Shipborne GNSS-R polar region ice region environment oil spill detection method

Abstract

The invention discloses a shipborne GNSS-R polar region ice region environment oil spill detection method. The method comprises the following steps: establishing a scattering transmission model of a polar region ice region on seawater, sea ice and oil film surfaces based on a GNSS-R signal; based on a coherent signal and incoherent signal accumulation processing method, an oil film existence area is identified; and carrying out feature extraction on the GNSS-R signal of the sea surface ice-oil mixture, and determining whether the polar region ice region environment spills oil or not. The shipborne GNSS-R polar region ice region environment oil spill detection technology provided by the invention has a wide application prospect. According to the ship-borne GNSS-R polar region ice region environment oil spill detection technology based on the dual-polarized antenna, the problem that targets such as sea ice blocks or oil spill are difficult to recognize can be solved, the detection and recognition accuracy of polar region navigation ice region oil spill can be improved, normalized monitoring of polar region navigation ice region oil spill and self-monitoring of navigation ships can be achieved in an assisted mode, and the detection and recognition efficiency of the polar region navigation ice region oil spill is improved. Technical guarantee is provided for polar region shipping safety, ice region strategic resource development and green and safe development of oil and gas storage and transportation in China, and the method has very important research value and popularization significance.

Description

Technical Field

[0001] The present invention relates to a ship navigation environment perception technology, and in particular to a shipborne GNSS-R (Global Navigation Satellite System Reflectometry) polar ice zone environment oil spill detection method. Background Art

[0002] At present, GNSS-R, as a new branch of global navigation satellite developed in the 1980s, provides the world with microwave remote sensing information with wide coverage, low cost, multi-source signal, high stability and long-term use through ocean reflection signals from hundreds of GNSS satellites such as GPS, GALLEO, GLONASS, COMPASS and Beidou satellite navigation system, and has advantages in marine accident monitoring. When GNSS-R signal is incident on the interface between two media, transmission and reflection phenomena will occur, which is mainly related to the wavelength, incident angle, electrical properties and surface roughness of the GNSS-R signal. In the early days, people drew a fuzzy image of the surface scattering coefficient in the delayed Doppler domain by measuring the delayed Doppler map (DDM), and proposed the imaging technology of the ocean surface based on the GNSS-R satellite receiver, which improved the detection capability of oil spills on the sea surface. Around 2000, the spatial integration method (SIA) was used to extract the scattering coefficient of sea surface oil spill detection from GNSS-R signals, which improved the accuracy of oil spill monitoring. Domestic scholars Li Ying and others from Dalian Maritime University have repeatedly used TM data, MODIS data and AVHRR data to achieve oil spill identification and monitoring in a multi-source satellite remote sensing environment. Zhu Xueyuan and others from Dalian Maritime University established a shipborne GNSS-R dual-base sea surface scattering information geometry model, explored the shipborne specular reflection point and mean surface slope (MSS) model, and developed a GNSS-R DDM generation algorithm based on the Jacobian determinant method to carry out oil spill DDM simulation and establish the relationship between signal characteristics and oil spill information.

[0003] In terms of ship navigation environment perception, the existing satellite-based and airborne platform oil spill detection methods have problems such as low spatial resolution, low monitoring frequency, and low recognition accuracy. GNSS-R detection technology has the advantages of multi-source signals, good real-time performance, wide detection range, and high resolution. Laser-induced fluorescence has the advantages of all-day, detecting oil films on the surface of ice and snow, and identifying oil types in different backgrounds. GNSS-R, as a microwave imaging sensor, detects marine targets by acquiring backscattered and reflected signals. From the perspective of physical chemistry and fluid mechanics, the presence of oil films on the sea surface will reduce the surface tension of seawater, damp the Bragg scattering effect of capillary waves and short gravity waves on the sea surface on the incident electromagnetic waves, reduce the roughness of the sea surface, and cause the echo signal received by the radar to weaken, thereby showing different scattering characteristics from other targets. The sea surface modulation effect of different substances is different, and the contribution level of characteristic parameters is different. On this basis, the simplified feature system for ice area oil spill detection can identify interference targets such as sea ice and seawater in complex background environments, and accurately and efficiently identify oil films in polar navigation ice areas. Summary of the invention

