A red tide identification method and device based on remote sensing data

Abstract

The present application provides a red tide identification method and device based on remote sensing data, belonging to the field of marine monitoring and remote sensing applications. The method provided by the present application includes: comparing the spectral reflectance curves of diatom algal solutions with different concentrations to determine the sensitive bands of red tide water bodies, where the sensitive bands include multiple ones; obtaining remote sensing data of a target sea area, and extracting the remote sensing reflectance corresponding to the sensitive bands in the remote sensing data; constructing a red tide inversion model; inputting the remote sensing reflectance under each of the sensitive bands into the red tide inversion model to calculate the real-time red tide index; and judging whether there is a red tide in the target sea area according to the numerical relationship between the real-time red tide index and the characteristic interval of the red tide index. The red tide identification method and device based on remote sensing data provided by the present application can more accurately reflect the differences between red tide and non-red tide water bodies, and improve the accuracy of identification.

Description

Technical Field

[0001] This application relates to the technical field of marine monitoring and remote sensing applications, and particularly to a red tide identification method and device based on remote sensing data. Background Art

[0002] A red tide is a natural phenomenon in which marine plankton reproduce explosively, causing the seawater to change color. In recent years, due to the increase in the use of agricultural fertilizers and the development of offshore aquaculture, organic nutrients such as nitrogen and phosphorus and soluble organic matter in the coastal seawater have become enriched, creating favorable conditions for the growth of algae, resulting in a significant increase in the frequency and scope of red tides. The frequent occurrence of red tides has caused serious damage to the marine ecosystem, leading to the massive death of fish, shrimps, and shellfish due to hypoxia or poisoning. Therefore, effective monitoring and research on red tides are of great practical significance for preventing red tides and reducing their harm.

[0003] Currently, satellite remote sensing technology has become an important means for red tide monitoring. Traditional remote sensing red tide identification methods mainly rely on a single spectral index, and its principle is to use the difference in reflectance between red tide organisms and normal seawater in specific spectral bands for judgment. This method can identify red tides to a certain extent and has the advantages of a wide monitoring range and high efficiency, being able to quickly obtain information on a large area of sea areas and providing preliminary red tide monitoring data for marine management departments.

[0004] However, this remote sensing red tide identification technology based on a single spectral index has obvious limitations. The marine environment is extremely complex, and the surrounding environmental factors have a huge impact on spectral reflectance. In the coastal sea areas, the concentration of suspended sediment is relatively high, and the reflection characteristics of sediment will interfere with the red tide signal, resulting in misjudgment. When encountering extreme weather such as storms and heavy rains, the optical properties of the water body will change violently, and at this time, a single spectral index is difficult to accurately distinguish red tide water bodies from normal water bodies affected by the weather. Moreover, the water quality, water depth, types and concentrations of plankton in different sea areas vary greatly, and a single spectral index cannot adapt to these complex changes, greatly reducing the accuracy of red tide identification. Summary of the Invention

[0005] In view of this, this application provides a red tide identification method and device based on remote sensing data, which can more accurately reflect the differences between red tide and non-red tide water bodies and improve the accuracy of identification.

[0006] Specifically, this application is implemented through the following technical solutions:

[0007] The first aspect of this application provides a red tide identification method based on remote sensing data, and the method includes:

[0008] Comparing the spectral reflectance curves of diatom algal solutions with different concentrations to determine the sensitive bands of red tide water bodies, and the sensitive bands include multiple;

[0009] Obtain remote sensing data of the target sea area, and extract the remote sensing reflectance corresponding to the sensitive bands in the remote sensing data;

[0010] Construct a red tide inversion model; wherein, the red tide inversion model is a function with the ratio of remote sensing reflectance as an independent variable, the independent variable of the function is the ratio of the sum of the first remote sensing reflectance and the second remote sensing reflectance to the third remote sensing reflectance, the function is obtained by taking the difference between the independent variable and a constant term, and the sensitive bands corresponding to the first remote sensing reflectance, the second remote sensing reflectance, and the third remote sensing reflectance are different;

[0011] Input the remote sensing reflectance at each of the sensitive bands into the red tide inversion model, and calculate the real-time red tide index, where the real-time red tide index is used to characterize the density of red tides in the target sea area;

[0012] Judge whether there is a red tide in the target sea area according to the numerical relationship between the real-time red tide index and the characteristic interval of the red tide index.

[0013] The second aspect of the present application provides a red tide recognition device based on remote sensing data, and the device includes a determination module, an extraction module, a construction module, a calculation module, and a judgment module;

[0014] Among them, the determination module is used to compare the spectral reflectance curves of diatom algal solutions with different concentrations, and determine the sensitive bands of red tide water bodies, and the sensitive bands include multiple;

[0015] The extraction module is used to obtain remote sensing data of the target sea area, and extract the remote sensing reflectance corresponding to the sensitive bands in the remote sensing data;

[0016] The construction module is used to construct a red tide inversion model; wherein, the red tide inversion model is a function with the ratio of remote sensing reflectance as an independent variable, the independent variable of the function is the ratio of the sum of the first remote sensing reflectance and the second remote sensing reflectance to the third remote sensing reflectance, the function is obtained by taking the difference between the independent variable and a constant term, and the sensitive bands corresponding to the first remote sensing reflectance, the second remote sensing reflectance, and the third remote sensing reflectance are different;

[0017] The calculation module is used to input the remote sensing reflectance at each of the sensitive bands into the red tide inversion model, and calculate the real-time red tide index, where the real-time red tide index is used to characterize the density of red tides in the target sea area;

[0018] The judgment module is used to judge whether there is a red tide in the target sea area according to the numerical relationship between the real-time red tide index and the characteristic interval of the red tide index.

[0019] The red tide identification method and device based on remote sensing data provided by this application establish a red tide inversion model in the form of a function with the remote sensing reflectance ratio as the independent variable according to the measured data, perform operations with the remote sensing reflectance of sensitive bands as the independent variable, can effectively integrate multi-band information, more accurately establish the correlation with the red tide density, and the calculated real-time red tide index can accurately quantify the severity of the red tide, ensuring the accuracy of red tide identification with a high-precision inversion model. In the first aspect, by comparing the spectral reflectance curves of diatom algal solutions with different concentrations, the sensitive bands of red tide water bodies are determined, ensuring the accuracy of the data input into the inversion model and its strong correlation with red tide water bodies, and being able to more accurately reflect the characteristics of red tide water bodies.

