A water content prediction method and system based on an airborne multi-spectral sensitive band combination

Abstract

The application discloses a kind of water content prediction method and system based on airborne multispectral sensitive waveband combination, it is related to agricultural remote sensing and image processing technical field, including steps: based on the response characteristics of soil and plant canopy reflectance spectrum to water, combined with the physical mechanism of O-H bond vibration in water molecule, and considering the influence of atmospheric water vapor on near infrared spectrum to determine the bandwidth of all center wavelengths;The reflectance spectrum of soil and vegetation canopy is generated, the relationship between center wavelength and bandwidth is used to generate spectral response function, the simulated spectrum is converted to water sensitive waveband, and reflectivity and spectral index are calculated;The reflectance spectrum image of soil and vegetation canopy at selected position is collected, and the reflectance spectrum image is input into the water content prediction model to predict water content.The present application can effectively solve the problem of insufficient waveband selection and low monitoring accuracy when the existing airborne multispectral sensor performs water monitoring task by optimizing waveband selection and bandwidth allocation.

Description

Technical Field

[0001] This invention relates to the field of agricultural remote sensing and image processing technology, and in particular to a water prediction method and system based on airborne multispectral sensitive band combination. Background Technology

[0002] With the intensification of global climate change and the increasing scarcity of water resources, the sustainable management of agricultural production and the ecological environment faces unprecedented challenges. Water, as a key factor affecting plant growth and soil health, is crucial for accurate monitoring to improve agricultural yields, optimize water resource use efficiency, cope with extreme weather events such as droughts, and protect the ecological environment. However, traditional water monitoring methods mainly rely on ground sampling and laboratory analysis. These methods are not only time-consuming and labor-intensive but also have limitations such as limited spatial coverage and insufficient real-time performance. These limitations significantly affect the efficiency and accuracy of agricultural management and environmental monitoring, especially in large-scale farmland or ecosystems.

[0003] Currently, water monitoring methods based on remote sensing technology are widely used. With the rapid development of remote sensing technology, water monitoring methods based on airborne sensors are gradually becoming an efficient alternative. Compared with ground-based monitoring, airborne sensing systems have advantages such as wide coverage, fast acquisition speed, and high flexibility. In particular, multispectral sensors mounted on UAV platforms can acquire reflectance data of large areas in a short time, enabling rapid monitoring of soil and vegetation moisture information. Furthermore, advancements in modern multispectral imaging technology, especially the precision of spectral band selection and the improvement of sensor resolution, have made detailed monitoring of soil and vegetation moisture status possible.

[0004] Under drought stress, the structure and water content of plant leaves undergo significant changes, leading to alterations in the response characteristics of the plant canopy reflectance spectrum. For example, when plants are water-deficient, chlorophyll synthesis decreases or decomposition accelerates, resulting in increased reflectance in the visible light region, accompanied by a red-edge blue shift. Water deficiency also leads to reduced leaf thickness, altered cell structure, and stomatal closure, thus affecting the leaf's reflectance and absorption characteristics in the near-infrared region. Furthermore, light passing through the canopy undergoes multiple scattering and reflections on the leaf and soil surfaces, meaning that the canopy reflectance spectrum not only contains information about the plant itself but can also be used to predict the soil moisture content in the root zone. For bare soil surfaces, the spectral reflectance characteristics are mainly influenced by factors such as soil particle structure, water content, and organic matter content. Changes in water content particularly significantly affect soil reflectance in the mid-infrared and short-wave infrared regions. Therefore, by combining the canopy spectrum and bare soil spectral response characteristics, a comprehensive prediction of the moisture content in the root zone and bare soil can be achieved.

[0005] However, most existing airborne multispectral sensors are concentrated in the visible and near-infrared bands. While these bands can provide some moisture information, their specificity and accuracy are limited. They lack sufficient spectral resolution and sensitivity, and cannot accurately capture the absorption characteristics of moisture, thus limiting their effectiveness in moisture monitoring applications. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of the prior art by providing a moisture prediction method and system based on airborne multispectral sensitive band combination, so as to solve the problem that the prior art lacks sufficient spectral resolution and sensitivity, cannot accurately capture the absorption characteristics of moisture, and thus has limited effectiveness in moisture monitoring applications.

[0007] The present invention specifically provides the following technical solution:

[0008] A moisture prediction method based on airborne multispectral sensitive band combination includes the following steps:

[0009] Based on the response characteristics of soil and plant canopy reflectance spectra to water, the center wavelengths of multiple water-sensitive bands were selected. According to the physical mechanism of O-H bond vibration in water molecules, the center wavelengths of multiple different water-sensitive bands were increased and selected, and the bandwidth of all center wavelengths was determined.

[0010] The reflectance spectra of soil and vegetation canopy are simulated and generated. A spectral response function is generated by using the relationship between the center wavelength and bandwidth of the multiple different water-sensitive bands. The simulated reflectance spectrum is converted to multiple water-sensitive bands through the spectral response function. The reflectance and spectral index of multiple water-sensitive bands are obtained. A water prediction model is constructed using the reflectance and spectral index.

[0011] All water-sensitive bands with a defined bandwidth are combined and integrated onto an airborne multispectral sensor to acquire reflectance spectral images of soil and vegetation canopy at selected locations. These reflectance spectral images are then input into a water prediction model to predict the water content of soil and vegetation canopy at the selected locations.

[0012] Preferably, the simulated generation of the reflectance spectra of soil and vegetation canopy includes:

[0013] The MARMIT2 radiative transfer model was used to simulate the reflectance spectrum of soil in the solar domain, and the soil moisture content (SMC) of the simulated soil reflectance spectrum was normalized as a dimensionless soil factor (psoil).

