Leaf inclination angle distribution inversion method and device, electronic equipment and storage medium

Abstract

The invention relates to the technical field of remote sensing data processing, and provides a leaf dip angle distribution inversion method and device, electronic equipment and a storage medium. Comprising the following steps: acquiring an actual relative brightness histogram of data to be inverted; wherein the to-be-inverted data is a multispectral image of the leaf; establishing a lookup table according to the sun angle and the observation angle; based on the lookup table, obtaining a theoretical relative brightness histogram of the to-be-inverted data; and performing inversion based on the actual relative brightness histogram and the theoretical relative brightness histogram, and obtaining a leaf dip angle distribution parameter of the data to be inverted. According to the method, more accurate large-range high-frequency near-surface blade inclination angle distribution data can be obtained.

Description

Technical Field

[0001] The invention relates to the technical field of remote sensing data processing, and in particular to a leaf pitch angle distribution inversion method, device, electronic equipment and storage medium. Background Art

[0002] Leaf Angle Distribution (LAD) is defined as the distribution of the angle between the normal direction of the leaf surface and the vertical direction. Leaf Angle Distribution is an important parameter to describe the structure of vegetation canopy, which has a significant impact on the transmission process of light in the vegetation canopy and the distribution of photosynthetically active radiation.

[0003] In the related art, the measurement methods of blade inclination angle are mainly divided into direct measurement method and indirect inversion method. The direct measurement method includes contact and non-contact measurement. The direct measurement method measures the blade angle based on the geometric structure of the blade. The indirect inversion method starts with other variables related to the blade inclination angle distribution (such as gap ratio and four-component ratio), and obtains the blade inclination angle distribution by inverting the relationship between variables. However, the direct measurement method and the indirect inversion method have the following problems: the direct measurement method has many restrictions during measurement, and due to manual measurement, the measurement results are not accurate enough and the measurement efficiency is low.

[0004] Therefore, how to improve the measurement accuracy and efficiency of blade inclination angle distribution data is a technical problem that needs to be solved urgently. Summary of the invention

[0005] In order to solve the above technical problem or at least partially solve the above technical problem, the present invention provides a blade pitch angle distribution inversion method, device, electronic device and storage medium.

[0006] In a first aspect, the present application provides a method for inverting blade tilt angle distribution, the method comprising:

[0007] Acquire an actual relative brightness histogram of the data to be inverted; the data to be inverted is a multispectral image of a leaf;

[0008] A lookup table is established according to the sun angle and the observation angle; the lookup table includes: a leaf brightness histogram corresponding to different leaf inclination angle distribution parameter values ​​under multiple sun angles;

[0009] Based on the lookup table, obtaining a theoretical relative brightness histogram of the data to be inverted;

[0010] Inversion is performed based on the actual relative brightness histogram and the theoretical relative brightness histogram to obtain leaf inclination angle distribution parameters of the data to be inverted.

[0011] Optionally, establishing a lookup table according to the sun angle and the observation angle includes:

[0012] At a preset observation angle, obtaining the sun angle at different times; the sun angle includes the sun zenith angle and the sun azimuth angle;

[0013] For the sun angles at different times, based on the dual-parameter Beta distribution, multiple preset leaf inclination angle distribution parameter combinations are used in turn to traverse and determine the leaf brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination; the preset leaf inclination angle distribution parameter combination includes: a first preset leaf inclination angle distribution parameter and a second preset leaf inclination angle distribution parameter;

[0014] A lookup table is established according to the sun angle at different times, a plurality of preset leaf inclination angle distribution parameter combinations, and the blade brightness histograms corresponding to each preset leaf inclination angle distribution parameter combination.

[0015] Optionally, obtaining a theoretical relative brightness histogram of the data to be inverted based on the lookup table includes:

[0016] Based on the target leaf inclination angle distribution parameter combination and the target sun angle, a theoretical relative brightness histogram of the data to be inverted is obtained in the lookup table.

[0017] Optionally, for the sun angles at different times, based on the two-parameter Beta distribution, traversing with multiple preset leaf inclination angle distribution parameter combinations in sequence to determine the leaf brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination includes:

[0018] Calculate the relative brightness of the leaves based on the leaf angle and the sun angle:

[0019] B i =|Ω s ·Ω l |

[0020]

[0021] Among them, B i represents the relative brightness of the i-th group of the histogram, the subscript s represents the sun, the subscript l represents the leaf, Ω represents the unit vector in the corresponding direction, and θ is They represent the zenith angle and azimuth angle of the vector direction in the spherical coordinate system respectively;

[0022] Get the leaf frequency corresponding to the relative brightness of the leaf:

[0023]

[0024] Among them, P(B i ) represents the relative brightness B i The blade frequency, g(θ l) represents the blade inclination angle distribution function, and the solid angle range ΔΩ corresponding to the blade normal i The corresponding zenith angle and azimuth angle ranges are Δθ i and Ω l The unit vector representing the normal direction of the blade, Ω v The unit vector representing the observation direction, |Ω l ·Ω v | indicates Ω l The unit area of ​​the blade in the direction of observation Ω v The projected area of

[0025] A plurality of preset leaf inclination angle distribution parameter combinations are traversed in sequence to determine the leaf brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination.

[0026] Optionally, the method further includes:

[0027] According to the ratio of the leaf reflectivity to the leaf transmittance, the relative brightness of the leaf is corrected:

[0028] B i =|Ω s ·Ω l |·f(Ω s ,Ω l )

[0029] Among them, B i represents the relative brightness of the i-th group of the histogram, subscript s represents the sun, subscript l represents the leaves, Ω represents the unit vector in the corresponding direction, and f(Ω s ,Ω l ) represents the ratio of leaf transmittance to reflectance;

[0030]

[0031] Where k is the ratio of the leaf transmittance to reflectance.

[0032] Optionally, the performing inversion based on the actual relative brightness histogram and the theoretical relative brightness histogram to obtain leaf inclination angle distribution parameters includes:

[0033] The KL divergence between the actual relative brightness histogram and the theoretical relative brightness histogram is calculated according to the KL divergence formula:

[0034]

[0035] Among them, P u,v represents the theoretical relative brightness histogram when the leaf inclination distribution parameters are u, v, Q represents the actual relative brightness histogram extracted from the data to be inverted, and B irepresents the relative brightness of the i-th group of the histogram, P u,v (B i ) and Q(B i ) respectively represent the frequency of the i-th group in the corresponding histogram, and n represents the number of groups in the histogram;

[0036] According to the KL divergence between the actual relative brightness histogram and the theoretical relative brightness histogram, the inverted leaf tilt angle distribution parameters are obtained:

[0037] (u * ,v * ) = argmin u,v D k (P u,v ||Q)

[0038] Among them, u * ,v * represent the inverted first leaf tilt angle distribution parameters and second leaf tilt angle distribution parameters respectively.

