Method and device for constructing rice grain radiation transmission model, equipment and medium

Abstract

The application relates to a rice grain radiation transmission model construction method, device, equipment and medium, the method comprising: based on a preset light ray tracing algorithm, initializing the geometric size parameters of a rice grain physical model and the optical parameters of incident light rays, uniformly sampling incident points on the surface of the rice grain physical model, iteratively tracing the propagation path of internal light rays of the rice grain physical model in sections, and calculating the energy attenuation value of each propagation path of the internal light rays; based on the energy attenuation value, determining the residual energy after energy attenuation of each propagation path, when the exit stage of the internal light rays of the rice grain physical model does not occur total reflection, distributing the transmission energy according to the exit direction of the internal light rays, accumulating the reflection energy and the transmission energy corresponding to all incident light rays, combining the energy conservation law to calculate the absorption energy of the light rays and correct the numerical error, and outputting the overall reflectivity, transmissivity and absorptivity of the rice grain physical model. The application realizes accurate spectral simulation of rice grains.

Description

Technical Field

[0001] This application relates to the field of agricultural production, and in particular to a method for constructing a rice grain radiative transfer model, a corresponding device, electronic equipment, and a computer-readable storage medium. Background Technology

[0002] As a core staple crop globally, rice's yield and quality directly impact food security and sustainable agricultural development. Accurate monitoring of rice's growth dynamics throughout its entire growth cycle is crucial for precise water and fertilizer regulation, yield prediction, and quality assessment. However, observations show that as rice enters its later growth stages, the contribution of grains to the canopy spectrum becomes increasingly significant with increasing maturity. This often overlooked spectral component can easily lead to biases in canopy spectral analysis, interfering with the accuracy of growth information retrieval. Therefore, targeted monitoring and analysis of rice grain spectral characteristics is a key prerequisite for eliminating interference from traditional monitoring methods and achieving precise rice monitoring.

[0003] In recent years, hyperspectral remote sensing technology has been widely used in crop monitoring. By constructing a correlation model between spectral and physiological parameters, non-destructive assessment can be achieved. Based on the modeling logic, hyperspectral monitoring models are divided into two categories: data-driven and physics-driven. Among them, data-driven models mainly explore the statistical correlation between spectral features and crop physiological parameters, and combine them with machine learning algorithms to realize their application in crop monitoring.

[0004] Crop radiative transfer models (RTMs) accurately simulate the propagation of light in crop organs such as leaves and canopy through physical mechanisms, revealing the fundamental principles of spectral signal formation. Compared to purely data-driven empirical models, RTMs offer stronger interpretability and cross-scenario portability, making them a core tool for crop optical modeling and remote sensing inversion. Based on the differences in organs and scales of the simulated objects, existing crop RTMs are mainly divided into leaf-scale and canopy-scale models, and most are based on the core assumption of "uniform mixing and planar stratification," forming a series of mature and classic models. At the leaf scale, the PROSPECT series of models simplifies the leaf as a "uniform flat plate," providing a core physical basis for the inversion of leaf biochemical parameters. Models such as PIOSL further improve the accuracy of leaf spectral simulation by optimizing parameterization schemes such as mesophyll structure. At the canopy scale, the SAIL series models simplify the canopy into a horizontally homogeneous multilayered medium, coupling leaf optical properties with canopy structural parameters such as leaf area index (LAI) or leaf tilt angle distribution (LAD) to achieve canopy-scale radiative transfer simulation. The PROSAIL model further completes the scale connection between leaves and canopy, becoming the mainstream tool for canopy spectral inversion and growth monitoring of densely planted field crops such as wheat, rice, and corn. However, existing traditional radiative transfer models centered on leaves and canopy are designed to accommodate the "flat, thin sheet-like" structure of leaves or the "layered, homogeneous" structure of canopies, making it difficult to directly meet the modeling requirements of grain optical properties. Grains, as storage carriers of photosynthetic products, often exhibit an ellipsoidal, three-dimensional, convex morphology, significantly different from the structural properties of leaves. After light incident on leaves, it is mainly reflected by the surface and scattered by a small amount of internal light multiple times, which is suitable for the homogeneous layer assumption. However, the surface curvature of ellipsoidal grains varies with the location, easily causing multi-angle scattering and refraction of light upon incident, and the heterogeneity of the distribution of internal components such as starch and protein further increases the complexity of the radiative transfer path. This difference in structure and optical mechanism makes it difficult to directly transfer traditional model frameworks to the optical simulation of grains.