[0014] The reflectance spectra of soils with the lowest and highest moisture content were used as inputs to the PROSAIL radiative transfer model to simulate the reflectance spectra of the vegetation canopy.

[0015] In this model, the MARMIT2 radiative transfer model simulates the soil reflectance spectrum within the solar domain. Based on the product of the input water layer thickness and surface water cover, the sigmoid function is used to fit the soil moisture content (SMC) representing the soil reflectance spectrum. The specific expression is as follows:

[0016] BRF Soil (λ)=f MARMIT2 (λ,L,ε,δ);

[0017] Among them, BRF Soil (λ) represents soil reflectance, and f represents the soil moisture content (SMC) characterizing the soil reflectance spectrum. MARMIT2 (·) is the operational function of the MARMIT2 radiative transfer model, λ is the wavelength, L is the water layer thickness, ε is the surface water coverage, and δ is the soil particle volume fraction.

[0018] In the PROSAIL radiative transfer model, the equivalent water thickness is obtained, and the equivalent water thickness of the vegetation canopy and the normalized soil moisture content are represented by the equivalent water thickness and psoil, respectively; the specific expressions are as follows:

[0019] CEWT = LAI × C w ;

[0020] DHR Leaf (λ)=f PROSPECT (λ,θ DHR );

[0021] BRF Canopy (λ)=f SAIL (λ,DHR(λ),θ BRF );

[0022] Where CEWT is the canopy equivalent water thickness, LAI is the leaf area index, and C w For equivalent water thickness, DHR Leaf (λ) is the value of f using the PROSPECT model. PROSPECT The simulated wavelength is λ, and the reflectivity of the blade's directional hemisphere is θ. DHR Input parameters for the PROSPECT model; BRF Canopy (λ) represents the value obtained by using the 4SAIL model, through f SAIL Coupled DHR Leaf (λ), the bidirectional reflectivity of the canopy at a simulated wavelength of λ, θ BRF These are the input parameters for the 4SAIL model.

[0023] Preferably, a spectral response function is generated using the relationship between the center wavelength and bandwidth of the multiple different moisture-sensitive bands. The simulated reflectance spectrum is then converted to multiple moisture-sensitive bands using the spectral response function to obtain the reflectance and spectral indices of the multiple moisture-sensitive bands, including:

[0024] Based on the center wavelength and bandwidth of the selected moisture-sensitive band, the spectral response function is approximated using a Gaussian function; the specific expression is:

[0025]

[0026] Where SRF(λ) is the spectral index at wavelength λ, λ0 is the center wavelength, and σ is the standard deviation, which is calculated by the bandwidth FWMH to control the width of the spectral response function curve;

[0027] The soil reflectance spectrum and the canopy reflectance spectrum are converted to obtain multiple selected combinations of water-sensitive bands, as well as the reflectance of these combinations; the specific expression is as follows:

[0028]

[0029] Where R band λ is the reflectance after resampling, λ1 and λ2 are the wavelength range of this band, and BRF(λ) is the reflectance when the simulated wavelength is λ.

[0030] By combining multiple moisture-sensitive bands and inputting them into the spectral response function, the spectral index is obtained.

[0031] Preferably, before inputting the reflectance spectral image into the moisture prediction model, the method further includes:

[0032] We collected real soil and vegetation canopy reflectance spectral data, and used spectral response functions to convert the real soil and vegetation canopy reflectance spectral data into multiple water-sensitive bands, obtaining the reflectance and spectral index of water-sensitive bands under multiple real conditions.

[0033] The water prediction model is evaluated by nested cross-validation using reflectance and spectral indices of water-sensitive bands under multiple real-world conditions. The water prediction model is then optimized using the results of the nested cross-validation to obtain the optimal water prediction model.

[0034] Preferably, the step of using the reflectance and spectral indices of multiple moisture-sensitive bands under real conditions to perform nested cross-validation of the moisture prediction model's prediction results includes:

[0035] The outer cross-validation layer of nested cross-validation randomly divides the data into 10 folds, retaining 1 fold as the test set and the remaining 9 folds as the training set, and evaluates the optimal model selected in the inner cross-validation layer. The inner cross-validation layer divides the training set of the outer cross-validation layer into 10 folds, optimizes and selects the hyperparameters of the moisture prediction model.

[0036] Preferably, based on the response characteristics of soil and plant canopy reflectance spectra to water, the center wavelengths of multiple water-sensitive bands are selected, and according to the physical mechanism of O-H bond vibration in water molecules, multiple center wavelengths of different water-sensitive bands are added and selected, including:

[0037] Based on the response characteristics of soil and plant canopy reflectance spectra to water, 560nm, 705nm, 750nm, 800nm, 900nm, 970nm, 1100nm, 1200nm, 1650nm and 2200nm were selected as the center wavelengths of the water-sensitive band.

[0038] Based on the physical mechanism of O-H bond vibration in water molecules, 450m and 660nm were selected as the center wavelengths of the water-sensitive band.

[0039] Preferably, determining the bandwidth of all center wavelengths includes:

[0040] By selecting bandwidth-balanced spectral information and spatiotemporal resolution in the moisture-sensitive bands, the selected bandwidths are 30nm, 30nm, 30nm, 10nm, 15nm, 40nm, 40nm, 10nm, 50nm, 50nm, 50nm and 100nm.

[0041] This invention provides a moisture prediction system based on an airborne multispectral sensitive band combination, comprising:

[0042] The selection module is used to select the center wavelength of multiple water-sensitive bands based on the response characteristics of soil and plant canopy reflectance spectra to water, and to add and select the center wavelengths of multiple different water-sensitive bands according to the physical mechanism of O-H bond vibration in water molecules, and to determine the bandwidth of all center wavelengths.