[0020] In the second aspect, according to the numerical relationship between the real-time red tide index and the characteristic interval of the red tide index, it is judged whether there is a red tide. Compared with the traditional single spectral index identification method, the misjudgment caused by complex marine environmental factors (such as suspended sediment, weather changes, etc.) is greatly reduced, and the accuracy of red tide identification is significantly improved. Brief Description of the Drawings

[0021] Figure 1 It is a flowchart of the first embodiment of the red tide identification method based on remote sensing data provided by this application;

[0022] Figure 2 It is the reflection spectral curves of diatoms with different concentrations at different wavelengths shown by this application;

[0023] Figure 3 It is an image shown by this application without red tide index calculation;

[0024] Figure 4 It is an image shown by this application after red tide index calculation;

[0025] Figure 5 It is a schematic structural diagram of the second embodiment of the red tide identification device based on remote sensing data provided by this application. Detailed Description of the Embodiment

[0026] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0027] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms "a", "the", and "said" used in this application are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0028] It should be understood that although the terms first, second, third, etc. may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of this application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "when" or "while" or "in response to a determination".

[0029] Specific embodiments are given below to introduce the technical solutions of this application in detail.

[0030] Embodiment 1:

[0031] Figure 1 This is a flowchart of Embodiment 1 of the red tide identification method based on remote sensing data provided for this application. Please refer to Figure 1 , the method provided in this embodiment may include:

[0032] S101. Compare the spectral reflectance curves of diatom algal solutions with different concentrations to determine the sensitive bands of the red tide water body, and the sensitive bands include multiple ones.

[0033] It should be noted that diatoms are a common type of planktonic algae and are widely present in the marine ecosystem. In the red tide phenomenon, diatoms are one of the main triggering organisms. Their cell structure and pigment composition cause specific optical property changes under different environments, and these changes are closely related to the occurrence and development of red tides. The spectral reflectance curve reflects the ability of an object to reflect light of different wavelengths. For a diatom algal solution, the spectral reflectance curve depicts the intensity change of the reflected light of the algal solution under the irradiation of light with different wavelengths. Diatom algal solutions with different concentrations have different absorption and scattering characteristics of light due to different cell numbers and physiological states, and thus have spectral reflectance curves with different shapes. The sensitive bands of the red tide water body refer to those specific wavelength ranges of light bands that respond significantly to the changes in the red tide water body. At these bands, the spectral reflectance differences between the red tide water body and normal seawater are obvious, and by monitoring the reflectance changes in these bands, the occurrence and development of red tides can be effectively identified.

[0034] It should be noted that diatom algal solutions with different concentrations can be obtained through experimental cultivation. Specifically, natural seawater is collected from the target red tide area and filtered through a 0.45 µm cellulose acetate membrane to remove particulate impurities. Then, nutrients are added according to the f / 2 medium formula to prepare the medium. After autoclaving the medium, it is placed in an incubator at 25 °C, with the light intensity set at 5500 lux and a light-dark cycle of 12 hours for cultivation. After the algae grow stably, 5, 10, 20, and 100 ml of the algal solution are respectively added to four conical flasks, and then made up to 100 ml with sterilized seawater to obtain diatom algal solution samples with different concentrations.

[0035] After obtaining the diatom algal solution samples with different concentrations, a professional spectral measurement instrument, such as a spectrometer, can be used to measure the different concentrations of diatom algal solution samples to obtain the reflectance data of each sample at different wavelengths. Then, with the wavelength as the abscissa and the reflectance as the ordinate, the spectral reflectance curves of diatom algal solutions with different concentrations are plotted. Furthermore, carefully compare the characteristics of the spectral reflectance curves of diatom algal solutions with different concentrations, and find those wavelength regions that show obvious differences in terms of curve shape, reflectance value, etc. with the change of algal solution concentration, that is, determine the wavelength region where the distribution gap of the ordinate values of the curves at different wavelengths at the same wavelength is greater than the preset value. The bands corresponding to these regions are the sensitive bands of the red tide water body.

[0036] Specifically, diatoms, as the main organisms causing red tides, contain photosynthetic pigments such as chlorophyll in their cells. At different concentrations, the number and distribution of diatom cells in the algal solution are different, and the absorption and scattering of light are also different. For example, at certain wavelengths, chlorophyll strongly absorbs light, resulting in a decrease in reflectance; while at other wavelengths, the scattering effect of light dominates and the reflectance increases. By comparing the spectral reflectance curves of algal solutions with different concentrations, those wavelength regions that are sensitive to the change of diatom concentration, that is, the sensitive bands, can be found. The change in reflectance of these sensitive bands can reflect the change in the concentration of diatom algal solution, and thus reflect the development degree of the red tide.

[0037] Optionally, the determination of the sensitive band includes: selecting multiple sampling points in the target sea area, where the sampling points include waters with different turbidities, clear waters between specified latitudes and longitudes, and areas with high suspended sediment concentrations near the shore of the target sea area; obtaining the latitudes and longitudes and remote sensing reflectance data of the sampling points based on remote sensing data; cultivating the diatom samples collected from the target red tide area based on the latitude and longitude information, measuring the spectral reflectance of algal solutions with different concentrations, and combining the remote sensing reflectance data to determine the sensitive band of the red tide.

[0038] Figure 2 For the reflectance spectral curves of diatoms with different concentrations shown in this application, please refer to Figure 2, the red tide water body shows obvious two absorption peaks and two reflection peaks. Among them, the absorption peaks are located in the 440 - 490 nm and 650 - 670 nm bands; the reflection peaks appear in the 560 - 580 nm and 690 - 710 nm bands. Since the red tide water body shows high sensitivity in these specific bands, these bands are regarded as sensitive bands.

[0039] S102. Obtain the remote sensing data of the target sea area, and extract the remote sensing reflectance corresponding to the sensitive bands in the remote sensing data.

[0040] It should be noted that the remote sensing reflectance refers to the ratio of the reflected radiation brightness received by the sensor from the water body of the target sea area to the incident radiation brightness in remote sensing observation. It reflects the reflection ability of the water body of the target sea area to light of different bands and is an important parameter for measuring the optical properties of the water body. By analyzing the remote sensing reflectance, information such as the composition and state of the water body can be understood.