[0005] In summary, existing traditional radiative transfer models, which are based on leaves and canopies, are designed to accommodate the structural characteristics of leaves that are "flat and thin" or canopies that are "uniformly layered," making it difficult to directly meet the requirements for modeling the optical properties of grains. The applicant has made corresponding explorations to address this problem. Summary of the Invention

[0006] The purpose of this application is to solve the above-mentioned problems by providing a method for constructing a rice grain radiative transfer model, a corresponding device, electronic equipment, and a computer-readable storage medium.

[0007] To achieve the various objectives of this application, the following technical solution is adopted:

[0008] A method for constructing a rice grain radiative transfer model, proposed to meet one of the purposes of this application, includes:

[0009] Obtain the ellipsoidal space equation corresponding to the physical model of the target rice grain, and the ray parameter equation of the incident light of the target rice grain. Substitute the ray parameter equation into the ellipsoidal space equation to solve for the intersection point of the incident light and the physical model of the rice grain. Calculate the reflection direction and refraction direction of the light at the intersection point based on the laws of reflection and refraction, respectively, to complete the physical constraint of the propagation path of the incident light during the interaction between the incident light and the physical model of the rice grain.

[0010] The radiative transmission process of a single incident light beam within a rice grain physical model is analyzed. By combining the Lambert-Beer law, Fresnel's law, and the law of conservation of energy, analytical expressions for the total emissivity and total absorptivity when a single incident light beam interacts with the rice grain physical model are derived.

[0011] Using the geometric dimensions, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters, a two-stage optimization strategy combining global search and local optimization is adopted. With the measured reflectance spectrum and measured transmittance spectrum as constraints, the optimal combination of model parameters matching the measured reflectance spectrum and measured transmittance spectrum is solved in reverse to complete the calibration of the key input parameters of the model.

[0012] Based on a preset ray tracing algorithm, the geometric dimensions of the rice grain physical model and the optical parameters of the incident ray are initialized. The incident points are uniformly sampled on the surface of the rice grain physical model. The propagation path of the internal ray of the rice grain physical model is iteratively traced segment by segment, and the energy attenuation value of each segment of the propagation path of the internal ray is calculated.

[0013] Based on the energy attenuation value, the remaining energy after energy attenuation through each propagation path is determined. When total internal reflection does not occur during the emission stage of the light rays inside the rice grain physical model, the transmitted energy is allocated according to the emission direction of the light rays inside the model. The reflected energy and transmitted energy corresponding to all incident light rays are accumulated. The energy absorbed by the rice grain physical model is calculated by combining the law of conservation of energy and the numerical error is corrected. The overall reflectivity, transmittance and absorptivity of the rice grain physical model are output to complete the construction of the rice grain radiative transfer model.

[0014] Optionally, the step of substituting the ray parameter equation into the ellipsoidal space equation to solve for the intersection point of the incident ray and the physical model of the rice grain includes:

[0015] Substituting the incident ray parameter equation into the ellipsoidal space equation corresponding to the rice grain physical model, a quadratic equation for the ray propagation distance is derived, and the coefficients of the quadratic equation for the ray propagation distance are calculated.

[0016] The discriminant of the quadratic equation for the propagation distance of the light ray is calculated. Based on the discriminant, it is determined whether there is an intersection point between the incident light ray and the physical model of the rice grain. If the discriminant is greater than zero, then there is an intersection point. The positive root of the quadratic equation is taken, and the coordinates of the near-side intersection point corresponding to the near-side intersection point between the incident light ray and the physical model of the rice grain are calculated to complete the solution of the intersection point between the incident light ray and the physical model of the rice grain.

