A method and apparatus for simultaneous inversion of leaf reflectance and transmittance

Abstract

This invention discloses a method and apparatus for simultaneously inverting leaf reflectance and transmittance, comprising: decomposing canopy reflectance within the passive optical remote sensing band into primary scattering and multiple scattering; constructing a first relationship model between primary scattering and leaf reflectance and a second relationship model between multiple scattering and leaf scattering coefficient; obtaining an inversion model based on the first and second relationship models, wherein the inversion model expresses canopy reflectance as a binary function of leaf reflectance and leaf transmittance; and inputting multi-angle canopy reflectance observations into the inversion model to obtain the inversion results of leaf reflectance and leaf transmittance. This invention solves the problem of ill-conditioned inversion in the field of vegetation remote sensing.

Description

Technical Field

[0001] This invention belongs to the field of remote sensing technology, and in particular relates to a method and apparatus for simultaneously inverting the reflectivity and transmittance of a leaf. Background Technology

[0002] Canopy reflectance is the product of the combined effects of multiple factors. The main factors influencing canopy reflectance include ① canopy structure (leaf area, aggregation, number of trees, tree distribution, leaf tilt angle distribution, etc.), ② background type (i.e., reflectance of the understory surface), ③ illumination observation geometry (i.e., the position of the sensor and the sun), ④ topography, and ⑤ leaf characteristics (leaf biochemical parameter content, thickness, shape, etc.). When using remote sensing to retrieve leaf biochemical parameters (⑤), the influence of the other four factors must be eliminated from the canopy reflectance to accurately retrieve the leaf biochemical parameters; otherwise, the retrieval results will have significant errors (Knyazikhin et al. 2013). To address this, Zhang et al. (2008) proposed a two-step retrieval method. This method first downscales the canopy reflectance to leaf reflectance, and then uses the leaf reflectance to retrieve the leaf biochemical parameters. The influences of canopy structure, background, illumination observation geometry, and topography can be eliminated by establishing a lookup table using a mechanistic model. The principle is as follows:

[0003] Canopy reflectivity is a linear combination of four parts, which correspond to the contributions of the illuminated canopy, the shaded canopy, the illuminated background, and the shaded background to the canopy reflectivity, respectively. This can be expressed by the following formula:

[0004]

[0005] in, , , and These represent the probabilities of seeing the illuminated canopy, the shadowed canopy, the illuminated background, and the shadowed background, respectively. , , and These represent the reflectance of the illuminated canopy, the shaded canopy, the illuminated background, and the shaded background, respectively. To express the reflectance of the illuminated canopy (… ) converted to leaf reflectance ( They introduced a multiple scattering factor. The effects of the shaded canopy and the shaded background are included in this scattering factor, and finally the conversion relationship between canopy reflectance and leaf reflectance is obtained:

[0006]

[0007] For a forest canopy, if its structural parameters and background information are known, under any illumination observation geometry, , and Since it is a constant, the above formula is actually a linear relationship. If defined... equal to a It equals b. Therefore, formula (2) becomes...

[0008]

[0009] The parameters a and b for each band are obtained through a canopy BRDF mechanism model called a four-scale model, thereby realizing the conversion of canopy reflectivity to leaf reflectivity.

[0010] Two-step inversion methods have been applied to global vegetation chlorophyll and Vcmax mapping. However, the biochemical parameter products generated by this method have significant errors, and the inversion of seasonal changes in deciduous forests is not very accurate. There are many sources of error, such as the influence of non-photosynthetic substances and pathological inversion problems (i.e., different substances in the same spectrum), with pathological inversion being the main cause. The optical property truly related to leaf biochemical parameters is leaf absorptivity. Leaf reflectance, transmittance, and absorptivity have the following relationship:

[0011]

[0012] Since current technology can only obtain leaf reflectivity from canopy reflectivity, it is impossible to calculate leaf absorptivity. Summary of the Invention

[0013] Current technologies can only obtain leaf reflectance from canopy reflectance, but not leaf transmittance. Using only leaf reflectance can lead to the problem of different materials in the same spectrum. This invention provides a method and apparatus for simultaneously inverting leaf reflectance and transmittance, which solves the problem of ill-conditioned inversion in the field of vegetation remote sensing.

