A method and system for remotely identifying marine emulsified oil spills based on red shift of reflectance dominant peak wavelength

Abstract

The present application provides a kind of marine emulsified oil spill remote sensing identification method and system based on reflectivity dominant peak wavelength red shift, belong to emulsified oil spill remote sensing identification field.To solve the problem that the existing emulsified oil spill identification method focuses on the response change of single band or a few bands, pays insufficient attention to the relative change relationship between multiple characteristic peaks, and lacks the ability to use the change rule of dominant peak between multiple candidate bands to realize emulsified oil spill identification and concentration interval determination.The present application realizes the stable identification and concentration interval determination of emulsified oil spill by identifying the transition behavior of dominant peak between multiple discrete bands under different concentration conditions, guides the band configuration of remote sensing system, and improves the system identification capability.Experiments show that, compared with the fixed band method, the identification rate of low concentration area is improved by 30%, the identification rate of high concentration area is improved by 40%, the inversion error is controlled within ±15%, and the calculation process is simple, suitable for rapid processing of multispectral and hyperspectral remote sensing data.

Description

Technical Field

[0001] This invention relates to the field of remote sensing identification technology for emulsified oil spills, and more specifically, to a remote sensing identification method and system for marine emulsified oil spills based on reflectance-dominant peak wavelength redshift. Background Technology

[0002] Oil spills on the ocean surface are a significant form of pollution in the marine ecosystem. After crude oil enters the ocean, it undergoes processes such as diffusion, emulsification, oxidation, and biodegradation under the influence of various environmental factors, including waves, wind, temperature changes, and solar radiation. Emulsified oil spills, due to their high water content and stability, often persist on the sea surface for extended periods, with a diffusion area significantly larger than the initial oil film area, causing continuous impacts on marine ecosystems and resources. Therefore, rapid and accurate identification of emulsified oil spills is crucial for oil spill pollution monitoring and emergency response.

[0003] Remote sensing technology, with its advantages of wide coverage, fast acquisition speed, and ability to achieve large-scale continuous observation, has become an important technical means for monitoring oil spills on the sea surface. Based on different methods of utilizing spectral information, existing remote sensing identification methods for oil spills on the sea surface mainly include the following categories:

[0004] (1) Single-band threshold identification method

[0005] This method distinguishes oil films from seawater backgrounds by selecting a specific wavelength band and setting a threshold for its reflectivity or brightness value. For example, in the visible or near-infrared bands, wavelengths sensitive to oil films are selected, and target extraction is performed using empirical thresholds. This type of method is simple to implement and computationally efficient, but it is easily affected by changes in solar altitude angle, water turbidity, and sea surface wave conditions, resulting in relatively limited recognition stability.

[0006] (2) Multi-band ratio or combination criterion method

[0007] This method enhances the contrast between the oil film and the background by constructing ratios, differences, or linear combinations of two or more bands. For example, the ratio between near-infrared and short-wave infrared bands can improve oil spill detection capabilities to some extent. This type of method has better anti-interference capabilities than single-band methods, but its band combinations usually rely on empirical selection, and its adaptability to different concentrations and emulsification states remains limited.

[0008] (3) Identification methods based on characteristic bands or empirical models

[0009] This method typically selects several characteristic bands sensitive to oil film response based on experiments or experience, and combines them with statistical or regression models to achieve oil spill identification or concentration estimation. Under certain conditions, this type of method can achieve good identification results, but its selection of characteristic bands is mostly based on limited experimental data and lacks a systematic analysis of the response variation law of emulsified oil spills across multiple bands.

[0010] Existing studies have shown that emulsified oil spills typically exhibit multiple characteristic reflection peaks in the visible to short-wave infrared range, and the reflectance in different bands shows different trends with varying degrees of emulsification or oil concentration. However, most current technologies focus on the response changes in a single band or a few bands, paying insufficient attention to the relative changes between multiple characteristic peaks, and lacking a systematic study on the competitive enhancement phenomenon among multiple characteristic peaks with changes in concentration.