[0039] Optionally, obtaining an actual relative brightness histogram of the data to be inverted includes:

[0040] Obtaining data to be inverted;

[0041] Based on a preset leaf segmentation algorithm, the multispectral image of the leaf is subjected to single leaf segmentation processing to obtain the leaf brightness of each leaf;

[0042] The leaf brightness of each leaf is normalized to obtain an actual leaf brightness histogram of the data to be inverted; the actual leaf brightness histogram is used to reflect the area ratio occupied by leaves of various relative brightness sizes.

[0043] In a second aspect, the present application provides a blade pitch angle distribution inversion device, the device comprising:

[0044] An acquisition module, used to acquire an actual relative brightness histogram of the data to be inverted; the data to be inverted is a multispectral image of a leaf;

[0045] Establishing module, used to establish a lookup table according to the sun angle and the observation angle; the lookup table includes: a leaf brightness histogram corresponding to different leaf inclination angle distribution parameter values ​​under multiple sun angles;

[0046] A search module, used for obtaining a theoretical relative brightness histogram of the data to be inverted based on the lookup table;

[0047] An inversion module is used to perform inversion based on the actual relative brightness histogram and the theoretical relative brightness histogram to obtain leaf inclination angle distribution parameters of the data to be inverted.

[0048] Optionally, the establishing module includes:

[0049] An acquisition unit, used to acquire the sun angle at different times at a preset observation angle; the sun angle includes the sun zenith angle and the sun azimuth angle;

[0050] A determination unit is used to traverse the sun angles at different times in sequence with a plurality of preset leaf inclination angle distribution parameter combinations based on a two-parameter Beta distribution to determine a blade brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination; the preset leaf inclination angle distribution parameter combination includes: a first preset leaf inclination angle distribution parameter and a second preset leaf inclination angle distribution parameter;

[0051] The establishing unit is used to establish a lookup table according to the sun angle at different times, a plurality of preset leaf inclination angle distribution parameter combinations, and the leaf brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination.

[0052] Optionally, the search module is specifically used to:

[0053] Based on the target leaf inclination angle distribution parameter combination and the target sun angle, a theoretical relative brightness histogram of the data to be inverted is obtained in the lookup table.

[0054] Optionally, the determining unit is specifically configured to:

[0055] Calculate the relative brightness of the leaves based on the leaf angle and the sun angle:

[0056] B i =|Ω s ·Ω l |

[0057]

[0058] Among them, B i represents the relative brightness of the i-th group of the histogram, the subscript s represents the sun, the subscript l represents the leaf, Ω represents the unit vector in the corresponding direction, and θ is They represent the zenith angle and azimuth angle of the vector direction in the spherical coordinate system respectively;

[0059] Get the leaf frequency corresponding to the relative brightness of the leaf:

[0060]

[0061] Among them, P(B i ) represents the relative brightness B i The blade frequency, g(θ l ) represents the blade inclination angle distribution function, and the solid angle range ΔΩ corresponding to the blade normal i The corresponding zenith angle and azimuth angle ranges are Δθi and Ω l The unit vector representing the normal direction of the blade, Ω v The unit vector representing the observation direction, |Ω l ·Ω v | indicates Ω l The unit area of ​​the blade in the direction of observation Ω v The projected area of

[0062] A plurality of preset leaf inclination angle distribution parameter combinations are traversed in sequence to determine the leaf brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination.

[0063] Optionally, the device further comprises:

[0064] The correction module is used to correct the relative brightness of the blade according to the ratio of the blade reflectivity to the blade transmittance:

[0065] B i =|Ω s ·Ω l |·f(Ω s ,Ω l )

[0066] Among them, B i represents the relative brightness of the i-th group of the histogram, subscript s represents the sun, subscript l represents the leaves, Ω represents the unit vector in the corresponding direction, and f(Ω s ,Ω l ) represents the ratio of leaf transmittance to reflectance;

[0067]

[0068] Where k is the ratio of the leaf transmittance to reflectance.

[0069] Optionally, the inversion module is specifically used for:

[0070] The KL divergence between the actual relative brightness histogram and the theoretical relative brightness histogram is calculated according to the KL divergence formula:

[0071]

[0072] Among them, P u,v represents the theoretical relative brightness histogram when the leaf inclination distribution parameters are u, v, Q represents the actual relative brightness histogram extracted from the data to be inverted, and B i represents the relative brightness of the i-th group of the histogram, P u,v (B i ) and Q(B i ) respectively represent the frequency of the i-th group in the corresponding histogram, and n represents the number of groups in the histogram;

[0073] According to the KL divergence between the actual relative brightness histogram and the theoretical relative brightness histogram, the inverted leaf tilt angle distribution parameters are obtained:

[0074] (u * ,v * ) = argmin u,v D k (P u,v ||Q)

[0075] Among them, u * ,v * represent the inverted first leaf tilt angle distribution parameters and second leaf tilt angle distribution parameters respectively.

[0076] Optionally, the acquisition module is specifically used to:

[0077] Obtain data to be inverted;

[0078] Based on a preset leaf segmentation algorithm, the multispectral image of the leaf is subjected to single leaf segmentation processing to obtain the leaf brightness of each leaf;

[0079] The leaf brightness of each leaf is normalized to obtain an actual leaf brightness histogram of the data to be inverted; the actual leaf brightness histogram is used to reflect the area ratio occupied by leaves of various relative brightness sizes.

[0080] In a third aspect, the present application provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the blade pitch angle distribution inversion method provided in the first aspect of the present application when executing the computer program.

[0081] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the blade pitch angle distribution inversion method provided in the first aspect of the present application are implemented.

[0082] The technical effects achieved by the technical solution provided by this application are as follows:

[0083] The leaf inclination angle distribution inversion method provided in the present application obtains the actual relative brightness histogram of the data to be inverted; wherein the data to be inverted is a multispectral image of the leaf; a lookup table is established according to the sun angle and the observation angle; based on the lookup table, a theoretical relative brightness histogram of the data to be inverted is obtained; inversion is performed based on the actual relative brightness histogram and the theoretical relative brightness histogram to obtain the leaf inclination angle distribution parameters of the data to be inverted. Since the multispectral image of the leaf contains rich brightness information, and the lookup table includes: the leaf brightness histogram corresponding to different leaf inclination angle distribution parameter values ​​under multiple sun angles, the leaf inclination angle distribution is obtained through the information of the leaf brightness histogram in the image, which can be used for sensors of near-ground platforms such as drones and phenological cameras, and can meet the needs of more accurate large-scale and high-frequency near-ground leaf inclination angle distribution data acquisition. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0085] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0086] Figure 1 A flow chart of a blade tilt angle distribution inversion method provided by an embodiment of the present invention;

[0087] Figure 2 A schematic diagram of canopy brightness distribution characteristics for a single blade angle and different blade angles provided by an embodiment of the present invention;

[0088] Figure 3 A schematic diagram of the structure of a blade pitch angle distribution inversion device provided by an embodiment of the present invention;

[0089] Figure 4 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0090] In order to more clearly understand the above-mentioned objectives, features and advantages of the present invention, the scheme of the present invention will be further described below. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict.

