Method for estimating chlorophyll a concentration in inland waters using ohmic hyperspectral images (OHS)

Abstract

This invention relates to the field of chlorophyll a estimation technology, and specifically to a method for estimating chlorophyll a concentration in inland water bodies using ohmic hyperspectral images (OHS). The method includes the following steps: 1. Establishing a reparameterized QAA model to obtain the phytoplankton absorption coefficient α for each wavelength. ph 2. Based on the characteristics of eutrophic lakes, the Chl-a concentration is estimated. This invention constructs a reparameterized quasi-analysis algorithm targeting the optical characteristics of inland lake water; and utilizes next-generation OHS hyperspectral imagery to achieve remote sensing estimation of chlorophyll-a concentration in inland lake water.

Description

Technical Field

[0001] This invention relates to the field of chlorophyll a estimation technology, and more specifically, to a method for estimating chlorophyll a concentration in inland water bodies using ohmic hyperspectral images (OHS). Background Technology

[0002] Chl-a concentration can be accurately measured in-situ, but long-term and extensive lake monitoring is difficult due to the time and effort involved. In contrast, long-term and extensive measurements of Chl-a concentration in aquatic environments can be accomplished through satellite remote sensing. Many remote sensing algorithms have been developed to estimate Chl-a concentration in inland waters. These algorithms can be broadly categorized into empirical algorithms, semi-analytical algorithms, and machine learning algorithms. Examples include: red to near-red band ratio algorithms, three-band algorithms, fluorescence line height, normalized chlorophyll difference index, and limiting gradient boosting trees. However, the aforementioned studies primarily use empirical algorithms to estimate Chl-a, lacking research into the mechanisms underlying Chl-a. Chl-a is related to the phytoplankton absorption coefficient (PhEPCO). ph It is closely related because its changes depend on specific absorption, therefore a ph It can be used as a surrogate parameter to invert Chl-a. Some scholars have developed an improved quasi-analytical algorithm (QAA) called TC2, which uses a... ph To estimate Chl-a in Class II turbid water bodies, TC2 requires a large amount of measured data and its estimation results are poor in eutrophic lakes. Some researchers have developed a quasi-analysis algorithm (QAA) based on the radial transport model, which is now widely used to derive the intrinsic optical properties (IOPs) of marine water bodies. However, previous studies have shown that QAA does not produce satisfactory results in inland water bodies, and it is necessary to reselect the reference band and reparameterize the formula to obtain more reasonable results in mixed water bodies.

[0003] Orbita Hyperspectral (OHS) is a new generation of domestically developed spectral satellite imagery. It is currently the only in-orbit imagery product globally that integrates high spectral resolution (5nm), high spatial resolution (10m), and high temporal resolution (2d), possessing enormous potential and advantages in inland water quality monitoring. However, since the Orbita Hyperspectral satellite has only recently been put into operation, research using its data is still relatively limited, and current research focuses primarily on land cover classification, with less emphasis on inland water bodies and satellite performance evaluation.

[0004] In summary, the shortcomings of the existing technology are as follows: (1) The existing QAA algorithm is mainly based on the optical characteristics of ocean and near-shore waters. However, a large number of studies have shown that it is ineffective and inapplicable in estimating parameters such as absorption coefficient and scattering coefficient in inland lakes, mainly because the optical characteristics of inland lakes are different from those of ocean and near-shore waters; (2) The existing technology does not involve the construction of an effective semi-analysis algorithm and its application in suitable hyperspectral images to achieve the estimation of chlorophyll a concentration in inland waters. Summary of the Invention

[0005] The present invention provides a method for estimating chlorophyll a concentration in inland water bodies using ohmic hyperspectral images (OHS), which can overcome some or all the defects of the prior art.

[0006] The method for estimating chlorophyll a concentration in inland water bodies using ohmic hyperspectral imaging (OHS) according to the present invention comprises the following steps:

[0007] 1. Establish a reparameterized QAA model to obtain the phytoplankton absorption coefficient α for each wavelength. ph ;

[0008] II. Estimate Chl-a concentration based on the characteristics of eutrophic lakes.