[0011] Furthermore, as the concentration of emulsified oil spills gradually increases, the rate of change of response in different characteristic bands may differ significantly, resulting in a certain band exhibiting the strongest response within a specific concentration range. This strongest response band can be considered the dominant peak, and its position may change among multiple discrete bands with varying concentration. However, currently, there is a lack of a method to identify emulsified oil spills and determine their concentration ranges by utilizing the variation of the dominant peak among multiple candidate bands. Summary of the Invention

[0012] The technical problem to be solved by this invention is:

[0013] To address the problem that existing methods for identifying emulsified oil spills focus on response changes in a single band or a few bands, pay insufficient attention to the relative changes between multiple characteristic peaks, and lack the ability to identify emulsified oil spills and determine concentration ranges by utilizing the variation patterns of the dominant peak among multiple candidate bands.

[0014] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows:

[0015] This invention provides a remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift, comprising the following steps:

[0016] S100. By identifying the transition behavior of the dominant peak between multiple discrete bands under different concentration conditions, stable identification and concentration range determination of emulsified oil spills are achieved, and the redshift law of the dominant peak is obtained.

[0017] S200, Spectral data acquisition: Acquire hyperspectral remote sensing data of the target area and obtain spectral response values ​​of the target area in multiple candidate bands;

[0018] S300, Dominant Peak Identification: The dominant peak is determined by comparing the spectral response values ​​of each candidate band.

[0019] S400, concentration range determination: Based on the dominant peak red shift law obtained in step S100, the concentration range of emulsified oil spill is determined according to the dominant peak band.

[0020] S500, dynamic band selection and oil spill area extraction: select the corresponding sensitive band according to the concentration range, perform threshold segmentation on the target area, and extract the emulsified oil spill area.

[0021] S600, Quantitative Inversion: Under the condition that the position of the dominant peak is known, the established reflectivity-concentration formula and reflectivity-thickness formula are called to perform quantitative inversion of oil spill concentration and oil spill thickness.

[0022] S700 Output Results: Outputs oil spill space distribution map, concentration level estimation, and thickness results.

[0023] Further, in step S100, the dominant peak redshift pattern is as follows: for low concentration, the candidate wavelength range is 810nm–850nm, and the dominant peak position is 830nm; for medium-low concentration, the candidate wavelength range is 1040nm–1080nm, and the dominant peak position is 1060nm; for medium concentration, the candidate wavelength range is 1100nm–1140nm, and the dominant peak position is 1120nm; and for higher concentration, the candidate wavelength range is 1240nm–1280nm, and the dominant peak position is 1260nm.

[0024] Further, in step S200, the candidate wavelength bands are 810nm–850nm, 1040nm–1080nm, 1100nm–1140nm, and 1240nm–1280nm.

[0025] Furthermore, in step S200, the hyperspectral remote sensing data of the target area is smoothed using the Savitzky-Golay filtering method.

[0026] Further, in step S300, the dominant peak band is obtained using the following formula:

[0027]

[0028] in: Indicates wavelength λ i Spectral response value at λ; i Indicates candidate bands; Indicates the dominant peak band.

[0029] Further, in step S600, the following are included:

[0030] S610, Spectral response model construction: The Monte Carlo radiative transfer model is used to simulate the spectral response of the emulsified oil spill system;

[0031] S620, Regression Model Construction

[0032] A model relating reflectance and concentration was established based on simulation data:

[0033]

[0034] Where R is reflectance, C is volume concentration, and the correlation coefficient Rc is... 2 =0.97;

[0035] A model relating reflectivity and thickness was established based on simulation data:

[0036]

[0037] Where R is reflectivity, H is the thickness of the oil emulsion layer, and the correlation coefficient R0 is... 2 =098;

[0038] S630, concentration and thickness inversion,

[0039] For low-concentration areas, the reflectance value of the 830 nm band is used, and the concentration is inverted by combining the corresponding regression model.

[0040] For high-concentration areas, the reflectance value of the 1260 nm band is used, and the concentration is inverted by combining the 1260 nm regression model.

[0041] If the emulsion layer thickness needs to be inverted, the inversion can be performed using a 1060 nm band reflectance and thickness regression model under known concentration conditions.

[0042] Furthermore, in step S700, the inversion results are output in raster form.

