A method and device for identifying Sargassum based on remote sensing data

Abstract

The present application provides a method and device for identifying Sargassum based on remote sensing data, belonging to the field of marine remote sensing monitoring and algae identification. The method includes: screening Sargassum sampling points at the first order of magnitude; screening target bands based on the correlation between the index values of each Sargassum sampling point and the characteristic data of each band; determining the form of the model function to be screened according to the remote sensing data difference between Sargassum and Enteromorpha; fitting the reflectance data of Sargassum sampling points and Enteromorpha sampling points under each band combination in the target band to determine the inversion model; obtaining real-time remote sensing data; calculating the Sargassum index value based on the real-time remote sensing data and the inversion model; determining the time characteristics of the real-time remote sensing data, and determining the optimal discrimination method for distinguishing Sargassum and Enteromorpha based on the time characteristics; identifying the Sargassum index value based on the optimal discrimination method, distinguishing Sargassum and Enteromorpha, and obtaining the identification result. The present application can achieve efficient, accurate identification and dynamic monitoring of Sargassum.

Description

Technical Field

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

[0002] Sargassum is an important part of the marine ecosystem and has an important impact on the marine ecological environment, fishery resources, etc. Accurately identifying and monitoring the distribution of Sargassum is of great significance for marine ecological research, marine environmental protection, and the development and utilization of marine resources.

[0003] Traditional methods for identifying Sargassum mainly rely on manual field surveys. This method is not only inefficient but also difficult to implement when monitoring large areas of the ocean, and cannot meet the needs of real-time and dynamic monitoring. With the development of remote sensing technology, using remote sensing data for Sargassum identification has become a feasible and efficient means. However, there are still many problems with the current methods for identifying Sargassum based on remote sensing data: the characteristics of different satellite data vary greatly and are difficult to be universal; Sargassum is easily confused with similar algae such as Enteromorpha in remote sensing data, and there is a lack of accurate discrimination methods; the existing identification methods do not fully consider the influence of geographical location, time and other factors on the identification results, resulting in unstable identification accuracy, etc. Therefore, there is an urgent need for a method for identifying Sargassum based on remote sensing data that can improve the accuracy, applicability and stability of identification. Summary of the Invention

[0004] In view of this, this application provides a method and device for identifying Sargassum based on remote sensing data, which can achieve efficient, accurate identification and dynamic monitoring of Sargassum.

[0005] Specifically, this application is implemented through the following technical solutions:

[0006] The first aspect of this application provides a method for identifying Sargassum based on remote sensing data, and the method includes:

[0007] Screen sampling points of the first order of magnitude, where the climate characteristics and growth times corresponding to each sampling point are different, and the sampling points include Sargassum sampling points and Enteromorpha sampling points;

[0008] Separate the characteristic data of each band in the remote sensing data, and screen the target bands based on the correlation between the Sargassum index value of each sampling point and the characteristic data of each band, and there are multiple target bands;

[0009] Determine the form of the model function to be screened according to the remote sensing data differences between Sargassum and Enteromorpha;

[0010] Based on the functional forms of each model to be screened, fit the reflectance data of each sampling point under each band combination in the target band, compare the similarity between the fitting curve and the distribution of actual data points, and determine the inversion model;

[0011] Obtain real-time remote sensing data;

[0012] Calculate the Sargassum index value based on the real-time remote sensing data and the inversion model;

[0013] Determine the time characteristics of the real-time remote sensing data, and based on the time characteristics, determine the optimal discrimination method for distinguishing Sargassum and Enteromorpha; the optimal discrimination method at least includes: comparing the calculated Sargassum index with the index intervals corresponding to Sargassum and Enteromorpha respectively, and judging the category corresponding to the real-time remote sensing data;

[0014] Based on the optimal discrimination method, identify the Sargassum index value, distinguish Sargassum and Enteromorpha, and obtain the identification result.

[0015] The second aspect of the present application provides a Sargassum identification device based on remote sensing data, and the device includes a screening module, a determination module, an acquisition module, a calculation module, a discrimination module, and an identification module;

[0016] The screening module is used to screen sampling points of the first order of magnitude, and the climate characteristics and growth times corresponding to each sampling point are different. The sampling points include Sargassum sampling points and Enteromorpha sampling points;

[0017] The screening module is further used to separate the feature data of each band in the remote sensing data, and screen the target bands based on the correlation between the Sargassum index value of each sampling point and the feature data of each band. There are multiple target bands;

[0018] The determination module is used to determine the functional form of the model to be screened according to the differences in the remote sensing data of Sargassum and Enteromorpha;

[0019] The determination module is further used to fit the reflectance data of each sampling point under each band combination in the target band based on the functional forms of each model to be screened, compare the similarity between the fitting curve and the distribution of actual data points, and determine the inversion model;

[0020] The acquisition module is used to obtain real-time remote sensing data;

[0021] The calculation module is used to calculate the Sargassum index value based on the real-time remote sensing data and the inversion model;

[0022] The discrimination module is used to determine the time characteristics of the real-time remote sensing data, and determine the optimal discrimination method for distinguishing Sargassum and Enteromorpha based on the time characteristics; the optimal discrimination method at least includes: comparing the calculated Sargassum index with the index intervals corresponding to Sargassum and Enteromorpha respectively to determine the category corresponding to the real-time remote sensing data.

[0023] The recognition module is used to recognize the Sargassum index value based on the optimal discrimination method, distinguish Sargassum and Enteromorpha, and obtain the recognition result.

[0024] The Sargassum recognition method and device based on remote sensing data provided by this application provide an adaptive Sargassum recognition method, which mainly includes automatically adjusting the bands, sampling points, and inversion model forms according to different target regions, source satellite data, and recognition accuracies, thereby improving the recognition accuracy in various scenarios and under various accuracy requirements. Specifically, by analyzing the correlation between the index values of each Sargassum sampling point and the characteristic data of each band, the target bands can be screened, and multiple bands that are most critical for Sargassum recognition can be found, excluding the influence of irrelevant or interfering bands, and improving the efficiency of data processing and the accuracy of recognition. Obtaining real-time remote sensing data and calculating the Sargassum index value based on this and the inversion model can monitor and analyze the current marine conditions in a timely manner, providing the possibility to grasp the distribution and dynamic changes of Sargassum in real time, and helping to take corresponding measures in a timely manner to deal with the possible impacts brought by Sargassum. In addition, by comparing the calculated Sargassum index with the index intervals corresponding to Sargassum and Enteromorpha respectively to determine the category, the influence of time factors on algae recognition is fully considered, further improving the accuracy and reliability of Sargassum recognition. Description of the Drawings

[0025] Figure 1 It is a flowchart of the first embodiment of the Sargassum recognition method based on remote sensing data provided by this application;

[0026] Figure 2 It is a HY-1C / D CZI image showing undistinguished Sargassum and Enteromorpha provided by this application;

[0027] Figure 3 It is a HY-1C / D CZI image showing distinguished Sargassum and Enteromorpha (only Sargassum) provided by this application;

[0028] Figure 4 It is a schematic structural diagram of the second embodiment of the Sargassum recognition device based on remote sensing data provided by this application. Detailed Embodiments

[0029] Exemplary embodiments will be described in detail herein, and examples thereof are shown in the accompanying drawings. When the following description refers to the accompanying 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 the present application.