[0004] In view of the problems existing in the prior art, the present invention discloses a shipborne GNSS-R polar ice environment oil spill detection method, a shipborne GNSS-R polar ice environment oil spill detection method, comprising the following steps:

[0005] Establish a scattering transmission model of polar ice regions based on GNSS-R signals on the surface of seawater, sea ice, and oil film;

[0006] Based on the coherent signal and incoherent signal accumulation processing method, the oil film existence area is identified;

[0007] Feature extraction of GNSS-R signals of sea surface ice-oil mixture was performed to determine whether there was oil spill in the polar ice environment.

[0008] Further: the scattering transmission model of the polar ice area based on the GNSS-R signal on the surface of seawater, sea ice and oil film includes the Kubelka-Munk model, the Kirk-Munk model or the Gaussian mixture model; or a combination of any two of the above three models.

[0009] Further: the process of optimizing the coherent signal and the incoherent signal by using the accumulation processing method is as follows:

[0010] For coherent signals, the general noise model is optimized;

[0011] Decompose the coherent signal into frequency domain information of different scales;

[0012] A Gaussian filter is used to remove noise in high-frequency components;

[0013] For incoherent signals, the signal strength is improved by accumulating the square sum, ignoring the phase information of the signal and retaining only the modulus value. The influence of phase error and frequency offset is eliminated by accumulating and subtracting the overlapping parts.

[0014] Further: the process of extracting features from the GNSS-R signal of the sea ice-oil mixture to determine whether there is oil spill in the polar ice environment is as follows:

[0015] Differentiated signal feature analysis: Analyze the differentiated signal features of GNSS signals after reflection from oil film, seawater and sea ice and direct signals, and identify signal disturbances caused by medium changes through phase analysis and amplitude comparison; distinguish seawater scattering areas and sea ice scattering areas from GNSS signals reflected from the sea surface by comparing the difference between the complex dielectric constant of the seawater surface and the complex dielectric constant of the area covered with strong scatterers;

[0016] Feature set construction and optimization: Based on the analysis of differentiated signal characteristics, a comprehensive feature set based on dielectric constant, reflectivity, signal-to-noise ratio and polarization characteristic parameters is constructed; the variable field decomposition method is used to adjust and determine the optimal value range of the feature parameters in the comprehensive feature set data;

[0017] Quantitative analysis of the contribution of characteristic parameters: Apply statistical analysis and model predictive control methods to establish a multivariate regression model to quantitatively evaluate the contribution and sensitivity of each characteristic parameter in oil film detection in polar ice areas.

[0018] Further: the process of adjusting and determining the optimal value range of the characteristic parameters in the comprehensive characteristic set data by using the variable field decomposition method is as follows:

[0019] Each variable field is standardized using the normalization formula;

[0020] Orthogonal decomposition of multivariate fields is achieved by calculating the covariance matrix;

[0021] Different weights are assigned to dielectric constant, reflectivity, signal-to-noise ratio and polarization characteristics to optimize the comprehensive feature set.

[0022] A shipborne GNSS-R polar ice environment oil spill detection device, comprising:

[0023] Establishment module: used to establish the scattering transmission model of polar ice areas based on GNSS-R signals on the surface of seawater, sea ice and oil film;

[0024] Identification module: Identify the oil film area based on the coherent signal and incoherent signal accumulation processing method;

[0025] Determination module: used to extract features from the GNSS-R signal of the sea surface ice-oil mixture and determine whether there is oil spill in the polar ice environment.

[0026] A readable storage medium stores a program module, wherein the program module can implement any of the methods described above when executed in a processor.