[0043] For the oil spill spatial distribution map, the areas where the emulsion exists are marked;

[0044] The oil spill concentration grading chart divides oil spill concentrations into four levels: 0.05%-0.2%, 0.2%-3%, 3%-7%, and 7%-15%.

[0045] For the oil spill thickness distribution map: output the thickness estimate within the thickness-sensitive range, which is 1cm-10cm.

[0046] A remote sensing identification system for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift is provided. The system has program modules corresponding to the steps described above, and executes the steps in the above-described remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift during runtime.

[0047] A computer-readable storage medium storing a computer program configured to, when invoked by a processor, implement the steps of a remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift.

[0048] Compared with the prior art, the beneficial effects of the present invention are:

[0049] 1. High recognition accuracy: By dynamically selecting the recognition band using the redshift characteristics of the reflection peak, the 830nm band is used in the low concentration region and the 1260nm band is used in the high concentration region. Compared with the fixed band method, the recognition rate in the low concentration region is improved by 30% and the recognition rate in the high concentration region is improved by 40%.

[0050] 2. Strong quantitative capability: By constructing a regression model of reflectivity with concentration and thickness, the joint quantitative inversion of oil concentration and oil layer thickness is realized, with the inversion error controlled within ±15%;

[0051] 3. The method has a clear structure and good compatibility: The method of this invention only requires comparison and judgment of the spectral responses of multiple candidate bands, without the need for complex high-dimensional spectral analysis or large-scale training process. The calculation process is simple and suitable for rapid processing of multispectral and hyperspectral remote sensing data. It is easy to deploy in engineering environments such as satellite ground processing systems and UAV real-time processing systems. Attached Figure Description

[0052] Figure 1 This is a flowchart of a remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift, as described in an embodiment of the present invention.

[0053] Figure 2 This is a graph showing the spectral reflectance of emulsified oil spills at different concentrations in an embodiment of the present invention.

[0054] Figure 3 This is a graph showing the spectral reflectance of emulsified oil spills of different thicknesses in an embodiment of the present invention.

[0055] Figure 4 These are comparative images showing the identification effects of different bands for different concentration areas in embodiments of the present invention. (a) is an AVIRIS remote sensing image of the Gulf of Mexico oil spill obtained on May 17, 2010; (b) is the identification result of a low-concentration area using the 830 nm band; (c) is the identification result of a low-concentration area using the 1060 nm band; (d) is the identification result of a low-concentration area using the 1260 nm band; (e) is the identification result of a high-concentration area using the 830 nm band; (f) is the identification result of a high-concentration area using the 830 nm band; and (g) is the identification result of a high-concentration area using the 1260 nm band.

[0056] Figure 5 The following are regression diagrams showing the relationship between reflectance and oil concentration and thickness in embodiments of the present invention, wherein (a) is a regression diagram showing the relationship between reflectance and oil concentration, and (b) is a regression diagram showing the relationship between reflectance and oil thickness;

[0057] Figure 6 This is a diagram showing the redshift transition relationship of the dominant characteristic band of emulsified oil spill under different concentration conditions in an embodiment of the present invention. Detailed Implementation

[0058] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0059] Specific Implementation Plan 1: Combining Figure 1 As shown, this invention provides a remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift, comprising the following steps:

[0060] S100: By identifying the transition behavior of the dominant peak between multiple discrete bands under different concentration conditions, stable identification and concentration range determination of emulsified oil spills can be achieved.

[0061] Experiments have shown that marine emulsified oil spills exhibit multiple stable spectral characteristic peak regions in the visible to shortwave infrared range, including but not limited to:

[0062] 830nm, preferably 810nm–850nm;

[0063] 1060nm, preferably 1040nm–1080nm;

[0064] 1120nm, preferably 1100nm–1140nm;

[0065] 1260nm, preferably 1240nm–1280nm;

[0066] 1700nm, preferably 1650nm–1750nm;

[0067] 2200nm, preferably 2150nm–2250nm;

[0068] Under different oil emulsion volume fractions, the spectral response change rates of each characteristic peak showed significant differences: at low concentrations, the short-wavelength characteristic peaks had stronger responses; as the oil concentration increased, the response enhancement rate of the long-wavelength characteristic peaks was significantly higher than that of the short-wavelength peaks, leading to a competitive enhancement relationship between different characteristic peaks, and ultimately causing the dominant peak to gradually redshift with increasing concentration.