[0091] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present invention, rather than all of the embodiments.

[0092] In the related art, the measurement methods of blade inclination angle are mainly divided into direct measurement method and indirect inversion method. The direct measurement method includes contact measurement and non-contact measurement.

[0093] The contact measurement method is to measure the blade angle by contact using an instrument with recorded angle information. Traditional ground measurements use a compass and a protractor to measure the azimuth and inclination of the blades respectively. When measuring the inclination of the blades, the inclination of each leaf is recorded by attaching a protractor with a plumb line hanging from it to the surface of the leaf. When measuring longer and curved leaves, a single leaf may have multiple different inclination angles. During measurement, it is also necessary to record the size characteristics of the leaves at each inclination angle and obtain the leaf inclination angle distribution by area weighting. Although this method can accurately measure the leaf inclination angle, the number of measurement parameters is large and it requires contact with the vegetation. The operation is cumbersome and destructive. It is difficult to apply to large-scale measurements and its spatial representativeness is also limited.

[0094] The horizontal digital photography method manually selects the leaves whose orientation is perpendicular to the observation direction from the photo each time (that is, the leaves that appear to be nearly straight in length), and then extracts the inclination angles of these leaves through angle measurement software. This method initially required the instrument to be placed on a platform such as an observation tower with a pole, and was still limited when facing tall vegetation or far away from the platform. The method was then expanded to drone platforms, enhancing its potential for application in vast forest environments. However, this method requires a lot of human-computer interaction, which may result in different results for different operators, and it is still cumbersome when facing tall vegetation canopies. In addition, this method only uses nearly straight leaves for measurement, which results in a certain degree of data waste.

[0095] The three-dimensional point cloud reconstruction method can obtain the three-dimensional information of the target object through laser radar and photogrammetry technology to calculate the spatial angle. The laser radar calculates the corresponding normal vector by clustering, plane fitting or voxelizing the point cloud to obtain the blade angle. The photogrammetry method is based on the principle of geometric perspective and reconstructs the three-dimensional point cloud of the target from multiple two-dimensional photos with overlapping degrees. The leaf inclination angle distribution calculation method that obtains the three-dimensional information of the target by laser radar or photogrammetry has a high degree of automation, but is easily disturbed by wind and difficult to conduct long-term observations. When facing larger scenes such as forests, the time cost and calculation amount of data acquisition will also increase significantly. In other words, the three-dimensional reconstruction method requires time and economic cost to obtain high-precision data, and is more suitable for accurate measurements in a smaller range. When facing larger scenes, it is difficult to provide efficiency guarantees.

[0096] The gap ratio is the most typical variable related to leaf inclination distribution. The gap ratio refers to the proportion of gaps between plant leaves under certain light conditions, and is an indicator of the light transmittance of the plant canopy. The size of the gap ratio will affect the light distribution inside the plant canopy. A higher gap ratio means that more light can penetrate into the lower layer of the plant canopy, which helps to improve the photosynthesis efficiency of the entire canopy.

[0097] Canopy structure parameters can also be linked to the four-component ratio (illuminated soil, shaded soil, illuminated leaves, shaded leaves). Based on the bidirectional gap ratio model, an analytical relationship between vegetation structure parameters and the four-component ratio is established, and multi-angle photos are used to simultaneously optimize the inversion of leaf area index and leaf inclination angle distribution. The method based on gap ratio or four-component inversion involves multi-parameter coupling of the model (such as leaf area index, aggregation index, etc.), and it is difficult to accurately decouple the leaf inclination angle distribution from it.

[0098] In view of the above problems, an embodiment of the present invention provides a method for inverting blade tilt angle distribution.

[0099] like Figure 1 As shown, the leaf tilt angle distribution inversion method includes the following steps:

[0100] S11. Obtaining an actual relative brightness histogram of the data to be inverted.

[0101] The data to be inverted is a multispectral image of a leaf. A multispectral image of a leaf is an image containing information of multiple bands, and each band has a corresponding grayscale image, which contains scene brightness information generated by the sensitivity of the sensor of the band.

[0102] In some embodiments, the above step S11 (obtaining the actual relative brightness histogram of the data to be inverted) can be implemented as follows:

[0103] A. Obtain the data to be inverted.

[0104] Specifically, the data to be inverted can also be understood as remote sensing data, that is, using sensors from near-ground platforms such as drones and phenological cameras to obtain canopy leaf images. Remote sensing data refers to information about the earth's surface and its atmosphere obtained through remote sensing technology. Remote sensing technology uses sensors carried by platforms such as aircraft, satellites, and drones to observe and record the earth's surface from a distance. The plant canopy refers to the upper branch and leaf layer of a plant community, which includes the collection of all leaves, branches, and stems of trees, shrubs, herbaceous plants, or other vegetation types. The canopy is the main place for plants to exchange matter and energy with the atmosphere, and plays a vital role in ecosystem function and biogeochemical cycles.

[0105] B. Performing single leaf segmentation processing on the multispectral image of the leaf based on a preset leaf segmentation algorithm to obtain the leaf brightness of each leaf.

[0106] Among them, the preset segmentation algorithm can be but is not limited to a threshold segmentation algorithm, an edge detection algorithm, a region growing algorithm, a watershed algorithm, etc., and no specific limitation is made here.

[0107] C. Normalize the leaf brightness of each leaf to obtain an actual leaf brightness histogram of the data to be inverted.

[0108] The actual leaf brightness histogram is used to reflect the area ratio occupied by leaves of different relative brightness levels.

[0109] Specifically, a leaf image is obtained by shooting with a drone / camera, the leaf image is segmented based on a leaf segmentation algorithm, leaf pixels are extracted, and the leaf brightness is normalized to obtain an actual leaf brightness histogram.

[0110] S12. Create a lookup table based on the sun angle and the observation angle.

[0111] The lookup table includes: leaf brightness histograms corresponding to different leaf inclination angle distribution parameter values ​​under multiple sun angles. That is, it can be understood that the lookup table includes theoretical leaf brightness histograms under all possible leaf inclination angle distributions. The observation angle can be used as a constant.

[0112] Lookup table is an indexing technology that improves data retrieval speed through pre-calculated results and additional spatial resources. Compared with the more complex numerical optimization method, the lookup table is more efficient in finding the optimal matching value. Lookup table is widely used in vegetation remote sensing due to its simple and fast characteristics, and has better performance than other methods. Considering that the inversion model of this study is not complicated, the pre-calculation time and space required to establish the lookup table can be basically ignored, so this study also uses the inversion strategy of the lookup table.