[0027] The present application discloses a shipborne GNSS-R polar ice environment oil spill detection method, which integrates the phase characteristics of left-handed circularly polarized GNSS (Global Navigation Satellite System) reflection signals, and uses the dual polarization method to eliminate the distance delay difference phase to extract the residual phase of penetrating sea ice and oil spill. In addition, it is possible to carry out a study on the concentration and constraint evaluation method of scattered signals in the oil spill area, summarize the typical scattering model characteristics of the three main targets in the ice area: ice surface, water surface, and oil film, and determine the scattering model to which the target scattering point belongs based on the generalized likelihood ratio of the scattering point distribution and the typical scattering model, so as to effectively identify the target attributes. On this basis, the various measurement results of GNSS-R are integrated to establish a multi-dimensional coupling model between parameters such as sea surface dielectric constant, scattering power, scattering point distribution characteristics and ice area targets, so as to realize multi-variable joint decision-making and judgment of oil spills.

[0028] In view of the target distribution pattern in the complex sea environment, the research optimizes the coherent / incoherent signal accumulation and processing method to effectively improve the signal-to-noise ratio, enhance the detection capability of scattered oil films, and accurately identify the areas where oil films exist; the research on the adaptive time-frequency joint processing method of scattered signals based on prior information is carried out. Through the measurement data collected during navigation, the statistical values ​​of parameters such as seawater reflectivity, coarseness, and scattering point distribution are obtained, and they are used as prior information to adaptively adjust the time-frequency processing search area to improve data processing efficiency.

[0029] Due to the adoption of the above technical scheme, the shipborne GNSS-R polar ice zone environment oil spill detection technology provided by the present invention has broad application prospects. The shipborne GNSS-R polar ice zone environment oil spill detection technology based on dual-polarized antennas can overcome the problem of difficult identification of targets such as sea ice or oil spills, improve the accuracy of polar navigation ice zone oil spill detection and identification, and help realize the normalized monitoring of polar navigation ice zone oil spills and self-monitoring of navigable ships, providing technical guarantees for my country's polar shipping safety, ice zone strategic resource development, and green and safe development of oil and gas storage and transportation, and has very important research value and promotion significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0031] Figure 1A position relationship diagram of a satellite signal source, a ship-borne receiver, and a sea surface reflection point in the present invention;

[0032] Figure 2 This is a typical model of sea surface GNSS signal scattering in the present invention. DETAILED DESCRIPTION

[0033] In order to make the technical solutions and advantages of the present invention more clear, the technical solutions in the embodiments of the present invention are clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0034] Figure 1 A position relationship diagram of a satellite signal source, a ship-borne receiver, and a sea surface reflection point in the present invention;

[0035] A shipborne GNSS-R polar ice environment oil spill detection method comprises the following steps:

[0036] S1: Establish a scattering transmission model based on GNSS-R signals in polar ice areas on the surface of seawater, sea ice, and oil film; and determine the scattering model to which the target scattering point belongs based on the generalized likelihood ratio of the scattering point distribution and the typical scattering model, and effectively identify the target attributes;

[0037] S2: Based on the coherent signal and incoherent signal accumulation processing method, the signal-to-noise ratio is improved, laying the foundation for accurately identifying the oil film existence area;

[0038] S3: Extract features of the GNSS-R signal of the sea surface ice-oil mixture to determine whether there is oil spill in the polar ice environment.

[0039] Steps S1 / S2 / S3 are executed sequentially;

[0040] Figure 2 It is a typical model of sea surface GNSS signal scattering in the present invention;

[0041] Furthermore, the scattering transmission model of GNSS-R signals on seawater, sea ice, and oil film surfaces based on the polar ice environment is determined, including the Kubelka-Munk model, the Kirk-Munk model, or the Gaussian mixture model; or a combination of two of the above three models. Among them, the Kubelka-Munk model is more suitable for the presence of snow, the Kirk-Munk model is suitable for the ocean surface under non-rainfall conditions, and the Gaussian mixture model is widely applicable to most situations. According to the characteristics of the polar ice environment, appropriate adjustments or combined models may be required.

[0042] like Figure 1 The positions of the signal source, receiver, and reflection point shown are distributed according to the Fresnel reflection law. The specific position information of the reflection point can be obtained from the time information of the GNSS-R signal, laying the foundation for establishing the relationship between the reflection point and the dielectric constant.