[0030] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit the present 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" as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0031] 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 information of the same type from each other. For example, without departing from the scope of the present 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 determining".

[0032] Specific embodiments are given below to introduce the technical solutions of the present application in detail.

[0033] Embodiment 1:

[0034] Figure 1 It is a flowchart of Embodiment 1 of the Sargassum recognition method based on remote sensing data provided by the present application. Please refer to Figure 1 , the method provided in this embodiment may include:

[0035] S101. Screen Sargassum sampling points of the first order of magnitude, where the climate characteristics and growth times corresponding to each sampling point are different, and the sampling points include Sargassum sampling points and Enteromorpha prolifera sampling points.

[0036] It should be noted that the first order of magnitude refers to a relative quantity or a range of quantities. For example, in actual operations, 100 Sargassum sampling points (or Enteromorpha prolifera sampling points, taking Sargassum sampling points as an example here) may be selected, or the number of Sargassum sampling points between 100 and 1000 can also be selected. The climate characteristics and growth time of each sampling point are different. Among them, the climate characteristics include environmental factors such as temperature, light intensity, salinity, and precipitation that affect the growth of Sargassum. The differences in climate characteristics in different regions and seasons will affect the growth state of Sargassum, and thus affect its performance in remote sensing data. The growth time refers to the time points when Sargassum is in different growth stages, such as the juvenile stage and the mature stage. Sargassum in different growth stages has differences in physiological structure and material composition, and also presents different characteristics in remote sensing data.

[0037] Specifically, determining the first order of magnitude further includes: determining the target distribution area and target distribution time of the remote sensing data to be recognized, determining the initial order of magnitude based on the overlap degree between the time span of the target distribution time and the growth cycle distribution of the Sargassum, and correcting the initial order of magnitude based on the anisotropy coefficient of the associated environmental factors in the target distribution area to obtain the first order of magnitude, and the correction can be a product. The associated environment is the environmental parameters that affect the growth of the Sargassum, and the anisotropy coefficient is the degree of difference in environmental factors in the target distribution area.

[0038] It should be noted that the overlap degree between the target distribution time and the growth cycle of the Sargassum has a significant impact on the number of sampling points. If the target distribution time covers multiple growth stages of the Sargassum, in order to comprehensively obtain the characteristic data of different growth stages, more sampling points are required, that is, the initial order of magnitude is larger; if the target distribution time only covers a small segment of its growth cycle, the initial order of magnitude can be correspondingly reduced. Taking the area of monitoring the Sargassum from the juvenile stage to the mature stage as an example, since the Sargassum in different growth stages has large differences in physiological structure and material composition and presents different characteristics in remote sensing data, more sampling points are needed to accurately reflect these changes, and the initial order of magnitude will be larger; if only a certain short growth stage is monitored, the initial order of magnitude can be reduced.

[0039] The anisotropy coefficient of associated environmental factors (such as temperature, light intensity, salinity, precipitation, etc.) is used to further adjust the initial order of magnitude. The anisotropy coefficient reflects the degree of difference in these environmental factors within the target distribution area. When the anisotropy coefficient is large, it means that the environmental differences within the area are large and the growth states of Sargassum are diverse. To accurately reflect this diversity, the number of sampling points needs to be increased, and the first order of magnitude is increased through product correction; conversely, if the anisotropy coefficient is small, the environmental differences are small, the growth states of Sargassum are relatively consistent, the number of sampling points can be reduced, and the first order of magnitude is correspondingly reduced. When monitoring Sargassum in a vast sea area with large differences in environmental factors such as temperature and salinity, to comprehensively capture the characteristics of Sargassum under different environments, the number of sampling points should be increased, that is, the first order of magnitude is increased; in a small bay with relatively uniform environment, the anisotropy coefficient is small, and the first order of magnitude can be reduced.

[0040] Determining the first order of magnitude in this way can accurately match the number of sampling points with the actual monitoring requirements, ensuring that the collected data is comprehensive and representative. It comprehensively reflects the characteristics of Sargassum under different environmental conditions and growth stages, provides a high-quality data basis for subsequent screening of target bands and constructing inversion models, thereby improving the accuracy and reliability of Sargassum identification.

[0041] S102. Separate the characteristic data of each band in the remote sensing data, and screen the target bands based on the correlation between the Sargassum index values at each sampling point and the characteristic data of each band. There are multiple target bands.

[0042] It should be noted that the band characteristic data is the information corresponding to different wavelength ranges (bands) in the remote sensing data. For example, visible light bands (blue light, green light, red light, etc.), near-infrared bands, etc. Different bands are sensitive to different substances to different degrees, and there are differences in the reflectance characteristics of Sargassum and other ground objects in each band.

[0043] It also should be noted that under different climatic conditions and growth stages of Sargassum, its own physiological characteristics (such as pigment content, cell structure, etc.) will change, and these changes will cause changes in its characteristics such as reflectance in different bands. By calculating the correlation between the index values at the sampling points and the characteristic data of each band, the bands closely related to the growth state of Sargassum can be found. A high correlation means that the characteristic data of this band can better reflect the changes of Sargassum. Using these bands as the target bands, the inversion model constructed based on these bands can more accurately identify Sargassum.

[0044] Specifically, under diverse environments such as different sea areas and different seasons, a certain number (the first order of magnitude) of representative Sargassum sampling points are purposefully selected. For example, under different conditions such as winter in tropical sea areas and spring in temperate sea areas, Sargassum sampling points at different growth times are selected to ensure coverage of various climatic characteristics and growth stages.