[0113] In some embodiments, the above step S12 (establishing a search based on the sun angle and the observation angle) can be implemented as follows:

[0114] (1) Obtain the sun angle at different times at the preset observation angle.

[0115] The solar angle includes the solar zenith angle and the solar azimuth angle.

[0116] (2) For the sun angles at different times, based on the two-parameter Beta distribution, multiple preset leaf inclination angle distribution parameter combinations are traversed in turn to determine the leaf brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination.

[0117] The preset blade inclination angle distribution parameter combination includes: a first preset blade inclination angle distribution parameter and a second preset blade inclination angle distribution parameter.

[0118] Calculate the relative brightness of the leaves based on the leaf angle and the sun angle:

[0119] B i =|Ω s ·Ω l |

[0120]

[0121] Among them, B i represents the relative brightness of the i-th group of the histogram, the subscript s represents the sun, the subscript l represents the leaf, Ω represents the unit vector in the corresponding direction, and θ is They represent the zenith angle and azimuth angle of the vector direction in the spherical coordinate system respectively;

[0122] Get the leaf frequency corresponding to the relative brightness of the leaf:

[0123]

[0124] Among them, P(B i ) represents the relative brightness B i The blade frequency, g(θ l ) represents the blade inclination angle distribution function, and the solid angle range ΔΩ corresponding to the blade normal i The corresponding zenith angle and azimuth angle ranges are Δθ i and Ω l The unit vector representing the normal direction of the blade, Ω v The unit vector representing the observation direction, |Ω l ·Ω v | indicates Ω l The unit area of ​​the blade in the direction of observation Ω v The projected area of

[0125] A plurality of preset leaf inclination angle distribution parameter combinations are traversed in sequence to determine the leaf brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination.

[0126] Specifically, the brightness of the blade essentially reflects the amount of energy that is intercepted by the blade and scattered to the sensor. The energy emitted by the blade can be divided into single scattering and multiple scattering. The interaction mechanism between brightness and blade angle in the single scattering process is relatively simple and clear, and the contribution of multiple scattering in the visible light band is almost negligible.

[0127] Assuming that multiple scattering is not considered in the theoretical derivation, solar radiation is the most important energy source for the blades. The ability of the blades to intercept solar irradiance at different angles is closely related to the absolute value of the cosine of the angle between the sun direction and the blade direction, which can be specifically expressed as the following formula:

[0128]

[0129] Where, subscript s represents the sun, subscript l represents the leaf, Ω represents the unit vector in the corresponding direction, and θ is They represent the zenith angle and azimuth angle of the vector direction in the spherical coordinate system respectively.

[0130] By further assuming that the leaves of vegetation are Lambertian bodies, the process of the leaves scattering energy outward can be simplified, that is, the radiant brightness emitted by the leaves in all directions is the same. Then, when there is no difference between reflectivity and transmittance, the brightness of the leaves is proportional to the cosine of the angle between the leaves and the direction of the sun. When the plane of the leaves is perpendicular to the direction of the sun's incidence, the direct solar flux density received by the leaves is the largest (cosine is 1). When the plane of the leaves is parallel to the direction of the sun's incidence, the leaves do not receive any direct solar energy (cosine is 0). Among them, the Lambertian body refers to the phenomenon that when the incident energy is uniformly reflected in all directions, that is, the incident energy isotropically reflects energy in all directions in the entire hemispherical space with the incident point as the center, which is called diffuse reflection, also known as isotropic reflection. A complete diffuser is called a Lambertian body.

[0131] Based on this, the absolute value of the cosine in the above formula (1) is actually the relative brightness of the leaf (normalized to 0-1), that is, the relative brightness of the leaf B i It can be expressed as:

[0132] B i =|Ω s ·Ω l |#(2)

[0133] For example, refer to Figure 2 As shown, Figure 2 Schematic diagram of canopy brightness distribution characteristics for a single blade angle and different blade angles. Figure (a) and Figure (b) are respectively a canopy with a single blade angle (60° in the example in the figure) and a canopy with different blade angles, θ represents the solar zenith angle (30° in the example in the figure), and Figure (c) and Figure (d) are corresponding schematic diagrams of the distribution of bidirectional reflection factors of the blades. When all the blades in the canopy are in one direction, there will be no change in the distribution of blade brightness. It is precisely because the directions of the blades in the canopy are different that the blades in the canopy will show different brightness in different directions.

[0134] The relative brightness of the leaves in the canopy is related to the corresponding angle, so the relative brightness distribution can be calculated by the leaf inclination distribution. The relative brightness histogram of the canopy leaves in the sensor field of view reflects the area ratio of leaves of different relative brightness. The relative brightness is B i The frequency of the blade P(B i ) is related to the ratio of the number of relative brightness blades and the projection area of ​​the relative brightness blades in the observation direction.

[0135] Relative brightness is in B i The proportion of leaves in the histogram cell centered on the total number of leaves can be expressed as Where ΔΩ i Indicates the relative brightness in [B i -Δ i ,B i +Δ i ](B i The solid angle range corresponding to the blade normal of the histogram interval) is 2Δ. i ,Ω l The unit area of ​​the blade in the direction of observation Ω v The projected area is |Ω l ·Ω v |, therefore, the relative brightness B i The blade frequency P(B i ) can be expressed as:

[0136]

[0137] Among them, the solid angle range ΔΩ corresponding to the blade normal i The corresponding zenith angle and azimuth angle ranges are Δθ i and For any θ l ∈Δθ i , According to formula (3), we have:

[0138]

[0139] Therefore, expressing the solid angle in formula (3) in the form of corresponding zenith angle and azimuth angle, we can obtain:

[0140]

[0141] Among them, P(B i ) represents the relative brightness B i The blade frequency, g(θ l ) represents the blade inclination angle distribution function, and the solid angle range ΔΩ corresponding to the blade normal i The corresponding zenith angle and azimuth angle ranges are Δθi and Ω l The unit vector representing the normal direction of the blade, Ω v The unit vector representing the observation direction, |Ω l ·Ω v | indicates Ω l The unit area of ​​the blade in the direction of observation Ω v The projected area of ​​.

[0142] For any small interval in the brightness histogram, the histogram frequency corresponding to the interval can be obtained by formulas (1)-(3). In actual calculation, for each B i ,Will From 0 to 2π, θ l Traverse from 0 to π / 2 and find all the and θ l As Δθ i and By numerically integrating formula (5), the relative brightness histogram distribution of the leaves can be calculated by giving the leaf inclination angle distribution and the solar incident angle.