[0014] To achieve the above objectives, the present invention adopts the following technical solution:

[0015] A method for simultaneously inverting the reflectivity and transmittance of blades includes:

[0016] Step S1: Decompose the canopy reflectivity within the passive optical remote sensing band range into primary scattering and multiple scattering;

[0017] Step S2: Construct a first relationship model between primary scattering and blade reflectivity, and a second relationship model between multiple scattering and blade scattering coefficient;

[0018] Step S4: Based on the first relational model and the first relational model, an inversion model is obtained, wherein the inversion model expresses the canopy reflectivity as a binary function of leaf reflectivity and leaf transmittance;

[0019] Step S3: Input the multi-angle canopy reflectance observations into the inversion model to obtain the inversion results of leaf reflectance and leaf transmittance.

[0020] Preferably, the passive optical remote sensing band range is 400nm–2500nm.

[0021] Preferably, the canopy reflectivity From primary scattering and multiple scattering Composition, namely:

[0022]

[0023] in, Indicates the band. Represents illumination-observation geometry.

[0024] Preferably, the first relationship model between primary scattering and blade reflectivity, and the second relationship model between multiple scattering and blade scattering coefficient are respectively:

[0025]

[0026]

[0027] in, Indicates in band Leaf reflectance at nm A linear transformation model representing the relationship between leaf reflectivity and canopy primary scattering. A nonlinear transformation model representing the relationship between leaf scattering factor and canopy multiple scattering. The leaf scattering factor is:

[0028]

[0029] in, Indicates in band Leaf transmittance at nm.

[0030] Preferably, the inversion model is:

[0031]

[0032] in, .

[0033] The present invention also provides a device for simultaneously inverting the reflectivity and transmittance of blades, comprising:

[0034] The splitting module is used to split the canopy reflectivity within the passive optical remote sensing band into primary scattering and multiple scattering;

[0035] The first building module is used to build a first relationship model between primary scattering and blade reflectivity and a second relationship model between multiple scattering and blade scattering coefficient;

[0036] The second construction module is used to obtain an inversion model based on the first relational model and the first relational model. The inversion model represents the canopy reflectivity as a binary function of leaf reflectivity and leaf transmittance.

[0037] The inversion module is used to input multi-angle canopy reflectance observations into the inversion model to obtain the inversion results of leaf reflectance and leaf transmittance.

[0038] Preferably, the passive optical remote sensing band range is 400nm–2500nm.

[0039] Preferably, the canopy reflectivity From primary scattering and multiple scattering Composition, namely:

[0040]

[0041] in, Indicates the band. Represents illumination-observation geometry.

[0042] Preferably, the first relationship model between primary scattering and blade reflectivity, and the second relationship model between multiple scattering and blade scattering coefficient are respectively:

[0043]

[0044]

[0045] in, Indicates in band Leaf reflectance at nm A linear transformation model representing the relationship between leaf reflectivity and canopy primary scattering. A nonlinear transformation model representing the relationship between leaf scattering factor and canopy multiple scattering. The leaf scattering factor is:

[0046]

[0047] in, Indicates in band Leaf transmittance at nm.

[0048] Preferably, the inversion model is:

[0049]

[0050] in,

[0051] This invention decomposes canopy reflectivity into primary scattering and multiple scattering, and establishes the relationship between primary scattering and leaf reflectivity, and multiple scattering and leaf scattering coefficient (i.e., the sum of leaf reflectivity and transmittance) respectively. Canopy reflectivity can be expressed as a binary function of leaf reflectivity and transmittance. Through multi-angle observation of canopy reflectivity, leaf reflectivity and transmittance can be inverted simultaneously. Attached Figure Description

[0052] Figure 1 This is a flowchart of the method for simultaneously inverting the reflectivity and transmittance of blades according to an embodiment of the present invention;

[0053] Figure 2 This is a flowchart illustrating the method for simultaneously inverting blade reflectivity and transmittance using the geometric optics model GOST2, as described in an embodiment of the present invention.