[0043] The model building module is used to simulate and generate the reflectance spectra of soil and vegetation canopy. It generates a spectral response function by utilizing the relationship between the center wavelength and bandwidth of the multiple different water-sensitive bands. The simulated reflectance spectrum is converted to multiple water-sensitive bands through the spectral response function, and the reflectance and spectral index of multiple water-sensitive bands are obtained. The reflectance and spectral index are then used to build a water prediction model.

[0044] The prediction module is used to combine all water-sensitive bands with a defined bandwidth and integrate them into an airborne multispectral sensor to acquire reflectance spectral images of soil and vegetation canopy at selected locations. The reflectance spectral images are then input into a water prediction model to predict the water content of soil and vegetation canopy at the selected locations.

[0045] The present invention provides a computer device, including a memory and a processor. The memory stores a program, and when the program is executed by the processor, the processor performs the steps of the above-described method for predicting moisture based on a combination of airborne multispectral sensitive bands.

[0046] The present invention provides a storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method for predicting moisture based on a combination of airborne multispectral sensitive bands.

[0047] Compared with the prior art, the present invention has the following significant advantages:

[0048] This invention utilizes the response characteristics of soil and plant canopy reflectance spectra to water, selecting the center wavelengths of multiple water-sensitive bands. Furthermore, based on the physical mechanism of O-H bond vibrations in water molecules, it adds and selects center wavelengths of several different water-sensitive bands and allocates corresponding bandwidths. This effectively solves the problems of insufficient band selection and low monitoring accuracy of existing multispectral sensors when performing water monitoring tasks. The invention also uses a spectral response function generated by the center wavelength and bandwidth to convert the simulated reflectance spectrum into water-sensitive bands. A water prediction model is then constructed using the converted reflectance and spectral index. Because the converted reflectance has sufficient spectral resolution and sensitivity, water prediction can be achieved by acquiring reflectance spectral images at any location using the water prediction model, thus improving prediction accuracy. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of typical vegetation and soil reflectance at the selected wavelength location for this invention; wherein... Figure 1 (a) shows the reflectance variation of different chlorophyll a+b contents in different wavelength bands. Figure 1 (b) shows the reflectance variation of different equivalent water thicknesses of the canopy in the band between 1000 and 2400. Figure 1 (c) shows the reflectance variation of different equivalent water thicknesses of the canopy in the band between 400 and 1000. Figure 1 (d) is a graph showing the reflectance variation of different soil moisture contents under different spectral bands;

[0050] Figure 2 This is a schematic diagram illustrating the stretching motion of the O-H bonds in water molecules during the present invention; wherein... Figure 2 (a) is a diagram of water molecules. Figure 2(b) is a diagram of the symmetric stretching vibration of water molecules. Figure 2 (c) is an asymmetric stretching vibration diagram of water molecules. Figure 2 (d) is a diagram of the bending vibration of water molecules;

[0051] Figure 3 This is a schematic diagram of the spectral response function of the moisture-sensitive band provided by the present invention;

[0052] Figure 4 This invention provides a schematic diagram of the reflectance of soil and vegetation under different moisture conditions in the moisture-sensitive wavelength band; where... Figure 4 (a) is a schematic diagram of the reflectance of vegetation under different moisture conditions in the moisture-sensitive wavelength band. Figure 4 (b) is a schematic diagram of the reflectance of soil under different moisture conditions in the moisture-sensitive band;

[0053] Figure 5 This is a flowchart of a water prediction method based on airborne multispectral sensitive band combination according to the present invention;

[0054] Figure 6 This is a cross-validation diagram provided for the present invention. Detailed Implementation

[0055] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0056] In traditional techniques, the NDVI (Normalized Difference Vegetation Index) is often used to assess vegetation cover and health. However, because it relies solely on red and near-infrared light, it struggles to accurately reflect differences in moisture content across different vegetation types or soils. Furthermore, when water molecules absorb radiant energy, the O-H bonds vibrate, triggering a combination of overtones involving symmetrical stretching, asymmetrical stretching, and bending motions, resulting in significant water absorption bands in multiple wavelengths above 1000 nm. These bandwidths typically include 970 nm, 1200 nm, 1650 nm, and 2200 nm.

[0057] like Figure 1 and Figure 5 As shown, this invention provides a moisture prediction method based on airborne multispectral sensitive band combination, specifically including the following steps:

[0058] Step S1: Based on the response characteristics of soil and plant canopy reflectance spectra to water, select the center wavelengths of multiple water-sensitive bands. According to the physical mechanism of O-H bond vibration in water molecules, increase and select the center wavelengths of multiple different water-sensitive bands, and consider the influence of atmospheric water vapor on near-infrared spectra to determine the bandwidth of all center wavelengths.