[0041] Specifically, a suitable remote sensing satellite or airborne remote sensing platform can be selected. Currently, there are many satellites dedicated to ocean observation, such as the ocean color satellite series and Ocean No. 1 satellite, etc., which can obtain ocean remote sensing data with different spatial resolutions and temporal resolutions. According to factors such as the scope of the target sea area and the monitoring frequency requirements, determine which satellite data source or sources to adopt. In specific implementation, apply for and download the remote sensing data of the target sea area for a specific period according to the data acquisition rules and processes of the satellite or airborne remote sensing platform. Generally, it can be operated through the official websites of each satellite data center or relevant data distribution platforms. Further, according to the previously determined sensitive bands of the red tide water body (such as 443 nm, 490 nm, 565 nm, etc.), screen out the data corresponding to the corresponding bands in the obtained remote sensing data. Using remote sensing data processing software (such as ENVI, Erdas, etc.), after performing preprocessing operations such as radiometric correction on the screened sensitive band data, extract the remote sensing reflectance values corresponding to each sensitive band according to the calculation formula of the remote sensing reflectance or the built-in function of the software.

[0042] It should be noted that the sensor on the remote sensing satellite or airborne platform emits or receives electromagnetic waves of different bands. When the electromagnetic waves irradiate the water surface of the target sea area, part of them is absorbed by the water body, and part of them is scattered and reflected. The reflected electromagnetic waves are received by the sensor, and the sensor converts the received radiation signal into an electrical signal or a digital signal for recording, forming remote sensing data. In these data, the intensity of the reflected signals of different bands reflects the reflection ability of the water body to the light of that band, that is, the remote sensing reflectance. By specific algorithms and processing procedures, the reflected signals of the sensitive bands are separated from the remote sensing data, and the corresponding remote sensing reflectance is calculated, which can provide key optical parameters for subsequent red tide analysis.

[0043] S103. Construct a red tide inversion model.

[0044] Among them, the red tide inversion model is a function with the remote sensing reflectance ratio as the independent variable. The independent variable of the function is the ratio of the sum of the first remote sensing reflectance and the second remote sensing reflectance to the third remote sensing reflectance. The function is obtained by subtracting the constant term from the independent variable. The sensitive bands corresponding to the first remote sensing reflectance, the second remote sensing reflectance, and the third remote sensing reflectance are different.

[0045] It should be noted that based on the relationship between red tide organisms and water body optical properties, a red tide inversion model can be established. The presence and concentration changes of red tide organisms (such as planktonic algae like diatoms) will change the absorption and scattering characteristics of the water body to light, thereby affecting the remote sensing reflectance. The red tide inversion model establishes a mathematical relationship between the remote sensing reflectance in sensitive bands and red tide-related parameters such as red tide organism concentration, and uses the known remote sensing reflectance data to inversely deduce the relevant characteristic information of the red tide. For example, through research, it is found that there is a certain functional relationship between the remote sensing reflectance ratio in certain sensitive bands and the red tide concentration. The red tide inversion model is constructed based on this relationship to achieve the inversion from optical data to red tide characteristics.

[0046] The core of constructing the red tide inversion model lies in using the relationship between red tide organisms and water body optical properties to establish a mathematical connection between the remote sensing reflectance in sensitive bands and red tide-related parameters, so as to achieve the inversion from optical data to red tide characteristics. First, based on the obtained remote sensing data and red tide-related parameters of the entire target sea area, fitting is carried out to construct a global overall model. Considering that there are differences in the interference characteristics and marine environmental conditions in different regions of the target sea area, in order to improve the model accuracy, a sub-region modeling method is also used to further optimize the model. The specific steps are as follows:

[0047] (1) Identify the boundary regions based on geographic information data, marine environmental characteristics, and historical interference data.

[0048] Boundary regions usually have unique geographical and environmental characteristics, being affected by both land and sea, with complex interference factors. Using geographic information data (such as coastline position, topography, etc.), marine environmental characteristics (such as water flow, salinity, etc.), and historical interference data (such as past storms, river inputs, etc.), these special boundary regions are identified.

[0049] (2) Divide the regions in the target sea area except the boundary regions into multiple sub-regions.

[0050] Specifically, dividing the regions in the target sea area except the boundary regions into multiple sub-regions includes:

[0051] (i)Collect the marine environmental characteristic data of the target sea area, where the marine environmental characteristic data includes at least water flow velocity field data, suspended matter concentration data, wind field data, temperature data, salinity data, and chlorophyll a concentration data.

[0052] It should be noted that the water flow velocity field data can reflect the flow of seawater, and different water flow velocities will affect the diffusion and aggregation of red tide organisms. The wind field data can reveal the wind-driven effect on the sea surface and regulate the vertical distribution and horizontal migration of red tide organisms by affecting the depth of the seawater mixed layer and the nutrient transport. The distribution data of the suspended matter concentration can reflect the turbidity of seawater and may affect the growth and reproduction of red tide organisms. The temperature data can characterize the thermal condition of seawater, directly affect the metabolism, reproduction, aggregation distribution, and species succession of red tide organisms, and is a key indicator for predicting the occurrence and development of red tides and identifying dominant species. The salinity data is an important indicator of the marine environment, and different salinity environments are suitable for the survival of different types of red tide organisms. The chlorophyll a concentration data is closely related to the biomass of phytoplankton and is an important reference for judging the occurrence of red tides.

[0053] (ii)Calculate the gradient changes of the water flow velocity, the spatial gradient changes of the suspended matter concentration, the spatial gradient changes of the wind speed and direction, the spatial gradient changes of the temperature, the spatial gradient changes of the salinity, and the spatial gradient changes of the chlorophyll a concentration, and divide the area other than the boundary area in the target sea area into multiple sub-areas according to the gradients of the comprehensive changes.

[0054] It should be noted that by calculating the gradient changes of the water flow velocity in different areas, the areas with large gradient changes indicate significant changes in the water flow, which have a greater impact on the distribution of red tide organisms. By evaluating the spatial gradient changes of the suspended matter concentration, the areas with significant concentration changes can be found, and the red tide conditions in these areas are different from those in other areas. By analyzing the spatial distribution characteristics of the wind field data, the seawater transport paths driven by the wind can be revealed, and there are significant differences in the diffusion range and aggregation form of red tides under different wind field conditions. By identifying the gradient changes of the temperature in different areas, the areas with large gradient changes indicate the existence of temperature mutation zones, which have a significant impact on the aggregation of red tide organisms. By studying the spatial distribution differences of the salinity, the areas with obvious salinity changes can be identified, and the red tide organism communities in different salinity areas are different. By considering the spatial gradient changes of the chlorophyll a concentration, the areas with large change rates are the key areas for the development or recession of red tides.

