A Method and System for Chlorophyll a Retrieval in Inland Lakes Based on Hyperspectral Data

Abstract

This disclosure relates to the field of chlorophyll a inversion technology for lakes, proposing a method and system for chlorophyll a inversion of inland lakes based on hyperspectral data. The method uses a multi-band index model and regression methods to invert the chlorophyll a content within the lake area. First, hyperspectral image data of the lake is acquired and preprocessed to remove noise and improve data quality. Then, the NDWI index is used to extract the water body extent, and band information from different sampling points is extracted. Based on the optical characteristics of lake water quality, a multi-band index model considering pigment absorption and remote sensing reflectance is constructed, and the inversion model is optimized by combining regression analysis. By using the above method and fusing hyperspectral remote sensing image data from only a small number of points, the constructed model is applied to the chlorophyll a distribution characteristics of lakes at the monitoring area scale, achieving accurate inversion of chlorophyll a concentration across the entire lake area, thus solving the problem of low accuracy when using only remote sensing data for inversion.

Description

Technical Field

[0001] This disclosure relates to the technical field of chlorophyll a inversion in lakes, specifically to a method and system for chlorophyll a inversion in inland lakes based on hyperspectral data. Background Technology

[0002] The statements in this section are merely background information relating to this disclosure and do not necessarily constitute prior art.

[0003] Under the combined influence of global warming and human activities, lake ecosystems face severe challenges, manifested in shrinking lake areas, intensified eutrophication, frequent algal blooms, water quality deterioration, and ecosystem function degradation. Satellite data monitoring results show that, as of the summer of 2012, over 63.1% of the 2058 lakes globally with an area greater than 25 km² were highly eutrophic, 39.4% were moderately eutrophic, and only 10.7% were oligotrophic. Therefore, monitoring lake eutrophication has become a critical need for the current environmental management of inland lakes. Chlorophyll-a is a common pigment in aquatic algae and is one of the important indicators for measuring the degree of eutrophication, as well as a key parameter in the global water carbon cycle. Obtaining long-term series data on lake chlorophyll-a content is of great significance for scientific research, socio-economic decision-making, and environmental management.

[0004] Traditional water quality monitoring methods primarily rely on field sampling and laboratory chemical analysis, assessing water conditions through data from isolated sampling points. However, limited by lake area, sampling density, and experimental costs, this method is not only extremely resource-intensive but also results in limited data coverage, making it difficult to comprehensively reflect the spatiotemporal variations of lake water quality. Compared to traditional monitoring methods, remote sensing technology offers advantages such as wide coverage, stable monitoring cycles, and low cost, enabling faster and more comprehensive acquisition of water quality information. In recent years, monitoring of chlorophyll a in lakes has mainly relied on ocean color satellites (such as MODIS, MERIS, and OLCI), using long-term series data for lake water quality change analysis. However, ocean color remote sensing satellites have low spatial resolution (300-1000m), suitable for monitoring large water areas but unable to meet the needs of refined monitoring of small lakes, bays, and small to medium-sized reservoirs. Furthermore, current remote sensing monitoring mainly relies on spaceborne multispectral remote sensing imagery, but multispectral remote sensing suffers from limited bands, lack of spectral information, and low spatial resolution, making it difficult to meet the requirements for high-precision water quality parameter retrieval. Meanwhile, spaceborne remote sensing is significantly affected by atmospheric conditions, and data is easily interfered with by clouds and fog, further limiting monitoring accuracy and data availability. Currently, water quality monitoring mainly relies on spaceborne multispectral remote sensing imagery as a data source. However, these images have several limitations: First, the spectral bands are limited and discontinuous, making it difficult to accurately identify the spectral characteristics of inland water bodies; second, the spatial resolution is relatively coarse, making it difficult to meet the needs of fine-grained quantitative remote sensing inversion. Furthermore, spaceborne remote sensing data acquisition cycles are long and easily affected by atmospheric conditions, often resulting in images containing a large amount of clouds and fog that are difficult to remove effectively. In addition, inland water bodies are typically small in area and fragmented, with complex optical characteristics, making them susceptible to the influence of weather, surrounding environment, and human activities. These factors further reduce the accuracy of water quality parameter inversion, leading to significant errors between the inversion results and the actual targets. Therefore, various difficulties exist in inverting water quality.

[0005] In remote sensing monitoring of lake water quality parameters, chlorophyll a retrieval currently employs three main models: empirical, analytical, and semi-analytical. Empirical methods are suitable for clean water bodies, but for turbid lakes (Class II water bodies), whose optical characteristics are complex and influenced by various factors such as chlorophyll a, total suspended solids, and yellow substances, relying solely on empirical methods cannot guarantee universality. Furthermore, empirical methods typically rely on data from specific years to build models, lacking applicability across time and regions. While analytical methods offer higher accuracy, they require a large amount of synchronous field observation data, which is difficult to obtain. Therefore, semi-analytical methods have become the mainstream approach, offering certain advantages in quantitative retrieval of water quality parameters. However, the complexity of lake water quality still limits the practical application of these methods.