[0059] Based on the response characteristics of soil and vegetation reflectance spectra to water, the center wavelength of the water-sensitive band is selected. For example... Figure 1 As shown, the soil reflectance spectrum generally decreases with increasing soil moisture content (SMC), while the vegetation canopy spectrum is relatively complex. In the visible light region, vegetation moisture has a relatively small impact on the canopy reflectance spectrum; the selection of wavelengths mainly needs to consider the effects of reduced chlorophyll synthesis and accelerated decomposition under water stress on the canopy reflectance spectrum. Due to the influence of chlorophyll absorption characteristics, the vegetation canopy reflectance spectrum forms a reflection peak in the green light region around 560 nm, while there are two significant absorption valleys in the blue light region around 450 nm and the red light region around 660 nm. Around 705-750 nm, the reflectance increases rapidly, forming a "red edge." When the chlorophyll content decreases, the reflectance in the visible light region increases, accompanied by a red-edge blue shift. In the near-infrared region, water shortage in plants leads to the shrinkage of plant cells and the wrinkling of leaves, resulting in thinner leaves and closed stomata, affecting the absorption and reflection characteristics of light in the near-infrared region. Influenced by plant water content and cell structure, the canopy reflectance spectrum exhibits several water absorption bands around 970 nm, 1200 nm, 1450 nm, 1900 nm, and 2500 nm, and several reflectance peaks around 1100 nm, 1300 nm, 1650 nm, and 2200 nm. The absorption valley around 1450 nm is solely influenced by leaf water content, while the three absorption valleys at 970, 1200, and 1900 nm are also affected by leaf starch and protein. Furthermore, considering the strong influence of water vapor in the 1300-1500 nm, 1750-2000 nm, and 2300-2500 nm wavelength ranges, the 1550–1750 nm range and the area around 2200 nm are the most suitable wavelength bands for monitoring plant canopy water content. In addition, 800 nm and 900 nm can be used to construct the Normalized Difference Vegetation Index (NDVI) and the Water Index (WI). Therefore, 560nm, 705nm, 750nm, 800nm, 900nm, 970nm, 1100nm, 1200nm, 1650nm and 2200nm were selected as the center wavelengths of the moisture-sensitive band.

[0060] Based on the physical mechanism of O-H bond vibration in water molecules, the center wavelength of the water-sensitive band is selected. Water molecules absorb radiation to gain energy, resulting in rotational transitions, intermolecular vibrational transitions, intramolecular vibrational transitions, and electronic transitions. Rotational and intermolecular vibrational transitions determine absorption in the microwave and far-infrared ranges, electronic transitions determine absorption in the ultraviolet range, and intramolecular vibrational transitions determine absorption in the mid-infrared range. In the visible and near-infrared regions, water absorption is mainly due to stretching overtones and vibrational absorption of O-H bond vibrations in water molecules. In liquid water, each water molecule can form up to four hydrogen bonds with surrounding water molecules. The continuous breaking and recombination of hydrogen bonds restricts the rotation and vibration of water molecules. This restriction leads to vibrational modes of water molecules, including symmetric stretching vibrations, asymmetric stretching vibrations, and bending vibrations, such as... Figure 2 The stretching and contracting vibrations of the O-H bonds mainly occur in the high-frequency region (approximately 3400 cm⁻¹). -1 Bending vibrations mainly occur in the mid-frequency region (approximately 1600 cm). -1 Because the energies required for symmetrical and asymmetrical stretching are very similar, the absorption peaks merge into a large absorption peak around 3000 nm. Bending vibrations occur at lower energy levels, resulting in an absorption peak around 6000 nm. Harmonics of these vibrations occur at higher energy levels (i.e., shorter wavelengths), increasing the required energy and decreasing the absorption intensity. The harmonics of bending and stretching vibrations cause shoulder peaks in the absorption spectrum of water in the visible and near-infrared range. Since the absorption intensity of continuous overtone transitions in water decreases by 10 to 20 times with increasing stretching quantum number, overtone transitions with more than 5 stretching quanta have no significant effect on reflectivity. The positions of vibrational absorption in liquid water are shown in Table 1.

[0061] Table 1. Locations of vibration absorption by liquid water

[0062] Wavelength (nm) <![CDATA[Wave number (cm -1 )]]> explain 1940 5260 av1 + v2 + bv3 (a + b = 1) 1450 6800 av1 + bv3(a + b = 2) 1190 8330 av1 + v2 + bv3 (a + b = 2) 970 10310 av1 + bv3 (a + b = 3) 836 11960 av1 + v2 + bv3 (a + b = 3) 750 13330 av1 + bv3(a + b = 4) 660 15150 av1 + v2 + bv3 (a + b = 4) 605 16500 av1 + bv3(a + b = 5) 514 19460 av1 + bv3(a + b = 6) 449 22270 av1 + bv3(a + b = 7) 401 24940 av1 + bv3(a + b = 8)

[0063] Therefore, 450m and 660nm were chosen as the center wavelengths of the moisture-sensitive band.

[0064] The bandwidth of the center wavelength of the moisture-sensitive band was determined based on factors such as spectral absorption characteristics, atmospheric transmittance, and practical feasibility. For example, 970nm is located in the water absorption band, and 705nm is located in the red edge region, requiring a narrower bandwidth to accurately capture changes in moisture content. In regions with low atmospheric transmittance, the bandwidth should be relatively wider to ensure sufficient signal acquisition, such as in the near-infrared regions of 800nm ​​and 900nm. Considering the flight speed and image resolution of the airborne platform, the bandwidth setting needs to strike a balance between capturing sufficient spectral information and ensuring spatiotemporal resolution. Based on these principles, the bandwidth of the selected moisture-sensitive band was chosen to balance spectral information and spatiotemporal resolution, with bandwidths of 30nm, 30nm, 30nm, 10nm, 15nm, 40nm, 40nm, 10nm, 50nm, 50nm, 50nm, and 100nm.

[0065] Step S2: Simulate and generate the reflectance spectra of soil and vegetation canopy, add Gaussian noise to the generated reflectance spectra to simulate real data, generate a spectral response function using the relationship between the center wavelength and bandwidth of multiple different water-sensitive bands, convert the simulated reflectance spectra to multiple water-sensitive bands using the spectral response function, obtain the reflectance and spectral index of multiple water-sensitive bands, and use the reflectance and spectral index to construct a water prediction model.