[0069] Typical redshift patterns of the dominant peak are shown in Table 1 below:

[0070] Table 1. Dominant Peak Redshift Pattern

[0071]

[0072] S200, spectral data acquisition and data preprocessing,

[0073] Acquire hyperspectral remote sensing data of the target area, including spectral response values ​​of the target area in multiple candidate bands, with a wavelength range of 380 nm to 2500 nm; perform radiometric calibration and atmospheric correction on the image to obtain corrected target spectral reflectance data;

[0074] The candidate wavebands include at least: 810nm–850nm, 1040nm–1080nm, 1100nm–1140nm, and 1240nm–1280nm;

[0075] Spectral response data can be obtained through experimental measurements and theoretical model simulations.

[0076] The extracted hyperspectral remote sensing data were smoothed using the Savitzky-Golay filtering method with a filter window width of 5 and a polynomial order of 2 to reduce the influence of random noise.

[0077] S300, dominant peak identification

[0078] By comparing the spectral response values ​​of each candidate band, the band with the maximum response is determined:

[0079]

[0080] in: Indicates wavelength λ i Spectral response value at λ; i Indicates candidate bands; Indicates the dominant peak band;

[0081] S400, concentration range determination

[0082] Based on the dominant peak redshift pattern obtained in step S100, the concentration range of emulsified oil spill is determined according to the dominant peak band:

[0083] If the dominant peak is located in the 810–850 nm range, it is considered a low concentration range.

[0084] If the dominant peak is located in the 1040–1080 nm range, it is determined to be in the low to medium concentration range.

[0085] If the dominant peak is located in the 1100–1140 nm range, it is determined to be a medium concentration range;

[0086] If the dominant peak is located in the 1240–1280 nm range, it is considered to be in the higher concentration range.

[0087] S500, dynamic band selection and oil spill area extraction

[0088] Select the corresponding sensitive band based on the concentration range, perform threshold segmentation on the target area, and extract the emulsified oil spill area;

[0089] Among them, based on the average reflectance of this band, the threshold range is set to the mean ± 0.5 times the standard deviation. Pixels with reflectance higher than the upper limit of the threshold are identified as oil emulsion areas.

[0090] S600, quantitative inversion,

[0091] Given the location of the dominant peak, the established reflectivity-concentration formula and reflectivity-thickness formula are used to quantitatively invert the oil spill concentration and thickness.

[0092] include,

[0093] S610. Spectral response model construction: The Monte Carlo radiative transfer model is used to simulate the spectral response of the emulsified oil spill system. The model input parameters include:

[0094] Oil droplet size distribution: log-normal distribution, with a median droplet size of 58 μm;

[0095] The concentration range is 0.1% to 10%;

[0096] The incident angle ranges from 0° to 60°;

[0097] The Monte Carlo radiative transfer model used in this study is employed to simulate the spectral response characteristics of emulsion oil spill systems. This model is based on the principle of random propagation of photons in multi-scattering media. By statistically analyzing a large number of photons' scattering and absorption processes in the medium, it can effectively describe the multiple scattering and absorption behavior of light in emulsion systems. The Monte Carlo radiative transfer method has been widely used in fields such as water remote sensing, marine optics, and spectral simulation of emulsion media. It can accurately reflect the law of spectral response variation with medium composition in multiphase media and has a clear physical basis.

[0098] During model parameter setting, a parameter range consistent with the actual characteristics of marine emulsified oil spills was adopted, including oil droplet size distribution, concentration range, and incident angle range, so that the model can reflect the typical spectral variation characteristics under actual application conditions. During the Monte Carlo model establishment process, the stability of the statistical results was ensured by increasing the number of photons and performing multiple repeated calculations. In a single simulation, the number of photons was set to 10. 7The model is designed to reduce the impact of random errors on the calculation results. Furthermore, the position and trend of the characteristic peaks output by the model are consistent with the typical spectral variation patterns of emulsified oil spills reported in existing literature, indicating that the constructed model has good physical consistency and reliability. Therefore, the spectral response data obtained based on the above model can provide reliable data support for the identification of the dominant peak and the determination of the concentration range described in this invention.