[0043] A mapping transmission model between scattered power and different targets in an ice-covered oil spill environment is established, and the dielectric constant of the sea surface is inverted using the relationship between the left-hand and right-hand components of the GNSS-R reflected signal and the polarization ratio of the direct signal. Under single polarization conditions, the dielectric constant of the reflection cutoff can be expressed as:

[0044]

[0045] Among them, θ is the satellite altitude angle, Г is the left-hand reflectivity (the right-hand signal component is weak after reflection from the sea surface, the signal-to-noise ratio is low, and it is not easy to extract), and ε is the dielectric constant. The dielectric constant of the reflecting surface medium can be calculated from the above formula. Generally speaking, the dielectric constant of seawater is 70-80, the dielectric constant of oil is 2-3, and the dielectric constant of sea ice is about 5-8. This is used as the criterion for detecting environmental targets such as oil spills and sea ice on the sea surface.

[0046] Furthermore, the process of optimizing the coherent signal and the incoherent signal by using the accumulation processing method is as follows:

[0047] For coherent signals, the general noise model is optimized;

[0048] Decompose the coherent signal into frequency domain information of different scales;

[0049] A Gaussian filter is used to optimize the denoising effect by removing noise in high-frequency components, thereby improving the accuracy and robustness of signal processing.

[0050] For incoherent signals, the signal strength is improved by accumulating the square sum calculation without relying on phase information. The phase information of the signal is ignored, only the modulus is retained, and the influence of phase error and frequency offset is eliminated by accumulating and subtracting the overlapping parts.

[0051] The process of optimizing the general noise model for the coherent signal is as follows:

[0052] Establish an appropriate noise model (such as Gaussian white noise or colored noise), then separate the signal and noise through filtering or transform domain methods, and estimate the statistical characteristics of the noise (such as variance and autocorrelation function). Select a suitable optimization method (such as minimum mean square error filtering, Kalman filtering or adaptive filtering), and achieve noise suppression through recursive updating and parameter adjustment. Then evaluate the optimization effect through indicators such as signal-to-noise ratio, mean square error and spectral characteristics, so as to achieve effective modeling and optimization of noise in coherent signals. At the same time, the adaptive time-frequency joint processing method of scattered signals based on prior information is used to adaptively adjust and optimize the noise reduction algorithm through the measurement data collected during navigation.

[0053] The multi-morphological distribution of ice and oil films will cause phase changes in the GNSS-R signal, reducing the coherence of the signal, which is an important means to judge the morphology of the ice-oil mixture on the sea surface. Under normal circumstances, the "oil above and ice below" mode or "ice above and oil below" has little effect on the signal coherence, while "ice wrapped in oil" has a greater impact on the signal coherence.

[0054] The process of extracting features from the GNSS-R signal of the sea ice-oil mixture to determine whether there is oil spill in the polar ice environment is as follows:

[0055] In order to improve the inversion accuracy of dielectric constant and suppress the influence of sea surface roughness on the accuracy, we proposed the polarization ratio method of dual-base polarization source.

[0056]

[0057] Where: θ is the incident angle of the GNSS signal, Y RL is the left-hand polarization component, Y RR is the right-hand polarization component.

[0058] The process of extracting features from the GNSS-R signal of the sea ice-oil mixture to determine whether there is oil spill in the polar ice environment is as follows:

[0059] S31: Differentiation signal feature analysis;

[0060] In GNSS-R applications, the power of the reflected signal collected by the receiver is closely related to the radar cross-section of different scattering points on the sea surface. This measurement is also called the "scattering coefficient" in ocean surface applications. It depends not only on the Fresnel reflection coefficient of the sea surface, but also on the roughness of the sea surface, which can be defined as:

[0061]

[0062] Where P is the probability density function, q is the horizontal component of the scattering vector, and q z is the vertical component of the scattering vector along the Z axis. The above formula explains the relationship between the scattering coefficient and the Fresnel reflection coefficient, surface roughness, and incident signal characteristics.

[0063] The scattering coefficient is a key parameter that can be used to distinguish seawater from other strong scattering media (sea ice) and to estimate the location of strong scattering areas on the sea surface.

