module is configured to calculate the similarity between the suspected target feature parameters and multiple feature parameters in the synthetic aperture radar feature time spectrum library of oil spill targets, and to take the average of the multiple similarities. When the average value exceeds a preset threshold, the suspected oil spill target is identified as an oil spill.

[0034] Furthermore, the process of acquiring the synthetic aperture radar characteristic time spectrum of the oil spill target includes:

[0035] Acquire multi-temporal synthetic aperture radar data of oil spill targets;

[0036] The multi-time-series synthetic aperture radar data is preprocessed;

[0037] Feature extraction is performed on the preprocessed multi-temporal synthetic aperture radar data;

[0038] Principal component analysis is used to select extracted features to obtain multiple features corresponding to the synthetic aperture radar data at each time step.

[0039] The information entropy of the oil spill target is calculated based on the principal components obtained by feature selection, and the clean pixel of the oil spill target with the minimum information entropy is obtained. Specifically, the probability density function of the pixel is calculated based on the principal components obtained by feature selection according to the following formula:

[0040] ;

[0041] in, Let be the probability density function of a pixel; Let be the gray value of the pixel in the i-th principal component; and The mean and variance of the pixels across multiple principal components;

[0042] The information entropy of a pixel is calculated using the following formula based on the pixel's probability density function. :

[0043] ;

[0044] in, Indicates the number of principal components;

[0045] Sort all oil spill pixels by information entropy and select the clean oil spill target pixels with the lowest information entropy.

[0046] The clean pixels of the oil spill target are mapped to multiple features obtained by feature selection, thereby obtaining the feature time series curves corresponding to each feature of the oil spill target. The multiple feature time series curves form the synthetic aperture radar feature time spectrum of the oil spill target.

[0047] Furthermore, the process of obtaining the feature parameters or suspected target feature parameters includes: extracting the feature parameters corresponding to the troughs on the feature time-series curve, wherein the feature parameters corresponding to the troughs include the trough position, trough width, trough symmetry, and trough depth; the trough position is the number of observation days corresponding to the minimum feature value; the trough width is the curve width at half the trough depth; and the trough symmetry refers to the degree of symmetry between the left and right parts of the vertical line passing through the trough position.

[0048] The beneficial technical effects of this invention are:

[0049] This invention proposes the establishment of a synthetic aperture radar (SAR) time-spectrum library (hereinafter referred to as the time-spectrum library) for marine oil spill targets, and utilizes the time-spectrum library to conduct research on oil spill target identification. This invention is the first to propose the concept of a SAR feature time-spectrum library, establishing a time-spectrum library of SAR features for marine oil spill targets. This expands the response characteristics of oil spill targets from a time dimension, serving as an important knowledge base and a fundamental database for marine oil spill target identification. Based on the time-series images of SAR features of oil spill targets, a method for extracting clean pixels of oil spill targets is proposed, ensuring the quality and reliability of the marine oil spill target SAR time-spectrum library. The time-varying SAR features of marine oil spill targets differ significantly from the time-varying features of other marine targets because oil spills not only physically spread, causing changes in thickness, but also weathering affects the physicochemical properties of oil spill targets. Utilizing the time-dimensional response characteristics of marine oil spill targets' SAR features can greatly improve the identification capability of oil spill targets. Attached Figure Description

[0050] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings. Several embodiments of the invention are illustrated in the drawings by way of example and not limitation, wherein:

[0051] Figure 1 This is a flowchart of a marine oil spill identification method based on a synthetic aperture radar feature time spectrum library, as described in an embodiment of the present invention.

[0052] Figure 2 This is a schematic diagram of the structure of a marine oil spill identification system based on a synthetic aperture radar feature time spectrum library, as described in an embodiment of the present invention. Detailed Implementation

[0053] The principles and spirit of the invention will now be described with reference to several exemplary embodiments. It should be understood that these embodiments are given merely to enable those skilled in the art to better understand and implement the invention, and are not intended to limit the scope of the invention in any way. Rather, these embodiments are provided to make this disclosure more thorough and complete, and to fully convey the scope of this disclosure to those skilled in the art.