[0045] After obtaining the remote sensing data, it is split according to different bands to obtain the characteristic data of each band, such as the reflectivity value corresponding to each band, etc. Specifically, the target bands are screened based on the correlation between the Sargassum index values of each sampling point and the characteristic data of each band, including:

[0046] (1) Based on the correlation calculation method, calculate the correlation coefficient between the Sargassum index value of each sampling point and the characteristic data of each band respectively.

[0047] It should be noted that by using the correlation calculation method (such as the Pearson correlation coefficient calculation method), calculate the correlation coefficient between the Sargassum index value of each sampling point and the characteristic data of each band respectively. For each sampling point, associate its Sargassum index value with the data of each band to calculate a coefficient representing the degree of association between the two. Specifically, for the principle of correlation calculation, please refer to the introduction of related technologies and will not be elaborated here.

[0048] (2) Compare the calculated correlation coefficients with a preset correlation coefficient threshold, and screen out the bands whose correlation coefficients are greater than or equal to the threshold.

[0049] It should be noted that the preset correlation coefficient threshold is a standard value set based on experience or previous experiments, used to judge whether the correlation between the band and the Sargassum index value of the sampling point is strong enough. Only when the calculated correlation coefficient is greater than or equal to this threshold, the corresponding band will be further considered for the subsequent screening steps. The setting of this threshold will affect the number and quality of the finally screened target bands. If the threshold is set too high, the number of screened bands may be small, but the correlation between these bands and the Sargassum index value is very strong; if the threshold is set too low, more bands will be screened out, but some bands with less tight correlation with the Sargassum index value may be included.

[0050] Compare all the calculated correlation coefficients with the preset correlation coefficient threshold (set as T ). For each band j , if there is at least one sampling point i such that rij ≥ T , then this band j is screened out. For example, the preset threshold T = 0.8. For a certain band B 3, if the correlation coefficients between the index values of some sampling points and B 3 are greater than or equal to 0.8, then B 3 will be screened out.

[0051] (3)Sort the selected bands in descending order according to the correlation coefficient, and select the top n bands as the target bands according to actual needs; where n is a positive integer greater than 1.

[0052] Specifically, sort the selected bands in descending order according to the average value of their correlation coefficients with the index values of all sampling points. For example, if the selected bands are B 1, B 3, B 5, calculate the average values of their correlation coefficients with the index values of all sampling points, which are r 1, r 3, r 5 respectively, and then sort the bands corresponding to these average values in descending order. Finally, select the top several bands as the target bands according to actual needs (for example, in order to ensure the simplicity and accuracy of the model, it is desired to select the first 2 bands).

[0053] Table 1 shows the correlation coefficients of single bands and band combinations of Sargassum in this application. Please refer to Table 1:

[0054] Table 1 Correlation Coefficients of Single Bands and Band Combinations of Sargassum

[0055]

[0056] Among them, B1, B2, B3, and B4 are the remote sensing reflectances of the blue, green, red, and near-infrared bands of the HY-1C / D CZI sensor respectively.

[0057] Combined with the above description, it can be seen that the correlation coefficient of the single band B4 is relatively high. Other bands are less sensitive to Sargassum. Therefore, in order to extract Sargassum more accurately, it is necessary to perform band combination to determine the sensitive band combination of Sargassum. Six multispectral band combination models B4 / B1, B4 - B2, B2*B4 / B3, B4 / B2 - B3, B2 / (B3 + B4), and (B4 - B2) / (B3 + B4) are selected from Table 1, and their R 2 values are all greater than 0.80 and are used to further determine the Sargassum inversion model.

[0058] S103. Determine the function form of the model to be screened according to the remote sensing data differences between Sargassum and Enteromorpha.

[0059] By obtaining the differences between the remote sensing data of Sargassum and Enteromorpha at the same location and the same time, fitting the differential change curve, and determining the function to be screened based on the shape of the curve. In practical applications, data at the same location and the same moment are extracted from a large number of synchronously observed remote sensing data of Sargassum and Enteromorpha and their differences are calculated. Then, these differential discrete points are connected into multiple broken lines as the differential change curve. If the curve shows a linear trend, a linear function can be considered; if it shows a power-law change characteristic, a power function may be more appropriate; if the change trend conforms to an exponential law, an exponential function is selected.

[0060] Among them, linear functions, power functions, and exponential functions can effectively describe the potential relationship between reflectance data and Sargassum index values. Different functions have different fitting abilities for the spectral feature changes of different ground objects. The reflectance differences between Sargassum and Enteromorpha in different bands are complex. Selecting multiple function forms can fit the data from different angles and increase the probability of finding the optimal model.

[0061] S104. Fit the reflectance data of the sampling points in each band combination of the target band based on each function form to be screened, compare the similarity between the fitting curve and the distribution of actual data points, and determine the inversion model.

[0062] It should be noted that the inversion model is a mathematical model used to invert the true value of a certain physical quantity or phenomenon based on observation information such as remote sensing data. In this application, the inversion model is obtained by fitting the reflectance data of Sargassum and Enteromorpha sampling points, and is used to calculate the Sargassum index value based on real-time remote sensing data, so as to distinguish Sargassum and Enteromorpha.

[0063] Specifically, determining the inversion model includes:

[0064] (1) Selecting multiple functions to be screened according to the differences in remote sensing data of Sargassum and Enteromorpha.

[0065] Referring to S103, specifically, obtaining the differences between the remote sensing data of Sargassum and Enteromorpha at the same location and the same time; fitting the differential change curve based on multiple differences; determining the function to be screened based on the shape of the differential change curve. Here, the differential change curve is linear formed by connecting multiple differential discrete points with straight lines, that is, multiple broken lines are connected end to end.

[0066] It should be noted that there are differences in the remote sensing data of Sargassum and Enteromorpha, and these differences are reflected in their reflectances in different bands. By analyzing these differences, some common function forms (such as linear functions, power functions, and exponential functions) can be selected as the function forms of the model to be screened. These function forms can describe the potential relationship between reflectance data and Sargassum index values.

[0067] (2)For each function form to be screened, use the reflectance data of each sampling point under each band combination in the target band for fitting.

[0068] It should be noted that the fitting process is to adjust the parameters of the function so that the function curve is as close as possible to the actual data points. For each function form to be screened, the reflectance data of the Sargassum sampling points and the Enteromorpha sampling points under each band combination in the target band are used as inputs, and fitting algorithms such as the least squares method are used to adjust the parameters of the function, thereby obtaining the fitting curve.