[0099] S620, Regression Model Construction

[0100] A model for the relationship between reflectance and concentration was established based on simulation data (1260 nm, incident angle 0°):

[0101]

[0102] Where R is reflectance, C is volume concentration (%), and the correlation coefficient Rc is... 2 =0.97;

[0103] A model relating reflectivity to thickness was established based on simulation data (1060 nm, concentration 0.5%, incident angle 0°):

[0104]

[0105] Where R is reflectivity, H is the thickness of the oil emulsion layer (cm), and the correlation coefficient R0 is... 2 =098;

[0106] S630, concentration and thickness inversion,

[0107] For low-concentration regions (dominant peak at 830 nm), use the reflectance value of the 830 nm band and combine it with the corresponding regression model to retrieve the concentration.

[0108] For high-concentration regions (dominant peak at 1260 nm), the reflectance value of the 1260 nm band should be used first, and the concentration should be inverted by combining the 1260 nm regression model.

[0109] If it is necessary to invert the emulsion layer thickness, under the condition of known concentration, the inversion can be performed using a 1060 nm band reflectance and thickness regression model;

[0110] S700, Result Output

[0111] Output an oil spill spatial distribution map and concentration level estimation results;

[0112] Output the inversion results as a raster image:

[0113] Oil spill spatial distribution map: indicates the area where the emulsion is present;

[0114] Oil spill concentration grading chart: The concentration is divided into four levels: 0.05% ~ 0.2%, 0.2% ~ 3%, 3% ~ 7%, and 7% ~ 15%.

[0115] Oil spill thickness distribution map: Outputs thickness estimates within the thickness-sensitive range (1 cm ~ 10 cm).

[0116] The inversion results were compared with field measurement data or visual interpretation results of high-resolution images. In areas with known concentrations, the relative error between the inverted concentration and the measured concentration was within ±15%; the thickness inversion error was within ±1.5 cm. Compared with traditional single-band inversion methods that do not employ redshift features, this invention improves the identification accuracy in low-concentration and high-concentration areas by 30% and 40%, respectively.

[0117] The method of this invention can be deployed on ground-based remote sensing data processing workstations, satellite ground receiving station processing systems, and UAV-borne hyperspectral real-time processing terminals. Required data sources include: reflectance data acquired by spaceborne (e.g., AVIRIS, PRISMA, EnMAP), airborne, or UAV-borne hyperspectral imagers, as well as corresponding geometric and atmospheric correction data.

[0118] This invention is not dependent on a specific oil type. While the specific positions of each characteristic peak may shift under different crude oil types, the pattern of transitions between multiple discrete bands in the dominant peak remains consistent. Therefore, in practical applications, the candidate band range can be calibrated based on the characteristics of the oil in the target sea area, and the identification methods and processing procedures fall within the scope of this invention.

[0119] The remote sensing identification and quantitative inversion method for emulsified oil spills proposed in this invention can be widely applied in marine environmental monitoring, oil spill emergency response, and oil spill pollution assessment. This method can utilize existing or future hyperspectral / multispectral remote sensing data to achieve high-precision identification of oil spill areas and quantitative inversion of concentrations, providing a scientific basis for oil spill disaster assessment and mitigation decisions. Simultaneously, the band optimization method proposed in this invention can provide theoretical guidance for the band design of novel marine remote sensing payloads.

[0120] Specific Implementation Scheme 2: The present invention provides a remote sensing identification system for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift. This system has program modules corresponding to the above steps, and executes the steps in the above-mentioned remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift when running.

[0121] The other combinations and connections in this implementation scheme are the same as in Specific Implementation Scheme 1.

[0122] Specific Implementation Scheme 3: The present invention provides a computer-readable storage medium storing a computer program configured to, when called by a processor, implement the steps of a remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift.