[0017] Optionally, the steps of calculating the reflection direction and refraction direction of the light ray at the intersection point based on the laws of reflection and refraction, respectively, to complete the physical constraint of the propagation path of the incident light ray during the interaction with the physical model of the rice grain, include:

[0018] The normal vector of the surface of the rice grain physical model at the near-side intersection is calculated and normalized. Based on the law of reflection, the reflection direction of the light at the near-side intersection is calculated by the incident light direction vector and the normalized normal vector. The normal vector of the surface of the rice grain physical model represents the vector at any near-side intersection of the surface of the rice grain physical model constructed based on the idealized ellipsoidal structure of the rice grain, which is perpendicular to the tangent plane of the intersection point and points to the outside of the rice grain physical model.

[0019] Based on the law of refraction and the relationship between the refractive index of the internal medium, the angle of incidence and the angle of refraction, the direction of refraction of the light at the near intersection is calculated to complete the physical constraint of the propagation path of the incident light during the interaction between the incident light and the physical model of rice grains.

[0020] Optionally, the step of analyzing the radiative transport process of a single incident light beam within a rice grain physical model, and combining the Lambert-Beer law, Fresnel's law, and the law of conservation of energy to derive analytical expressions for the total emissivity and total absorptivity when the single incident light beam interacts with the rice grain physical model, includes:

[0021] The radiative transmission process of a single incident light beam in a physical model of a rice grain was analyzed, and the energy attenuation law and single-pass transmittance of the single incident light beam in the internal medium of the target rice grain were determined by combining the Lambert-Beer law.

[0022] Based on Fresnel's law, the reflectivity and transmittance of a single incident ray at different medium interfaces are determined. The energy distribution relationship of reflected energy, transmitted energy, and absorbed energy of the ray within the model is clarified by the law of conservation of energy, so as to derive analytical expressions for the total emissivity and total absorptivity when a single incident ray interacts with the physical model of rice grains.

[0023] Optionally, using the geometric dimensions, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters, and employing a two-stage optimization strategy combining global search and local optimization, with measured reflectance and transmittance spectra as constraints, the step of inversely solving for the optimal combination of model parameters matching the measured reflectance and transmittance spectra includes:

[0024] Using the geometric dimensions, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters, objective functions are constructed with the normalized squared error between the simulated reflectance spectrum and the measured reflectance spectrum, and between the simulated transmittance spectrum and the measured transmittance spectrum as error terms, and regularization terms are introduced to constrain the range of parameter values.

[0025] The objective function is solved by a two-stage optimization strategy that combines global search with local optimization. The differential evolution algorithm is used to traverse the global range of parameters to initially locate the optimal parameter interval.

[0026] Using the aforementioned optimal parameter range as initial values, local refinement adjustments are made using the limited-memory Broyden-Fletcher-Goldfarb-Shannon band boundary constraint algorithm. Each spectral band is independently inverted, and the goodness of fit is evaluated using the coefficient of determination. The parameter result with the highest goodness of fit is selected as the parameter value for the corresponding band. The optimal model parameter combination that matches the measured reflectance spectrum and measured transmittance spectrum of the target rice grain is obtained by reverse solving.

[0027] Optionally, based on a preset ray tracing algorithm, the steps of initializing the geometric parameters of the rice grain physical model and the optical parameters of the incident ray, uniformly sampling the incident points on the surface of the rice grain physical model, iteratively tracing the propagation path of the internal ray of the rice grain physical model segment by segment, and calculating the energy attenuation value of each segment of the propagation path of the internal ray include:

[0028] The preset ray tracing algorithm is invoked to initialize the geometric dimension parameters and incident ray optical parameters of the rice grain physical model and form a parameter dataset. Incident points are uniformly sampled on the ellipsoidal surface of the rice grain physical model at a preset density and an incident point coordinate set is generated.

[0029] Based on the parameter dataset and the incident point coordinate set, the propagation path of the internal light rays formed after the incident light rays enter the rice grain physical model is iteratively tracked segment by segment. At the same time, the energy attenuation value corresponding to each segment of the internal light ray propagation path is calculated in combination with the Lambert-Beer law to construct an energy attenuation dataset.

[0030] Optionally, based on the energy attenuation value, the remaining energy after energy attenuation through each propagation path is determined. When total internal reflection does not occur during the emission stage of the light rays inside the rice grain physical model, the transmitted energy is allocated according to the emission direction of the internal light rays. The reflected and transmitted energy corresponding to all incident light rays are accumulated. The energy absorbed by the rice grain physical model for light is calculated using the law of conservation of energy, and numerical errors are corrected. The overall reflectivity, transmittance, and absorptivity of the rice grain physical model are then output. This process includes:

[0031] The energy decay values ​​of the energy decay dataset are called, and the remaining energy dataset after the internal light rays of the rice grain physical model have decayed through each propagation path is determined by iterative calculation.