[0054] Figure 3 A schematic diagram showing the relationship between canopy primary scattering and leaf reflectance under various illumination-observation geometry;

[0055] Figure 4 A schematic diagram showing the relationship between canopy primary scattering and leaf scattering factor under various illumination-observation geometry;

[0056] Figure 5 A band-by-band comparison of canopy reflectance simulated by GOST2 and observed by unmanned aerial vehicles (UAVs);

[0057] Figure 6 This is a schematic diagram illustrating the absolute value of the relative error in the GOST2 simulation; the gray bars represent the absolute value of the relative error. The scatter plots represent the model simulation (GOST2), lookup table simulation (LUT), and UAV-measured canopy reflectivity.

[0058] Figure 7 This study analyzes the trend of the proportion of canopy primary scattering in canopy reflectivity as a function of wavelength, where the dashed line represents the coefficient of determination R for canopy reflectivity and leaf reflectivity. 2 ;

[0059] Figure 8 A schematic diagram showing the relative errors of blade reflectance and transmittance calculated for different angle combinations;

[0060] Figure 9The blade reflectance and transmittance are obtained by inversion under the preferred angle combination; wherein, the relative error of the blade reflectance obtained by inversion under the preferred angle combination is within... Between 10%;

[0061] Figure 10 A schematic diagram showing the variation of the absolute error of canopy reflectivity and the ratio of canopy multiple scattering in the Chinese GOST2 simulation with illumination-observation geometry for each band. Detailed Implementation

[0062] 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 embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0064] Example 1:

[0065] like Figure 1 As shown, this embodiment of the invention provides a method for simultaneously inverting the reflectivity and transmittance of a blade, comprising:

[0066] Step S1: Decompose the canopy reflectivity within the passive optical remote sensing band range into primary scattering and multiple scattering;

[0067] Step S2: Construct a first relationship model between primary scattering and blade reflectivity, and a second relationship model between multiple scattering and blade scattering coefficient;

[0068] Step S3: Based on the first relational model and the first relational model, an inversion model is obtained, wherein the inversion model expresses the canopy reflectivity as a binary function of leaf reflectivity and leaf transmittance;

[0069] Step S4: Input the multi-angle canopy reflectance observations into the inversion model to obtain the inversion results of leaf reflectance and leaf transmittance.

[0070] In one embodiment of the present invention, in step S1, assuming that the structure and background information of a canopy are known, its reflectance is only affected by leaf characteristics (i.e., leaf reflectance and leaf transmittance) and the geometry of light observation, expressed as follows: In the wavelength range of 400nm–2500nm, which is of interest in passive optical remote sensing, canopy reflectivity consists of two parts: primary scattering (…). ) and multiple scattering ( ),Right now:

[0071]

[0072] in, Indicates the band. Represents illumination-observation geometry.

[0073] In one embodiment of the present invention, in step S2, single scattering refers to the portion of sunlight that is reflected once by the leaves after reaching the canopy and is then received by the sensor; therefore, it is directly related to the leaf reflectivity. Multiple scattering, on the other hand, refers to sunlight that is scattered once by the leaves but not directly received by the sensor; instead, it propagates to other leaves for further scattering before finally being received by the sensor. This portion of radiation is related to the leaf's scattering coefficient (…). The scattering coefficient of a leaf (i.e., the sum of its reflectance and transmissivity) is closely related to its properties. The first relationship model between primary scattering and leaf reflectance, and the second relationship model between primary scattering and leaf scattering coefficient, are as follows:

[0074]

[0075] in, Indicates in band Leaf reflectance at nm; This represents a linear transformation model from leaf reflectance to canopy primary scattering, which is determined by the probability of seeing sun-exposed leaves. The nonlinear transformation model representing the relationship between leaf scattering factors and canopy multiple scattering is closely related to canopy structure and background complexity. , It can be determined using a geometric optics model or a machine learning model; This represents the leaf scattering factor, and its value is:

[0076]

[0077] As one embodiment of the present invention, in step S3, combining formulas (5)-(8), the inversion model is: that is:

[0078]

[0079] Where G represents the relationship model between canopy reflectivity and leaf reflectivity and transmittance.