[0123] The other combinations and connections in this implementation scheme are the same as in Specific Implementation Scheme 1.

[0124] Example 1

[0125] Figure 2 This is a graph showing the spectral reflectance of water-in-oil emulsion spills under different oil concentrations. The horizontal axis represents wavelength (nm), and the vertical axis represents reflectance. Figure 2 The image shows the reflectance spectra of oil concentrations of 0.1%, 0.5%, 1%, 2%, 5%, and 10%. The arrows indicate that as the concentration increases, the maximum reflectance peak undergoes discrete transitions from approximately 830 nm to 1060 nm, 1120 nm, and 1260 nm, reflecting the enhanced competition among multiple characteristic peaks and the migration of the dominant peak.

[0126] Example 2

[0127] Figure 3 The spectral reflectance curves of oil emulsions (concentration C = 0.5%) at different thicknesses are shown. Two peaks in reflectance are observed near 830 nm and 1060 nm, with the main peak located at 1060 nm. Unlike the effect of crude oil concentration, the position of the main reflectance peak does not exhibit a significant redshift with increasing oil emulsion thickness, remaining consistently near 1060 nm. The reflectance change is most pronounced near the 1060 nm wavelength band with increasing oil emulsion thickness.

[0128] The area proportion method was used to evaluate the recognition effect.

[0129] In the low concentration region (Run 9), the emulsion area identified using 831.52 nm accounted for 0.678% of the total area, while using 1263.26 nm it was only 0.243%, with the latter missing approximately 64%.

[0130] In the high-concentration region (Run 14), the recognition area ratio was 0.058% when using 831.52 nm, which increased to 0.249% when using 1263.26 nm, with the former missing about 77%.

[0131] The verification results show that selecting the identification band based on the red shift state of the reflection peak can significantly improve the identification integrity of different concentration regions.

[0132] Example 3

[0133] Figure 4The comparison charts show the effectiveness of using different wavelengths for identification in different concentration regions. The results indicate that in the low concentration region (Run 9), the identified emulsion area accounted for 0.678% of the total area when using 831.52 nm, but only 0.243% when using 1263.26 nm, with the latter missing approximately 64%. In the higher concentration region (Run 14), the identified area accounted for 0.058% when using 831.52 nm, but this increased to 0.249% when using 1263.26 nm, with the former missing approximately 77%. These results demonstrate that based on the dominant peak redshift pattern, the identification completeness in different concentration regions can be significantly improved.

[0134] Example 4

[0135] Figure 5 The graph shows the regression relationship between reflectance and oil concentration. The horizontal axis represents oil concentration and thickness, and the vertical axis represents reflectance. The fitted relationship is marked on the graph.

[0136] Example 5

[0137] Combination Figure 6 As shown, with increasing emulsion concentration, the dominant reflection peak transitions between multiple candidate wavelength bands. At low concentrations (approximately 0.1%), the dominant peak is located around 830 nm; as the concentration increases to approximately 0.5%, the dominant peak transitions to approximately 1060 nm; as the concentration further increases to approximately 5%, the dominant peak moves to approximately 1120 nm; and at even higher concentrations (approximately 10%), the dominant peak further transitions to approximately 1260 nm. This transition relationship indicates that the wavelength position of the dominant peak can serve as an important basis for determining the emulsion concentration range, thus providing a foundation for subsequent quantitative estimation of oil spills.

[0138] Table 2. Correspondence between dominant characteristic bands and emulsion concentration ranges

[0139]

[0140] Based on the spectral simulation results under different emulsion concentration conditions, the dominant characteristic bands corresponding to each concentration range were summarized and the correspondence between the dominant bands and concentration ranges is shown in Table 2. In the low concentration range (0.05%–0.2%), the dominant characteristic peak is mainly located around 830 nm; as the concentration increases to the 0.2%–1% range, the dominant peak gradually shifts to around 1060 nm; when the concentration further increases to the 1%–7% range, the dominant peak stably appears around 1120 nm; in the higher concentration range (7%–15%), the dominant peak further shifts to around 1260 nm. Therefore, in the actual remote sensing identification process, by detecting the maximum response band of the target area among multiple candidate bands and combining it with the interval correspondence in Table 1, the concentration range of the emulsion can be quickly determined.