[0066] The spectral response function describes the sensor's sensitivity to light of different wavelengths, and simulating the spectral response function is a crucial step in designing multispectral sensors. According to this invention, the center wavelength and bandwidth of the moisture-sensitive band are selected, and a Gaussian function is used to approximate the spectral response function. Specifically:

[0067]

[0068] Where SRF(λ) is the spectral index at wavelength λ, λ0 is the center wavelength, and σ is the standard deviation, calculated using the bandwidth FWMH, which controls the width of the spectral response function curve. The spectral response function is as follows: Figure 3 As shown.

[0069] Based on the approximate spectral response function, the soil reflectance spectrum and the canopy reflectance spectrum are converted to obtain a selection of multiple water-sensitive band combinations (12 water-sensitive band combinations selected in this invention), and the reflectance of multiple water-sensitive band combinations, specifically:

[0070]

[0071] Where R bandThis represents the resampled reflectance, where λ1 and λ2 are the wavelength ranges for this band, and BRF(λ) is the reflectance at a simulated wavelength of λ. Gaussian noise at a 10% level is added to the resampled reflectance data to simulate real data. The resampling results are as follows... Figure 4 .

[0072] By combining multiple moisture-sensitive bands and inputting them into the spectral response function, the spectral index is obtained.

[0073] The MARMIT2 model was used to simulate the reflectance spectrum of soil within the solar domain. The SMC of the simulated soil reflectance spectrum was normalized and used as the dimensionless soil factor psoil. The reflectance spectra of soils with the lowest and highest water content were used as inputs to the PROSAIL radiative transfer model to simulate the canopy reflectance spectrum.

[0074] Approximately 100,000 soil reflectance spectra were simulated using the MARMIT2 model. Specifically, after simulating the reflectance spectra of soil within the solar domain using the MARMIT2 model, the soil moisture content (SMC) representing the soil reflectance spectra was fitted using the sigmoid function based on the product of the input water layer thickness and surface water cover. The details are as follows:

[0075] BRF Soil (λ)=f MARMIT2 (λ,L,ε,δ);

[0076] Among them, BRF Soil (λ) represents soil reflectance, and f represents the soil moisture content (SMC) characterizing the soil reflectance spectrum. MARMIT2 (·) is the operational function of the MARMIT2 radiative transfer model, λ is the wavelength, L is the water layer thickness, ε is the surface water coverage, and δ is the soil particle volume fraction.

[0077] The PROSAIL radiative transfer model was used to simulate 50,000 canopy reflectance spectra, and the equivalent water thickness was obtained. The canopy equivalent water thickness (CEWT) and normalized soil moisture content were represented by psoil, respectively. Specifically:

[0078] CEWT = LAI × C w ;

[0079] DHR Leaf (λ)=f PROSPECT (λ,θ DHR );

[0080] BRF Canopy (λ)=f SAIL (λ,DHR(λ),θ BRF );

[0081] Where CEWT is the canopy equivalent water thickness, LAI is the leaf area index, and C w For equivalent water thickness, DHR Leaf (λ) is obtained by using the PROSPECT model through f PROSPECT The simulated directional hemispherical reflectance (DHR) of the blade at a wavelength of λ, θ DHR The input parameters for the PROSPEC T model; BRF Canopy (λ) represents the value obtained by using the 4SAIL model, through f SAIL Coupled DHR Leaf (λ), the bidirectional reflectance factor (BRF) of the canopy at a simulated wavelength of λ, θ BRF These are the input parameters for the 4SAIL model. The range of input parameters is shown in Table 2:

[0082] Table 2. Input parameter ranges for the MARMIT2 and PROSAIL radiative transfer models.

[0083] Input parameters Physical meaning unit scope L Water layer thickness cm 0-0.05 ε Surface water coverage Dimensionless 0-1 δ Soil particle volume fraction Dimensionless 0-0.25 N Blade structural parameters Dimensionless 1.5 Cab Chlorophyll a+b content <![CDATA[μg / cm 2 ]]> 30-70 Car Carotenoid content <![CDATA[μg / cm 2 ]]> Cab / 4 Cw Equivalent water thickness cm 0-0.05 Cm Dry matter content <![CDATA[g / cm 2 ]]> 0.012 LAI Leaf area index <![CDATA[m 2 / m 2 ]]> 1-5 ALA Mean blade tilt angle ° 20-75 Hotspot Hotspot effect parameters Dimensionless 0.5 / LAI tts Solar zenith angle ° 25-50 tto Observation zenith angle ° 0 psi relative azimuth ° 0-360 psoil Soil factors Dimensionless 0-1

[0084] Vegetation indices were calculated using the constructed dataset, with the formulas for spectral indices shown in Tables 3-1 and 3-2:

[0085] Table 3-1 Vegetation Index Formula

[0086] Vegetation Index formula RVI R800 / R660 NDVI (R800-R660) / (R800+R660) WI R900 / R970 SR2 R560 / R660 mNDVI (R750-R705) / (R750+R705) WI / mNDVI WI / mNDVI WI / NDVI WI / NDVI RDVI (R800-R660) / sqrt(R800+R660) GNDVI (R800-R560) / (R800+R560) RGRI R660 / R560 OSAVI 1.16*(R800-R660) / (R800+R660+0.16) TCARI 3*((R705-R660)-0.2*(R705-R560)*R705 / R660) OSAVI / TCARI OSAVI / TCARI MCARI TCARI / 3 MCARI1 1.2*(2.5*(R800-R660)-1.3*(R800-R560)) MSR (R800 / R660-1) / sqrt(R800 / R660+1) MTVI3 1.2*(1.2*(R800-R560)-2.5*(R660-R560)) <![CDATA[mSR 705 ]]> (R750-R450) / (R705-R450) <![CDATA[mNDVI 705 ]]> (R750-R705) / (R750+R705-2*R450) SIPI (R800-R450) / (R800+R660) TVI 0.5*(120*(R750-R560)-200*(R660-R560)) NDRE (R800-R750) / (R800+R750) VARI (R560-R660) / (R560+R660-R450)

[0087] In the formula, R450, R560, R660, R705, R750, R800, R900, R970, R1100, R1200, R1650, and R2200 represent the reflectivity at 450nm, 560nm, 660nm, 705nm, 750nm, 800nm, 900nm, 970nm, 1100nm, 1200nm, 1650nm, and 2200nm, respectively.