[0055] To comprehensively consider these factors and calculate the gradient of the comprehensive change, a weighted sum calculation form is adopted. Let the gradient change of the water flow velocity field be G1, the gradient change of the suspended matter concentration be G2, the gradient change of the wind field be G3, the gradient change of the temperature be G4, the gradient change of the salinity be G5, and the gradient change of the chlorophyll a concentration be G6. The corresponding weights are w1, w2, w3, w4, w5, and w6 respectively. The weights are related to the distribution characteristics of the parameters corresponding to each gradient change (such as water flow velocity, suspended matter concentration, wind field, temperature, salinity, chlorophyll a concentration) in the target sea area. For example, in areas with more significant characteristics, the corresponding gradient change weights are relatively larger. The calculation formula for the comprehensive change gradient G is: G = w1G1 + w2G2 + w3G3 + w4G4 + w5G5 + w6G6. According to the calculated comprehensive change gradient G, the area in the target sea area except the boundary area is divided into multiple sub-areas.

[0056] Furthermore, according to the modeling parameter sets determined for each sub-area, corresponding inversion models are established respectively. Specifically, for each sub-area, according to its unique environmental characteristics and data features, appropriate modeling parameter sets are determined, and then these parameter sets are used to establish their respective inversion models to ensure that the models of each sub-area can accurately reflect the red tide situation in that area.

[0057] (3)Based on the historical remote sensing data and red tide concentration of the sub-areas, establish the sub-area inversion models.

[0058] The historical remote sensing data of each sub-area contains the remote sensing reflectance data of sensitive bands in this area at different times. Combining with the corresponding red tide concentration data, through appropriate mathematical methods (such as regression analysis, machine learning algorithms, etc.), establish the mathematical relationship between the remote sensing reflectance of sensitive bands and parameters such as red tide concentration in this sub-area, that is, the sub-area inversion model.

[0059] (4)Based on the historical remote sensing data and historical interference parameters of the boundary area, establish the boundary area inversion model.

[0060] Specifically, based on the historical remote sensing data and historical interference parameters of the boundary area, establish the boundary area inversion model, including:

[0061] (i)According to the geographic information data, marine environmental characteristics and historical interference data, determine the geographical location of the boundary area. The boundary area includes at least the designated area of the coastline and the estuary area.

[0062] It should be noted that the designated area of the coastline and the estuary area are common boundary areas. These areas are greatly affected by land runoff, human activities, etc., and the red tide situation is relatively complex.

[0063] (ii)Denoise the historical remote sensing data of the boundary area.

[0064] Historical remote sensing data may be affected by factors such as atmospheric interference and sensor noise, resulting in noise. By using denoising methods such as filtering and smoothing, these noises are removed to improve the data quality.

[0065] (iii) Use the denoised historical remote sensing data of the boundary area and the historical interference parameters for fitting to establish an inversion model for the boundary area.

[0066] An appropriate mathematical model (such as multiple linear regression, non-linear model, etc.) can be adopted to establish the mathematical relationship between the remote sensing reflectance of sensitive bands, historical interference parameters and red tide concentration and other parameters within the boundary area.

[0067] (5) Summarize the sub-region inversion model and the boundary area inversion model to obtain a red tide inversion model.

[0068] Integrate the sub-region inversion models and the boundary area inversion model together to form a complete red tide inversion model (calibrated red tide inversion model). In practical applications, according to the remote sensing data of different positions in the target sea area input, determine the sub-region or boundary area to which the position belongs, and then call the calibrated red tide inversion model to invert the red tide characteristics. Specifically, when implementing, according to the remote sensing data of different positions in the target sea area input, first determine whether the position belongs to the boundary area. If it belongs to the boundary area, call the boundary area inversion model to invert the red tide characteristics, and compare the inversion result with the inversion result of the calibrated red tide inversion model in this boundary area, and calculate the calibration margin (the calibration margin can be measured by the absolute difference, relative error percentage or standardized root mean square error of the two inversion results, and the typical threshold is set to 15%). If the calibration margin calculated in the boundary area is greater than the preset calibration margin threshold, it indicates that the accuracy of the calibrated red tide inversion model in some areas needs to be improved and further calibration is required. For the sub-region, compare the coefficients of the sub-region inversion model and the calibrated red tide inversion model. Screen out the sub-regions where the coefficients of the sub-region inversion model and the calibrated red tide inversion model differ by more than the threshold. For these screened sub-regions, call the corresponding sub-region inversion model to invert the red tide characteristics, and use the results to calibrate the inversion results of the calibrated red tide inversion model in this sub-region.

[0069] It should be noted that such zoning modeling takes into account the environmental differences in different regions, can more accurately reflect the red tide situation in each region, and has higher accuracy compared to a single overall model. In addition, establishing a special model for special areas such as the boundary area can effectively cope with complex interference factors and improve the applicability of the model in complex environments. By separately modeling and summarizing different regions, the model can more comprehensively and accurately reflect the red tide characteristics of the target sea area, enhancing the reliability and stability of the model.

[0070] In addition, it should be noted that constructing a red tide inversion model also includes:

[0071] (1) Cultivate diatom samples collected from the target red tide area, measure the spectral reflectance of diatom algal solutions with different concentrations under the same cultivation conditions, and combine the remote sensing reflectance data corresponding to the diatom samples to determine the sensitive bands of red tides; the target red tide area is the area where the red tide occurrence frequency is greater than a preset value.

[0072] Specifically, collect diatom samples in the target red tide area, record the location information of the sampling points at the same time, cultivate the collected diatom samples, measure the spectral reflectance of diatom algal solutions with different concentrations under the same cultivation conditions, and combine the remote sensing reflectance data corresponding to the diatom samples (obtained from satellites or other remote sensing platforms), analyze the change characteristics of the spectral reflectance curve, and determine multiple bands sensitive to the changes in red tide water bodies.

[0073] (2) Use a preset number of sampling points randomly selected from the sampling points of diatom samples collected from the target red tide area to obtain the measured red tide concentrations corresponding to these sampling points, and perform a correlation analysis between the measured red tide concentrations and the band combinations of the sensitive bands.