[0080] If the form of G can be determined, a system of two equations can be established using canopy reflectance observations from two angles to solve for the angle-invariant leaf reflectance and transmittance. The form of G is determined through an inversion model. The specific steps for determining G are as follows: First, input the canopy structure, background information, and illumination observation geometry into the inversion model. Then, input a set of simulated leaf reflectance and transmittance data into the inversion model to obtain canopy primary and multiple scattering. Regressing the simulated leaf reflectance and transmittance against the simulated canopy primary and multiple scattering yields the initial form of G. and Finally, we get G.

[0081] Experimental Example 1:

[0082] The field sampling site was Saihanba National Forest Park in Chengde, Hebei Province, in August 2018. The data obtained included UAV canopy reflectivity data from eight angles (Table 1), canopy structure parameters, background reflectivity, slope gradient, and leaf reflectivity and transmittance (Table 2).

[0083] Table 1

[0084]

[0085] 1. Angle between the sun and the sensor; 2. Solar azimuth angle; 3. Solar zenith angle; 4. Observation azimuth angle; 5. Observation zenith angle

[0086] Table 2

[0087]

[0088]

[0089] This invention utilizes the geometric optics model GOST2 to simultaneously invert the reflectivity and transmittance of the blades, such as... Figure 2 As shown, the specific steps are as follows:

[0090] 1) Use GOST to create a lookup table

[0091] First, a simulated leaf spectral dataset was established using the PROSPECT-D leaf radiative transfer model, with parameter settings as shown in Table 2. A total of 1250 leaf reflectance and transmittance curves were simulated. Measured canopy structure parameters (leaf area index, aggregation index, canopy radius, diameter at breast height, tree height, tree distribution, tree density, number of quadrats, trunk height, canopy height, and leaf projected area), background information (background reflectance), topographic data (slope and aspect), and illumination-observation geometry (solar altitude angle and azimuth angle, observation altitude angle and azimuth angle) acquired by UAV were input into the GOST2 model (parameters detailed in Table 2). Then, the leaf reflectance and transmittance simulated by PROSPECT-D were individually substituted into the GOST2 model to obtain canopy reflectance, canopy primary scattering, and canopy multiple scattering under various illumination-observation geometries. Regression analysis was performed between the simulated canopy primary scattering and multiple scattering and the simulated leaf reflectance and transmittance by PROSPECT-D to obtain the relationship between canopy primary scattering and leaf reflectance, and between multiple scattering and leaf scattering factor under various illumination-observation geometries. The final relationships between the various bands are shown in Figures 1 and 2. This can be represented as:

[0092]

[0093] in, , This represents the conversion coefficient from primary scattering by the canopy to the reflectance of the leaves. , , This represents the conversion coefficient from canopy multiple scattering to leaf scattering coefficient. These coefficients vary with waveband and angle, and can be obtained by establishing a lookup table using a geometric optics model or a machine learning model. From formulas (10) and (11), the relationship between leaf reflectivity and canopy primary scattering, and between leaf scattering factor and canopy multiple scattering, can be obtained:

[0094]

[0095] The final relationship between canopy reflectance, leaf reflectance, and transmittance is as follows:

[0096]

[0097] In the above relationship, and These are variables that need to be estimated from measured canopy reflectivity; they do not change with illumination-observation geometry, while other variables do change with illumination-observation geometry. Therefore, with just two observations from different angles, the solution can be obtained. and ,Right now

[0098]

[0099] In the near-infrared band, the leaf scattering factor Since the first term on the right side of equation (11) is relatively small, it can be ignored. Therefore, it can be simplified to...

[0100]

[0101] Accordingly, formula (14) can be simplified to

[0102]

[0103] In the actual process of establishing the lookup tables, regressions were performed in different bands to create two tables: one for canopy primary scattering and the other for multiple scattering coefficients. The specific method is as follows: First, the measured canopy structure parameters, background reflectance, topography, and illumination observation geometry were input into GOST2. In this embodiment, these parameters are all measured data, and corresponding remote sensing products are available for future global mapping. After completing this step, the simulated 1250 leaf reflectance and transmittance spectra were input into the GOST2 model to simulate 1250 canopy reflectance spectra, 1250 canopy primary scattering spectra, and 1250 canopy multiple scattering spectra. Linear regression analysis was performed band-by-band on canopy primary scattering and leaf reflectance to obtain the slope and intercept for each band. Simultaneously, regression was performed band-by-band on the logarithmic form of canopy multiple scattering and the leaf scattering factor. For the near-infrared band, linear regression was performed to obtain the slope and intercept for each band; for other bands, quadratic regression was performed to obtain three coefficients. Regression analysis always uses the least squares method to obtain regression coefficients.

