A chlorophyll a and suspended matter matching collaborative inversion method based on Sentinel 2A

Abstract

The present invention discloses a chlorophyll a and suspended matter matching collaborative inversion method based on Sentinel 2A, the method comprising the following steps: step 1) reading the remote sensing reflectance of band 1 to band 5 of Sentinel 2A image; step 2) setting different continuous chlorophyll a concentration values ​​and suspended matter concentration values, and simulating the remote sensing reflectance value by using a 2SecColor model; step 3) constructing a concentration-reflectance lookup table; step 4) matching the remote sensing reflectance on the Sentinel 2A image and the simulated remote sensing reflectance to obtain the concentrations of chlorophyll and suspended matter.

Description

Technical Field

[0001] The present invention relates to the technical field of water environment remote sensing monitoring, and in particular to a chlorophyll a and suspended matter matching collaborative inversion method based on Sentinel-2A. Background Art

[0002] Ground-based water environment survey and monitoring basically adopts conventional processes such as ground survey, manual sampling, laboratory identification and analysis. Although it has certain advantages such as accurate acquisition, the monitoring process requires a lot of manpower and material resources, the overall monitoring efficiency is low, and there are problems such as scattered information, complex collection, and untimely analysis and acquisition. Limited by the objective limitations of limited spatial distribution, there are technical shortcomings in comprehensively reflecting the spatial distribution of the water ecological environment. Satellite remote sensing monitoring has the advantages of large-scale, high-efficiency, quantitative, and traceable monitoring. As the "eye in the sky" for water environment monitoring, it is an indispensable tool for monitoring the dynamic changes of the macro water environment, an important means to achieve refined environmental protection and information management, and can provide a reliable, effective and accurate source of information for water ecological protection. It is one of the most effective means to strengthen the supervision of sensitive water ecological environment areas such as rivers, lakes, and wetlands.

[0003] Chlorophyll a is an effective indicator of phytoplankton biomass and can reflect the primary productivity and eutrophication level of water bodies (Pan Yingyang et al., 2017). Its content is directly related to the species and quantity of algae in the water body, and is of great significance to the maintenance of the aquatic ecosystem. The vertical stratification, temperature and nutrient supply of the water body all have a certain impact on the production efficiency of chlorophyll. One-half of the oxygen in the atmosphere comes from the photosynthesis of phytoplankton, and chlorophyll a is the main pigment used by phytoplankton in the water for photosynthesis, and its concentration has an important influence on the spectrum of the water body.

[0004] Suspended matter refers to the residue after being filtered through a filter membrane with a pore size of 0.45 μm. In nearshore waters, the main components are silt particles suspended in the water. Its concentration is the main factor affecting the spectral characteristics of the waters, directly affecting the transparency of the water body, and can also enrich pollutants and cause local pollution of the water quality environment. This is of great significance to the biological, chemical and physical processes in the water body, and is of great value to the study of silt transport, sludge analysis and estuary delta evolution. It is also an important reference factor for beach reclamation and coastal facility construction (Yang SL et al., 2002).

[0005] Monitoring of chlorophyll a and suspended matter in water bodies by satellite remote sensing is an important means of obtaining water quality. The current mainstream medium and high spatial resolution multispectral data (Landsat MSS, Landsat TM, SPOT HRV, RS-1C, NOAA / AVHRR, MODIS, GF-1 / WFV, Sentinel-2, etc.) and hyperspectral data (AVIRIS, CASI, AISA, CIS, etc.) have all developed and tested a variety of chlorophyll a and suspended matter inversion algorithms. Among them, Sentinel-2 has obvious advantages in spatial resolution and band composition and covers most of China's sea areas and coastal areas. At present, there are fewer algorithms for synchronous inversion of multi-component water color elements of this satellite. The present invention proposes a chlorophyll a and suspended matter matching collaborative inversion method based on Sentinel-2A satellite for the first time, which can effectively evaluate the ability of synchronous inversion of multi-component water color elements of Sentinel-2A satellite, and provide effective scientific basis and data support for water environment supervision. Summary of the invention

[0006] Aiming at the problem that Sentinel-2 satellite lacks a synchronous inversion algorithm for water color elements, the present invention proposes a chlorophyll a and suspended matter matching collaborative inversion method based on Sentinel-2A satellite, which can realize high-precision synchronous inversion of water color elements.

[0007] The purpose of the present invention is achieved through the following technical steps:

[0008] Step 1) Read the remote sensing reflectance of Sentinel-2A image band 1 to band 5;

[0009] Step 2) setting different continuous chlorophyll a concentration values ​​and suspended solids concentration values, and using the 2SecColor model to simulate the remote sensing reflectance value;

[0010] Step 3) Constructing a concentration-reflectivity lookup table;

[0011] Step 4) Match the remote sensing reflectance on the Sentinel-2A image with the simulated remote sensing reflectance to obtain the concentrations of chlorophyll and suspended matter.