[0064] The core concept of this method is based on the fact that different reflecting media have different complex dielectric constants, which leads to different scattering coefficients of GNSS signals reflected by these media and different powers of reflected signals. Therefore, by comparing the difference between the complex dielectric constant of the seawater surface and the complex dielectric constant of the area covered with strong scatterers (sea surface), the seawater scattering area and the sea ice scattering area can be distinguished from the GNSS signal reflected by the sea surface. Based on the above principle, the differentiated signal characteristics of the GNSS signal after reflection from the oil film, seawater and sea ice and the direct signal are deeply analyzed (compared with the direct reflection signal difference). Through advanced signal processing techniques, such as phase analysis and amplitude comparison, the signal disturbance caused by medium changes can be accurately identified.

[0065] S32. Feature set construction and optimization: Based on the signal feature analysis, a comprehensive feature set including parameters such as dielectric constant, reflectivity, signal-to-noise ratio and polarization characteristics is developed. They can be regarded as different variable fields, and the different variable fields are regarded as a multivariate field after being combined. In the subsequent analysis, the structure of the multivariate field is decomposed based on the method of univariate field decomposition, and the main characteristic components are extracted from the data, aiming to reduce the interference of non-uniform information in each field.

[0066] Considering that the above variable fields can change simultaneously, and each variable field with the same number of truth points established under a unified scale, they can be processed together. Since the units of each variable field are different, in order to explore the relationship between variable fields, it is necessary to standardize each variable field. The normalization formula is as follows:

[0067]

[0068] in and σ i are the mean and standard value of the variable domain, is the normalized and feature set normalized value.

[0069] Then, the orthogonal decomposition of the multivariate field is achieved by calculating the covariance matrix. At the same time, the matrix F can be transformed into a new matrix T, where S is a matrix composed of a series of orthogonal bases arranged in rows. After the above processing, the covariance matrix F can be calculated by the following formula:

[0070] D=TT T

[0071] =(SF)(SF) T

[0072] =SCS T

[0073] After orthogonal decomposition, parameters such as dielectric constant, reflectivity, signal-to-noise ratio and polarization characteristics can independently affect our observation of sea ice and oil in ice, but their weights are different. Based on general empirical measurements, we give the following weights: dielectric constant: 0.52, reflectivity: 0.21, signal-to-noise ratio: 0.05 and polarization characteristics: 0.12. It can be seen that dielectric constant is the main parameter affecting the detection of sea ice and oil in ice, followed by reflectivity and polarization characteristics, and signal-to-noise ratio has the least impact.

[0074] In addition, machine learning algorithms can be used to optimize the feature set to adjust and determine the optimal value range of the feature parameters in the experimental data.

[0075] S33. Quantitative analysis of the contribution of characteristic parameters: Apply statistical analysis and model predictive control methods to quantitatively evaluate the contribution and sensitivity of each characteristic parameter in oil film detection in polar ice areas. This step is completed by establishing a multivariate regression model to improve the accuracy of the characteristic parameter adjustment process and detection accuracy.

[0076] Based on the results of feature analysis, we adopt iterative optimization research methods to develop and implement mathematical models to accurately analyze the complex coupling relationship between targets and features. At the same time, we study and develop joint processing algorithms for GNSS direct and reflected signals to optimize the signal decoding and analysis process.

[0077] Design optimization and hardware adaptability of GNSS-R receivers: Based on the requirements of the model and signal processing algorithm of this application, the design of the GNSS-R receiver is optimized to ensure that the hardware can efficiently process and analyze complex signals in polar environments.

[0078] The present invention makes the following improvements on the basis of general GNSS-R reception:

[0079] (1) Adding a power amplifier gain module: The power amplifier gain module mainly uses the low noise amplifier, filter and frequency converter of the RF front end to amplify, filter and convert the received direct and reflected signals into intermediate frequency signals. For the reception of GPS satellite signals, the RF front end is required to set up precise frequency conversion, amplification, filtering and gain control circuits. The on-chip phase-locked loop generates a 2.456MHz local oscillator signal, which is mixed with the received 1575.42MHz signal to generate an 880.58MHz signal, which is then mixed with the 927MHz local oscillator signal to generate a 46.42MHz analog intermediate frequency signal.

[0080] (2) Adding a pulse code modulation module: The pulse code modulation module uses a high-speed A / D converter to sample, quantize and encode the above intermediate frequency signal, and converts the analog intermediate frequency signal into a digital signal. In addition, guide the adaptive upgrading and transformation of the hardware end to meet the operating requirements under extreme environmental conditions.