[0069] (3)Calculate the goodness-of-fit index values and error index values under different function fittings, and select the function form with the largest goodness-of-fit index value and the smallest error index value during the fitting process of the Sargassum remote sensing data, and at the same time, the discrimination degree is greater than the preset value during the fitting process of the Enteromorpha remote sensing data as the optimal function form of the target band combination.

[0070] It should be noted that the goodness-of-fit index is an index used to measure the closeness of the fitting curve to the actual data points. Common goodness-of-fit indexes include the coefficient of determination ( R 2 ), R 2 The closer it is to 1, the better the fitting effect, that is, the function curve can well explain the changes in the actual data. The error index value is an index reflecting the error size between the fitting result and the actual data. For example, the root mean square error (RMSE), which is the square root of the average of the squares of the deviations between the predicted value and the actual value. The smaller the RMSE, the smaller the fitting error.

[0071] In this step, calculate the goodness-of-fit index values and error index values under different function fittings respectively. For the Sargassum remote sensing data, preferentially select the function with the largest goodness-of-fit index value and the smallest error index value, which means that this function can well explain the changes in the Sargassum reflectance data, and the deviation between the predicted value and the actual value is small. At the same time, for the Enteromorpha remote sensing data, it is required that the discrimination degree of this function is greater than the preset value, that is, this function can effectively distinguish the reflectance characteristics of Sargassum and Enteromorpha. By comprehensively considering these two aspects, the optimal function form is selected.

[0072] (4)Determine the function coefficients based on the optimal function form, and construct an inversion model with the determined target band combination, optimal function form, and function coefficients.

[0073] After determining the optimal function form, further accurately determine the coefficients of the function. For example, for the linear function y = ax + b, determine the specific values of a and b. Then, combined with the target band combinations selected previously, construct an inversion model using the target band combinations, the optimal function form, and the determined function coefficients. This model can calculate the corresponding Sargassum index value based on real-time remote sensing data (the reflectance of the target bands), thereby being used to determine whether the area is a Sargassum distribution area or an Enteromorpha distribution area.

[0074] Table 2 shows the Sargassum band combination inversion model of this application. Please refer to Table 2:

[0075] Table 2 Sargassum Band Combination Inversion Model

[0076]

[0077] Among them, B1, B2, B3, and B4 are the remote sensing reflectances of the blue, green, red, and near-infrared bands of the HY-1C / D CZI sensor respectively.

[0078] Combined with the above description, it can be known that the finally constructed inversion model is in the form of a linear function, and the inversion model takes the difference between the near-infrared band and the green band combination as the independent variable. The sum of the product of the independent variable and the independent variable coefficient and the constant term obtains the linear function. That is, X = B4 - B2, HSI = 1.2142X - 0.0258, where B2 and B4 are the remotely sensed data of the green and near-infrared bands after atmospheric correction, and HSI is the Sargassum index value.

[0079] S105. Obtain real-time remote sensing data.

[0080] Currently, there are mainly two methods: satellite remote sensing and aerial remote sensing. Satellite remote sensing uses sensors carried by artificial satellites to obtain data, such as the Landsat series in the United States and the high-resolution satellite series in China. Aerial remote sensing obtains data by carrying remote sensing equipment on aircraft and other aircraft. It is necessary to select a suitable data source according to specific needs. For example, for large-area and macroscopic monitoring data, satellite remote sensing is more suitable; if high-resolution monitoring of a small area is required, aerial remote sensing may be better.

[0081] The data types and qualities obtained by different sensors are different. Optical sensors mainly obtain electromagnetic wave information in visible light, near-infrared and other bands, and can be used to identify features such as the color and shape of ground objects; radar sensors use microwaves for detection, are not restricted by weather and day and night, and are suitable for obtaining data under bad weather conditions. Select the corresponding sensor according to the monitoring object and purpose. For example, when monitoring the distribution of Sargassum in the ocean, optical sensors can obtain information by using the spectral reflection differences between Sargassum and seawater; and when the weather is bad, radar sensors can also provide certain references.

[0082] By establishing a connection with relevant data centers or platforms, data is received according to certain protocols and processes. Some satellite data can be downloaded from official data distribution websites. For example, the China National Satellite Meteorological Center website provides satellite data download services for meteorological satellites; aerial remote sensing data may need to be obtained in cooperation with professional aerial remote sensing agencies.

[0083] Specifically, sensors obtain information by detecting the electromagnetic waves reflected or emitted by the target ground objects. Different ground objects have different absorption, reflection, and emission characteristics of electromagnetic waves due to differences in their material composition, structure, etc. For example, there are obvious differences in the reflectivity of Sargassum and seawater to electromagnetic waves in different bands. The sensor receives these different electromagnetic wave signals, converts them into electrical signals or digital signals, and after a series of processing and transmission, finally forms remote sensing data available for analysis.

[0084] In addition, it should be noted that before obtaining real-time remote sensing data, it also includes determining the satellite type corresponding to the real-time remote sensing data, specifically including:

[0085] (1) Obtain the spectral response function and radiation resolution of the target satellite.

[0086] It should be noted that different satellites have differences in aspects such as orbit, sensor configuration, and data acquisition capabilities. For example, some satellites are low-orbit satellites, which can obtain high-resolution images but have a limited coverage area; high-orbit satellites are the opposite, with a large coverage area but relatively low resolution. There are also various sensors carried by satellites, such as optical sensors, microwave sensors, etc., which will affect characteristics such as the band range and radiation resolution of the data.

[0087] The spectral response functions and radiation resolutions of different satellites are different. Obtaining these parameters can understand the sensing ability of the satellite sensor to electromagnetic waves of different wavelengths and the ability to distinguish radiation differences. For example, the spectral response function of a hyperspectral satellite can cover a narrower and more refined wavelength range, and the radiation resolution is also relatively high, enabling the acquisition of richer and more accurate ground object spectral information.

[0088] (2) Conduct a correlation analysis between the Sargassum index values of the sampling points and the characteristic data of each band of the target satellite data, and select the target bands sensitive to Sargassum.

[0089] Since the reflectivity and other characteristics of Sargassum are different in different bands, through correlation analysis, the bands closely related to the Sargassum index values of the sampling points are found. These bands are more sensitive to the characteristics of Sargassum. For example, if the correlation coefficient between a certain band and the Sargassum index value of the sampling point is high, it means that the information of this band can better reflect the characteristics of Sargassum. Selecting it as the target band can improve the accuracy of subsequent identification.