[0143] There are many forms of expression for leaf inclination angle distribution, among which the most influential are the six typical leaf inclination angle distribution forms given by De Wit, in addition to the β distribution, ellipsoid distribution and other forms. Although the six typical leaf inclination angle distributions are widely used, it is difficult to cover all leaf inclination angle distributions that exist in nature. The β distribution is a distribution function that can represent any probability distribution between 0 and 1. When applied to the leaf inclination angle distribution, it can be expressed as follows:

[0144]

[0145] Among them, g(θ l ,u,v) is the leaf tilt angle distribution function, θ l is the blade inclination angle, u and v are respectively the first preset blade inclination angle parameter and the second preset blade inclination angle parameter of the β distribution, and u and v jointly control the shape of the distribution.

[0146] The β distribution can help to more precisely characterize the leaf inclination angle distribution of vegetation and characterize a wider range of leaf inclination angle distribution types. Therefore, this study selected the β distribution as the inverted leaf inclination angle distribution form. The β distribution parameters and statistics corresponding to the six typical leaf inclination angle distributions are shown in Table 1.

[0147] Table 1

[0148] Leaf inclination angle distribution type u v Mean(°) variance Flat type 2.770 1.172 26.76 342.4 Vertical 1.172 2.770 63.24 342.4 Ball Type 1.101 1.930 57.3 464.8 Uniform 1 1 45 675 Inclined 3.326 3.326 45 264.6 Extreme 0.433 0.433 45 1085.2

[0149] In some embodiments, the relative brightness of the blade is corrected according to the ratio of the blade reflectivity to the blade transmittance:

[0150] B i =|Ω s ·Ω l |·f(Ω s ,Ω l )

[0151] Among them, B i represents the relative brightness of the i-th group of the histogram, subscript s represents the sun, subscript l represents the leaves, Ω represents the unit vector in the corresponding direction, and f(Ω s ,Ω l ) represents the ratio of leaf transmittance to reflectance;

[0152]

[0153] Where k is the ratio of the leaf transmittance to reflectance.

[0154] It should be noted that the above derivation is obtained under the assumption that there is no difference in reflectivity and transmittance of the leaves. In fact, there is a difference in reflectivity and transmittance of the leaves. Therefore, the relative brightness formula (2) also needs to introduce the ratio of the reflectivity and transmittance of the leaves for correction, as shown in formulas (7) and (8).

[0155] B i =|Ω s ·Ω l |·f(Ω s ,Ω l )#(7)

[0156]

[0157] Among them, the sun direction refers to the direction corresponding to the vector pointing to the sun. If the product of the cosine of the angle between the sun direction and the blade direction and the cosine of the angle between the observation direction and the blade direction is greater than 0, it means that the incident direction and the observation direction are on the same side of the blade plane, that is, the field of view receives the energy reflected by the blade. If the product of the cosine of the angle between the sun direction and the blade direction and the cosine of the angle between the observation direction and the blade direction is less than or equal to 0, it means that the incident direction and the observation direction are on different sides of the blade plane, that is, the field of view receives the energy projected by the blade, and k represents the ratio of the blade transmittance to the reflectance. Without loss of generality, it is assumed that the blade transmittance is less than the reflectance, that is, the value range of k is 0 to 1. If the blade transmittance is greater than the reflectance, 1 and k in formula (8) can be exchanged.

[0158] (3) A lookup table is established based on the sun angle at different times, a plurality of preset leaf inclination angle distribution parameter combinations, and the leaf brightness histograms corresponding to each preset leaf inclination angle distribution parameter combination.

[0159] Specifically, since the preset leaf tilt angle distribution parameter combination includes a first preset leaf tilt angle distribution parameter u and a second preset leaf tilt angle distribution parameter v, and the value range of u and v can be 0.5-3.5 (traversed every 0.1), it is equivalent to establishing a 35*35 table, and each u and v represents a probability density curve of a certain shape. Among them, u and v are related to the leaf tilt angle probability density distribution function. The β distribution can characterize a probability density distribution curve of any shape by controlling the two parameters u and v, and a lookup table can be established to obtain the probability density distribution function (relative brightness histogram) of all possible leaf tilt angle distributions.

[0160] S13. Based on the lookup table, obtain a theoretical relative brightness histogram of the data to be inverted.

[0161] In some embodiments, the above step S13 (obtaining the theoretical relative brightness histogram of the data to be inverted based on the lookup table) can be implemented as follows:

[0162] Based on the target leaf inclination angle distribution parameter combination and the target sun angle, a theoretical relative brightness histogram of the data to be inverted is obtained in the lookup table.

[0163] S14. Perform inversion based on the actual relative brightness histogram and the theoretical relative brightness histogram to obtain leaf inclination angle distribution parameters of the data to be inverted.

[0164] In some embodiments, the above step S14 (performing inversion based on the actual relative brightness histogram and the theoretical relative brightness histogram to obtain the leaf inclination angle distribution parameters of the data to be inverted) can be implemented as follows:

[0165] a. Calculate the KL divergence between the actual relative brightness histogram and the theoretical relative brightness histogram according to the KL divergence formula:

[0166]

[0167] Among them, P u,v represents the theoretical relative brightness histogram when the leaf inclination distribution parameters are u, v, Q represents the actual relative brightness histogram extracted from the data to be inverted, and B i represents the relative brightness of the i-th group of the histogram, P u,v (B i ) and Q(B i ) respectively represent the frequencies of the i-th group in the corresponding histogram, and n represents the number of groups in the histogram.

[0168] Specifically, the inversion method based on the lookup table needs to find u and v in the lookup table that make the measured brightness distribution and the predicted brightness distribution most similar, that is, the characteristic index used for inversion is the distribution histogram of relative brightness. There are many ways to measure the similarity between two distributions. Here, Kullback-Leibler divergence (hereinafter referred to as KL divergence) is used as an indicator of the similarity between different histograms. KL divergence is a method to measure the relative entropy of two probability distributions. It measures the degree of information loss when one distribution is incorrectly interpreted as another distribution. For discrete random variables, its mathematical expression is:

[0169]

[0170] In the formula, P and Q are two probability distributions defined on the same sample space χ, and P(x) is the probability of observing event x in P. Q(x) is the probability of observing event x in Q, and log usually represents the logarithm with the natural constant e as the base. D_k(P||Q) represents the KL divergence from P to Q. Substituting the relative brightness histogram predicted by the model and extracted from the data to be inverted into the formula, the loss function used in this study is:

[0171]

[0172] Among them, P u,v represents the theoretical relative brightness histogram when the leaf inclination distribution parameters are u, v, Q represents the actual relative brightness histogram extracted from the data to be inverted, and B i represents the relative brightness of the i-th group of the histogram, P u,v (B i ) and Q(B i ) respectively represent the frequencies of the i-th group in the corresponding histogram, and n represents the number of groups in the histogram.