[0074] It should be noted that the band combinations are formed by selecting different bands from the determined sensitive bands to form multiple different combination schemes. Different band combinations may have different abilities to reflect the red tide concentration. Through correlation analysis, the band combinations with higher correlations with the red tide concentration are screened out.

[0075] Specifically, using a preset number of sampling points randomly selected from the sampling points of diatom samples collected from the target red tide area to obtain the measured red tide concentrations corresponding to these sampling points, and performing a correlation analysis between the measured red tide concentrations and the band combinations of the sensitive bands includes: selecting different bands from the determined sensitive bands for combination to form multiple band combination schemes; associating the red tide concentration of each sampling point among the preset number of sampling points with the remote sensing reflectance data corresponding to each band combination respectively; using a preset correlation calculation method to calculate the correlation coefficients between the real-time red tide concentration of each sampling point and the remote sensing reflectance of each band combination respectively; sorting all the calculated correlation coefficients and screening out the band combinations with correlation coefficients greater than a preset threshold.

[0076] Among them, the correlation coefficient is a value used to measure the degree of linear correlation between the red tide concentration of the sampling point and the remote sensing reflectance of each band combination. Its value range is usually between -1 and 1. The closer the absolute value is to 1, the stronger the correlation; the closer the absolute value is to 0, the weaker the correlation.

[0077] (3)Based on the results of the correlation analysis, select the n band combinations with the highest correlation. For the n band combinations, select multiple functions to be screened.

[0078] Among them, the function forms of the multiple functions to be screened may include linear functions, exponential functions, and Gaussian functions. Specifically, the above three function forms are only examples. For example, it may also include ratio functions, etc., which are not shown one by one here.

[0079] (4)For each function form to be screened, use the remote sensing reflectance data of each sampling point under the selected band combination and the corresponding red tide concentration data for fitting.

[0080] Specifically, for each function form to be screened, use the remote sensing reflectance data of each sampling point under the selected band combination and the corresponding red tide concentration data, and perform fitting through mathematical methods to determine the parameters in the function so that the function can fit the data points as much as possible.

[0081] (5)Determine the red tide inversion model according to the fitting results.

[0082] It should be noted that after using the remote sensing reflectance data of each sampling point under the selected band combination and the corresponding red tide concentration data for fitting for each function form to be screened, it includes:

[0083] (1)Calculate the goodness-of-fit index value and error index value of the functions after fitting for different function forms to be screened, and select the function form with the largest goodness-of-fit index value and the smallest error index value as the optimal function form of the band combination.

[0084] It should be noted that the goodness-of-fit index is used to evaluate the goodness of fit of the function to the data. The larger the goodness-of-fit index value, the better the fitting effect of the function to the data, that is, the function can more accurately describe the relationship between the red tide concentration and the remote sensing reflectance. The error index value is a numerical value reflecting the error between the function fitting result and the actual data. The smaller the value, the closer the fitting result is to the actual data, and the higher the accuracy of the model.

[0085] Specifically, the goodness-of-fit index value (coefficient of determination R²) and the error index value (root mean square error RMSE) are used together to evaluate the goodness of fit of the model between the red tide concentration and the remote sensing reflectance of the band combination. The coefficient of determination R² is used to directly measure the strength of the linear correlation between the red tide concentration and the remote sensing reflectance of the band combination, and its value ranges from 0 to 1. When R² = 0, it indicates no linear correlation; when R² = 1, it indicates a perfect linear correlation. The specific determination criteria are as follows: R² > 0.75 (strong correlation), 0.5 < R² ≤ 0.75 (medium correlation), R² ≤ 0.5 (weak correlation); RMSE reflects the prediction accuracy of the model, and the smaller its value, the lower the inversion error (RMSE < 0.5 is considered high precision). Calculate the goodness-of-fit index value and the error index value of the functions after fitting different functions to be screened, and compare the fitting effects of each function to be screened on the red tide concentration data. According to the comprehensive evaluation results of the goodness-of-fit index and the error index, select the optimal function form as the final form of the target band combination (select the function form with the smallest error index value from the function forms with the largest goodness-of-fit index value), and this function form can most accurately describe the relationship between the red tide concentration and the remote sensing reflectance.

[0086] (2) Determine the function coefficients based on the optimal function form, and construct a red tide inversion model with the determined band combination, optimal function form, and function coefficients.

[0087] Based on the selected optimal function form, further determine the coefficients in the function so that the function can accurately reflect the actual data.

[0088] It should also be noted that the red tide inversion model will obtain a function value after calculation, and this function value is the real-time red tide index.

[0089] Specifically, when implementing, extract the remote sensing reflectance data corresponding to each sensitive band from the previously obtained remote sensing data of the target sea area. Ensure that these data are accurate, complete, and have been subjected to necessary preprocessing (such as radiometric correction, etc.), and then determine that the function form of the red tide inversion model is a function with the independent variable including the ratio of remote sensing reflectances (the independent variable of the function is the ratio of the sum of the first remote sensing reflectance and the second remote sensing reflectance to the third remote sensing reflectance, and the function is obtained by taking the difference between the independent variable and the constant term), that is, determine the sensitive bands corresponding to the first remote sensing reflectance, the second remote sensing reflectance, and the third remote sensing reflectance respectively, and the specific value of the constant term. These parameters are usually determined through the analysis and fitting of a large amount of historical data during the construction of the red tide inversion model. Furthermore, according to the definition of the above function, calculate the sum of the first remote sensing reflectance and the second remote sensing reflectance, and then divide this sum by the third remote sensing reflectance to obtain the value of the independent variable of the function. Finally, substitute the calculated value of the independent variable into the function and subtract the constant term to obtain the final real-time red tide index.

[0090] Specifically, the red tide inversion model is: Red Tide Index = (First Remote Sensing Reflectance + Second Remote Sensing Reflectance) / Third Remote Sensing Reflectance - Constant Term. Among them, the first, second, and third remote sensing reflectances are the remote sensing reflectances corresponding to the three most sensitive bands with the highest correlation with the red tide concentration. For example, the first, second, and third remote sensing reflectances can be Rrs443, Rrs565, and Rrs490 respectively (Rrs443, Rrs490, and Rrs565 are the remote sensing reflectances at wavelengths of 443 nm, 490 nm, and 565 nm). In practical applications, based on the results of correlation analysis, three bands with the strongest correlation with the red tide concentration can be selected from multiple sensitive bands, and the corresponding remote sensing reflectances are substituted into the above model to calculate the red tide index of the target sea area, providing strong support for the monitoring and early warning of red tides.