[0104] 2) Synchronous inversion and model error handling

[0105] After establishing the lookup table, the relationships between canopy primary and secondary scattering with respect to leaf reflectance and transmittance are determined using specific wavelengths and illumination observation geometry. Adding these two relationships together yields the relationships between canopy reflectance, leaf reflectance, and transmittance under specific wavelengths and illumination observation geometry. Two different sets of binary equations, namely formulas (15) and (16), can be obtained using two different illumination observation geometries. Finally, the leaf reflectance and transmittance are simultaneously inverted.

[0106] In the above implementations, it is assumed that the inversion model is accurate and error-free. Under these circumstances, the established relationship between canopy reflectivity, leaf reflectivity, and transmittance (i.e., formula (14)) is correct. Substituting the observed canopy reflectivity at the two angles into formulas (15) and (16) will yield the correct leaf reflectivity and transmittance. However, in reality, the model will inevitably contain errors. If the error between the canopy reflectivity simulated by the inversion model and the measured canopy reflectivity is expressed as... And taking this into account in the inversion model, equations (15) and (16) can be written as:

[0107]

[0108] In this situation, relying solely on solving the equations for the two formulas mentioned above would result in very large errors, or even no solution at all. This embodiment of the invention defines an objective equation for optimization. This equation is the sum of squares of the errors in the inversion model, i.e.

[0109]

[0110] The reflectance and transmittance of the blades were obtained by optimizing the above equations. The initial values ​​were set to (0.49, 0.49) in the near-infrared band and (0.2, 0.2) in the visible light band. The constraints are as follows: as well as < Optimization can be achieved using fmincon in MATLAB.

[0111] 3) Accuracy Analysis

[0112] The accuracy of the inversion results is evaluated using relative error, which is calculated using the formula:

[0113]

[0114] Where x' and x represent the simulated value and the measured value, respectively.

[0115] In some cases, such as evaluating the accuracy of the GOST2 model or assessing the impact of model errors on inversion results, the relative absolute value of error (ARE) and absolute error (AE) are used. Their calculation formulas are as follows:

[0116]

[0117] The technical effects of Experiment Example 1 are as follows:

[0118] 1. The simulation accuracy of GOST2

[0119] The GOST2 geometric optics model used in this embodiment of the invention demonstrates excellent simulation results, providing a guarantee for the multi-angle synchronous inversion of leaf reflectance and transmittance proposed in this invention. The band-by-band simulation results of GOST2 show that the simulated canopy reflectance and the canopy reflectance measured by the UAV are highly consistent. In the figure, the scatter plot of canopy reflectance simulated by GOST2 and the data observed by the UAV intersect at a point on the third or fourth angle to the right of the hot spot. This indicates that the model underestimates the reflectance in the direction of the hot spot, while overestimating it in the direction away from the hot spot. This error may be due to the model not fully considering the influence of non-photochemical components such as the tree trunk. The relative error statistics for each angle show that the error is smallest in the third and fourth illumination-observation geometries (e.g., ...). Figure 6 (As shown). The angles with the largest errors are the 6th, 7th, and 8th illumination-observation geometries. At these observation angles, the absolute value of the relative error in the visible light band is close to or even exceeds 30%. If these angles are used to invert the reflectivity and transmittance of the blade, there will be a large error.