[0006] In summary, accurate chlorophyll a retrieval cannot be achieved using single remote sensing data. Existing chlorophyll a retrieval methods are mainly designed for marine water types. Due to the unique characteristics of lake environments, water bodies are affected by multiple factors, such as suspended solids, phytoplankton, and dissolved organic matter. These factors work together to create complex optical properties. In addition, inland lakes often have high turbidity, which increases light scattering and absorption, reducing the quality of remote sensing signals. This makes it difficult to retrieve water quality parameters based solely on remote sensing signals, and when applied to lake waters, the retrieval results are often unsatisfactory. Furthermore, the complexity of lake water quality limits the practical application of existing methods. Summary of the Invention

[0007] To address the aforementioned issues, this disclosure proposes a method and system for inverting chlorophyll a in inland lakes based on hyperspectral data. This method, specifically for chlorophyll a retrieval in inland lakes, enables inversion analysis of complex inland lake water bodies and fully reflects the spatial distribution characteristics of water quality parameters across the entire lake. It enhances the applicability and accuracy of the inversion method, providing a scientific basis for inland lake water quality monitoring, and also offering theoretical support and reference for the treatment and improvement of inland lake water quality.

[0008] To achieve the above objectives, the present disclosure adopts the following technical solution:

[0009] One or more embodiments provide a method for inverting chlorophyll a in inland lakes based on hyperspectral data, comprising the following steps:

[0010] By collecting lake water samples in the field, a measured dataset of chlorophyll a in lakes was constructed.

[0011] Acquire a dataset of lake remote sensing reflectance and extract hyperspectral satellite imagery data from the measured sampling time for preprocessing;

[0012] The preprocessed hyperspectral satellite image data is used to extract the water body range through NDWI index calculation, and band information is extracted to obtain the band data and remote sensing images of the lake area corresponding to each measured point.

[0013] A multi-band index model was constructed to consider the effects of pigment absorption and remote sensing reflectance. The multi-band index model and the regression model were then integrated to construct an inversion model. The data from the measured dataset were used as the output, and the band data from each measured point and the remote sensing images of the lake area were used as the input. The constructed inversion model was then iteratively optimized.

[0014] Obtain hyperspectral satellite imagery data of the location to be measured, and input it into the optimized inversion model to obtain the inversion result of chlorophyll a at the location to be measured.

[0015] One or more embodiments provide an inland lake chlorophyll a retrieval system based on hyperspectral data, comprising:

[0016] The field measurement module is configured to construct a field measurement dataset of chlorophyll a in lakes by collecting lake water samples in the field.

[0017] The first remote sensing data processing module is configured to acquire a lake remote sensing reflectance dataset and extract hyperspectral satellite image data from the measured sampling time for preprocessing.

[0018] The second remote sensing data processing module is configured to extract the water body range from the preprocessed hyperspectral satellite image data through NDWI index calculation, extract band information, and obtain the band data and remote sensing images of the lake area corresponding to each measured point.

[0019] The model optimization module is configured to construct a multi-band index model considering the effects of pigment absorption and remote sensing reflectance, and to construct an inversion model by fusing the multi-band index model and the regression model. The inversion model is constructed using the data from the measured dataset as output and the band data from each measured point and the remote sensing images of the lake area as input. The inversion model is iteratively optimized.

[0020] The inversion module is configured to acquire hyperspectral satellite image data of the location to be measured, and input it into the optimized inversion model to obtain the inversion result of chlorophyll a at the location to be measured.

[0021] An electronic device includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, which, when executed by the processor, perform the steps in the above-described method for inverting chlorophyll a in inland lakes based on hyperspectral data.

[0022] A computer-readable storage medium for storing computer instructions, which, when executed by a processor, complete the steps in the above-described method for inverting chlorophyll a in inland lakes based on hyperspectral data.

[0023] Compared with the prior art, the beneficial effects of this disclosure are as follows:

[0024] This disclosed method utilizes hyperspectral satellite imagery combined with measured data to invert chlorophyll a content within a lake using a multi-band index model and regression methods. First, hyperspectral imagery of the lake is acquired and preprocessed to remove noise and improve data quality. Then, the NDWI index is used to extract the water body extent, and band information from different sampling points is extracted. Based on the optical characteristics of lake water quality, a multi-band index model considering pigment absorption and remote sensing reflectance is constructed, and the inversion model is optimized using regression analysis. Finally, the optimized model is used to process the hyperspectral imagery of the area to be monitored. By using the above method, which combines measured data from only a small number of points with hyperspectral remote sensing imagery, and employing a model constructed based on bio-optical principles and statistical methods, the distribution characteristics of chlorophyll a in the lake at the monitoring regional scale are accurately inverted, achieving precise inversion of chlorophyll a concentration across the entire lake area. This enables comprehensive chlorophyll a water quality monitoring across the entire lake region, achieving both regional and macro-level analysis, and solving the problem of low accuracy when using only remote sensing data for inversion.