[0091] Table 1 shows multiple models fitted in this application. Please refer to Table 1:

[0092] Table 1 Multiple Fitted Models

[0093]

[0094] It should be noted that in the experiment, various band combination forms were used for comparison, and only the relatively precise function forms with better effects are listed in the table. Among them, for the fitting result of the Gaussian function form, "\ " represents unfitted.

[0095] According to Table 1, their correlations are compared through the root mean square error (RMSE) and the coefficient of determination (R²). Correlation analysis shows that the model established with a function whose independent variable includes the ratio of remote sensing reflectances (Model 1 in Table 1) has the highest accuracy (R² = 0.8885, RMSE = 0.3122). Therefore, the red tide inversion model is:

[0096] HABI = (Rrs443 + Rrs565) / Rrs490 - 2;

[0097] Among them, HABI is the red tide index, and Rrs443, Rrs490, and Rrs565 are the remote sensing reflectances at wavelengths of 443 nm, 490 nm, and 565 nm.

[0098] S104. Input the remote sensing reflectances under each of the sensitive bands into the red tide inversion model to calculate the real-time red tide index, and the real-time red tide index is used to characterize the density of red tides in the target sea area.

[0099] It should be noted that the real-time red tide index is a numerical index used to quantify the red tide density in the target sea area. Through this index, the current severity of the red tide in the target sea area can be intuitively understood. In addition, the higher the red tide distribution density, the higher the corresponding red tide index. It should also be noted that in addition to red tides, thick clouds and atmospheric conditions (such as haze, precipitation, etc.) will affect the calculation accuracy of the red tide index. Therefore, in order to ensure accurate calculation under various interference factors, images with clear skies and no thick clouds can be selected for analysis.

[0100] S105. Determine whether there is a red tide in the target sea area according to the numerical relationship between the real-time red tide index and the characteristic interval of the red tide index.

[0101] It should be noted that the characteristic interval of the red tide index is a specific numerical interval determined based on the historical remote sensing data and historical red tide status of the target sea area under preset seasons and preset meteorological conditions. By inputting the historical remote sensing data into the red tide inversion model to calculate the historical red tide index, and then based on the historical red tide status and the historical red tide index, this interval reflects the change range of the red tide index in the target sea area under different red tide states.

[0102] Specifically, before determining whether there is a red tide in the target sea area according to the numerical relationship between the real-time red tide index and the characteristic interval of the red tide index, it includes: [[ID=X]] [[ID=Y]]

[0103] (1) Obtain the historical remote sensing data and historical red tide status of the target sea area under preset seasons and preset meteorological conditions.

[0104] It should be noted that determine the preset season (such as the high-incidence season of red tides in summer, etc.) and preset meteorological conditions (such as stable meteorological conditions such as sunny days and small wind speeds), collect the historical remote sensing data of the target sea area under these preset conditions, covering remote sensing observation information at different time points. At the same time, obtain the historical red tide status records of the target sea area during the corresponding time period, which can be obtained through channels such as on-site monitoring reports and statistical data of relevant marine environmental monitoring departments.

[0105] (2) Input the historical remote sensing data into the red tide inversion model to calculate the historical red tide index corresponding to the historical red tide status.

[0106] Specifically, inputting the historical remote sensing data into the red tide inversion model to calculate the historical red tide index corresponding to the historical red tide status includes: identifying the boundary area; obtaining the historical remote sensing data of the boundary area; inputting the historical remote sensing data into the corresponding boundary area inversion model to calculate the red tide index of the boundary area, and replacing the real-time red tide index.

[0107] It should be noted that, based on geographical information data (such as coastline position, estuary topography, etc.), marine environmental characteristics (such as water velocity field, salinity distribution, etc.) and historical interference data (such as past pollutant emission records, storm influence range, etc.), boundary regions in the target sea area are determined. Common boundary regions include designated areas near the coastline and estuary areas, etc. Then, remote sensing data information corresponding to the boundary regions is extracted from the collected historical remote sensing data. Furthermore, the historical remote sensing data of the boundary regions is input into the corresponding boundary region inversion model according to its specific location and characteristics. The inversion model of the boundary region is constructed based on the geographical information data, marine environmental characteristic parameters and historical interference data sets of this region. Finally, the red tide inversion model of the boundary region is used to calculate the input data, and the red tide index of each boundary region is obtained. The red tide index of these boundary regions is used to replace the real-time red tide index to calibrate the calculation result of the red tide index of the boundary region, making the calculation result more in line with the complex actual situation of the boundary region.

[0108] (3) Determine the characteristic interval of the red tide index based on the historical red tide state and the historical red tide index.

[0109] It should be noted that a comprehensive analysis is carried out on the calculated historical red tide index and the corresponding historical red tide state. For example, the historical data is classified according to whether a red tide occurs or not, and the value range of the red tide index under different categories is statistically analyzed. The key thresholds that can distinguish different states of the red tide are determined, thereby delimiting the characteristic interval of the red tide index. For example, it is set that when the red tide index is greater than a certain threshold, it is the red tide occurrence interval, and when it is less than the threshold, it is the non-red tide occurrence interval, etc.

[0110] Furthermore, the current real-time red tide index of the target sea area is calculated (in the same way as inputting the remote sensing reflectance of sensitive bands into the red tide inversion model before), and the real-time red tide index is compared with the determined characteristic interval of the red tide index. According to the comparison result, it is judged whether there is a red tide in the target sea area currently. If the real-time red tide index falls within the red tide occurrence interval, it is judged that there is a red tide; if it falls within the non-occurrence interval, it is judged that there is no red tide.

[0111] Figure 3 This is an image shown in this application without red tide index calculation. Figure 4 This is an image shown in this application after red tide index calculation. Please refer to Figure 3 and Figure 4 , it can be seen that the recognition accuracy is higher after red tide index calculation. Among them, the higher the red tide distribution density, the closer the color is to red in the figure.