[0090] (3) For the new target band combination, use multiple functions for fitting, determine the function form with the largest goodness-of-fit index value and the smallest error index value, solve the function coefficients, and construct a new inversion model.

[0091] For the newly determined target band combination, use multiple functions (such as linear functions, power functions, exponential functions, etc.) for fitting. By calculating the goodness-of-fit index values (such as the coefficient of determination R 2 ) and error index values (such as the root mean square error RMSE) under different function fittings, select the function form with the largest goodness-of-fit index value and the smallest error index value as the optimal function form. Then determine the coefficients of this function form, thereby constructing a new inversion model applicable to the satellite data.

[0092] (4) At the preset time, collect the data of the Sargassum and Enteromorpha sampling points under the target satellite data, calculate the index values, and re-determine the Sargassum index interval and the Enteromorpha index interval.

[0093] Further, after calculating the index values using the newly constructed inversion model, by statistically analyzing these index values, such as calculating the mean, standard deviation, etc., re-determine the Sargassum index interval and the Enteromorpha index interval. Since the characteristics of different satellite data are different, the corresponding algae index intervals may also be different. Re-determining can ensure the accuracy of identification. Specifically, the methods for determining the Sargassum index interval and the Enteromorpha index interval are described below and will not be elaborated here.

[0094] (5) Based on the re-determined Sargassum index interval and Enteromorpha index interval, judge the distribution area of Sargassum.

[0095] Based on the re-determined index intervals, compare the Sargassum index values calculated from the target satellite real-time data with these two intervals. If the index value is within the Sargassum index interval, it is determined as the Sargassum distribution area; if it is within the Enteromorpha index interval, it is determined as the Enteromorpha distribution area. In this way, the distribution area of Sargassum can be accurately judged according to the target satellite data.

[0096] In this way, by performing targeted processing for different satellite types, considering their unique spectral response functions and radiation resolutions and other characteristics, selecting appropriate target bands and constructing exclusive inversion models, Sargassum can be identified more accurately and misjudgments can be reduced. In addition, this Sargassum identification method can also be applied to multiple satellite data. Different satellites have their own advantages and limitations in ocean monitoring. Through this process, the data of different satellites can be fully utilized. Whether it is a low-orbit high-resolution satellite or a high-orbit wide-coverage satellite, it can be used for Sargassum identification, expanding the application scope of the method.

[0097] S106. Calculate the Sargassum index value based on the real-time remote sensing data and the inversion model.

[0098] It should be noted that different landforms (such as Sargassum, Enteromorpha, seawater, etc.) have different reflection and radiation characteristics for electromagnetic waves in different bands. In previous studies, by analyzing the reflectivity data of a large number of Sargassum sampling points and Enteromorpha sampling points in various bands, the target band combination that is sensitive to the characteristics of Sargassum was screened out, and the appropriate function form was determined to describe the relationship between these band data and the characteristics of Sargassum, and then the inversion model was constructed. When there is real-time remote sensing data, the target band data can be processed using the inversion model, and the information contained in these band data can be converted into a quantitative Sargassum index value. This index value can reflect the relevant information of Sargassum in the area. For example, the size of the index value may be related to the density and biomass of Sargassum.

[0099] S107, determining the time characteristics of the real-time remote sensing data, and determining the optimal discrimination method for distinguishing Sargassum and Enteromorpha based on the time characteristics; the optimal discrimination method includes: comparing the calculated Sargassum index with the index intervals corresponding to Sargassum and Enteromorpha, respectively, to determine the category corresponding to the real-time remote sensing data.

[0100] It should be noted that the time characteristics of real-time remote sensing data (i.e., algae growth cycle characteristics in this application) refer to the time information corresponding to the remote sensing data acquired in real time, including specific dates, seasons, time periods of the day, etc. At different time points, the lighting conditions on the earth's surface, the growth status of the objects, etc. will be different, and these factors will affect the spectral characteristics of the objects in the remote sensing data. For example, in different seasons, the growth stages of Sargassum and Enteromorpha are different, and their reflectivity and other characteristics in the remote sensing data will also be different.

[0101] It should be noted that the optimal discrimination method for distinguishing Sargassum and Enteromorpha is determined based on the time characteristics. Specifically, the optimal discrimination method under different time periods is determined according to the time distribution characteristics of the growth cycles of the Sargassum and Enteromorpha, and the target time period is determined based on the time characteristics corresponding to the currently identified remote sensing data, and the optimal discrimination method corresponding to the target time period is used as the discrimination method corresponding to the currently identified remote sensing data. When there is only Sargassum or only Enteromorpha, the judgment can be made directly. Among them, when only Sargassum exists, Sargassum is usually distributed in dots, small blocks, strips or clusters, and may form a floating raft-like structure. Sargassum patches are highly dispersed and cover a small area, but under certain conditions (such as spring), they may also form a large-scale aggregation area in the open sea. When only Enteromorpha exists, Enteromorpha is usually distributed in blocks or sheets, covers a large area, and is distributed more concentratedly, usually forming a dense cushion structure in the coastal area.

[0102] When Sargassum and Enteromorpha exist simultaneously, a relatively large aggregation area may form in the open water. At this time, the optimal discrimination method adopted includes: comparing the calculated Sargassum index with the index intervals corresponding to Sargassum and Enteromorpha respectively to determine the category corresponding to the real-time remote sensing data. Before obtaining the index intervals corresponding to Sargassum and Enteromorpha respectively, it is also necessary to determine the endpoints of the index intervals corresponding to Sargassum and Enteromorpha respectively. Specifically, the determination of the endpoints of the index intervals corresponding to Sargassum and Enteromorpha respectively includes:

[0103] (1) Distinguish algae from seawater according to prior knowledge.

[0104] Among them, the seawater value is less than 0, and the algae value is greater than 0. According to the existing prior knowledge, the algae and seawater are initially distinguished by using the index value. Due to the essential differences in the composition, structure, etc. between seawater and algae, the seawater value is less than 0 and the algae value is greater than 0 are presented significantly in the index value obtained through specific remote sensing data processing and calculation.

[0105] (2) Count the index values of a preset number of Sargassum sampling points and Enteromorpha sampling points, calculate the mean and standard deviation of the Sargassum index values of the sampling points and the mean and standard deviation of the corresponding Enteromorpha index values of the sampling points, and set the corresponding upper and lower limits based on the mean and standard deviation to obtain the Sargassum index interval and the Enteromorpha index interval; the Sargassum index interval and the Enteromorpha index interval are independent of each other and have no overlap.