[0025] The advantages of this disclosure, as well as its additional advantages, will be described in detail in the following specific embodiments. Attached Figure Description

[0026] The accompanying drawings, which form part of this disclosure, are used to provide a further understanding of this disclosure. The illustrative embodiments of this disclosure and their descriptions are used to explain this disclosure and do not constitute a limitation thereof.

[0027] Figure 1 This is a flowchart of the inversion method of Embodiment 1 of this disclosure;

[0028] Figure 2 This is a schematic diagram of the accuracy evaluation results of the chlorophyll inversion model based on ZY-1 02E satellite hyperspectral data in the experimental example of Embodiment 1 of this disclosure;

[0029] Figure 3 This is a chlorophyll inversion result image based on ZY-1 02E satellite hyperspectral data in the experimental example of Embodiment 1 of this disclosure; Detailed Implementation

[0030] The present disclosure will be further described below with reference to the accompanying drawings and embodiments.

[0031] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of this disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains.

[0032] It should be noted that the terminology used herein is for descriptive purposes only and is not intended to limit the exemplary embodiments according to this disclosure. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof. It should be noted that, without conflict, the various embodiments and features within those embodiments can be combined with each other. The embodiments will now be described in detail with reference to the accompanying drawings.

[0033] Example 1

[0034] In one or more of the technical solutions disclosed in the embodiments, such as Figures 1 to 3 As shown, the method for inverting chlorophyll a in inland lakes based on hyperspectral data includes the following steps:

[0035] Step 1: Collect lake water samples in the field to construct a measured dataset of chlorophyll a in the lake and obtain a remote sensing reflectance dataset of the lake;

[0036] Step 2: Extract hyperspectral satellite imagery data from the lake remote sensing reflectance dataset at the measured sampling time and perform preprocessing.

[0037] Step 3: Extract the water body range from the preprocessed hyperspectral satellite image data using NDWI index calculation, extract band information, and obtain the band data and remote sensing images of the lake area corresponding to each measured point.

[0038] Step 4: Considering the influence of pigment absorption and remote sensing reflectance, construct a multi-band index model, integrate the multi-band index model and the regression model to construct an inversion model, take the data of the measured dataset as the output, and take the band data of each measured point and the remote sensing image of the lake area as the input to iteratively optimize the constructed inversion model.

[0039] Step 5: Obtain hyperspectral satellite imagery data of the location to be measured, and input it into the optimized model to obtain the inversion result of chlorophyll a at the location to be measured.

[0040] In this embodiment, the method utilizes hyperspectral satellite imagery data combined with measured data to invert chlorophyll a content within a lake using a multi-band index model and regression methods. First, hyperspectral imagery data of the lake is acquired and preprocessed to remove noise and improve data quality. Then, the NDWI index is used to extract the water body extent and extract band information from different sampling points. Based on the optical characteristics of lake water quality, a multi-band index model considering pigment absorption and remote sensing reflectance is constructed, and the inversion model is optimized using regression analysis. Finally, the optimized model is used to process the hyperspectral imagery of the area to be measured. By fusing hyperspectral remote sensing imagery data from only a small number of points using the above method, the constructed model is further applied to the chlorophyll a distribution characteristics of the lake at the monitoring area scale. This enables accurate inversion of chlorophyll a concentration across the entire lake area, achieving chlorophyll a water quality detection for the entire lake region. This approach integrates regional and macroscopic data analysis, solving the problem of low accuracy when using only remote sensing data for inversion.

[0041] The algorithm flow of this embodiment has greater applicability, ensuring better results when applied to complex inland lakes, and enabling the detection of spatiotemporal dynamic changes in overall lake surface water quality for different lakes. This applicability makes the present invention suitable not only for inland high-turbidity lakes but also for high water quality monitoring accuracy, which is of great significance for the future development of satellite data monitoring.

[0042] The above method first performs multi-point sampling in the field for the area to be detected (such as a lake). Based on the sampled measured data and data points corresponding to the locations to be detected in satellite remote sensing data, the trained inversion model is optimized. Finally, based on the optimized model using data from the area to be detected, a comprehensive detection and analysis of the area is performed. When the lake to be detected is changed, the same method is used to re-optimize the model and perform detection again.

[0043] In step 1, lake water samples are collected in the field to obtain lake chlorophyll a dataset and lake remote sensing reflectance dataset. The lake chlorophyll a dataset is used as the output of the inversion model.

[0044] The lake chlorophyll a measured dataset includes the coordinates of the measured chlorophyll a locations in the lake, the concentration and spectral data of chlorophyll a at the corresponding locations;

[0045] The coordinate information is used to provide image data for the corresponding location in step 2 as input to the inversion model, the chlorophyll a concentration is used as output to the inversion model, and the spectral data provides a reference for the range of the inversion model iteration.

[0046] The lake remote sensing reflectance dataset is image data obtained through hyperspectral satellites, including data such as reflectance corresponding to location coordinates; the remote sensing image data is hyperspectral satellite image data, and in this embodiment, the images are collected from the Lake Resource-1 hyperspectral satellite.