[0054] Those skilled in the art will recognize that embodiments of the present invention can be implemented as a system, apparatus, device, method, or computer program product. Therefore, this disclosure can be specifically implemented in the following forms: entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software. It should be understood herein that any number of elements in the accompanying drawings is for illustrative purposes only and not as a limitation, and any naming is for distinction only and has no limiting meaning.

[0055] This invention proposes a method for identifying marine oil spills based on a synthetic aperture radar (SAR) feature time spectrum library. This method comprises two parts: establishing a SAR feature time spectrum library for oil spill targets and identifying marine oil spill targets. For example... Figure 1 As shown, the method specifically includes the following steps:

[0056] Step 1: Establish a synthetic aperture radar (SAR) feature time spectrum library for oil spill targets; wherein, the SAR feature time spectrum library for oil spill targets includes the SAR feature time spectrum of oil spill targets and feature parameters obtained from the SAR feature time spectrum;

[0057] Step 2: Acquire multi-temporal synthetic aperture radar data of suspected oil spill targets;

[0058] Step 3: Extract the suspected target feature parameters corresponding to the multi-temporal synthetic aperture radar data of suspected oil spill targets;

[0059] Step 4: Calculate the similarity between the suspected target feature parameters and multiple feature parameters in the synthetic aperture radar feature time spectrum library of oil spill targets, and take the average of the multiple similarities. When the average exceeds a preset threshold, the suspected oil spill target is identified as an oil spill.

[0060] In step one, a synthetic aperture radar (SAR) feature time spectrum library for oil spill targets is established. This library includes the SAR feature time spectrum of oil spill targets, its feature parameters, and metadata. The process of acquiring the SAR feature time spectrum of oil spill targets includes:

[0061] Acquire multi-temporal synthetic aperture radar (SAP) data of the oil spill target; preprocess the multi-temporal SAP data; extract features from the preprocessed multi-temporal SAP data; select extracted features using principal component analysis (PCA) to obtain multiple features corresponding to the SAP data at each time step; calculate the information entropy of the oil spill target based on the principal components obtained by feature selection, and obtain the clean pixel of the oil spill target with the smallest information entropy; map the clean pixel of the oil spill target to the multiple features obtained by feature selection, thereby obtaining the feature time-series curve corresponding to each feature of the oil spill target, and the multiple feature time-series curves constitute the SAP feature time spectrum of the oil spill target.

[0062] The process of obtaining feature parameters or suspected target feature parameters includes: removing the envelope of the feature time series curve; extracting the feature parameters corresponding to the troughs on the feature time series curve, wherein the feature parameters corresponding to the troughs include the trough position, trough width, trough symmetry, and trough depth; the trough position is the number of observation days corresponding to the minimum feature value, where the number of observation days refers to the day after the accident, with the date of the accident as the first day; the trough width is the curve width at half the trough depth; and the trough symmetry refers to the degree of symmetry between the left and right parts of the vertical line passing through the trough position.

[0063] According to an embodiment of the present invention, 1) for acquiring multi-temporal synthetic aperture radar (SAR) data: selecting a ship oil spill accident or an offshore platform oil spill accident is beneficial for tracking the oil spill target and acquiring multiple SAR images of the accident waters, i.e., acquiring multi-temporal SAR images covering the accident waters. The SAR data here can be data in single-polarization, dual-polarization, or full-polarization operating modes.

[0064] 2) For data preprocessing: The specific preprocessing content of synthetic aperture radar data acquired by different satellites, including data from different operating modes, varies, but generally includes radiometric calibration and filtering.

[0065] 3) Feature extraction: For single-polarization SAR images, texture features are mainly extracted. For single-polarization, dual-polarization, and fully polarization SAR images, in addition to texture features, polarization features can also be extracted. This also includes other features derived from texture or polarization features. Among these, based on oil film scattering characteristics and the oil spill monitoring mechanism of polarization features, polarization features can be divided into three categories: (a) Oil film scattering features, including backscattering energy and parameters reflecting the scattering mechanism, such as polarization entropy H / anisotropy A / scattering angle. (b) Parameters defined based on polarization channel correlation, such as the order of magnitude of co-polarization correlation. The real part of the same polarization cross product (c) Polarization parameter combination, similar to band combination in optical remote sensing.