[0012] The specific method of step 1) is:

[0013] a) Perform atmospheric correction and other preprocessing on Sentinel-2A images; b) Read the remote sensing reflectance of each pixel point in the first band (443nm), second band (490nm), third band (560nm), fourth band (665nm) and fifth band (705nm) of Sentinel-2A images, and establish the remote sensing reflectance image set of Sentinel-2A bands 1 to 5.

[0014] The specific method of step 2) is:

[0015] a) Chlorophyll a concentration value setting range from 10 0 μg / L to 10 2 μg / L, the setting range of suspended solids concentration is from 10 0 mg / L to 10 2 mg / L, the two values ​​increase exponentially, with the increment being 0.01 each time the exponent increases; b) The remote sensing reflectance value is simulated based on the 2SecColor model, and 200*200 data are calculated for each band.

[0016]

[0017]

[0018] μ w =cosθ′ s

[0019]

[0020]

[0021] In the formula, is the directional hemispherical reflectivity of the semi-infinite medium;

[0022] x is the total backscattering coefficient of the water body b b (Unit: m -1 ) and the total absorption coefficient a (unit: m -1 );

[0023] μ w Refers to the cosine value of the solar zenith angle under a flat water surface; if the solar zenith angle on the water surface is θ s , then θ′ s =arcsin(sinθ s / n w ), where n w is the refractive index of water, which takes a value of 1.33;

[0024] R(0 - ) is the radiance reflectance below the horizontal plane;

[0025] Q is the ratio of the radiance reflectance of the horizontal plane upward to the radiance reflectance of the horizontal plane downward, and the value is 3.25;

[0026] R rs is the reflectivity off water.

[0027] The specific method of step 3) is:

[0028] Based on step 2), the chlorophyll a concentration value, the suspended matter concentration value and the remote sensing reflectance R of Sentinel-2A bands 1 to 5 are established with the columns as chlorophyll a concentration values, the rows as suspended matter concentration values ​​and the simulated remote sensing reflectance results in the middle. rs The lookup table.

[0029] The specific method of step 4) is:

[0030] a) Calculate the root mean square error between the remote sensing reflectance obtained at each pixel point on the image in each band and the remote sensing reflectance from the lookup table; b) Find the minimum root mean square error and determine the chlorophyll a concentration and suspended matter concentration values ​​that match it in the lookup table; c) Repeat the same processing for each pixel point to obtain the spatial distribution map of chlorophyll a and suspended matter.

[0031] The formula for calculating the root mean square error is:

[0032]

[0033] Where i refers to the wavelength, which are 443nm, 490nm, 560nm, 665nm and 705nm respectively;

[0034] It is the off-water reflectivity in different bands simulated by the 2SeaColor model;

[0035] It is the water reflectance in different bands read on the image. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A schematic diagram of a method flow chart of an embodiment of the present invention;

[0037] Figure 2 This is a partial lookup table result diagram;

[0038] Figure 3 To match the collaborative inversion of chlorophyll a and suspended matter distribution results. DETAILED DESCRIPTION

[0039] The present invention "a chlorophyll a and suspended matter matching collaborative inversion method based on Sentinel 2A" is further explained below in conjunction with the accompanying drawings.

[0040] (I) Setting of chlorophyll a concentration and suspended matter concentration and calculation of simulated remote sensing reflectance

[0041] The absorption coefficient in water takes into account the four components in the water, namely the sum of the absorption coefficients of water molecules, chlorophyll a, suspended matter and yellow matter. The absorption coefficient of pure water molecules is a function of the wavelength λ, and the absorption coefficient of yellow matter is obtained based on actual measurements.

[0042] The calculation formula of the absorption coefficient of chlorophyll a is:

[0043]

[0044] a chla (λ)=(a0(λ)+a1(λ)ln a Chla (440))× a chla (440)

[0045] In the formula, C chla is the chlorophyll a concentration, and the range is set to 10 in the calculation 0 μg / L to 10 2 μg / L; a0(λ) and a1(λ) are empirical parameters, a chl (440) is the spectral absorption coefficient of Taihu chlorophyll a at a wavelength of 440nm.

[0046] The calculation formula of the absorption coefficient of suspended matter is:

[0047]

[0048] Where C s is the suspended matter concentration, the calculation range is set from 10 0 mg / L to 10 2 mg / L; a s (440) is the spectral absorption coefficient of Taihu suspended matter at a wavelength of 440 nm, is the normalized unit absorption coefficient of suspended matter.