[0032] Based on the remaining energy dataset, the total internal reflection state of the internal light during the emission stage is determined. If it is determined that no total internal reflection has occurred, the transmission energy is quantified and allocated according to the emission direction of the internal light and a transmission energy dataset is generated.

[0033] The reflected energy dataset corresponding to all incident rays is collected synchronously and accumulated with the transmitted energy dataset. Based on the law of conservation of energy, the absorbed energy of the rice grain physical model is calculated and numerical error correction is performed on various energy data. Finally, the reflectivity, transmittance and absorptivity of the rice grain physical model are calculated and output based on the corrected energy data.

[0034] A rice grain radiative transfer model construction device provided for another purpose of this application includes:

[0035] The physical interaction path constraint module is configured to obtain the ellipsoidal space equation corresponding to the physical model of the target rice grain, and the ray parameter equation of the incident light of the target rice grain. The ray parameter equation is substituted into the ellipsoidal space equation to solve for the intersection point of the incident light and the physical model of the rice grain. Based on the laws of reflection and refraction, the reflection direction and refraction direction of the light at the intersection point are calculated respectively to complete the physical constraint of the propagation path of the incident light during the interaction between the incident light and the physical model of the rice grain.

[0036] The optical properties analysis module is set to analyze the radiative transmission process of a single incident light beam in the physical model of rice grains. Combining the Lambert-Beer law, Fresnel law and the law of conservation of energy, analytical expressions for the total emissivity and total absorptivity when a single incident light beam interacts with the physical model of rice grains are derived.

[0037] The key parameter calibration module is configured to use the geometric size parameters, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters. It adopts a two-stage optimization strategy combining global search and local optimization, and uses the measured reflectance spectrum and measured transmittance spectrum as constraints to solve in reverse the optimal combination of model parameters that matches the measured reflectance spectrum and measured transmittance spectrum, so as to complete the calibration of the key input parameters of the model.

[0038] The energy attenuation calculation module is configured to initialize the geometric parameters of the rice grain physical model and the optical parameters of the incident light rays based on a preset ray tracing algorithm, uniformly sample the incident points on the surface of the rice grain physical model, iteratively trace the propagation path of the internal light rays of the rice grain physical model segment by segment, and calculate the energy attenuation value of each segment of the propagation path of the internal light rays.

[0039] The radiative transfer model construction module is configured to determine the remaining energy after energy attenuation through each propagation path based on the energy attenuation value. When total internal reflection does not occur during the emission stage of light rays inside the rice grain physical model, the transmitted energy is allocated according to the emission direction of the internal light rays. The reflected and transmitted energy corresponding to all incident light rays are accumulated. The energy absorption energy of the rice grain physical model is calculated by combining the law of conservation of energy and the numerical error is corrected. The overall reflectivity, transmittance and absorptivity of the rice grain physical model are output to complete the construction of the rice grain radiative transfer model.

[0040] An electronic device provided for another purpose of this application includes a central processing unit and a memory, wherein the central processing unit is used to invoke and run a computer program stored in the memory to perform the steps of the rice grain radiative transfer model construction method of this application.

[0041] A computer-readable storage medium is provided for another purpose of this application, which stores, in the form of computer-readable instructions, a computer program implemented according to the method for constructing the rice grain radiative transfer model, which, when called by a computer, executes the steps included in the corresponding method.