[0112] Optionally, when using other satellite data for red tide identification, it further includes:

[0113] Obtain information on the spectral characteristics, spatial resolution, temporal resolution, and atmospheric correction method of the target satellite sensor; based on the determined sensitive bands for red tides, compare the channel settings of the target satellite, and screen out the channels closest to the wavelength range of the sensitive bands as the target channels; re-determine the band combination method and calculation relationship based on the remote sensing reflectance of the target channels, and adjust the inversion model; use the historical data of the target sea area obtained by the target satellite and the corresponding measured red tide data, and calculate the red tide index in combination with the adjusted inversion model.

[0114] The method provided in this embodiment can accurately find the bands sensitive to the changes in red tide water bodies by culturing diatom algal solutions with different concentrations and analyzing the spectral reflectance curves, and can keenly capture the differences between red tide water bodies and normal seawater, providing a basis for accurately identifying red tides subsequently. By adopting a zoning modeling strategy, comprehensively considering the characteristics of the marine environment, dividing multiple sub-regions, establishing inversion models for the sub-regions and boundary regions respectively, and then summarizing them to form a complete red tide inversion model. This method fully considers the interference factors and environmental differences in different regions, can more accurately reflect the red tide situation in each region, and greatly improves the model accuracy. In addition, by determining the characteristic interval of the red tide index based on the historical data of the target sea area and comparing the real-time red tide index with it, it can reliably judge whether there is a red tide in the target sea area. This process fully considers the relationship between the historical red tide state and the index, sets a reasonable threshold to distinguish whether a red tide occurs, and reduces misjudgment. Finally, for different satellite data, it can screen appropriate channels according to the characteristics of the satellite sensor, adjust the inversion model and calculate the red tide index, broadening the scope of technical application, improving the adaptability and accuracy of red tide identification under different satellite data, and ensuring effective monitoring of red tides in various satellite monitoring scenarios.

[0115] In addition, by continuously obtaining remote sensing data of the target sea area, continuously updating the remote sensing reflectance values of the sensitive bands, and then calculating the red tide index in real time, dynamic monitoring of the red tide situation in the target sea area can be realized. Once the real-time red tide index exceeds the normal range and enters the warning interval for the occurrence of red tides, it can help issue an alarm in a timely manner. In addition, as the continuously accumulated remote sensing data and on-site monitoring data are obtained, the red tide inversion model can be continuously optimized using these new data. On the one hand, the determination of the sensitive bands can be further refined and adjusted to make it more in line with the red tide characteristics of different sea areas and different seasons; on the other hand, by re-analyzing the data, the characteristic interval of the red tide index can be optimized to improve the accuracy of judgment.

[0116] Embodiment 2:

[0117] Corresponding to the foregoing embodiment of a red tide identification method based on remote sensing data, the present application also provides an embodiment of a red tide identification device based on remote sensing data.

[0118] Figure 5Schematic diagram of Embodiment 2 of the red tide identification device based on remote sensing data provided by this application. Please refer to Figure 5 , the device provided in this embodiment includes a determination module 510, an extraction module 520, a construction module 530, a calculation module 540, and a judgment module 550;

[0119] Among them, the determination module 510 is used to compare the spectral reflectance curves of diatom algal solutions with different concentrations to determine the sensitive bands of red tide water bodies, and the sensitive bands include multiple;

[0120] The extraction module 520 is used to obtain remote sensing data of the target sea area and extract the remote sensing reflectance corresponding to the sensitive bands in the remote sensing data;

[0121] The construction module 530 is used to construct a red tide inversion model; among them, the red tide inversion model is a function with the ratio of remote sensing reflectance as an independent variable, the independent variable of the function is the ratio of the sum of the first remote sensing reflectance and the second remote sensing reflectance to the third remote sensing reflectance, and the function is obtained by taking the difference between the independent variable and a constant term, and the sensitive bands corresponding to the first remote sensing reflectance, the second remote sensing reflectance, and the third remote sensing reflectance are different;

[0122] The calculation module 540 is used to input the remote sensing reflectance under each sensitive band into the red tide inversion model to calculate the real-time red tide index, and the real-time red tide index is used to characterize the density of red tide in the target sea area;

[0123] The judgment module 550 is used to judge whether there is a red tide in the target sea area according to the numerical relationship between the real-time red tide index and the numerical range of the red tide index characteristic interval.

[0124] The device in this embodiment can be used to execute Figure 1 the steps of the method embodiment shown, and the specific implementation principle and process are similar, so details are not described here.

[0125] For the implementation process of the functions and roles of each unit in the above device, please refer to the implementation process of the corresponding steps in the above method for details, and will not be elaborated here.

[0126] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiments described above are only illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this application. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0127] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the scope of protection of the present application.

Claims

1. A red tide recognition method based on remote sensing data, characterized in that The method includes: Comparing the spectral reflectance curves of diatom algal solutions with different concentrations to determine the sensitive bands of red-tide water bodies, where the sensitive bands include multiple ones; Obtaining remote sensing data of the target sea area and extracting the remote sensing reflectance corresponding to the sensitive bands in the remote sensing data; Constructing a red-tide inversion model; wherein, the red-tide inversion model is a function with the ratio of remote sensing reflectance as the independent variable, the independent variable of the function is the ratio of the sum value of the first remote sensing reflectance and the second remote sensing reflectance to the third remote sensing reflectance, the function is obtained by taking the difference between the independent variable and the constant term, and the sensitive bands corresponding to the first remote sensing reflectance, the second remote sensing reflectance, and the third remote sensing reflectance are different; Inputting the remote sensing reflectance under each of the sensitive bands into the red-tide inversion model to calculate the real-time red-tide index, and the real-time red-tide index is used to characterize the density of red-tide in the target sea area; Judging whether there is a red-tide in the target sea area according to the numerical relationship between the real-time red-tide index and the characteristic interval of the red-tide index; The constructing of the red-tide inversion model includes: Identifying boundary regions based on geographical information data, marine environmental characteristics, and historical interference data; Dividing the area in the target sea area except the boundary regions into multiple sub-regions; Establishing a sub-region inversion model based on the historical remote sensing data and red-tide concentration of the sub-regions; Establishing a boundary-region inversion model based on the historical remote sensing data and historical interference parameters of the boundary regions; Summarizing the sub-region inversion model and the boundary-region inversion model to obtain the red-tide inversion model; The dividing of the area in the target sea area except the boundary regions into multiple sub-regions includes: Collecting the marine environmental characteristic data of the target sea area, where the marine environmental characteristic data at least includes water flow velocity field data, suspended matter concentration data, wind field data, temperature data, salinity data, and chlorophyll a concentration data; Calculating the spatial gradient changes of water flow velocity, suspended matter concentration, wind speed and direction, temperature, salinity, and chlorophyll a concentration, and dividing the area in the target sea area except the boundary regions into multiple sub-regions according to the gradients of the comprehensive changes.