[0066] 4) Feature selection: Feature extraction yields a large number of feature images, which may contain redundant information. If all of these are used to build the temporal spectrum library, it will be difficult to reflect the characteristics of the oil spill target; moreover, the computational load will be large. This invention utilizes PCA (Principal Component Analysis) to first identify the principal components containing 99% of the information (the principal components are images, and there may be 3 to 6 images, or even more). It then further identifies which feature images contribute significantly to the principal components containing 99% of the information, and these feature images are used as the feature images for building the temporal spectrum library.

[0067] It should be noted that the characteristic images used to establish the spectral library may be different for SAR data in different operating modes (single polarization, dual polarization, and full polarization) or even for SAR data acquired by different satellites.

[0068] 5) For determining the pure pixels of the oil spill target: calculate the information entropy based on the principal components that contain 99% of the information obtained from feature selection, and obtain the pixel with the minimum information entropy. Specifically: calculate the mean and variance of the image based on the principal components to obtain the information entropy. The smaller the information entropy, the purer the pixel.

[0069] The probability density function of a pixel is:

[0070] ;

[0071] in Let be the probability density function of a pixel; Let be the gray value of the pixel in the i-th principal component; and Let be the mean and variance of the pixel among several principal components; exp be an exponential function with the natural constant e as the base; after obtaining the probability density function, the information entropy of the pixel can be calculated by the following formula:

[0072] ;

[0073] in, Indicates the number of principal components. The information entropy of a pixel represents the amount of information contained in the pixel. The calculation results are sorted, and the pure pixels with the smallest information content are selected (the pixel with the smallest 3% entropy can be selected) for subsequent acquisition of the feature information of the pixel.

[0074] 6) Map the obtained clean pixels to multiple feature images obtained by feature selection, that is, obtain the values ​​of multiple features of the oil spill target at that moment.

[0075] Repeating steps 1) to 6) will obtain the characteristics of the oil spill target at different times on the multi-time series SAR images. That is, corresponding to multiple feature images, multiple time series values ​​of the oil spill target are obtained, and then the feature time series curves corresponding to each feature are obtained. Multiple feature time series curves form the synthetic aperture radar feature time spectrum of the oil spill target.

[0076] For multiple feature images, the SAR feature time spectrum of the oil spill target consists of multiple curves, each curve corresponding to a feature. The horizontal axis of the curve is the number of days the SAR image was observed (the day the oil spill occurred is recorded as 1, and so on, as day 1, day 2, day 3, day 4, ..., day n), and the vertical axis is the pixel value of the clean oil spill target on the feature image.

[0077] For a curve, there are valleys and peaks in its shape. Feature parameter extraction is to extract the feature parameters of the valleys. Borrowing the concept from optical remote sensing, the valley here is equivalent to absorption. The extracted features include: valley position (P), valley width (W), valley symmetry (S), and valley depth (H). The feature parameters of the time spectrum curve are all calculated using the curve after removing the envelope. The valley position P is the number of observation days at the point where the feature value is the minimum. The valley width W is defined as the curve width at half the valley depth.

[0078] ;

[0079] in and These represent the times corresponding to half the depth of the left and right troughs, respectively. It should be noted that since the horizontal axis represents days... and It may not be an integer value; it has no actual time significance and is only used to describe the shape of the curve.

[0080] The symmetry S of a trough refers to the degree of symmetry between the left and right parts of the vertical line passing through the trough position. It can be expressed as the ratio of the wavelength width at half the depth of the left (right) trough to the trough width.

[0081] ;

[0082] In this embodiment, preferably, establishing a time spectrum library also requires recording metadata related to the data. This metadata includes observation target information (such as the type, shape, and size of the target object), observation equipment information (such as radar model and operating parameters), observation location information (such as the geographical location and environmental conditions of the observation point), observation time information (such as the observation start time and duration), and observation environment information (such as atmospheric conditions and noise levels). Recording metadata helps to identify whether the synthetic aperture radar data of the target to be identified is consistent with the synthetic space radar data in the time spectrum library. Therefore, in practical applications, if the data is rich enough, time spectrum libraries can be established separately based on synthetic aperture radar data from different sources.