[0049] Since the backscattering effect of yellow matter on light is very weak, the calculation of the backscattering coefficient only considers three components in the water body, namely water molecules, chlorophyll and suspended matter, among which the scattering coefficient of pure water molecules is a function of wavelength λ.

[0050] Among them, the calculation formula of the scattering coefficient of chlorophyll a is:

[0051]

[0052] In the formula, is the normalized unit absorption coefficient of chlorophyll a.

[0053] The calculation formula of the scattering coefficient of suspended matter is:

[0054] b s (λ)=0.03(λ / 550) -1.15 ·C S

[0055] (II) Lookup table construction

[0056] The first band (443nm), the second band (490nm), the third band (560nm), the fourth band (665nm) and the fifth band (705nm) correspond to 40,000 remote sensing reflectance data respectively, and the chlorophyll a concentration value, suspended matter concentration value and remote sensing reflectance R of Sentinel-2A band 1 to band 5 are established. rs Each table has 200*200 data items, a total of 5 tables, some of the lookup table results are as follows Figure 2 shown.

[0057] 3. Collaborative inversion based on reflectivity matching

[0058] Calculate the root mean square error between the remote sensing reflectance of each pixel point from band 1 to band 5 read by Sentinel-2A image and the remote sensing reflectance from the lookup table. When the root mean square error is the smallest, the chlorophyll a concentration and suspended matter concentration that match it are the chlorophyll a concentration value and suspended matter concentration value of the pixel point. Repeat the same process for each pixel point to obtain the spatial distribution map of chlorophyll a and suspended matter ( Figure 3 ), the matching results of some pixels are shown in Table 1.

[0059] Table 1

[0060]

Claims

1. A chlorophyll a and suspended matter matching collaborative inversion method based on Sentinel 2A, the method comprising the following steps: Step 1) Read the remote sensing reflectance of band 1 to band 5 of Sentinel-2A image; a) Perform atmospheric correction preprocessing on Sentinel-2A images; b) Read the remote sensing reflectance of each pixel point in the first band 443nm, second band 490nm, third band 560nm, fourth band 665nm and fifth band 705nm of Sentinel-2A images, and establish the remote sensing reflectance image set of Sentinel-2A bands 1 to 5; Step 2) setting different continuous chlorophyll a concentration values ​​and suspended solids concentration values, and using the 2SeaColor model to simulate the remote sensing reflectance value; a) The chlorophyll a concentration value setting range is from 100μg / L to 102μg / L, and the suspended solids concentration value setting range is from 100mg / L to 102mg / L. The two values ​​increase exponentially, and the increment is 0.01 each time the exponent increases; b) The remote sensing reflectance value is simulated based on the 2SeaColor model, and 200*200 data are calculated for each band; m w =cosθ s ′ In formula (1), is the directional hemispherical reflectivity of the semi-infinite medium; x is the total backscattering coefficient of the water body b b The ratio of the total absorption coefficient a; μ w Refers to the cosine value of the solar zenith angle under a flat water surface; if the solar zenith angle on the water surface is θ s , then θ′ s =arcsin(sinθ s / n w ), where n w is the refractive index of water, which is 1.33; R(0 - ) is the radiance reflectance under the horizontal plane; Q is the ratio of the radiance reflectance upward to the radiance reflectance downward, which is 3.25; R rs is the off-water reflectivity; Step 3) Constructing a concentration-reflectivity lookup table; Based on step 2), the chlorophyll a concentration values, suspended matter concentration values ​​and remote sensing reflectance R values ​​of Sentinel 2A bands 1 to 5 are established with the columns as chlorophyll a concentration values, the rows as suspended matter concentration values ​​and the middle as simulated remote sensing reflectance results. rs Lookup table for Step 4) Match the remote sensing reflectance on the Sentinel-2A image with the simulated remote sensing reflectance to obtain the concentrations of chlorophyll and suspended matter; a) Calculate the root mean square error between the remote sensing reflectance obtained at each pixel point on the image in each band and the remote sensing reflectance from the lookup table; b) Find the minimum root mean square error and determine the chlorophyll a concentration and suspended matter concentration values ​​that match it in the lookup table; c) Repeat the same process for each pixel point to obtain the spatial distribution map of chlorophyll a and suspended matter; The formula for calculating the root mean square error is: In formula (2), i refers to the wavelength, which are 443nm, 490nm, 560nm, 665nm and 705nm respectively; It is the off-water reflectivity in different bands simulated by the 2SeaColor model; It is the water reflectance in different bands read on the image.

Images