[0173] b. According to the KL divergence between the actual relative brightness histogram and the theoretical relative brightness histogram, the inverted leaf tilt angle distribution parameters are obtained:

[0174] (u * ,v * ) = argmin u,v D k (P u,v ||Q)

[0175] Among them, u * ,v * represent the inverted first leaf tilt angle distribution parameters and second leaf tilt angle distribution parameters respectively.

[0176] When the difference between the two distributions is greater, the KL divergence is greater, so the matching target can be expressed as:

[0177] (u * ,v * ) = argmin u,v D k (P u,v ||Q)#(11)

[0178] In the formula, u * ,v * These are the parameters of the inverted leaf tilt angle distribution.

[0179] Generally speaking, it is assumed that the leaf tilt angle distribution of the canopy will not change much at different times of the day. Therefore, by increasing the sun angle and introducing multi-phase data of illumination changes, more information can be provided for the inversion of leaf tilt angle distribution parameters. Therefore, the sun angle can be introduced when establishing the lookup table, and the law of the plane radiation flux changing with the sun during the day can be used to provide more information for the inversion of leaf tilt angle distribution parameters.

[0180] This method obtains the inclination angle distribution of the blades by inverting the distribution of blade brightness, that is, the information on the blade inclination angle is inferred based on the amount of solar energy received by the blades through the sensor. It can be used for sensors on near-ground platforms such as drones and phenological cameras, and can meet the needs of obtaining more accurate large-scale and high-frequency near-ground leaf inclination angle distribution data.

[0181] In order to verify the effect of the method of inverting leaf inclination angle distribution through leaf brightness distribution, the results are verified based on simulation data and actual measurement data. The real leaf inclination angle distribution of simulation data can be pre-set as a scene parameter, and the leaf inclination angle distribution of measured data reference is obtained through the existing method based on mobile phone 3D reconstruction and manual measurement.

[0182] The LESS (Large-Scale Emulation System) model interface is easy to understand, simple to operate, and fast to simulate. It can distinguish between illumination and shadow components according to whether the target receives direct light or scattered light, and then synchronously output the four-component category to which each pixel belongs. Therefore, the present invention uses LESS to generate simulation data. The input parameters of LESS mainly include the three-dimensional model of vegetation, component spectrum, illumination and observation geometry. By controlling these parameters, multi-spectral images of vegetation canopy simulation with different leaf inclination distribution characteristics under different suns and observation angles can be generated.

[0183] The present invention uses the SimpleCrownGenerator Python script provided by the LESS (Large-Scale Emulation System) model to generate six typical uniform canopy scenes with leaf inclination angle distribution. Based on this simulation and considering the change of the sun angle from 8:00 to 12:00 in the morning on a certain day in late August in Beijing, 65°, 54°, 44°, 35°, and 30° are used as the solar zenith angles to simulate multispectral images at five moments. The resolution of the above simulation data is 0.5 cm, the observation zenith angle is 0°, and the component spectrum adopts the default value of LESS. Tables 2 and 3 show the specific component spectrum values ​​and specific scene parameter settings. Among them, Table 2 is the leaf and soil component spectrum parameters, and Table 3 is the uniform canopy scene parameters.

[0184] Table 2

[0185] Wavelength(nm) Blue (450:16) Green (560:16) Red (650:16) Near infrared (840:26) Leaf reflectivity 0.045 0.096 0.039 0.475 Leaf transmittance 0.045 0.096 0.039 0.475 Soil reflectivity 0.047 0.150 0.229 0.336

[0186] Table 3

[0187] Input parameter types Specific settings Blade shape Round Leaf area(m2) 0.01 Number of blades 10000 Leaf inclination angle distribution type Flat, vertical, spherical, uniform, inclined, extreme Canopy size (m) 6×6 Canopy height (m) 3 Scene size (m) 8×8 Image pixels 1500×1500 SZA(°) 65 / 54 / 44 / 35 / 30 VZA(°) 0 Whether to consider multiple scattering yes Sky Diffuse Ratio 0.15

[0188] Considering the strong multiple scattering in the near-infrared band, the visible light band is used for inversion. Table 4 summarizes the β distribution parameters inverted from the simulated data and the deviations from the true values. Table 4 shows the inversion results of the simulated data. Except for the extreme type, the brightness histogram-based method can invert the results that are close to the actual leaf tilt angle distribution. The average deviations of the average leaf tilt angle inverted from the blue, green, and red bands in the leaf tilt angle distribution type are 1.1°, 3.8°, and 4.14°. The error of the extreme type is large. This is because the three-dimensional model required for the simulated data is obtained by sampling the β distribution corresponding to the six typical leaf tilt angle distributions. When u or v of the β distribution is less than 1, the function will not converge at the boundary. Since it is difficult to effectively sample the part where the function tends to infinity, there is a certain difference between the leaf tilt angle distribution of the actual three-dimensional model and the leaf tilt angle distribution used to calculate the predicted relative brightness histogram. In addition, the numerical integration strategy adopted in this study also leads to certain accuracy problems in calculating the defect integral. In actual statistics of canopy leaf angles, non-convergence rarely occurs, so the part with β distribution parameter less than 1 can be deleted in the lookup table for inversion, thereby obtaining a more accurate result.

[0189] Table 4

[0190]

[0191] Exemplarily, the method of the present invention was tested in an orchard and a corn plot near a certain place. The plot information is shown in Table 5, which is the basic information of the measured plot. The multi-phase spectral data used for inversion was collected by DJI Phant4 Multispectral (DJI Technology Co., Ltd. Shenzhen, China) (referred to as P4M), and the green light band was used for inversion, and the observation zenith angle was 10°. The flight information is shown in Table 6, which is a record of the multi-phase flight test of the UAV. The reference leaf inclination angle distribution of the fruit tree was obtained by the SfM-PCNN method based on three-dimensional reconstruction of mobile phone digital photos by Qi et al. The reference leaf inclination angle distribution of corn was obtained by manual measurement.

[0192] Table 5

[0193] serial number Tree species type Average tree height (m) Leaf Area Index Underlying surface conditions 1 Fruit trees 4.2 2.26 Soil with a few weeds 2 corn 1.88 2.72 soil

[0194] Table 6

[0195]

[0196] The method of the present invention obtained a leaf inclination angle distribution type between the uniform type and the inclined type in the fruit tree sample plots, and a leaf inclination angle distribution type close to the spherical distribution in the corn sample plots. The deviations of the inversion results from the average leaf inclination angle were 1.8° and 3.4° in the fruit tree and corn sample plots, respectively.

[0197] The leaf inclination angle distribution inversion method provided in the present application obtains the actual relative brightness histogram of the data to be inverted; wherein the data to be inverted is a multispectral image of the leaf; a lookup table is established according to the sun angle and the observation angle; based on the lookup table, a theoretical relative brightness histogram of the data to be inverted is obtained; inversion is performed based on the actual relative brightness histogram and the theoretical relative brightness histogram to obtain the leaf inclination angle distribution parameters of the data to be inverted. Since the multispectral image of the leaf contains rich brightness information, and the lookup table includes: the leaf brightness histogram corresponding to different leaf inclination angle distribution parameter values ​​under multiple sun angles, the leaf inclination angle distribution is obtained through the information of the leaf brightness histogram in the image, which can be used for sensors of near-ground platforms such as drones and phenological cameras, and can meet the needs of more accurate large-scale and high-frequency near-ground leaf inclination angle distribution data acquisition.