[0083] The above steps established a synthetic aperture radar (SAR) signature time spectrum library for oil spill targets.

[0084] In step two, multi-temporal synthetic aperture radar (MAP) data of suspected oil spill targets are acquired; in step three, suspected target feature parameters corresponding to the multi-temporal MAP data of suspected oil spill targets are extracted.

[0085] According to an embodiment of the present invention, referring to the steps of establishing a synthetic aperture radar feature time spectrum library for oil spill targets, multiple time-series SAR images of any marine oil spill accident to be identified are acquired, and the following steps are completed in sequence: preprocessing of multiple time-series SAR images, feature extraction, feature selection, clean pixel acquisition, and finally obtaining the feature time-series values ​​of suspected oil spill targets, thereby obtaining the feature parameters of the feature time spectrum of suspected oil spill targets.

[0086] In step four, the similarity between the suspected target feature parameters and multiple feature parameters in the synthetic aperture radar feature time spectrum library of oil spill targets is calculated, and the average of the multiple similarities is taken. When the average exceeds a preset threshold, the suspected oil spill target is identified as an oil spill.

[0087] According to an embodiment of the present invention, the feature parameters of the time spectrum of a suspected oil spill target are compared with the time spectrum feature parameters of the corresponding SAR data and corresponding features in the time spectrum library. The similarity is calculated using the Pearson correlation coefficient. If the average similarity exceeds a preset threshold (e.g., set to 0.5), the suspected oil spill target can be considered an oil spill, and the sample is labeled. Here, since a suspected oil spill target may have time spectra of multiple features, multiple sets of feature parameters will be obtained. Therefore, when comparing with the time spectrum feature parameters of the corresponding SAR data and corresponding features in the time spectrum library, multiple Pearson correlation coefficients will be obtained. Therefore, it is necessary to average the similarity.

[0088] Another embodiment of the present invention proposes a marine oil spill identification system based on a synthetic aperture radar feature time spectral library, such as... Figure 2 As shown, the system includes:

[0089] The feature time spectrum library establishment module 210 is configured to establish a synthetic aperture radar feature time spectrum library for oil spill targets; wherein, the synthetic aperture radar feature time spectrum library for oil spill targets includes synthetic aperture radar feature time spectrum of oil spill targets and feature parameters obtained from synthetic aperture radar feature time spectrum;

[0090] The suspected oil spill target data acquisition module 220 is configured to acquire multi-temporal synthetic aperture radar data of suspected oil spill targets.

[0091] The suspected oil spill target feature extraction module 230 is configured to extract the suspected target feature parameters corresponding to the multi-temporal synthetic aperture radar data of the suspected oil spill target.

[0092] The oil spill identification module 240 is configured to calculate the similarity between the suspected target feature parameters and multiple feature parameters in the synthetic aperture radar feature time spectrum library of the oil spill target, and take the average of the multiple similarities. When the average value exceeds a preset threshold, the suspected oil spill target is identified as an oil spill.

[0093] In this embodiment, preferably, the process of obtaining the synthetic aperture radar characteristic time spectrum of the oil spill target includes:

[0094] Acquire multi-temporal synthetic aperture radar data of oil spill targets;

[0095] The multi-time-series synthetic aperture radar data is preprocessed;

[0096] Feature extraction is performed on the preprocessed multi-temporal synthetic aperture radar data;

[0097] Principal component analysis is used to select extracted features to obtain multiple features corresponding to the synthetic aperture radar data at each time step.