[0088] Table 3-2 Vegetation Index Formula

[0089] ARVI (R800-(2*R660-R450)) / (R800+(2*R660-R450)) <![CDATA[VARI rededge ]]> (R705-R660) / (R705+R660) DVI R800-R660 BNDVI (R800-R450) / (R800+R450) <![CDATA[RDVI green ]]> (R750-R560) / (R750+R560) <![CDATA[DVI green ]]> (R750-R560) CAR (R800-(1*R660+0.5*R450)) / (R800+(1*R660+0.5*R450)) GCI R800 / R450-1 EVI 2.5*(R800-R660) / (R800+6*R660-7.5*R450+1) LWVI1 (R1100-R900) / (R1100+R900) LWVI2 (R1100-R1200) / (R1100+R1200) DWSI (R800+R560) / (R1650+R660) DWSI1 R800 / R1650 DWSI2 R1650 / R560 DWSI3 R1650 / R660 DWSI4 (R800-R560) / (R1650+R660) SR R1650 / R2200

[0090] Step S3: Combine all the water-sensitive bands with determined bandwidth and integrate them into the airborne multispectral sensor. Collect reflectance spectral images of the soil and vegetation canopy at the selected location, and input the reflectance spectral images into the water prediction model to predict the water content of the soil and vegetation canopy at the selected location.

[0091] Before inputting reflectance spectral images into the moisture prediction model, the following steps are also included:

[0092] We collected real soil and vegetation canopy reflectance spectral data, and used spectral response functions to convert the real soil and vegetation canopy reflectance spectral data into multiple water-sensitive bands, obtaining the reflectance and spectral index of the water-sensitive bands under multiple real conditions.

[0093] The water prediction model's prediction results were subjected to nested cross-validation using reflectance and spectral indices of water-sensitive bands under multiple real-world conditions. The model was then optimized based on the results of this nested cross-validation to obtain the optimal water prediction model. Details are as follows:

[0094] Based on reflectance data and spectral indices of moisture-sensitive bands, a random forest regression model was established using nested cross-validation to predict soil moisture content (SMC), canopy equivalent water thickness (CEWT), and soil factor ps oil, and to evaluate the rationality of band selection for soil and vegetation moisture monitoring.

[0095] In this nested cross-validation process, the outer cross-validation layer randomly divides the data into 10 folds, retaining one fold as the test set and the remaining 9 folds as the training set to evaluate the optimal model selected in the inner cross-validation layer. The inner cross-validation layer divides the training set from the outer cross-validation layer into 10 folds to optimize and select the hyperparameters of the random forest regression model (water prediction model). Specifically... Figure 6 As shown.

[0096] By using independent inner and outer layer validation, information leakage during the model selection process is reduced, and a more accurate estimate of the model's generalization ability is provided.

[0097] The predicted performance for CEWT and psoil in the examples is as follows:

[0098] Table 4 Predicted performance of CEWT and psoil

[0099]

[0100] in, It is the mean of the coefficients of determination (RMSE) of the prediction results from 10 nested cross-validations. ncv and nRMSE ncv represents the root mean square error and the mean normalized root mean square error of the results of 10 nested cross-validations, respectively.

[0101] The prediction results show that the combination of noisy simulated reflectance and spectral index can effectively predict the three moisture parameters, indicating the theoretical rationality of the band selection for soil and vegetation moisture monitoring in this invention. The rationality of the band selection is verified using real datasets of soil and vegetation canopy reflectance spectra.

[0102] Table 5 Data Source

[0103] type Dataset Name Sample size Predicting targets Collection area soil Bablet_2016 106 SMC Tunisia, France soil Dupiau_2020 72 SMC Chile, Tunisia, the United States, China soil Humper_2015 455 SMC Tunisia soil Lesaignoux_2008 190 SMC France soil Liu_2002 367 SMC France soil Lobell_2002 41 SMC USA soil Marcq_2012 258 SMC Germany, China, Tunisia soil Philpot_2014 405 SMC USA vegetation ISRAEL_2004 207 CEWT Israel vegetation ISRAEL_2005 124 CEWT Israel

[0104] The soil reflectance spectral dataset used to validate the rationality of band selection was collected from multiple countries and regions worldwide, demonstrating good representativeness. Eight soil datasets were merged, and soil monograms (SMCs) were predicted to validate the rationality of the band selection. The vegetation canopy reflectance spectral dataset provides canopy reflectance spectra for wheat throughout its entire growth period, including dry weight (g / m³). 2 ), moisture content (%) and leaf area index (m²) 2 / m 2 The parameters are:

[0105]

[0106] Where WC is the leaf water content of the plant, DW is the dry weight of the leaf per unit area, and W... f For the fresh weight of the leaves, W d Let A be the leaf dry weight and A be the sampled plant area. The prediction target for the vegetation canopy reflectance spectral dataset is CEWT, which is equal to the product of LAI and EWT. EWT can be calculated using the following formula:

[0107]

[0108] Where, ρ w The density of water. Due to the lack of measured LAI data for some samples in the vegetation canopy reflectance spectrum dataset, the sample size is only 150 and 96. The target variable was predicted using modeling methods to verify the rationality of the band selection; the results are shown in the table below.