[0198] In one embodiment, a blade pitch angle distribution inversion device is provided, such as Figure 3 As shown, the blade pitch angle distribution inversion device 300 includes:

[0199] The acquisition module 310 is used to acquire the actual relative brightness histogram of the data to be inverted; the data to be inverted is a multispectral image of the blade;

[0200] Establishing module 320, for establishing a lookup table according to the sun angle and the observation angle; the lookup table includes: a leaf brightness histogram corresponding to different leaf inclination angle distribution parameter values ​​under multiple sun angles;

[0201] A lookup module 330, configured to obtain a theoretical relative brightness histogram of the data to be inverted based on the lookup table;

[0202] The inversion module 340 is used to perform inversion based on the actual relative brightness histogram and the theoretical relative brightness histogram to obtain the leaf inclination angle distribution parameters of the data to be inverted.

[0203] Optionally, the establishing module includes:

[0204] An acquisition unit, used to acquire the sun angle at different times at a preset observation angle; the sun angle includes the sun zenith angle and the sun azimuth angle;

[0205] A determination unit is used to traverse the sun angles at different times in sequence with a plurality of preset leaf inclination angle distribution parameter combinations based on a two-parameter Beta distribution to determine a blade brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination; the preset leaf inclination angle distribution parameter combination includes: a first preset leaf inclination angle distribution parameter and a second preset leaf inclination angle distribution parameter;

[0206] The establishing unit is used to establish a lookup table according to the sun angle at different times, a plurality of preset leaf inclination angle distribution parameter combinations, and the leaf brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination.

[0207] Optionally, the search module is specifically used to:

[0208] Based on the target leaf inclination angle distribution parameter combination and the target sun angle, a theoretical relative brightness histogram of the data to be inverted is obtained in the lookup table.

[0209] Optionally, the determining unit is specifically configured to:

[0210] Calculate the relative brightness of the leaves based on the leaf angle and the sun angle:

[0211] B i =|Ω s ·Ω l |

[0212]

[0213] Among them, B i represents the relative brightness of the i-th group of the histogram, the subscript s represents the sun, the subscript l represents the leaf, Ω represents the unit vector in the corresponding direction, and θ is They represent the zenith angle and azimuth angle of the vector direction in the spherical coordinate system respectively;

[0214] Get the leaf frequency corresponding to the relative brightness of the leaf:

[0215]

[0216] Among them, P(B i ) represents the relative brightness B i The blade frequency, g(θ l ) represents the blade inclination angle distribution function, and the solid angle range ΔΩ corresponding to the blade normal i The corresponding zenith angle and azimuth angle ranges are Δθ i and Ω l The unit vector representing the normal direction of the blade, Ω v The unit vector representing the observation direction, |Ω l ·Ω v | indicates Ω l The unit area of ​​the blade in the direction of observation Ω v The projected area of

[0217] A plurality of preset leaf inclination angle distribution parameter combinations are traversed in sequence to determine the leaf brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination.

[0218] Optionally, the device further comprises:

[0219] The correction module is used to correct the relative brightness of the blade according to the ratio of the blade reflectivity to the blade transmittance:

[0220] B i =|Ω s ·Ω l |·f(Ω s ,Ω l )

[0221] Among them, B i represents the relative brightness of the i-th group of the histogram, subscript s represents the sun, subscript l represents the leaves, Ω represents the unit vector in the corresponding direction, and f(Ω s ,Ω l ) represents the ratio of leaf transmittance to reflectance;

[0222]

[0223] Where k is the ratio of the leaf transmittance to reflectance.

[0224] Optionally, the inversion module is specifically used for:

[0225] The KL divergence between the actual relative brightness histogram and the theoretical relative brightness histogram is calculated according to the KL divergence formula:

[0226]

[0227] Among them, P u,v represents the theoretical relative brightness histogram when the leaf inclination distribution parameters are u, v, Q represents the actual relative brightness histogram extracted from the data to be inverted, and B i represents the relative brightness of the i-th group of the histogram, P u,v (B i ) and Q(B i ) respectively represent the frequency of the i-th group in the corresponding histogram, and n represents the number of groups in the histogram;

[0228] According to the KL divergence between the actual relative brightness histogram and the theoretical relative brightness histogram, the inverted leaf tilt angle distribution parameters are obtained:

[0229] (u * ,v * ) = argmin u,v D k (P u,v ||Q)

[0230] Among them, u * ,v * represent the inverted first leaf tilt angle distribution parameters and second leaf tilt angle distribution parameters respectively.

[0231] Optionally, the acquisition module is specifically used to:

[0232] Obtaining data to be inverted;

[0233] Based on a preset leaf segmentation algorithm, the multispectral image of the leaf is subjected to single leaf segmentation processing to obtain the leaf brightness of each leaf;

[0234] The leaf brightness of each leaf is normalized to obtain an actual leaf brightness histogram of the data to be inverted; the actual leaf brightness histogram is used to reflect the area ratio occupied by leaves of various relative brightness sizes.

[0235] The leaf inclination angle distribution inversion device provided in the present application obtains the actual relative brightness histogram of the data to be inverted; wherein the data to be inverted is a multispectral image of the leaf; a lookup table is established according to the sun angle and the observation angle; based on the lookup table, a theoretical relative brightness histogram of the data to be inverted is obtained; inversion is performed based on the actual relative brightness histogram and the theoretical relative brightness histogram to obtain the leaf inclination angle distribution parameters of the data to be inverted. Since the multispectral image of the leaf contains rich brightness information, and the lookup table includes: the leaf brightness histogram corresponding to different leaf inclination angle distribution parameter values ​​under multiple sun angles, the leaf inclination angle distribution is obtained through the information of the leaf brightness histogram in the image, which can be used for sensors of near-ground platforms such as drones and phenological cameras, and can meet the needs of more accurate large-scale and high-frequency near-ground leaf inclination angle distribution data acquisition.