[0009] As a preferred option, the specific steps in step one are as follows:

[0010] Step 1: Based on the remote sensing reflectance R of the water surface rs (λ) Calculate the remote sensing reflectance r of the lower surface of the water body. rs (λ);

[0011] r rs (λ)=R rs (λ) / (0.52+1.7×R rs (λ))

[0012] λ represents wavelength;

[0013] Step 2: Based on the radiative transfer model, the remote sensing reflectance r below the water surface can be obtained. rs (λ) is a function of the ratio of the backscattering coefficient of water to the sum of the absorption coefficient and the backscattering coefficient of water, and its formula is as follows:

[0014] r rs (λ)=g0(λ)×u(λ)+g1(λ)×[u(λ)] 2

[0015] in

[0016] u(λ)=b b (λ) / (a(λ)+b b(λ))

[0017] Where g0 and g1 are empirical parameters, a(λ) represents the total absorption coefficient of the water body, and b b (λ) represents the total backscattering coefficient of the water body, and u is an intermediate variable; the above equation yields:

[0018]

[0019] Step 3: via a w (715) and Δa(λ) are used to calculate a(λ0), as follows:

[0020]

[0021] In the formula, a w (715) is the absorption coefficient a of pure water at λ0=715nm, and Δa(λ) is the sum of the absorption coefficients of all optical components of the water.

[0022] Step 4: After calculating a(λ0), the backscattering coefficient b of the water body bp The following results were obtained at 715nm:

[0023]

[0024] In the formula b bw (715) is the backscattering coefficient of pure water at 715 nm.

[0025] Step 5: Calculate the value of η using the reflectance ratio. The value of η is the probability of the particulate backscattering coefficient.

[0026]

[0027] Step 6: Calculate the particle backscattering coefficient b based on the reference wavelength. bp (λ0) Calculate the backscattering b of particles at each wavelength. bp (λ), whose formula is an exponential function:

[0028]

[0029] Step 7: Use b bp a(λ) and u(λ) are used to calculate a(λ) at each wavelength:

[0030]

[0031] Step 8: Calculate the empirical parameters ζ, S, and ξ to derive the phytoplankton absorption coefficient a. ph :

[0032]

[0033]

[0034] Step 9: Based on a and a w Calculate a dg a ph By removing a w and a dg The contributions were recognized as follows:

[0035]

[0036] a dg (λ)=a dg (443)e -S(λ-443) ,a ph (λ)=a(λ)-a dg (λ)-a w (λ)

[0037] Based on the above equations, a and a at each wavelength... ph and b bp By calculating a and b at the reference wavelength bp get.

[0038] Preferably, in step two, an empirical function is used to generate a as a function of Chl-a concentration. ph (λ), the formula is:

[0039] C Chl-a ∝a ph (λ)

[0040] The Chl-a concentration is obtained from aph(λ).

[0041] The beneficial effects of this invention are as follows:

[0042] (1) To construct a reparameterized quasi-analysis algorithm for the optical properties of inland lake water;

[0043] (2) Using the new generation of OHS hyperspectral imagery, remote sensing estimation of chlorophyll a concentration in inland lakes was achieved. Attached Figure Description

[0044] Figure 1 This is a flowchart illustrating a method for estimating chlorophyll a concentration in inland water bodies using ohmic hyperspectral images (OHS) in an embodiment.

[0045] Figure 2(a) is a schematic diagram of the estimation and verification of chlorophyll a by QAA_V6 at 490nm, 565nm and 665nm in the embodiment;

[0046] Figure 2(b) is a schematic diagram of TC2 estimation and verification of chlorophyll a at 490 nm, 565 nm and 665 nm in the embodiment;

[0047] Figure 2(c) shows the QAA at 490nm, 565nm and 665nm in the embodiment. OWM A schematic diagram illustrating the estimation and verification of chlorophyll a;

[0048] Figure 2(d) is a schematic diagram of the estimation and verification of chlorophyll a by OAA_Z at 490nm, 565nm and 665nm in the embodiment;

[0049] Figure 3(a) is a schematic diagram of the estimated Chl-a concentration in the OHS image after atmospheric correction using the FLAASH model in the example;

[0050] Figure 3(b) is a schematic diagram of the estimated Chl-a concentration in the OHS image after atmospheric correction using the QUAC model in the example;

[0051] Figure 3(c) is a schematic diagram of the estimated Chl-a concentration in the OHS image after atmospheric correction using the 6S model in the embodiment;

[0052] Figure 3(d) is a schematic diagram of the estimated Chl-a concentration in the OHS image after atmospheric correction using the dark pixel model in the embodiment. Detailed Implementation

[0053] To further understand the content of this invention, a detailed description of the invention will be provided in conjunction with the accompanying drawings and embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention.