[0106] It should be noted that based on the calculated mean and standard deviation, the upper and lower limits of the Sargassum index interval and the Enteromorpha index interval are set. A common setting method is to take the mean as the center and determine the interval range by adding and subtracting a certain multiple of the standard deviation. For example, usually, the mean plus 1.5 times the standard deviation can be taken as the upper limit, and the mean minus 1.5 times the standard deviation can be taken as the lower limit (1.5 times here is a suitable multiple determined according to the actual situation and experience, aiming to cover most of the normal sampling point data), to obtain the Sargassum index interval and the Enteromorpha index interval. And it is necessary to ensure that these two index intervals are independent of each other and have no overlap, so as to accurately judge whether it is Sargassum or Enteromorpha according to the calculated Sargassum index value in the follow-up. If there is overlap, further adjustment is required.

[0107] S108. Identify the Sargassum index value based on the optimal discrimination method, distinguish Sargassum from Enteromorpha, and obtain the identification result.

[0108] It should be noted that identifying the Sargassum index value based on the optimal discrimination method, distinguishing Sargassum from Enteromorpha, and obtaining the identification result includes:

[0109] (1) Calculate the Sargassum index value according to the inversion model.

[0110] It should be noted that according to the previously constructed inversion model, the band data related to the model in the remotely sensed data obtained in real time is input into the model. The inversion model will process and calculate these band data according to its specific functional form and the determined coefficients, and finally output a value, that is, the Sargassum index value.

[0111] (2) Compare the calculated Sargassum index value with the Sargassum index range and the Enteromorpha index range. If the Sargassum index value is within the Sargassum index range, determine that the target area is the Sargassum distribution area; if the Sargassum index value is within the Enteromorpha index range, determine that the target area is the Enteromorpha distribution area.

[0112] Specifically, it is judged whether the index value is between the lower limit and the upper limit of the Sargassum index range, and at the same time, it is also judged whether it is between the lower limit and the upper limit of the Enteromorpha index range. If the Sargassum index value is within the Sargassum index range, it means that the spectral characteristics of this area conform to the typical spectral characteristics of Sargassum, so it can be determined that the target area is the Sargassum distribution area; if the Sargassum index value is within the Enteromorpha index range, it indicates that the spectral characteristics of this area are more in line with the characteristics of Enteromorpha, thus determining that the target area is the Enteromorpha distribution area. If the index value is not within these two ranges, it may mean that there are other ground objects in this area, or there are abnormalities in the data, and further analysis and processing are required.

[0113] Figure 2 The HY-1C / D CZI image showing undifferentiated Sargassum and Enteromorpha for this application Figure 3 The HY-1C / D CZI image showing differentiated Sargassum and Enteromorpha (only Sargassum) for this application. Please refer to Figure 2 and Figure 3 , it can be seen that based on the optimal discrimination method to identify the Sargassum index value and distinguish between Sargassum and Enteromorpha, the obtained recognition result is more accurate.

[0114] In addition, after determining whether the target area is the Sargassum distribution area, it includes:

[0115] (1) Collect remotely sensed data within the target area and the preset range, including HY-1C / D satellite CZI data under different times and different weather conditions.

[0116] It should be noted that the HY-1C / D satellite is a marine satellite in China, and HY-1C and HY-1D are different models of this series of satellites. These satellites are mainly used for marine water color and temperature environment monitoring and can obtain various information on the ocean surface, including reflectance data in different bands. CZI data is the data obtained by the Coastal Zone Imager. The Coastal Zone Imager is a sensor on the satellite used to image and observe the coastal zone area and obtain multi-band remotely sensed data.

[0117] Under different times (such as different seasons, different months, different dates, etc.) and different weather conditions (sunny, cloudy, overcast, etc.), the spectral characteristics of ground objects will change. Collecting such data can provide a more comprehensive understanding of the performance of Sargassum and other ground objects under various circumstances.

[0118] (2) Conduct statistical analysis on the reflectance data of Sargassum and other similar ground objects in the target band in the collected data to obtain the variation range and distribution characteristics of the reflectance in the target band.

[0119] Conduct statistical analysis on the reflectance data of Sargassum and other similar ground objects (such as other algae, some aquatic plants, etc.) in the target band in the collected data. By calculating statistical quantities such as the maximum value, minimum value, mean value, and standard deviation of the reflectance, obtain the variation range and distribution characteristics of the reflectance in the target band. For example, analyze the mean value and fluctuation range of the reflectance of Sargassum in the target band, as well as the corresponding characteristics of the reflectance of other similar ground objects, in order to find the differences between them.

[0120] (3) Based on the statistical results, combined with the spectral differences between Sargassum and other ground objects in different environments, adjust the Sargassum index range.

[0121] It should be noted that if it is found that the reflectance characteristics of Sargassum have changed in certain cases, resulting in the original index range being unable to accurately reflect the presence of Sargassum, then the range of the index range needs to be expanded or reduced accordingly. For example, if statistics show that the reflectance of Sargassum generally increases in a certain season, then the lower limit of the Sargassum index range can be appropriately increased.

[0122] (4) Based on the adjusted Sargassum index range, screen the areas of unidentifiable pixel points and re-identify Sargassum.

[0123] Use the adjusted Sargassum index range to screen the previous areas of unidentifiable pixel points. Among them, the areas of unidentifiable pixel points refer to the areas where the pixel points that cannot be clearly judged whether Sargassum exists are located during the previous process of identifying Sargassum using the CZI data of HY-1C / D satellites. These areas may be difficult to accurately identify due to reasons such as data quality and unclear spectral characteristics. Judge the remote sensing data of these areas according to the new index range, calculate the corresponding index values and compare them with the adjusted range to re-determine whether these areas are Sargassum distribution areas, so as to achieve the re-identification of these areas.

[0124] After determining whether the target area is a Sargassum distribution area, it further includes:

[0125] (1)Collect other satellite remote sensing data sources except the CZI data of HY-1C / D satellites, and extract the feature information related to Sargassum identification for different data sources.

[0126] It should be noted that for different data sources, the feature information related to Sargassum identification is extracted because different satellite data sources have different characteristics and advantages. By collecting multiple data sources, more comprehensive information can be obtained, improving the accuracy of Sargassum identification.