2. The method according to claim 1, characterized in that, Before judging whether there is a red-tide in the target sea area according to the numerical relationship between the real-time red-tide index and the characteristic interval of the red-tide index, the method further includes: Obtaining the historical remote sensing data and historical red-tide status of the target sea area under preset seasons and preset meteorological conditions; Inputting the historical remote sensing data into the red-tide inversion model to calculate the historical red-tide index corresponding to the historical red-tide status; Determining the characteristic interval of the red-tide index based on the historical red-tide status and the historical red-tide index.

3. The method according to claim 1, wherein The establishing of the boundary-region inversion model based on the historical remote sensing data and historical interference parameters of the boundary regions includes: Determining the geographical locations of the boundary regions according to the geographical information data, marine environmental characteristics, and historical interference data, where the boundary regions at least include the designated areas of the coastline and the estuary areas; Performing denoising processing on the historical remote sensing data of the boundary regions; Performing fitting using the denoised historical remote sensing data of the boundary regions and the historical interference parameters to establish the boundary-region inversion model.

4. The method according to claim 2, wherein Inputting the historical remote sensing data into the red tide inversion model to calculate the historical red tide index corresponding to the historical red tide state includes: Identifying the boundary area; Obtaining the historical remote sensing data of the boundary area; Inputting the historical remote sensing data into the corresponding boundary area inversion model to calculate the red tide index of the boundary area and replacing the real-time red tide index.

5. The method according to claim 1, wherein The construction of the red tide inversion model further includes: Culturing diatom samples collected from the target red tide area, measuring the spectral reflectance of diatom algal solutions with different concentrations under the same culture conditions, and combining the remote sensing reflectance data corresponding to the diatom samples to determine the sensitive bands of red tides; the target red tide area is the area where the red tide occurrence frequency is greater than a preset value; Using a preset number of sampling points randomly selected from the sampling points where diatom samples are collected in the target red tide area to obtain the measured red tide concentrations corresponding to these sampling points, and performing a correlation analysis on the measured red tide concentrations and the band combinations of the sensitive bands; Based on the correlation analysis results, selecting the n band combinations with the highest correlation, and selecting multiple functions to be screened for the n band combinations; For each form of the function to be screened, using the remote sensing reflectance data of each sampling point under the selected band combination and the corresponding red tide concentration data for fitting; Determining the red tide inversion model according to the fitting results.

6. The method according to claim 5, characterized in that After using the remote sensing reflectance data of each sampling point under the selected band combination and the corresponding red tide concentration data for fitting for each form of the function to be screened, it includes: Calculating the goodness-of-fit index value and error index value of the functions after fitting with different forms of the function to be screened, and selecting the function form with the largest goodness-of-fit index value and the smallest error index value as the optimal function form of the band combination; Determining the function coefficients based on the optimal function form, and constructing the red tide inversion model with the determined band combination, optimal function form, and function coefficients.

7. The method according to claim 5, wherein The step of using a preset number of sampling points randomly selected from the sampling points where diatom samples are collected in the target red tide area to obtain the measured red tide concentrations corresponding to these sampling points, and performing a correlation analysis on the measured red tide concentrations and the band combinations of the sensitive bands includes: Selecting different bands from the determined sensitive bands for combination to form multiple band combination schemes; Associating the real-time red tide concentration of each sampling point among the preset number of sampling points with the remote sensing reflectance data corresponding to each band combination respectively; Using a preset correlation calculation method to calculate the correlation coefficients between the real-time red tide concentration of each sampling point and the remote sensing reflectance of each band combination respectively; Sorting all the calculated correlation coefficients, and screening out the band combinations with correlation coefficients greater than a preset threshold.

8. A red tide recognition device based on remote sensing data, characterized in that, The device includes a determination module, an extraction module, a construction module, a calculation module, and a judgment module; Among them, the determination module is used to compare the spectral reflectance curves of diatom algal solutions with different concentrations to determine the sensitive bands of red tide water bodies, and the sensitive bands include multiple; The extraction module is used to obtain the remote sensing data of the target sea area and extract the remote sensing reflectance corresponding to the sensitive bands in the remote sensing data; The building module is used to build a red tide inversion model; wherein, the red tide inversion model is a function with the remote sensing reflectance ratio as the independent variable, the independent variable of the function is the ratio of the sum of the first remote sensing reflectance and the second remote sensing reflectance to the third remote sensing reflectance, the function is obtained by taking the difference between the independent variable and the constant term, and the sensitive bands corresponding to the first remote sensing reflectance, the second remote sensing reflectance, and the third remote sensing reflectance are different; The calculation module is used to input the remote sensing reflectance under each of the sensitive bands into the red tide inversion model to calculate the real-time red tide index, and the real-time red tide index is used to characterize the density of red tide in the target sea area; The judgment module is used to judge whether there is a red tide in the target sea area according to the numerical relationship between the real-time red tide index and the red tide index characteristic interval; The building of the red tide inversion model includes: Identifying the boundary area based on the geographic information data, marine environmental characteristics, and historical interference data; Dividing the area in the target sea area except the boundary area into multiple sub-areas; Establishing a sub-area inversion model based on the historical remote sensing data and red tide concentration in the sub-area; Establishing a boundary area inversion model based on the historical remote sensing data and historical interference parameters in the boundary area; Summarizing the sub-area inversion model and the boundary area inversion model to obtain the red tide inversion model; The dividing of the area in the target sea area except the boundary area into multiple sub-areas includes: Collecting the marine environmental characteristic data of the target sea area, and the marine environmental characteristic data at least includes the water flow velocity field data, suspended matter concentration data, wind field data, temperature data, salinity data, and chlorophyll a concentration data; Calculating the spatial gradient changes of the water flow velocity, the suspended matter concentration, the wind speed and direction, the temperature, the salinity, and the chlorophyll a concentration, and dividing the area in the target sea area except the boundary area into multiple sub-areas according to the comprehensive change gradient.

Images