[0081] High-fidelity simulation and field demonstration are used to verify the scientificity, effectiveness and operability of the technical solution. This step ensures that all theoretical analysis and hardware transformations meet the expected results in practical applications. At the same time, the present invention evaluates the detection error, and the accuracy rate of oil in ice recognition is over 90%.

[0082] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A shipborne GNSS-R polar ice environment oil spill detection method, characterized by: The following steps are involved: Establish a scattering transmission model of polar ice regions based on GNSS-R signals on the surface of seawater, sea ice, and oil film; Based on the coherent signal and incoherent signal accumulation processing method, the oil film existence area is identified; Feature extraction of GNSS-R signals of sea surface ice-oil mixture was performed to determine whether there was oil spill in the polar ice environment.

2. The method for constructing a polar ice region sea surface scattering model according to claim 1, further comprising: The determination is based on a scattering transmission model of the polar ice area based on the GNSS-R signal on the surface of seawater, sea ice, and oil film, including the Kubelka-Munk model, the Kirk-Munk model, or the Gaussian mixture model; or a combination of any two of the above three models.

3. The method for optimizing coherent / incoherent signal accumulation processing according to claim 1, further characterized in that: The process of optimizing the coherent signal and the incoherent signal by using the accumulation processing method is as follows: For coherent signals, the general noise model is optimized; Decompose the coherent signal into frequency domain information of different scales; A Gaussian filter is used to remove noise in high-frequency components; For incoherent signals, the signal strength is improved by accumulating the square sum calculation, ignoring the phase information of the signal and retaining only the modulus value. The influence of phase error and frequency offset is eliminated by accumulating and subtracting the overlapping parts.

4. The method for extracting GNSS-R signal characteristics of sea ice-oil mixture according to claim 1, further characterized by: The process of extracting features from the GNSS-R signal of the sea ice-oil mixture to determine whether there is oil spill in the polar ice environment is as follows: Differentiated signal feature analysis: Analyze the differentiated signal features of GNSS signals after being reflected by oil film, seawater and sea ice and the direct signal, and identify the signal disturbance caused by medium changes through phase analysis and amplitude comparison; distinguish the seawater scattering area and the sea ice scattering area from the GNSS signal reflected by the sea surface by comparing the difference between the complex dielectric constant of the seawater surface and the complex dielectric constant of the area covered with strong scatterers; Feature set construction and optimization: Based on the analysis of differentiated signal characteristics, a comprehensive feature set based on dielectric constant, reflectivity, signal-to-noise ratio and polarization characteristic parameters is constructed; the variable field decomposition method is used to adjust and determine the optimal value range of the feature parameters in the comprehensive feature set data; Quantitative analysis of the contribution of characteristic parameters: Apply statistical analysis and model predictive control methods to establish a multivariate regression model to quantitatively evaluate the contribution and sensitivity of each characteristic parameter in oil film detection in polar ice areas.

5. The method for extracting GNSS-R signal characteristics of sea ice-oil mixture according to claim 4, further characterized by: The process of adjusting and determining the optimal value range of the characteristic parameters in the comprehensive characteristic set data by using the variable field decomposition method is as follows: Each variable field is standardized using the normalization formula; Orthogonal decomposition of multivariate fields is achieved by calculating the covariance matrix; Different weights are assigned to dielectric constant, reflectivity, signal-to-noise ratio and polarization characteristics to optimize the comprehensive feature set.

6. A shipborne GNSS-R polar ice environment oil spill detection device, characterized by: include: Establishment module: used to establish the scattering transmission model of polar ice areas based on GNSS-R signals on the surface of seawater, sea ice and oil film; Identification module: Identify the oil film area based on the coherent signal and incoherent signal accumulation processing method; Determination module: used to extract features from the GNSS-R signal of the sea surface ice-oil mixture and determine whether there is oil spill in the polar ice environment.

7. A readable storage medium storing a program module, characterized in that: The program module is executed in a processor to implement the method according to any one of claims 1 to 5.

Images