[0047] In this step, the acquired remote sensing image data corresponds to the actual sampling time in step 1, and is the sampling data within the same time period;

[0048] Furthermore, methods for preprocessing the acquired remote sensing image data include the following:

[0049] 1) Perform radiometric calibration processing, converting the DN value into a radiance image based on the radiometric calibration parameters. The calculation formula is as follows:

[0050] L = G·DN + B;

[0051] Where L is radiance; DN is the original digital quantization value; G is the gain coefficient; and B is the bias.

[0052] 2) Convert the radiometrically calibrated data to BIL storage format and perform atmospheric correction;

[0053] Specifically, for ZY-1 02E satellite data, the sensor altitude is set to 778km, the water vapor inversion is performed using the 940nm band, the number of bands for spectral polishing is set to 3, and the FLAASH atmospheric correction algorithm can be used for atmospheric correction.

[0054] In this embodiment, different image processing methods are used for remote sensing data. Preprocessing such as radiometric calibration, atmospheric correction, and geometric correction is performed on the remote sensing images to ensure the accuracy of information extraction for each type of data and to extract the details of image features, thereby improving the accuracy of subsequent chlorophyll a inversion.

[0055] In step 3, the water body range and band information are extracted from the preprocessed remote sensing image data to obtain the data of each band in the remote sensing image data and the remote sensing image of the water area of ​​the lake to be measured.

[0056] Step 31: Water body range extraction

[0057] 31.1) Calculate the NDWI index for the obtained remote sensing image data. The calculation formula is as follows:

[0058]

[0059] Wherein, Green represents the reflectance of the green band in the remote sensing image data, and NIR is the reflectance of the red band in the remote sensing image data;

[0060] 31.2) Set the threshold to zero, and define the area with an NDWI index greater than the threshold as a water body area by comparing the NDWI index with the threshold.

[0061] In this embodiment, when NDWI>0, it indicates that the area is a water body. Areas with higher NDWI values ​​usually represent water bodies, while vegetation, soil, etc. are non-water body areas.

[0062] 31.3) Crop the remote sensing image to retain water areas and remove non-water areas.

[0063] In the above scheme, the water body information is enhanced by calculating the normalized difference between the green band and the near-infrared band, while suppressing the influence of vegetation and soil, thus achieving high-precision water body range extraction.

[0064] Step 32: Extract band information based on the measured locations of chlorophyll a data of the lake, and use it as input for the inversion model; that is, extract the band information of the corresponding measured locations in the remote sensing image based on the measured locations of chlorophyll a data of the lake.

[0065] 32.1) For the extracted remote sensing image data of the water body area, perform image reprojection to match the coordinate system of the measured point data, so that the coordinate system of the remote sensing image is consistent with the measured point.

[0066] 32.2) Obtain the measured point location and extract the band raster value of the remote sensing image corresponding to the coordinate position of the measured point location as the extracted band information;

[0067] In this embodiment, the band information extraction is performed by extracting the raster value of each band of the image corresponding to the measured lake chlorophyll a sample point based on the coordinate position of the sample point. This value is used as the remote sensing reflectance corresponding to the measured lake sample point.

[0068] Raster value refers to the numerical value of a pixel in a remote sensing image, which usually represents the reflectance or radiance of that pixel in a certain band. In remote sensing data processing, each pixel has corresponding geographic coordinates (latitude and longitude or projected coordinates) and values ​​for multiple bands. These values ​​are used to analyze surface features.

[0069] In step 4, for the constructed inversion model, the parameter settings of the multi-band exponential model are determined by combining the measured spectral data and the iterative algorithm, and the optimal inversion model is determined by screening through various regression models.

[0070] Optionally, the multi-band index model can adopt a three-band index model or a four-band index model. Preferably, this embodiment adopts a three-band index model.

[0071] Furthermore, the method for constructing the inversion model includes the following steps:

[0072] Step 41: Considering the effects of pigment absorption and remote sensing reflectance, select the optimal band;

[0073] Step 42: Construct a multi-band index model based on the selected optimal band and calculate the three-band index;

[0074] In this embodiment, considering the influence of pigment absorption and remote sensing reflectance, the specific form of the three-band index model is as follows:

[0075] x=(R rs (λ1) -1 -R rs (λ2) -1 )*R rs (λ3)

[0076] Where λ1 represents the first optimal band, λ2 represents the second optimal band, and λ3 represents the third optimal band; R rs This indicates the reflectivity of the corresponding waveband;

[0077] When constructing a three-band index model, the method for selecting the optimal band includes the following steps:

[0078] 4.1) The first optimal band λ1 is selected as the chlorophyll a absorption peak band;

[0079] Specifically, the first optimal band λ1 is selected as the position where pigment absorption has the greatest impact on remote sensing reflectance, and the absorption of xanthine and non-pigment particles and the total backscattering have a smaller impact on remote sensing reflectance, namely the chlorophyll a absorption peak in the red band.