[0127] (2)According to the characteristics and data types of different data sources, select a fusion algorithm, use the fused data as supplementary information, and screen the areas of unidentifiable pixel points again for Sargassum identification.

[0128] It should be noted that different data sources may have different resolutions, data formats, and statistical characteristics. Therefore, a fusion algorithm that can give full play to the advantages of each data source needs to be selected. For example, if the resolutions of the data sources vary greatly, a fusion algorithm based on wavelet transform can be selected to better retain the detailed information of the image; if there is a strong correlation between the data sources, a fusion algorithm based on principal component analysis can be selected to reduce data redundancy.

[0129] In this way, by collecting multiple satellite remote sensing data sources and performing data fusion, the advantages of each data source can be fully utilized to obtain more comprehensive and accurate information. This helps to more accurately identify Sargassum, reduce misjudgment and missed judgment caused by the limitations of a single data source, and improve the accuracy and reliability of identification. In addition, different satellite data sources have differences in time, space, and spectral resolutions. Through data fusion, these differences can be compensated, improving the usability of the data. By re-identifying the areas of unidentifiable pixel points, the gaps in the previous identification process can be filled, making the identification results more complete and accurate.

[0130] The method provided in this embodiment calculates the correlation between the index values and the feature data of each band by collecting the sampling point data with different climate characteristics and growth times, screens out the target bands, excludes irrelevant or interfering bands, focuses on the bands crucial for Sargassum identification, and improves the data processing efficiency and identification accuracy. Distinguish algae and seawater based on prior knowledge, and statistically calculate the mean and standard deviation of the index values of Sargassum and Enteromorpha sampling points, and set independent and non-overlapping index intervals. Comparing the calculated Sargassum index values with the index intervals can accurately determine whether the target area is a Sargassum or Enteromorpha distribution area, reducing misjudgment. Analyze the time characteristics of real-time remote sensing data, and determine the optimal discrimination method according to the growth stages and spectral feature differences of Sargassum and Enteromorpha at different times, improving the accuracy and reliability of identification under different time conditions.

[0131] In addition, obtaining real-time remote sensing data and calculating the Sargassum index value in combination with the inversion model can timely grasp the distribution and dynamic changes of Sargassum, provide real-time data support for marine ecological research, environmental protection and resource development, and help take timely measures to address issues such as Sargassum outbreaks.

[0132] Embodiment 2:

[0133] Corresponding to the foregoing embodiment of a method for identifying Sargassum based on remote sensing data, the present application also provides an embodiment of a device for identifying Sargassum based on remote sensing data.

[0134] Figure 4 It is a schematic structural diagram of Embodiment 2 of the device for identifying Sargassum based on remote sensing data provided by the present application. Please refer to Figure 4 The device provided in this embodiment includes a screening module 410, a determination module 420, an acquisition module 430, a calculation module 440, a discrimination module 450, and an identification module 460;

[0135] The screening module 410 is used to screen sampling points of the first order of magnitude, where the climate characteristics and growth times corresponding to each sampling point are different, and the sampling points include Sargassum sampling points and Enteromorpha prolifera sampling points;

[0136] The screening module 410 is further used to separate the characteristic data of each band in the remote sensing data, and screen the target bands based on the correlation between the Sargassum index value of each sampling point and the characteristic data of each band, and there are multiple target bands;

[0137] The determination module 420 is used to determine the form of the model function to be screened according to the remote sensing data differences between Sargassum and Enteromorpha prolifera;

[0138] The determination module 420 is further used to fit the reflectance data of each sampling point under each band combination in the target band based on each form of the model function to be screened, compare the similarity between the fitting curve and the distribution of actual data points, and determine the inversion model;

[0139] The acquisition module 430 is used to acquire real-time remote sensing data;

[0140] The calculation module 440 is used to calculate the Sargassum index value based on the real-time remote sensing data and the inversion model;

[0141] The discrimination module 450 is used to determine the time characteristics of the real-time remote sensing data, and determine the optimal discrimination method for distinguishing Sargassum and Enteromorpha prolifera based on the time characteristics; the optimal discrimination method at least includes: comparing the calculated Sargassum index with the index intervals corresponding to Sargassum and Enteromorpha prolifera respectively, and judging the category corresponding to the real-time remote sensing data;

[0142] The recognition module 460 is configured to recognize the sargassum index value based on the optimal discrimination method, distinguish sargassum from enteromorpha, and obtain a recognition result.

[0143] The device of this embodiment can be used to execute Figure 1 the steps of the method embodiment shown. The specific implementation principle and process are similar and will not be elaborated here.

[0144] 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.

[0145] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can be referred to the partial description of the method embodiment. The device embodiment described above is only illustrative. The unit described as a separated component may or may not be physically separated, and the component shown as a unit may or may not be a physical unit, that is, it may be located in one place or 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. A person of ordinary skill in the art can understand and implement it without creative efforts.

[0146] The above are only the preferred embodiments of this application and are not intended to limit this application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of this application shall be included in the scope of protection of this application.

Claims

1. A method for identifying Sargassum based on remote sensing data, characterized in that, The method includes: Screening sampling points of the first order of magnitude, where the climate characteristics and growth times corresponding to each sampling point are different, and the sampling points include Sargassum sampling points and Enteromorpha prolifera sampling points; Separating the characteristic data of each band in the remote sensing data, and screening target bands based on the correlation between the Sargassum index values of each sampling point and the characteristic data of each band, where there are multiple target bands; Determining the form of the model function to be screened according to the remote sensing data differences between Sargassum and Enteromorpha prolifera; Fitting the reflectance data of each sampling point under each band combination in the target band based on each form of the model function to be screened, comparing the similarity between the fitting curve and the distribution of actual data points, and determining the inversion model; Obtaining real-time remote sensing data; Calculating the Sargassum index value based on the real-time remote sensing data and the inversion model; Determining the time characteristics of the real-time remote sensing data, and determining the optimal discrimination method for distinguishing Sargassum and Enteromorpha prolifera based on the time characteristics; the optimal discrimination method at least includes: comparing the calculated Sargassum index with the index intervals corresponding to Sargassum and Enteromorpha prolifera respectively, and judging the category corresponding to the real-time remote sensing data; Identifying the Sargassum index value based on the optimal discrimination method, distinguishing Sargassum and Enteromorpha prolifera, and obtaining the identification result; The determining the inversion model includes: Selecting multiple model functions to be screened according to the remote sensing data differences between Sargassum and Enteromorpha prolifera; For each form of the model function to be screened, fitting using the reflectance data of each sampling point under each band combination in the target band; Calculating the goodness-of-fit index value and the error index value under different function fittings, and selecting the function form with the largest goodness-of-fit index value and the smallest error index value during the fitting process of Sargassum remote sensing data, and at the same time, the discrimination degree greater than the preset value during the fitting process of Enteromorpha prolifera remote sensing data as the optimal function form of the target band combination; Determining the function coefficients based on the optimal function form, and constructing an inversion model with the determined target band combination, optimal function form and function coefficients.