[0141] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift, characterized in that, Includes the following steps: S100. By identifying the transition behavior of the dominant peak between multiple discrete bands under different concentration conditions, stable identification and concentration range determination of emulsified oil spills are achieved, and the redshift law of the dominant peak is obtained. S200. Spectral data acquisition: Acquire hyperspectral remote sensing data of the target area and obtain spectral response values ​​of the target area in multiple candidate bands. S300, Dominant Peak Identification: The dominant peak is determined by comparing the spectral response values ​​of each candidate band. S400, concentration range determination: Based on the dominant peak red shift law obtained in step S100, the concentration range of emulsified oil spill is determined according to the dominant peak band. S500, dynamic band selection and oil spill area extraction: select the corresponding sensitive band according to the concentration range, perform threshold segmentation on the target area, and extract the emulsified oil spill area. S600, Quantitative Inversion: Under the condition that the position of the dominant peak is known, the established reflectivity-concentration formula and reflectivity-thickness formula are called to perform quantitative inversion of oil spill concentration and oil spill thickness. include, S610, Spectral response model construction: The Monte Carlo radiative transfer model is used to simulate the spectral response of the emulsified oil spill system; S620, Regression Model Construction A model relating reflectance and concentration was established based on simulation data: Where R is reflectance, C is volume concentration, and the correlation coefficient Rc is... 2 =0.97; A model relating reflectivity and thickness was established based on simulation data: Where R is reflectivity, H is the thickness of the oil emulsion layer, and the correlation coefficient R0 is... 2 =098; S630, concentration and thickness inversion, For low-concentration areas, the reflectance value of the 830 nm band is used, and the concentration is inverted by combining the corresponding regression model. For high-concentration areas, the reflectance value of the 1260 nm band is used, and the concentration is inverted by combining the 1260 nm regression model. If it is necessary to invert the emulsion layer thickness, under the condition of known concentration, the 1060 nm band reflectance and thickness regression model is used for inversion; S700 Output Results: Outputs oil spill space distribution map, concentration level estimation, and thickness results.

2. The remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift as described in claim 1, characterized in that: In step S100, the dominant peak redshift pattern is as follows: for low concentration, the candidate wavelength range is 810nm–850nm, and the dominant peak position is 830nm; for medium-low concentration, the candidate wavelength range is 1040nm–1080nm, and the dominant peak position is 1060nm; for medium concentration, the candidate wavelength range is 1100nm–1140nm, and the dominant peak position is 1120nm; and for higher concentration, the candidate wavelength range is 1240nm–1280nm, and the dominant peak position is 1260nm.

3. The remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift as described in claim 2, characterized in that: In step S200, the candidate wavelength bands are 810nm–850nm, 1040nm–1080nm, 1100nm–1140nm, and 1240nm–1280nm.

4. The remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift as described in claim 3, characterized in that: In step S200, the hyperspectral remote sensing data of the target area is smoothed using the Savitzky-Golay filtering method.

5. The remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift as described in claim 4, characterized in that: In step S300, the dominant peak band is obtained using the following formula: in: Indicates wavelength λ i Spectral response value at λ; i Indicates candidate bands; Indicates the dominant peak band.

6. The remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift as described in claim 5, characterized in that: In step S700, the inversion results are output in raster form. For the oil spill spatial distribution map, the areas where the emulsion exists are marked; The oil spill concentration grading chart divides oil spill concentrations into four levels: 0.05%-0.2%, 0.2%-3%, 3%-7%, and 7%-15%. For the oil spill thickness distribution map: output the thickness estimate within the thickness-sensitive range, which is 1 cm-10 cm.

7. A remote sensing identification system for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift, characterized in that: The system has a program module corresponding to the steps of any one of the claims 1-6 above, and executes the steps in the above-described remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift when it is run.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program configured to, when invoked by a processor, implement the steps of any one of claims 1-6, a remote sensing identification method for marine emulsified oil spills based on reflectivity-dominated peak wavelength redshift.

Images