[0080] 4.2) The second optimal band λ2 was selected as the chlorophyll a fluorescence peak band;

[0081] Specifically, the second optimal band λ2 is selected as the position near λ1 where the absorption of chlorophyll a is relatively small, that is, the chlorophyll a fluorescence peak in the red light band can meet the selection requirements of λ2.

[0082] 4.3) The third optimal band λ3 is selected as the pure water absorption band;

[0083] Specifically, the selection criteria for the third optimal band λ3 are: a band whose total absorption coefficient is much greater than the backscattering coefficient and is not affected by the absorption of chlorophyll, non-pigment particles and xanthine, that is, the remote sensing reflectance is mainly affected by the absorption of pure water.

[0084] In this embodiment, an innovative selection condition is proposed, in which three optimal bands are selected and an index is constructed to improve the inversion accuracy, which is particularly suitable for water quality monitoring in high turbidity waters.

[0085] Furthermore, after selecting the optimal bands of the three-band index model, the determination of the three optimal bands λ1, λ2, and λ3 is achieved through iterative optimization using the characteristic band range as the final optimal bands. This involves iteratively performing correlation analyses with lake chlorophyll a concentration, including the following steps:

[0086] Step 4-1: Based on the spectral data obtained from the lake water samples collected in Step 1, preliminarily determine the characteristic band ranges of the three optimal bands, which will serve as the preliminary band ranges for λ1, λ2, and λ3.

[0087] For example, in this embodiment, based on the spectral data of a certain lake, the characteristic band range is used to determine the three optimal bands λ1, λ2 and λ3 according to the measured spectral curve. The value range of λ1 is 650-680nm, the value range of λ2 is 680-720nm, and the value range of λ3 is 700-770nm.

[0088] Step 4-2: Based on the preliminary band range corresponding to the corresponding satellite, determine the band range corresponding to the corresponding satellite. Based on the turbidity, broaden the band selection range determined by the second optimal band λ2 and the third optimal band λ3 to obtain the band selection ranges of λ1, λ2 and λ3.

[0089] Based on the value range of the three bands, the equivalent value range of the ZY-1 02E satellite hyperspectral data is as follows: the value range of λ1 is B32-B36, the value range of λ2 is B36-B42, and the value range of λ3 is B38-B46. Iteration is performed within the selected range of the three bands respectively.

[0090] Step 4-3: Within the range of λ1, λ2 and λ3 bands, fix any two bands and iterate the remaining band within the range of band selection. In each iteration, calculate the correlation between the corresponding band data and the concentration of chlorophyll a, and identify the band with the highest correlation as the output band to obtain the optimal band of the three bands.

[0091] Specifically, starting with λ1, bands λ2 and λ3 are fixed as B39 and B42, respectively. Iterations are performed within the iteration range of λ1. The band with the highest correlation in the results is taken as the output band for λ1. Then, iterations are performed for λ2. After completing the iterations of λ2, iterations of λ3 continue until all three bands have been iterated. λ1 is iterated again. If the resulting band is the same as in the first iteration, then the three selected bands are considered the optimal bands; otherwise, iterations continue until the same iteration result is obtained.

[0092] Optionally, the Pearson simple correlation coefficient can be used as a reference index during the iteration process, or the root mean square error can be selected as an iteration index.

[0093] Furthermore, in step 4-2, based on the turbidity of the lake area sampled in the field, the band selection range determined by the second optimal band λ2 and the third optimal band λ3 of the three optimal bands is broadened to serve as the selected band selection range.

[0094] Specifically, the formula for broadening the band selection range is as follows:

[0095]

[0096] in, To determine the maximum suspended solids concentration in the lake during season t within a given time period, Let C be the minimum suspended solids concentration of the lake in season t within a given time period, and let E be the average suspended solids concentration of the lake in that water area.

[0097] Optionally, the time period can be any period of years, months, etc. For example, if the time period is set to three years, the suspended solids concentration corresponding to season t for each year can be taken, and the maximum value within the three years can be used as the time period. Minimum value as

[0098] For waters with high turbidity, the response of remote sensing reflectance to chlorophyll a concentration is less obvious due to the high concentration of suspended matter. Therefore, the range of characteristic bands for the three optimal bands of the three-band index for waters with high turbidity should be broadened and improved.

[0099] This embodiment broadens the band selection range. As water turbidity increases, the range of band selection can achieve comprehensive coverage. The broadened band range encompasses more spectral features, capturing reflectance changes caused by suspended matter, thereby improving the sensitivity of remote sensing data to suspended matter. The broadened band selection range obtained through the broadening formula in this embodiment reduces instability in the inversion model caused by insufficient spectral variation, especially in high turbidity regions. In this case, the inversion model can extract water quality information more stably, reducing outliers and biases, and enhancing the model's predictive ability.