[0042] Compared to existing technologies, this application addresses the problems of traditional radiative transfer models centered on leaves and canopies, whose design logic is often adapted to the "flat, thin sheet-like" structure of leaves or the "uniformly layered" structure of canopies, making it difficult to directly meet the modeling requirements for the optical properties of grains. This application offers the following advantages, including but not limited to:

[0043] Firstly, existing crop radiative transfer models (RTMs) all rely on the core assumptions of "flat, thin leaf" and "uniformly layered canopy," which are only suitable for optical simulations at the leaf and canopy scales. However, rice grains have an ellipsoidal, three-dimensional, convex structure with surface curvature varying with location, leading to multi-angle scattering and refraction of incident light. The heterogeneity of internal component distribution also complicates the radiative transfer path, making traditional model frameworks unsuitable for direct application. This application abstracts the rice grain into an ideal ellipsoidal structure. By constructing ellipsoidal spatial equations and ray parameter equations, it solves for the intersection points of light rays and grains and constrains the propagation path based on the laws of reflection and refraction. This specifically adapts to the three-dimensional structural characteristics of the grain, achieving for the first time a dedicated modeling of radiative transfer at the organ scale of rice grains. This fills the technical gap in grain-specific modeling and eliminates spectral simulation biases caused by structural mismatches in traditional models.

[0044] Secondly, this application analyzes the radiative transmission process of a single beam of light within the grain, integrating the Lambert-Beer law (energy decay), Fresnel's law (interface transmission and reflection), and the law of conservation of energy (energy distribution) to derive analytical expressions for the total emissivity and total absorptivity of the interaction between the single beam of light and the grain. This reveals the formation principle of the spectral signal of rice grains from a physical perspective, upgrading grain spectral simulation from "statistical correlation" to "physical mechanism driven," significantly improving the model's cross-variety and cross-scenario portability, and avoiding excessive reliance on samples by data-driven models.

[0045] Thirdly, this application proposes a two-stage optimization strategy combining global search and local optimization, using grain geometry, internal absorption coefficient, and medium refractive index as key input parameters. First, a differential evolution algorithm is used to traverse the global range of parameters to locate the optimal interval. Then, a limited-memory Broyden-Fletcher-Goldfarb-Shannon band boundary constraint algorithm (L-BFGS-B algorithm) is used for local refinement. Simultaneously, measured reflectance and transmittance spectra are used as constraints to independently invert each spectral band, and the goodness of fit is evaluated using the coefficient of determination. This strategy avoids the problem of single local optimization algorithms easily getting trapped in local optima, and ensures the matching between parameters and actual grains through measured spectral constraints. It achieves accurate joint inversion of the core optical parameters of the grain, laying a high-precision parameter foundation for subsequent spectral simulations.

[0046] Fourth, this application relies on a ray tracing algorithm. By initializing the grain and ray parameters, uniformly sampling the incident point, iteratively tracing the internal ray path segment by segment and calculating energy attenuation, combining the remaining energy to determine the total internal reflection state and distributing the transmitted energy according to the emission direction, and finally accumulating the transmitted and reflected energy of all rays, calculating the absorbed energy, and correcting numerical errors, it achieves accurate statistical calculation of the overall optical properties of the grain. This method avoids the limitations of analytical methods. Through refined simulation on a ray-by-ray and path-by-path basis, it significantly reduces spectral simulation errors. The simulated spectra of different grain varieties are highly consistent with the peak and valley shapes and band fluctuations of the measured spectra. The RMSE values ​​are generally at a low level, and the simulated RMSE of transmittance is even as low as 0.001, which fully demonstrates the accuracy and reliability of the spectral simulation of the model in this application.

[0047] Fifth, this application achieves precise simulation of rice grain spectra, effectively analyzing the contribution ratio of grains to the canopy spectrum, eliminating monitoring interference caused by this spectral component, and making canopy spectrum analysis more accurate in the later stages of rice growth. At the same time, the optical properties of grains are closely related to quality parameters such as grain maturity, starch, and protein content. The model of this application can serve as the core physical basis for non-destructive inversion of grain quality parameters, providing key technical support for precise monitoring of rice throughout its entire growth period and efficient evaluation of yield and quality, and contributing to precise water and fertilizer regulation in rice cultivation and sustainable agricultural development. Attached Figure Description

[0048] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0049] Figure 1 This is a flowchart illustrating the method for constructing a rice grain radiative transfer model in the embodiments of this application.

[0050] Figure 2 This is a diagram illustrating the effect of the rice grain radiative transfer model construction method in the embodiments of this application;

[0051] Figure 3 This is a schematic diagram illustrating the assumption that the target rice grains in this embodiment are ideally ellipsoidal structures.