[0120] 2. Trends in the proportion of primary and multiple scattering from the canopy with different wavelengths

[0121] The figure illustrates the proportion of primary scattering in canopy reflectivity under different illumination-observation geometries. Generally, in the visible light band (412 nm–677 nm), primary scattering accounts for a relatively large proportion, especially in the strong absorption bands of 412 nm, 442 nm, 466 nm, 645 nm, 666 nm, and 677 nm, where the proportion exceeds 0.9. In the green peaks (530 nm, 547 nm, and 554 nm), although the proportion of primary scattering in canopy reflectivity is relatively weak and fluctuates significantly, it still exceeds 0.7. In contrast, in the near-infrared band (747 nm–936 nm), the proportions of primary and multiple scattering are roughly equal, but their proportions vary with angle; the proportion of primary scattering is higher in hot spots and lower in directions away from the hot spots.

[0122] The figure also plots the coefficients of determination R for canopy reflectance and leaf reflectance. 2 The results showed that the correlation trend between canopy reflectance and leaf reflectance was in very good agreement with the trend of primary scattering proportion. R 2 The highest spectral density is found in the blue and red valleys, indicating that primary scattering accounts for a very large proportion in these bands, while the effect of multiple scattering is negligible. In the green peak, the proportion of multiple scattering increases, thus weakening the correlation between canopy reflectance and leaf reflectance, but it remains greater than 0.9. In the near-infrared band, the correlation between the two weakens sharply and remains stable throughout the entire near-infrared band; in these bands, the effect of multiple scattering is comparable to that of primary scattering and cannot be ignored.

[0123] 3. Synchronous inversion results

[0124] The figure shows the relative errors of leaf reflectance and transmittance obtained from simultaneous inversion of canopy reflectance under different angle combinations. In the near-infrared band, all angle combinations can simultaneously retrieve relatively accurate leaf reflectance and transmittance, with relative errors all within the range of [missing information]. Within 10%. However, in the visible light band, the relative error of leaf reflectance is significantly smaller than that of transmittance. Leaf transmittance inversion in the visible light band involves significant uncertainty, especially when using models to simulate observation angles with large errors. In strong absorption bands such as the blue valley (412 nm-487 nm) and red valley (645 nm-677 nm), the relative error of transmittance inverted by the algorithm of this invention is very large, often exceeding 100%. However, the transmittance in these bands is inherently very weak, essentially close to zero, therefore the absolute error is very small (as shown in the figure).

[0125] 4. The Influence of Model Error

[0126] According to formula (9), the inversion of leaf reflectance depends on canopy multiple scattering. To investigate how model accuracy affects the inversion of leaf transmittance, this embodiment of the invention analyzes the relationship between the absolute error of canopy reflectance simulated by the GOST2 model and canopy multiple scattering. Figure 1 shows the ratio of the model's absolute error to canopy multiple scattering as a function of illumination-observation geometry. In the visible light band, the model's simulation error far exceeds that of canopy multiple scattering, with the ratio of the model's absolute error to canopy multiple scattering exceeding 40. Only at certain observation angles and in certain bands can the model's absolute error be comparable to that of canopy multiple scattering, such as in the green peak band (530 nm, 547 nm, 554 nm) under the third illumination-observation geometry. However, in the near-infrared band (747 nm – 936 nm), the influence of model error weakens rapidly, not exceeding half of that of canopy multiple scattering. This is why the accuracy of leaf transmittance inverted in the near-infrared band is higher.

[0127] Example 2:

[0128] This invention also provides a device for simultaneously inverting the reflectivity and transmittance of blades, comprising:

[0129] The splitting module is used to split the canopy reflectivity within the passive optical remote sensing band into primary scattering and multiple scattering;

[0130] The first building module is used to build a first relationship model between primary scattering and blade reflectivity and a second relationship model between multiple scattering and blade scattering coefficient;

[0131] The second construction module is used to obtain an inversion model based on the first relational model and the first relational model. The inversion model represents the canopy reflectivity as a binary function of leaf reflectivity and leaf transmittance.

[0132] The inversion module is used to input multi-angle canopy reflectance observations into the inversion model to obtain the inversion results of leaf reflectance and leaf transmittance.

[0133] As one embodiment of the present invention, the passive optical remote sensing band range is 400nm–2500nm.

[0134] As one embodiment of the present invention, the canopy reflectivity From primary scattering and multiple scattering Composition, namely:

[0135]

[0136] in, Indicates the band. Represents illumination-observation geometry.