[0100] In the above scheme, a three-band index model is constructed. Based on the measured spectral morphology characteristics, an appropriate band range is selected and the correlation with the measured chlorophyll a concentration is calculated using an iterative algorithm. The above process constructs a three-band index model. The following step is to establish a chlorophyll a remote sensing inversion model suitable for lakes through regression model fusion processing.

[0101] Step 43: Using the three-band index calculated by the multi-band index model as the independent variable and the concentration of chlorophyll a in the measured dataset of chlorophyll a as the output, establish multiple regression models for fitting.

[0102] Due to the high suspended matter content in lakes, the relationship between chlorophyll a content and the three-band index is sometimes not merely linear. Different regions or different satellite data may employ different regression models after screening. Therefore, five regression models—linear, quadratic polynomial, cubic polynomial, exponential, and power—were established to determine a closer relationship between the two. Specific forms include:

[0103] y = ax + b;

[0104] y = ax 2 +bx+c;

[0105] y = ax 3 +bx 2 +cx+d;

[0106] y = ae bx ;

[0107] y = (ax) b ;

[0108] Where y is the concentration of chlorophyll a, x is the three-band index calculated by the multi-band index model, and a, b and c are coefficients that can be obtained after multiple fittings.

[0109] Step 44: Set evaluation indicators, perform leave-one-out cross-validation and error assessment on the fitting results, select the optimal regression model, and fuse it with the constructed multi-band exponential model as the final inversion model.

[0110] The Leave-One-Out Cross Validation (LOOCV) regression model was used to validate its fit accuracy. Specifically, one sample out of N samples was used as a separate validation set, and the remaining N-1 samples were used for modeling. This process was repeated N times. Accuracy testing was then performed to examine the model's precision and stability, allowing for the selection of a more accurate and stable model later.

[0111] The evaluation indicators constructed for error assessment may include: coefficient of determination (R2), root mean square error (RMSE), and mean absolute error percentage (MAPE). The larger the R2, the smaller the RMSE and the smaller the MAPE, the higher the model accuracy and the better the stability. The regression model with the highest accuracy and the best stability is selected as the final model and fused with the multi-band exponential model to form the inversion model.

[0112] In step 5, the hyperspectral satellite image data to be measured is acquired. After preprocessing in step 2, the three-band index is calculated by the multi-band index model in the inversion model and used as the input of the regression model. The optimal regression model selected in the inversion model is used for inversion calculation to obtain the concentration of chlorophyll a.

[0113] This embodiment proposes a remote sensing inversion method for lake chlorophyll a based on remote sensing imagery by combining field-collected lake water sample data with hyperspectral remote sensing imagery. First, a measured dataset of lake chlorophyll a is obtained through field water sample collection, providing reliable reference data for the subsequent inversion model. Second, hyperspectral satellite imagery from the Lake Resources-1 satellite is used, and preprocessed with ENVI software for radiometric calibration, atmospheric correction, and geometric correction to ensure the preservation of image data accuracy and detail. Based on this, the water body extent is extracted using NDWI index calculation, and further band information is extracted to accurately obtain image data related to chlorophyll a concentration. Finally, a three-band index model is constructed, and an iterative algorithm is used to optimize the correlation between band selection and chlorophyll a concentration, thereby achieving the construction of a high-precision remote sensing inversion model for lake chlorophyll a. This technical solution can effectively improve the accuracy of chlorophyll a concentration inversion, providing a scientific basis for lake ecological monitoring and water quality assessment.

[0114] To illustrate the effectiveness of the above method, the inversion method of this embodiment was applied to ZY-1 02E hyperspectral remote sensing data to verify the technical effect;

[0115] (1) The experimental area of ​​this example is a lake. A synchronous experiment with the ZY-1 02E (Ziyuan-1 02E satellite) was carried out in the water area of ​​the experimental lake. Water surface sampling points were set up, and spectral measurement and chlorophyll a determination of surface water samples were completed on site and the location information was recorded. A total of 20 sampling points were collected.

[0116] (2) Obtain ZY-1 02E hyperspectral data on the day of the experiment. The image quality was good and there was no cloud cover over the water body in the study area. Based on the radiometric calibration parameters, the DN values ​​were converted into radiance images and then into BIL storage format. Atmospheric correction was then performed using ENVI's FLAASH module, and finally, geometric correction of the images was performed.

[0117] (3) The NDWI index was calculated using ENVI software. A threshold was set so that when NDWI > 0, the area was considered a body of water. The remote sensing image was then cropped. The calculation formula for NDWI band settings for the ZY-1 02E satellite is as follows:

[0118]

[0119] Based on the location information recorded during the synchronous experiment in the experimental lake water area, band information was extracted from the processed remote sensing image data.

[0120] (4) Construct a three-band index model. Based on the measured spectral curves, the optimal bands λ1, λ2 and λ3 are determined using the characteristic band range. The value range of λ1 is 650-680nm, the value range of λ2 is 680-720nm, and the value range of λ3 is 700-770nm. Based on the value range of the three bands, the equivalent value range of the ZY-1 02E satellite hyperspectral data is: the value range of λ1 is B32-B36, the value range of λ2 is B36-B42, and the value range of λ3 is B38-B46. Iteration is performed within the three band selection ranges respectively. The final iteration results are λ1 is B34, λ2 is B39, and λ3 is B42.