[0236] For the specific definition of the blade tilt angle distribution inversion device, please refer to the definition of the blade tilt angle distribution inversion method mentioned above, which will not be repeated here. Each module in the above-mentioned blade tilt angle distribution inversion device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the processor in the computing device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0237] In one embodiment, an electronic device is provided, the electronic device may be an onboard computer, and its internal structure diagram may be as shown in FIG. Figure 4 As shown. The electronic device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. When the computer program is executed by the processor, a method for inverting the distribution of blade inclination angles is implemented. The display screen of the electronic device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the electronic device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0238] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0239] In one embodiment, the blade pitch angle distribution inversion device provided in the present application can be implemented in the form of a computer, and the computer program can be used in the following manner: Figure 4 The memory of the electronic device may store various program modules constituting the blade pitch angle distribution inversion device of the electronic device, for example, Figure 3 The acquisition module, establishment module, search module, and inversion module shown in the computer program composed of various program modules enable the processor to execute the steps of the blade pitch angle distribution inversion method of the electronic device of each embodiment of the present application described in this specification.

[0240] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media that include computer-usable program code.

[0241] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor, or any conventional processor, etc.

[0242] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0243] Computer readable media include permanent and non-permanent, removable and non-removable storage media. Storage media can be implemented by any method or technology to store information, and the information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0244] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0245] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments described herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for inverting blade inclination angle distribution, characterized in that: include: Obtaining the actual relative brightness histogram of the data to be inverted; The data to be inverted is a multispectral image of a blade; A lookup table is established according to the sun angle and the observation angle; the lookup table includes: a leaf brightness histogram corresponding to different leaf inclination angle distribution parameter values ​​under multiple sun angles; Based on the lookup table, obtaining a theoretical relative brightness histogram of the data to be inverted; Inversion is performed based on the actual relative brightness histogram and the theoretical relative brightness histogram to obtain leaf inclination angle distribution parameters of the data to be inverted.

2. The blade inclination angle distribution inversion method according to claim 1, characterized in that: The step of establishing a lookup table according to the sun angle and the observation angle comprises: At a preset observation angle, obtaining the sun angle at different times; the sun angle includes the sun zenith angle and the sun azimuth angle; For the sun angles at different times, based on the dual-parameter Beta distribution, multiple preset leaf inclination angle distribution parameter combinations are used in turn to traverse and determine the leaf brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination; the preset leaf inclination angle distribution parameter combination includes: a first preset leaf inclination angle distribution parameter and a second preset leaf inclination angle distribution parameter; A lookup table is established according to the sun angle at different times, a plurality of preset leaf inclination angle distribution parameter combinations, and the blade brightness histograms corresponding to each preset leaf inclination angle distribution parameter combination.

3. The blade inclination angle distribution inversion method according to claim 1, characterized in that: The step of obtaining a theoretical relative brightness histogram of the data to be inverted based on the lookup table comprises: Based on the target leaf inclination angle distribution parameter combination and the target sun angle, a theoretical relative brightness histogram of the data to be inverted is obtained in the lookup table.

4. The blade inclination angle distribution inversion method according to claim 2, characterized in that: The method of traversing the sun angles at different times in sequence using a plurality of preset leaf inclination angle distribution parameter combinations based on a dual-parameter Beta distribution to determine a leaf brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination includes: Calculate the relative brightness of the leaves based on the leaf angle and the sun angle: B i =|Ω s ·Oh l | Among them, B i represents the relative brightness of the i-th group of the histogram, the subscript s represents the sun, the subscript l represents the leaf, Ω represents the unit vector in the corresponding direction, and θ is They represent the zenith angle and azimuth angle of the vector direction in the spherical coordinate system respectively; Get the leaf frequency corresponding to the relative brightness of the leaf: Among them, P(B i ) represents the relative brightness B i The blade frequency, g(θ l ) represents the blade inclination angle distribution function, and the solid angle range ΔΩ corresponding to the blade normal i The corresponding zenith angle and azimuth angle ranges are Δθ i and Ω l The unit vector representing the normal direction of the blade, Ω v The unit vector representing the observation direction, |Ω l ·Ω v | represents Ω l The unit area of ​​the blade in the direction of observation Ω v The projected area of A plurality of preset leaf inclination angle distribution parameter combinations are traversed in sequence to determine the leaf brightness histogram corresponding to each preset leaf inclination angle distribution parameter combination.

5. The blade inclination angle distribution inversion method according to claim 4, characterized in that: The method further comprises: According to the ratio of the leaf reflectivity to the leaf transmittance, the relative brightness of the leaf is corrected: B i =|Ω s ·Oh l |·f(Ω s ,Oh l ) Among them, B i represents the relative brightness of the i-th group of the histogram, subscript s represents the sun, subscript l represents the leaves, Ω represents the unit vector in the corresponding direction, and f(Ω s ,Ω l ) represents the ratio of leaf transmittance to reflectance; Where k is the ratio of the leaf transmittance to reflectance.

6. The blade inclination angle distribution inversion method according to claim 1, characterized in that: The inversion based on the actual relative brightness histogram and the theoretical relative brightness histogram to obtain leaf inclination angle distribution parameters includes: The KL divergence between the actual relative brightness histogram and the theoretical relative brightness histogram is calculated according to the KL divergence formula: Among them, P u,v represents the theoretical relative brightness histogram when the leaf inclination distribution parameters are u, v, Q represents the actual relative brightness histogram extracted from the data to be inverted, and B i represents the relative brightness of the i-th group of the histogram, P u,v (B i ) and Q(B i ) respectively represent the frequency of the i-th group in the corresponding histogram, and n represents the number of groups in the histogram; According to the KL divergence between the actual relative brightness histogram and the theoretical relative brightness histogram, the inverted leaf tilt angle distribution parameters are obtained: (u * ,v * )=argmin u,v D k (P u,v ||Q) Among them, u * ,v * represent the inverted first leaf tilt angle distribution parameters and second leaf tilt angle distribution parameters respectively.

7. The blade inclination angle distribution inversion method according to claim 1, characterized in that: The step of obtaining an actual relative brightness histogram of the data to be inverted comprises: Obtaining data to be inverted; Based on a preset leaf segmentation algorithm, the multispectral image of the leaf is subjected to single leaf segmentation processing to obtain the leaf brightness of each leaf; The leaf brightness of each leaf is normalized to obtain an actual leaf brightness histogram of the data to be inverted; the actual leaf brightness histogram is used to reflect the area ratio occupied by leaves of various relative brightness sizes.

8. A blade inclination angle distribution inversion device, characterized in that: include: An acquisition module, used for acquiring an actual relative brightness histogram of the data to be inverted; The data to be inverted is a multispectral image of a blade; Establishing module, used to establish a lookup table according to the sun angle and the observation angle; the lookup table includes: a leaf brightness histogram corresponding to different leaf inclination angle distribution parameter values ​​under multiple sun angles; A search module, used for obtaining a theoretical relative brightness histogram of the data to be inverted based on the lookup table; An inversion module is used to perform inversion based on the actual relative brightness histogram and the theoretical relative brightness histogram to obtain leaf inclination angle distribution parameters of the data to be inverted.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the blade pitch angle distribution inversion method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the blade pitch angle distribution inversion method according to any one of claims 1 to 7 are implemented.

Images