[0137] As one embodiment of the present invention, the first relationship model between primary scattering and blade reflectivity, and the second relationship model between multiple scattering and blade scattering coefficient are respectively:

[0138]

[0139]

[0140] in, Indicates in band Leaf reflectance at nm A linear transformation model representing the relationship between leaf reflectivity and canopy primary scattering. A nonlinear transformation model representing the relationship between leaf scattering factor and canopy multiple scattering. The leaf scattering factor is:

[0141]

[0142] in, Indicates in band Leaf transmittance at nm.

[0143] As one embodiment of the present invention, the inversion model is as follows:

[0144]

[0145] in, .

[0146] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for simultaneously inverting the reflectivity and transmittance of blades, characterized in that, include: Step S1: Decompose the canopy reflectivity within the passive optical remote sensing band range into primary scattering and multiple scattering; Step S2: Construct a first relationship model between primary scattering and leaf reflectivity, and a second relationship model between multiple scattering and leaf scattering factor; Step S3: Based on the first relational model and the second relational model, an inversion model is obtained, in which the canopy reflectivity is expressed as a binary function of leaf reflectivity and leaf transmittance. Step S4: Input the multi-angle canopy reflectance observations into the inversion model to obtain the inversion results of leaf reflectance and leaf transmittance; The canopy reflectivity From primary scattering and multiple scattering Composition, namely: in, Indicates the band. Representing illumination-observation geometry; The first relationship model between primary scattering and blade reflectivity, and the second relationship model between multiple scattering and blade scattering factor are respectively: in, Indicates in band Leaf reflectance at nm A linear transformation model representing the relationship between leaf reflectivity and canopy primary scattering. A nonlinear transformation model representing the relationship between leaf scattering factor and canopy multiple scattering. The leaf scattering factor is: in, Indicates in band Leaf transmittance at nm; The measured leaf area index, aggregation index, canopy radius, diameter at breast height (DBH), tree height, tree distribution, tree density, number of quadrats, trunk height, canopy height, leaf projected area, background reflectance, slope and aspect, as well as the solar altitude angle and azimuth angle, observation altitude angle and azimuth angle obtained by UAV, were input into the GOST2 model. Then, the leaf reflectance and transmittance simulated by PROSPECT-D were successively substituted into the GOST2 model to obtain the canopy reflectance, canopy primary scattering, and canopy multiple scattering under various illumination-observation geometry. The simulated canopy primary scattering and multiple scattering were then subjected to regression analysis with the leaf reflectance and transmittance simulated by PROSPECT-D to obtain the relationship between canopy primary scattering and leaf reflectance, and multiple scattering and leaf scattering factor under various illumination-observation geometry. The relationship between leaf reflectivity and canopy primary scattering, and between leaf scattering factor and canopy multiple scattering: 。 2. The method for simultaneously inverting blade reflectivity and transmittance as described in claim 1, characterized in that, The passive optical remote sensing band range is 400nm–2500nm.

3. The method for simultaneously inverting blade reflectivity and transmittance as described in claim 2, characterized in that, The inversion model is as follows: in, A model representing the relationship between canopy reflectivity and leaf reflectivity and transmittance.

4. A device for simultaneously inverting the reflectivity and transmittance of blades, implementing the method for simultaneously inverting the reflectivity and transmittance of blades as described in claim 1, characterized in that, include: The splitting module is used to split the canopy reflectivity within the passive optical remote sensing band into primary scattering and multiple scattering; The first building module is used to build a first relationship model between primary scattering and leaf reflectivity and a second relationship model between multiple scattering and leaf scattering factor. The second construction module is used to obtain an inversion model based on the first relational model and the second relational model. The inversion model expresses the canopy reflectivity as a binary function of leaf reflectivity and leaf transmissivity. The inversion module is used to input multi-angle canopy reflectance observations into the inversion model to obtain the inversion results of leaf reflectance and leaf transmittance.

5. The device for simultaneously inverting blade reflectivity and transmittance as described in claim 4, characterized in that, The passive optical remote sensing band range is 400nm–2500nm.

6. The device for simultaneously inverting blade reflectivity and transmittance as described in claim 5, characterized in that, The inversion model is as follows: in, .

Images