[0121] Correlation was assessed using the Pearson simple correlation coefficient method. A larger absolute value of the correlation coefficient indicates a stronger correlation; a correlation coefficient closer to 1 or -1 indicates a stronger correlation; and a correlation coefficient closer to 0 indicates a weaker correlation. Using the three-band model obtained from the above iterations as the independent variable and chlorophyll a content as the dependent variable, regression fitting was employed. Due to the influence of high suspended matter in lakes, the relationship between chlorophyll a content and the three-band index sometimes exceeds a linear one. Therefore, five regression models—linear, quadratic polynomial, cubic polynomial, exponential, and power-law—were established to explore a closer relationship between the two. The chlorophyll a inversion model is shown below:

[0122] Chla = 92.97 * (R) rs (B34) -1 -R rs (B39) -1 )*R rs (B42)+10.18

[0123] (4) The optimal chlorophyll a remote sensing inversion model established above was applied to the hyperspectral remote sensing images of the study area from the domestic ZY-1 satellite, and the chlorophyll a content remote sensing inversion map of the lake water area was obtained as follows. Figure 2 As shown, from Figure 2 As can be seen, the calculated chlorophyll a values ​​retrieved by the model have a high degree of fit with the measured values. The coefficient of determination R² = 0.91 indicates that the model has strong explanatory power. The root mean square error (RMSE) is 1.01, and the mean absolute percentage error (MAPE) is 8.52%, indicating that the model has small errors and high accuracy in estimating chlorophyll a concentration. In addition, most of the data points are distributed near the 1:1 reference line, indicating that the model has good reliability in retrieving chlorophyll a content in lake waters and can be used for regional water quality monitoring and ecological environment assessment.

[0124] Specifically, the inversion results are as follows: Figure 3As shown in the figure, it can be seen that after optimizing the inversion model by sampling from multiple measurement points, the water quality of the entire lake area can be accurately calculated.

[0125] Example 2

[0126] Based on Example 1, this example provides an inland lake chlorophyll a retrieval system based on hyperspectral data, including:

[0127] The field measurement module is configured to construct a field measurement dataset of chlorophyll a in lakes by collecting lake water samples in the field.

[0128] The first remote sensing data processing module is configured to acquire a lake remote sensing reflectance dataset and extract hyperspectral satellite image data from the measured sampling time for preprocessing.

[0129] The second remote sensing data processing module is configured to extract the water body range from the preprocessed hyperspectral satellite image data through NDWI index calculation, extract band information, and obtain the band data and remote sensing images of the lake area corresponding to each measured point.

[0130] The model optimization module is configured to construct a multi-band index model considering the effects of pigment absorption and remote sensing reflectance, and to construct an inversion model by fusing the multi-band index model and the regression model. The inversion model is constructed using the data from the measured dataset as output and the band data from each measured point and the remote sensing images of the lake area as input. The inversion model is iteratively optimized.

[0131] The inversion module is configured to acquire hyperspectral satellite image data of the location to be measured, and input it into the optimized inversion model to obtain the inversion result of chlorophyll a at the location to be measured.

[0132] It should be noted that each module in this embodiment corresponds one-to-one with each step in embodiment 1, and their specific implementation process is the same, so it will not be repeated here.

[0133] Example 3

[0134] This embodiment provides an electronic device, including a memory and a processor, as well as computer instructions stored in the memory and running on the processor. When the computer instructions are executed by the processor, they complete the steps in the inland lake chlorophyll a inversion method based on hyperspectral data in Embodiment 1.

[0135] Example 4

[0136] This embodiment provides a computer-readable storage medium for storing computer instructions, which, when executed by a processor, complete the steps in the inland lake chlorophyll a inversion method based on hyperspectral data in Embodiment 1.

[0137] The above description is merely a preferred embodiment of this disclosure and is not intended to limit this disclosure. Various modifications and variations can be made to this disclosure by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

[0138] While the specific embodiments of this disclosure have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of this disclosure. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of this disclosure are still within the scope of protection of this disclosure.