[0098] The information entropy of the oil spill target is calculated based on the principal components obtained by feature selection, and the clean pixel of the oil spill target with the minimum information entropy is obtained. Specifically, the probability density function of the pixel is calculated based on the principal components obtained by feature selection according to the following formula:

[0099] ;

[0100] in, Let be the probability density function of a pixel; Let be the gray value of the pixel in the i-th principal component; and The mean and variance of the pixels across multiple principal components;

[0101] The information entropy of a pixel is calculated using the following formula based on the pixel's probability density function. :

[0102] ;

[0103] in, Indicates the number of principal components;

[0104] Sort all oil spill pixels by information entropy and select the clean oil spill target pixels with the lowest information entropy.

[0105] The clean pixels of the oil spill target are mapped to multiple features obtained by feature selection, thereby obtaining the feature time series curves corresponding to each feature of the oil spill target. The multiple feature time series curves form the synthetic aperture radar feature time spectrum of the oil spill target.

[0106] In this embodiment, preferably, the process of obtaining the feature parameters or suspected target feature parameters includes: removing the envelope of the feature time series curve; extracting the feature parameters corresponding to the troughs on the feature time series curve, wherein the feature parameters corresponding to the troughs include the trough position, trough width, trough symmetry, and trough depth; the trough position is the number of observation days corresponding to the minimum feature value; the trough width is the curve width at half the trough depth; and the trough symmetry refers to the degree of symmetry between the left and right parts of the vertical line passing through the trough position.

[0107] The functionality of the marine oil spill identification system based on a synthetic aperture radar feature time spectrum library in this embodiment of the invention can be described by the aforementioned marine oil spill identification method based on a synthetic aperture radar feature time spectrum library. Therefore, for the parts not described in detail in the system embodiment, please refer to the above method embodiment, and they will not be repeated here.

[0108] It should be noted that although several units, modules, or sub-modules are mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of the present invention, the features and functions of two or more modules described above can be embodied in one module. Conversely, the features and functions of one module described above can be further divided and embodied by multiple modules.

[0109] Furthermore, although the operations of the method of the present invention are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0110] While the spirit and principles of the invention have been described with reference to several specific embodiments, it should be understood that the invention is not limited to the disclosed specific embodiments, and the division of aspects does not imply that features in these aspects cannot be combined for benefit; such division is merely for ease of description. The invention is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A method for identifying marine oil spills based on a synthetic aperture radar feature time spectrum library, characterized in that, The method comprises the following steps: Step one, establishing an oil spill target synthetic aperture radar feature time spectrum library; wherein the oil spill target synthetic aperture radar feature time spectrum library comprises an oil spill target synthetic aperture radar feature time spectrum and feature parameters obtained from the synthetic aperture radar feature time spectrum; the acquisition process of the oil spill target synthetic aperture radar feature time spectrum comprises: acquiring multi-time sequence synthetic aperture radar data of an oil spill target; preprocessing the multi-time sequence synthetic aperture radar data; extracting features from the preprocessed multi-time sequence synthetic aperture radar data; selecting the extracted features by using principal component analysis to obtain a plurality of features corresponding to the synthetic aperture radar data at each time point; calculating the information entropy of the oil spill target according to the principal components obtained by feature selection, and obtaining pure pixels of the oil spill target with the minimum information entropy; the calculation of the probability density function of the pixel according to the principal components obtained by feature selection is as follows: ； wherein, is the probability density function of the pixel; is the gray value of the pixel in the i-th principal component; and is the mean and variance of the pixel in the principal components. The information entropy of the pixel is calculated according to the probability density function of the pixel according to the following formula : ； wherein, represents the amount of the main component; sorting the information entropy of all oil spill pixels and selecting a plurality of pure pixels of the oil spill target with the minimum information entropy; mapping the pure pixels of the oil spill target to the plurality of features obtained by feature selection, and then obtaining feature time sequence curves corresponding to each feature of the oil spill target, and a plurality of feature time sequence curves form the oil spill target synthetic aperture radar feature time spectrum; Step two, acquiring multi-time sequence synthetic aperture radar data of a suspected oil spill target; Step three, extracting feature parameters corresponding to the multi-time sequence synthetic aperture radar data of the suspected oil spill target; Step four, calculating the similarity of the feature parameters of the suspected target and a plurality of feature parameters in the oil spill target synthetic aperture radar feature time spectrum library, and taking the average of a plurality of similarities, and when the average exceeds a preset threshold, identifying the suspected oil spill target as an oil spill.