[0054] Example

[0055] like Figure 1 As shown, this embodiment provides a method for estimating chlorophyll a concentration in inland water bodies using Orbital Hyperspectral Image (OHS), which includes the following steps:

[0056] 1. Establish a reparameterized QAA model to obtain the phytoplankton absorption coefficient α for each wavelength. ph ;

[0057] II. Estimate Chl-a concentration based on the characteristics of eutrophic lakes.

[0058] Step one, the specific steps are as follows:

[0059] Step 1: Based on the remote sensing reflectance R of the water surface rs (λ) Calculate the remote sensing reflectance r of the lower surface of the water body. rs (λ);

[0060] r rs (λ)=R rs (λ) / (0.52+1.7×R rs(λ))

[0061] λ represents wavelength;

[0062] Step 2: Based on the radiative transfer model, the remote sensing reflectance r below the water surface can be obtained. rs (λ) is a function of the ratio of the backscattering coefficient of water to the sum of the absorption coefficient and the backscattering coefficient of water, and its formula is as follows:

[0063] r rs (λ)=g0(λ)×u(λ)+g1(λ)×[u(λ)] 2

[0064] in

[0065] u(λ)=b b (λ) / (a(λ)+b b (λ))

[0066] Where g0 and g1 are empirical parameters, a(λ) represents the total absorption coefficient of the water body, and b b (λ) represents the total backscattering coefficient of the water body, and u is an intermediate variable; the above equation yields:

[0067]

[0068] Step 3: via a w (715) and Δa(λ) are used to calculate a(λ0), as follows:

[0069]

[0070] In the formula, a w (715) is the absorption coefficient a of pure water at λ0=715nm, and Δa(λ) is the sum of the absorption coefficients of all optical components of the water.

[0071] Step 4: After calculating a(λ0), the backscattering coefficient b of the water body bp The following results were obtained at 715nm:

[0072]

[0073] In the formula b bw (715) is the backscattering coefficient of pure water at 715 nm.

[0074] Step 5: Calculate the value of η using the reflectance ratio. The value of η is the probability of the particulate backscattering coefficient.

[0075]

[0076] Step 6: Calculate the particle backscattering coefficient b based on the reference wavelength. bp(λ0) Calculate the backscattering b of particles at each wavelength. bp (λ), whose formula is an exponential function:

[0077]

[0078] Step 7: Use b bp a(λ) and u(λ) are used to calculate a(λ) at each wavelength:

[0079]

[0080] Step 8: Calculate the empirical parameters ζ, S, and ξ to derive the phytoplankton absorption coefficient a. ph :

[0081]

[0082]

[0083] Step 9: Based on a and a w Calculate a dg a ph By removing a w and a dg The contributions were recognized as follows:

[0084]

[0085] a dg (λ)=a dg (443)e -S(λ-443) ,a ph (λ)=a(λ)-a dg (λ)-a w (λ)

[0086] Based on the above equations, a and a at each wavelength... ph and b bp By calculating a and b at the reference wavelength bp The results are shown in Table 1.

[0087] Table 1. Steps for reparameterized QAA remote sensing estimation

[0088]

[0089]

[0090] In step two, for the Chl-a concentration, Chl-a can be expressed as:

[0091]

[0092] In the formula, a* ph(λ) is the unit absorption coefficient of phytoplankton. Based on previous research, a* ph The value of a*ph(λ) is related to cell size, changes in intracellular pigment concentration, and local light adaptation; therefore, accurate calculation of a*ph(λ) is difficult and requires large-scale global or regional studies. The measurement. However, Chl-a and a ph (λ) has a strong correlation, therefore an empirical function is used to generate a as a function of Chl-a concentration. ph (λ). The formula is:

[0093] C Chl-a ∝a ph (λ)

[0094] Chl-a concentration from a ph( λ) is obtained. And a ph( λ) can be obtained from R in the reparameterized QAA model rs get.

[0095] This invention constructs a reparameterized quasi-analysis algorithm based on the optical characteristics of inland water bodies, and uses next-generation OHS hyperspectral imagery to achieve remote sensing estimation of chlorophyll a concentration in inland water bodies.