2. The method according to claim 1, wherein The screening of the target band based on the correlation between the Sargassum index values of each sampling point and the characteristic data of each band includes: Based on the correlation calculation method, calculating the correlation coefficient between the Sargassum index value of each sampling point and the characteristic data of each band respectively; Comparing the calculated correlation coefficient with the preset correlation coefficient threshold, and screening out the bands with the correlation coefficient greater than or equal to the threshold; For the screened bands, sorting them in descending order of the correlation coefficient, and selecting the top n bands in the ranking as the target bands according to actual needs; n is a positive integer greater than 1.

3. The method according to claim 1, wherein The inversion model is in the form of a linear function. The inversion model takes the difference between the near-infrared band and the green band combination as the independent variable, and the sum of the product of the independent variable and the independent variable coefficient and the constant term gives the linear function.

4. The method according to claim 1, wherein The identifying the Sargassum index value based on the optimal discrimination method, distinguishing Sargassum and Enteromorpha prolifera, and obtaining the identification result includes: Calculating the Sargassum index value according to the inversion model; Compare the calculated Sargassum index value with the Sargassum index range and the Enteromorpha index range. If the Sargassum index value is within the Sargassum index range, determine that the target area is a Sargassum distribution area. If the Sargassum index value is within the Enteromorpha index range, determine that the target area is an Enteromorpha distribution area.

5. The method according to claim 1, wherein After determining whether the target area is a Sargassum distribution area, it includes: Collect remote sensing data within the target area and the preset range, including HY-1C / D satellite CZI data under different times and weather conditions; Statistically analyze the reflectance data of Sargassum and other similar ground objects in the target band in the collected data, and obtain the change range and distribution characteristics of the reflectance in the target band; Based on the statistical results, combine the spectral differences between Sargassum and other ground objects in different environments to adjust the Sargassum index range; Based on the adjusted Sargassum index range, screen the area of pixels that cannot be recognized, and re-identify Sargassum.

6. The method according to claim 1, wherein After determining whether the target area is a Sargassum distribution area, it also includes: Collect other satellite remote sensing data sources in addition to HY-1C / D satellite CZI data, and extract characteristic information related to Sargassum identification for different data sources; According to the characteristics and data types of different data sources, select a fusion algorithm, use the fused data as supplementary information, and screen the area of pixels that cannot be recognized again for Sargassum identification.

7. The method according to claim 1, wherein The determination of the index range endpoints corresponding to Sargassum and Enteromorpha respectively includes: Distinguish algae and seawater according to prior knowledge; Statistically analyze the index values of a preset number of Sargassum sampling points and Enteromorpha sampling points, calculate the mean and standard deviation of the Sargassum index values corresponding to the sampling points and the mean and standard deviation of the Enteromorpha index values corresponding to the sampling points, and set corresponding upper and lower limits based on the mean and standard deviation to obtain the Sargassum index range and the Enteromorpha index range; the Sargassum index range and the Enteromorpha index range are independent of each other and do not overlap.

8. The method according to claim 1, wherein Before obtaining real-time remote sensing data, it also includes determining the satellite type corresponding to the real-time remote sensing data, specifically including: Obtain the spectral response function and radiation resolution of the target satellite; Conduct a correlation analysis between the Sargassum index values of the sampling points and the characteristic data of each band of the target satellite data, and select the target band sensitive to Sargassum; For the new target band combination, use multiple function fittings to determine the function form with the largest goodness-of-fit index value and the smallest error index value, and solve the function coefficients to construct a new inversion model; At the preset time, collect the data of Sargassum and Enteromorpha sampling points under the target satellite data, calculate the index values, and re-determine the Sargassum index range and the Enteromorpha index range; Based on the re-determined Sargassum index range and Enteromorpha index range, judge the Sargassum distribution area.

9. A Sargassum identification device based on remote sensing data, characterized in that, The device includes a screening module, a determination module, an acquisition module, a calculation module, a distinction module, and an identification module; The screening module is used to screen sampling points of the first order of magnitude, and the climate characteristics and growth times corresponding to each sampling point are different. The sampling points include Sargassum sampling points and Enteromorpha sampling points; The screening module is further configured to separate the feature data of each band in the remote sensing data, and screen the target bands based on the correlation between the Sargassum index values of each sampling point and the feature data of each band, where there are multiple target bands; The determination module is configured to determine the form of the model function to be screened according to the difference in remote sensing data between Sargassum and Enteromorpha; The determination module is further configured to fit the reflectance data of each sampling point under each band combination in the target band based on each form of the model function to be screened, compare the similarity between the fitted curve and the distribution of actual data points, and determine the inversion model; The acquisition module is configured to acquire real-time remote sensing data; The calculation module is configured to calculate the Sargassum index value based on the real-time remote sensing data and the inversion model; The discrimination module is configured to determine the time characteristics of the real-time remote sensing data, and determine the optimal discrimination method for distinguishing Sargassum and Enteromorpha based on the time characteristics; The optimal discrimination method at least includes: comparing the calculated Sargassum index with the index intervals corresponding to Sargassum and Enteromorpha respectively to determine the category corresponding to the real-time remote sensing data; The identification module is configured to identify the Sargassum index value based on the optimal discrimination method, distinguish between Sargassum and Enteromorpha, and obtain the identification result; The determination of the inversion model includes: Selecting multiple functions to be screened according to the difference in remote sensing data between Sargassum and Enteromorpha; For each form of the function to be screened, fitting is performed using the reflectance data of each sampling point under each band combination in the target band; Calculating the goodness-of-fit index value and the error index value under different function fittings, and selecting the function form with the largest goodness-of-fit index value and the smallest error index value during the fitting of Sargassum remote sensing data, and at the same time, the discrimination degree greater than the preset value during the fitting of Enteromorpha remote sensing data as the optimal function form of the target band combination; Determining the function coefficients based on the optimal function form, and constructing an inversion model with the determined target band combination, optimal function form, and function coefficients.

Images