2. The method according to claim 1, wherein, The acquisition process of the feature parameters or the suspected target feature parameters comprises removing the envelope of the feature time sequence curve; extracting feature parameters corresponding to the wave troughs on the feature time sequence curve, wherein the feature parameters corresponding to the wave troughs comprise a wave trough position, a wave trough width, a wave trough symmetry degree and a wave trough depth; the wave trough position is the observation date corresponding to the minimum feature value; the wave trough width is the curve width at half the wave trough depth; and the wave trough symmetry degree refers to the symmetry degree of the left and right parts of the vertical line passing through the wave trough position.

3. The method according to claim 1, wherein, The oil spill target synthetic aperture radar feature time spectrum library further comprises metadata, and the metadata comprises observation target information, observation device information, observation location information, observation time information and observation environment information.

4. The method according to claim 1, wherein, The synthetic aperture radar data comprises radar data collected by the radar in a single polarization, dual polarization or full polarization working mode; the preprocessing comprises radiation calibration and filtering; and the features extracted in the feature extraction comprise texture features and polarization features.

5. The method according to claim 1, wherein, The similarity is calculated by using the Pearson correlation coefficient in step four.

6. A marine oil spill identification system based on a synthetic aperture radar feature time library, characterized in that, The method comprises: a feature time spectrum library establishing module configured to establish an oil spill target synthetic aperture radar feature time spectrum library; wherein the oil spill target synthetic aperture radar feature time spectrum library comprises an oil spill target synthetic aperture radar feature time spectrum and feature parameters obtained from the synthetic aperture radar feature time spectrum; the acquisition process of the oil spill target synthetic aperture radar feature time spectrum comprises: Obtain multi-time synthetic aperture radar data of an oil spill target; Preprocess the multi-time synthetic aperture radar data; Extract features from the preprocessed multi-time synthetic aperture radar data; Select the extracted features by using principal component analysis to obtain a plurality of features corresponding to the synthetic aperture radar data at each time point; Calculate the information entropy of the oil spill target according to the principal components obtained by feature selection, and obtain pure pixels of the oil spill target with the minimum information entropy; specifically, the probability density function of the pixel is calculated according to the principal components obtained by feature selection according to the following formula: ； wherein, is the probability density function of the pixel; is the gray value of the pixel in the i-th principal component; and is the mean and variance of the pixel in the plurality of principal components; The information entropy of the pixel is calculated according to the probability density function of the pixel according to the following formula : ； wherein, represents the amount of the main component; Sort the information entropy of all oil spill pixels, and select a plurality of pure pixels of the oil spill target with the minimum information entropy; Map the pure pixels of the oil spill target to the plurality of features obtained by feature selection, and further obtain a feature time series curve corresponding to each feature of the oil spill target, and a plurality of feature time series curves constitute a synthetic aperture radar feature time spectrum of the oil spill target; A suspected oil spill target data acquisition module configured to obtain multi-time synthetic aperture radar data of a suspected oil spill target; A suspected oil spill target feature extraction module configured to extract suspected target feature parameters corresponding to the multi-time synthetic aperture radar data of the suspected oil spill target; An oil spill identification module configured to calculate the similarity between the suspected target feature parameters and a plurality of feature parameters in an oil spill target synthetic aperture radar feature time spectrum library, and take the average of a plurality of similarities, and when the average exceeds a preset threshold, identify the suspected oil spill target as an oil spill.

7. The system according to claim 6, wherein, The feature parameters or suspected target feature parameters are obtained by: removing the envelope of the feature time series curve; extracting feature parameters corresponding to the wave troughs on the feature time series curve, wherein the feature parameters corresponding to the wave troughs include wave trough position, wave trough width, wave trough symmetry, and wave trough depth; the wave trough position is the observation date corresponding to the minimum feature value; the wave trough width is the curve width at half the wave trough depth; and the wave trough symmetry refers to the symmetry degree of the left and right parts of the vertical line passing through the wave trough position.

Images