[0096] Validation of the reparameterized QAA model (OAA_Z)

[0097] To demonstrate the applicability and uncertainty of the reparameterized QAA model (OAA_Z), it is compared with QAA_V6, TC2, and QAA. OWM A comparison was made. In Figures 2(a), 2(b), 2(c), and 2(d), the estimation performance of QAA_V6 was very poor, with a significant underestimation (MAPD of 58.45%–76.24%, RMSD of 133m). -1 ~1.76m -1 ); TC2 performs better than OAA_V6 in estimation across each band, but still exhibits significant underestimation (MAPD 37.02%–52.41%, RMSD 0.98m). -1 ~2.05m -1 ); QAA OWM The estimation results are significantly improved, especially for 665m, with accuracy twice that of QAA_V6. However, the results still underestimate the accuracy (MAPD 35.04%–46.71%, RMSD 0.90m). -1 ~1.76m -1 In this embodiment, OAA_Z has the lowest MAPD and RMSD (MAPD is 22.15%–29.77%, RMSD is 0.61m). -1 ~0.87m-1 The results were basically on the 1:1 line, and the deviation from the measured data was within an acceptable range, achieving more satisfactory results than other OAA methods.

[0098] After reparameterization, a is calculated using OAA_Z. ph Analyze a ph The correlation between chl-a concentration and the highest correlation coefficient was selected for chl-a. ph A semi-analytical model for Chl-a estimation was established. The influence of different atmospheric correction models on the Chl-a semi-analytical model was evaluated using the FLAASH model, QUAC model, 6S model, and dark pixel model, as shown in Figures 3(a), 3(b), 3(c), and 3(d). The FLAASH model validated the best results (MAPD = 24.02%, RMSD = 14.31 ug / L) and can be applied to OHS imagery. The QUAC model overestimated Chl-a concentration, while the dark pixel model underestimated it. The bias in Chl-a concentration estimation may be due to the low SNR and high NE of the OHS imagery. chl-a Caused by.

[0099] The present invention and its embodiments have been described above illustratively. This description is not restrictive, and the figures shown are only one embodiment of the present invention; the actual structure is not limited thereto. Therefore, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the present invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for estimating chlorophyll a concentration in inland water bodies using ohmic hyperspectral images (OHS), characterized by: Includes the following steps:

1. Establish a reparameterized QAA model to obtain the phytoplankton absorption coefficient α for each wavelength. ph ; II. Estimate Chl-a concentration based on the characteristics of eutrophic lakes; Step one, the specific steps are as follows: Step 1: Based on the remote sensing reflectance R of the water surface rs (λ) Calculate the remote sensing reflectance r of the lower surface of the water body. rs (λ); r rs (λ)=R rs (λ) / (0.52+1.7×R rs (l)) λ represents wavelength; Step 2: Based on the radiative transfer model, the remote sensing reflectance r below the water surface can be obtained. rs (λ) is a function of the ratio of the backscattering coefficient of water to the sum of the absorption coefficient and the backscattering coefficient of water, and its formula is as follows: r rs (λ)=g0(λ)×u(λ)+g1(λ)×[u(λ)] 2 in u(λ)=b b (λ) / (a(λ)+b b (l)) Where g0 and g1 are empirical parameters, a(λ) represents the total absorption coefficient of the water body, and b b (λ) represents the total backscattering coefficient of the water body, and u is an intermediate variable; the above equation yields: Step 3: via a w (715) and Δa(λ) are used to calculate a(λ0), as follows: α=-0.649, β=1.149, γ=0.037 In the formula, a w (715) is the absorption coefficient a of pure water at λ0=715nm, and Δa(λ) is the sum of the absorption coefficients of all optical components of the water body; Step 4: After calculating a(λ0), the backscattering coefficient b of the water body bp The following results were obtained at 715nm: In the formula b bw (715) is the backscattering coefficient of pure water at 715 nm; Step 5: Calculate the value of η using the reflectance ratio. The value of η is the probability of the particulate backscattering coefficient. Step 6: Calculate the particle backscattering coefficient b based on the reference wavelength. bp (λ0) Calculate the backscattering b of particles at each wavelength. bp (λ), whose formula is an exponential function: Step 7: Use b bp a(λ) and u(λ) are used to calculate a(λ) at each wavelength: Step 8: Calculate the empirical parameters ζ, S, and ξ to derive the phytoplankton absorption coefficient a. ph : Step 9: Based on a and a w Calculate a dg a ph By removing a w The absorption coefficient α of colored dissolved organic matter and fragments dg The contributions were recognized as follows: a dg (λ)=a dg (443)e -S(λ-443) ,a ph (λ)=a(λ)-a dg (l)-a w (l) Based on the above equations, a and a at each wavelength... ph and b bp By calculating a and b at the reference wavelength bp get; In step two, an empirical function is used to generate a, which serves as a function of Chl-a concentration. ph (λ), the formula is: C Chl-a ∝a ph (l) Chl-a concentration from a ph( λ) is obtained.

Images