[0109] The prediction results show that the combination of reflectance and spectral index constructed based on real datasets can predict CWET and SMC well, which indicates that the band selection for soil and vegetation moisture monitoring in this invention also has high performance in practical use.

[0110] Table 6 verifies the rationality of the band selection.

[0111]

[0112] Table 7 Texture Index Formula

[0113]

[0114] For the actual acquired image data, the texture index of the grayscale image for each band and vegetation index can be further calculated. The texture index can be extracted using the gray-level co-occurrence matrix method, and the calculation formula is shown in Table 7.

[0115] In the formula: i and j represent the gray values ​​at the i-th row and j-th column of the image, respectively; P(i,j) represents the probability of the corresponding gray value appearing in the i-th row and j-th column of the matrix; N is the gray level of the image; Mean is the mean of the gray-level co-occurrence matrix, σ i and σ j The standard deviation is denoted as .

[0116] The embodiments in this description provide the selection criteria and water prediction methods for using airborne multispectral water-sensitive band combinations. By establishing a machine learning inversion model, accurate and efficient inversion of soil and crop water information in agricultural areas can be achieved.

[0117] Based on the above method, the present invention provides a moisture prediction system based on airborne multispectral sensitive band combination, including: a selection module, a model building module and a prediction module.

[0118] The selection module is used to select the center wavelengths of multiple water-sensitive bands based on the response characteristics of soil and plant canopy reflectance spectra to water. According to the physical mechanism of O-H bond vibration in water molecules, it adds and selects the center wavelengths of multiple different water-sensitive bands and determines the bandwidth of all center wavelengths. The model building module is used to simulate and generate the reflectance spectra of soil and vegetation canopy. It generates a spectral response function using the relationship between the center wavelengths and bandwidths of multiple different water-sensitive bands. The simulated reflectance spectrum is converted to multiple water-sensitive bands through the spectral response function to obtain the reflectance and spectral index of multiple water-sensitive bands. The reflectance and spectral index are then used to build a water prediction model. The prediction module is used to combine all water-sensitive bands with determined bandwidths and integrate them into an airborne multispectral sensor. It collects reflectance spectral images of soil and vegetation canopy at selected locations and inputs the reflectance spectral images into the water prediction model to predict the water content of soil and vegetation canopy at selected locations.

[0119] The present invention also provides a computer device, including a memory and a processor. The memory stores a program, and when the program is executed by the processor, the processor performs the steps of a moisture prediction method based on an airborne multispectral sensitive band combination.

[0120] According to the disclosed embodiments, the computer device can communicate with one or more external devices (e.g., keyboard, pointing device, Bluetooth communication, etc.) or with any device that enables the computing device to communicate with one or more other computing devices (e.g., router, demodulator, etc.).

[0121] The present invention also provides a storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of a moisture prediction method based on an airborne multispectral sensitive band combination.

[0122] According to the disclosed embodiments, the storage medium can be a non-volatile computer-readable storage medium, such as, but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this invention, the storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0123] The above description, in conjunction with specific preferred embodiments, provides a more detailed explanation of the present invention. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such deductions or substitutions should be considered to fall within the scope of protection of the present invention.

Claims

1. A moisture prediction method based on airborne multispectral sensitive band combination, characterized in that, Includes the following steps: Based on the response characteristics of soil and plant canopy reflectance spectra to water, the center wavelengths of multiple water-sensitive bands were selected. According to the physical mechanism of O-H bond vibration in water molecules, the center wavelengths of multiple different water-sensitive bands were increased and selected, and the bandwidth of all center wavelengths was determined. The reflectance spectra of soil and vegetation canopy are simulated and generated. A spectral response function is generated by using the relationship between the center wavelength and bandwidth of the multiple different water-sensitive bands. The simulated reflectance spectrum is converted to multiple water-sensitive bands through the spectral response function. The reflectance and spectral index of multiple water-sensitive bands are obtained. A water prediction model is constructed using the reflectance and spectral index. All water-sensitive bands with a defined bandwidth are combined and integrated onto an airborne multispectral sensor to acquire reflectance spectral images of soil and vegetation canopy at selected locations. These reflectance spectral images are then input into a water prediction model to predict the water content of soil and vegetation canopy at the selected locations.

2. The moisture prediction method based on airborne multispectral sensitive band combination as described in claim 1, characterized in that, The simulated reflectance spectra of soil and vegetation canopy include: The MARMIT2 radiative transfer model was used to simulate the reflectance spectrum of soil in the solar domain, and the soil moisture content (SMC) of the simulated soil reflectance spectrum was normalized as a dimensionless soil factor (psoil). The reflectance spectra of soils with the lowest and highest moisture content were used as inputs to the PROSAIL radiative transfer model to simulate the reflectance spectra of the vegetation canopy. In this model, the MARMIT2 radiative transfer model simulates the soil reflectance spectrum within the solar domain. Based on the product of the input water layer thickness and surface water cover, the sigmoid function is used to fit the soil moisture content (SMC) representing the soil reflectance spectrum. The specific expression is as follows: BRF Soil (λ)=f MARMIT2 (λ,L,e,d); Among them, BRF Soil (λ) represents soil reflectance, and f represents the soil moisture content (SMC) characterizing the soil reflectance spectrum. MARMIT2 (·) is the operational function of the MARMIT2 radiative transfer model, λ is the wavelength, L is the water layer thickness, ε is the surface water coverage, and δ is the soil particle volume fraction. In the PROSAIL radiative transfer model, the equivalent water thickness is obtained, and the equivalent water thickness of the vegetation canopy and the normalized soil moisture content are represented by the equivalent water thickness and psoil, respectively; the specific expressions are as follows: CEWT=LAI×C w ; DHR Leaf (λ)=f PROSPECT (λ,θ DHR ); BRF Canopy (λ)=f SAIL (λ,DHR(λ),θ BRF ); Where CEWT is the canopy equivalent water thickness, LAI is the leaf area index, and C w For equivalent water thickness, DHR Leaf (λ) is the value of f using the PROSPECT model. PROSPECT The simulated wavelength is λ, and the reflectivity of the blade's directional hemisphere is θ. DHR Input parameters for the PROSPECT model; BRF Canopy (λ) represents the value obtained by using the 4SAIL model, through f SAIL Coupled DHR Leaf (λ), the bidirectional reflectivity of the canopy at a simulated wavelength of λ, θ BRF These are the input parameters for the 4SAIL model.