Claims

1. A method for inverting chlorophyll a in inland lakes based on hyperspectral data, characterized in that, Includes the following steps: By collecting lake water samples in the field, a measured dataset of chlorophyll a in lakes was constructed. Acquire a dataset of lake remote sensing reflectance and extract hyperspectral satellite imagery data from the measured sampling time for preprocessing; The preprocessed hyperspectral satellite image data is used to extract the water body range through NDWI index calculation, and band information is extracted to obtain the band data and remote sensing images of the lake area corresponding to each measured point. A multi-band index model was constructed to consider the effects of pigment absorption and remote sensing reflectance. This model was then fused with a regression model to build an inversion model. The model was iteratively optimized using measured datasets as output and band data from various measured points and remote sensing images of the lake's water area as input. The construction method of the inversion model includes the following steps: Considering the effects of pigment absorption and remote sensing reflectance, the optimal band is selected. A multi-band index model is constructed based on the selected optimal band, and the three-band index is calculated. Using the three-band index calculated by the multi-band index model as the independent variable and the concentration of chlorophyll a in the measured chlorophyll a dataset as the output, various regression models were established for fitting. Evaluation indicators are set, and leave-one-out cross-validation and error assessment are performed on the fitting results. The optimal regression model is selected and combined with the constructed multi-band exponential model as the final inversion model. The specific form of the three-band index model, considering the influence of pigment absorption and remote sensing reflectance, is as follows: in, Indicates the first optimal band. Indicates the second optimal band. Indicates the third optimal band; This indicates the reflectivity of the corresponding waveband; After selecting the optimal band for the three-band index model, , and The determination of these three optimal bands involves iterative optimization using the characteristic band range to arrive at the final optimal bands, and includes the following steps: Preliminary determination based on spectral data obtained from lake water samples collected on-site. , and The initial band range; Based on the initial band range corresponding to the corresponding satellite, the band range corresponding to the corresponding satellite is determined, and the second optimal band is determined based on turbidity. and the third optimal band The defined band selection range is broadened to obtain , and The band selection range; where the formula for broadening the band selection range is: in, To determine the maximum suspended solids concentration in the lake during season t within a given time period, Let C be the minimum suspended solids concentration of the lake in season t within a given time period, and let E be the average suspended solids concentration of the lake in that water area. exist , and The band selection range is fixed by fixing any two bands and iterating the remaining band within the band selection range. In each iteration, the correlation between the corresponding band data and the concentration of chlorophyll a is calculated, and the band with the highest correlation is identified as the output band, thus obtaining the final optimal band of the three bands. Acquire hyperspectral satellite imagery data of the location to be measured, and input it into the optimized inversion model to obtain the inversion result of chlorophyll a at the location to be measured.

2. The inversion method for chlorophyll a in inland lakes based on hyperspectral data as described in claim 1, characterized in that, For the preprocessed remote sensing image data, water body extent extraction is performed, including: Calculate the NDWI index based on the obtained remote sensing image data; By setting the threshold to zero, the area with an NDWI index greater than the threshold is defined as a water body area. The remote sensing images are cropped to retain water areas and remove non-water areas.

3. The inversion method for chlorophyll a in inland lakes based on hyperspectral data as described in claim 1, characterized in that, Band information was extracted based on the measured locations of chlorophyll a data from lakes, including: The remote sensing image data of the extracted water body area is reprojected to match the coordinate system of the measured point data, so that the coordinate system of the remote sensing image is consistent with that of the measured point. Obtain the measured points and extract the band raster values ​​of the remote sensing image corresponding to the coordinates of the measured points as the extracted band information.

4. The inversion method for chlorophyll a in inland lakes based on hyperspectral data as described in claim 1, characterized in that: The multi-band index model uses either a three-band index model or a four-band index model.

5. The inversion method for chlorophyll a in inland lakes based on hyperspectral data as described in claim 1, characterized in that: For the three-band index model, the method for selecting the optimal band includes the following steps: First optimal band The chlorophyll a absorption peak band was selected. Second optimal band The chlorophyll a fluorescence peak band was selected. Third optimal band The pure water absorption band was selected.

6. The inversion method for chlorophyll a in inland lakes based on hyperspectral data as described in claim 1, characterized in that: Various regression models include linear, polynomial, exponential, and / or power regression models.

7. A chlorophyll a retrieval system for inland lakes based on hyperspectral data, characterized in that, Implementing the inland lake chlorophyll a inversion method based on hyperspectral data as described in any one of claims 1-6, comprising: The field measurement module is configured to construct a field measurement dataset of chlorophyll a in lakes by collecting lake water samples in the field. The first remote sensing data processing module is configured to acquire a lake remote sensing reflectance dataset and extract hyperspectral satellite image data from the measured sampling time for preprocessing. The second remote sensing data processing module is configured to extract the water body range from the preprocessed hyperspectral satellite image data through NDWI index calculation, extract band information, and obtain the band data and remote sensing images of the lake area corresponding to each measured point. The model optimization module is configured to construct a multi-band index model considering the effects of pigment absorption and remote sensing reflectance, and to construct an inversion model by fusing the multi-band index model and the regression model. The inversion model is constructed using the data from the measured dataset as output and the band data from each measured point and the remote sensing images of the lake area as input. The inversion model is iteratively optimized. The inversion module is configured to acquire hyperspectral satellite image data of the location to be measured, and input it into the optimized inversion model to obtain the inversion result of chlorophyll a at the location to be measured.

8. An electronic device, characterized in that, It includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, which, when executed by the processor, perform the steps in the inland lake chlorophyll a inversion method based on hyperspectral data as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, Used to store computer instructions, which, when executed by a processor, complete the steps in the inland lake chlorophyll a inversion method based on hyperspectral data as described in any one of claims 1-6.

Images