3. The moisture prediction method based on airborne multispectral sensitive band combination as described in claim 2, characterized in that, A spectral response function is generated using the relationship between the center wavelength and bandwidth of the multiple different moisture-sensitive bands. This spectral response function is then used to convert the simulated reflectance spectrum to the multiple moisture-sensitive bands, obtaining the reflectance and spectral indices of these bands, including: Based on the center wavelength and bandwidth of the selected moisture-sensitive band, the spectral response function is approximated using a Gaussian function; the specific expression is: Where SRF(λ) is the spectral index at wavelength λ, λ0 is the center wavelength, and σ is the standard deviation, which is calculated by the bandwidth FWMH to control the width of the spectral response function curve; The soil reflectance spectrum and the canopy reflectance spectrum are converted to obtain multiple selected combinations of water-sensitive bands, as well as the reflectance of these combinations; the specific expression is as follows: Where R band λ is the reflectance after resampling, λ1 and λ2 are the wavelength range of this band, and BRF(λ) is the reflectance when the simulated wavelength is λ. By combining multiple moisture-sensitive bands and inputting them into the spectral response function, the spectral index is obtained.

4. The moisture prediction method based on airborne multispectral sensitive band combination as described in claim 1, characterized in that, Before inputting the reflectance spectral image into the moisture prediction model, the process also includes: We collected real soil and vegetation canopy reflectance spectral data, and used spectral response functions to convert the real soil and vegetation canopy reflectance spectral data into multiple water-sensitive bands, obtaining the reflectance and spectral index of water-sensitive bands under multiple real conditions. The water prediction model is evaluated by nested cross-validation using reflectance and spectral indices of water-sensitive bands under multiple real-world conditions. The water prediction model is then optimized using the results of the nested cross-validation to obtain the optimal water prediction model.

5. The moisture prediction method based on airborne multispectral sensitive band combination as described in claim 4, characterized in that, The method of using reflectance and spectral indices of multiple moisture-sensitive bands under real-world conditions to perform nested cross-validation of the moisture prediction model results includes: The outer cross-validation layer of nested cross-validation randomly divides the data into 10 folds, retaining 1 fold as the test set and the remaining 9 folds as the training set, and evaluates the optimal model selected in the inner cross-validation layer. The inner cross-validation layer divides the training set of the outer cross-validation layer into 10 folds, optimizes and selects the hyperparameters of the moisture prediction model.

6. The moisture prediction method based on airborne multispectral sensitive band combination as described in claim 1, characterized in that, Based on the response characteristics of soil and plant canopy reflectance spectra to water, the center wavelengths of multiple water-sensitive bands are selected. According to the physical mechanism of O-H bond vibration in water molecules, multiple center wavelengths of different water-sensitive bands are added and selected, including: Based on the response characteristics of soil and plant canopy reflectance spectra to water, 560nm, 705nm, 750nm, 800nm, 900nm, 970nm, 1100nm, 1200nm, 1650nm and 2200nm were selected as the center wavelengths of the water-sensitive band. Based on the physical mechanism of O-H bond vibration in water molecules, 450m and 660nm were selected as the center wavelengths of the water-sensitive band.

7. The moisture prediction method based on airborne multispectral sensitive band combination as described in claim 1, characterized in that, Determining the bandwidth of all center wavelengths includes: By selecting bandwidth-balanced spectral information and spatiotemporal resolution in the moisture-sensitive bands, the selected bandwidths are 30nm, 30nm, 30nm, 10nm, 15nm, 40nm, 40nm, 10nm, 50nm, 50nm, 50nm and 100nm.

8. A moisture prediction system based on airborne multispectral sensitive band combination, characterized in that, include: The selection module is used to select the center wavelength of multiple water-sensitive bands based on the response characteristics of soil and plant canopy reflectance spectra to water, and to add and select the center wavelengths of multiple different water-sensitive bands according to the physical mechanism of O-H bond vibration in water molecules, and to determine the bandwidth of all center wavelengths. The model building module is used to simulate and generate the reflectance spectra of soil and vegetation canopy. It generates a spectral response function by utilizing the relationship between the center wavelength and bandwidth of the multiple different water-sensitive bands. The simulated reflectance spectrum is converted to multiple water-sensitive bands through the spectral response function, and the reflectance and spectral index of multiple water-sensitive bands are obtained. The reflectance and spectral index are then used to build a water prediction model. The prediction module is used to combine all water-sensitive bands with a defined bandwidth and integrate them into an airborne multispectral sensor to acquire reflectance spectral images of soil and vegetation canopy at selected locations. The reflectance spectral images are then input into a water prediction model to predict the water content of soil and vegetation canopy at the selected locations.

9. A computer device, characterized in that, The device includes a memory and a processor, wherein the memory stores a program that, when executed by the processor, causes the processor to perform the steps of a moisture prediction method based on an airborne multispectral sensitive band combination as described in any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the moisture prediction method based on the combination of airborne multispectral sensitive bands as described in any one of claims 1 to 7.

Images