[0052] Figure 4 This is a schematic diagram illustrating the interaction process between light rays and ellipsoidal rice grains based on the physical constraints of the propagation path in an embodiment of this application.

[0053] Figure 5 This is a schematic diagram showing the overall trend of the simulated and measured spectral curves of different varieties of seeds in the embodiments of this application;

[0054] Figure 6 This is a schematic diagram of the error distribution histogram of different optical properties of rice grains in the embodiments of this application;

[0055] Figure 7 This is a scatter plot showing the distribution characteristics of the root mean square error of reflectance, transmittance and absorptance in multiple different wavelength ranges and the number of samples in the embodiments of this application.

[0056] Figure 8 This is a schematic diagram of the rice grain radiative transfer model construction device in the embodiments of this application;

[0057] Figure 9 This is a schematic diagram of the structure of the computer device in the embodiments of this application. Detailed Implementation

[0058] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0059] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this application means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.

[0060] Those skilled in the art will understand that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0061] Those skilled in the art will understand that the terms "client," "terminal," and "terminal device" as used herein include both devices that receive wireless signals, devices that only possess wireless signal receiver capabilities without transmission capabilities, and devices with receiving and transmitting hardware, devices that have receiving and transmitting hardware capable of bidirectional communication over a bidirectional communication link. Such devices may include: cellular or other communication devices such as personal computers or tablets, having single-line displays, multi-line displays, or cellular or other communication devices without multi-line displays; PCS (Personal Communications Service) that can combine voice, data processing, fax, and / or data communication capabilities; PDAs (Personal Digital Assistants) that may include radio frequency receivers, pagers, internet / intranet access, web browsers, notebooks, calendars, and / or GPS (Global Positioning System) receivers; and conventional laptops and / or handheld computers or other devices that have and / or include radio frequency receivers. As used herein, "client," "terminal," and "terminal device" can be portable, transportable, installed in a means of transportation (air, sea, and / or land), or suitable and / or configured to operate locally and / or in a distributed manner, operating in any other location on Earth and / or in space. "Client," "terminal," and "terminal device" as used herein can also be a communication terminal, an internet access terminal, or a music / video playback terminal, such as a PDA, a MID (Mobile Internet Device), and / or a mobile phone with music / video playback capabilities, or a smart TV, set-top box, etc.

[0062] The hardware referred to by the names "server," "client," and "service node" in this application is essentially an electronic device with the equivalent capabilities of a personal computer. It is a hardware device with the necessary components revealed by the von Neumann architecture, such as a central processing unit (including an arithmetic logic unit and a control unit), memory, input devices, and output devices. The computer program is stored in its memory, and the central processing unit loads the program stored in the secondary storage into the main memory to run it, execute the instructions in the program, and interact with the input and output devices to complete specific functions.

[0063] It should be noted that the concept of "server" used in this application can also be extended to the case of server clusters. Based on the network deployment principles understood by those skilled in the art, the servers should be logically divided. Physically, these servers can be independent of each other but accessible through interfaces, or they can be integrated into a single physical computer or a computer cluster. Those skilled in the art should understand this flexibility and should not use it to constrain the implementation of the network deployment method in this application.

[0064] One or more of the technical features of this application, unless explicitly specified herein, can be deployed on a server and accessed by a client remotely calling the online service interface provided by the server, or can be directly deployed and run on a client to access the service.

[0065] Unless otherwise specified, the neural network models referenced or potentially referenced in this application may be deployed on a remote server and invoked remotely on the client, or deployed on a client with the capability to invoke directly. In some embodiments, when running on the client, the corresponding intelligence may be acquired through transfer learning in order to reduce the requirements on the client's hardware resources and avoid excessive consumption of the client's hardware resources.

[0066] Unless otherwise specified, all data involved in this application may be stored remotely on a server or on a local terminal device, as long as it is suitable for use by the technical solution of this application.

[0067] Those skilled in the art will understand that although the various methods in this application are described based on the same concept and thus present commonality among them, they can be performed independently unless otherwise specified. Similarly, the various embodiments disclosed in this application are all based on the same inventive concept; therefore, concepts expressed in the same way, as well as concepts that are appropriately changed for convenience but are expressed differently, should be understood equivalently.

[0068] Unless otherwise expressly stated, the various embodiments disclosed in this application can be combined in a cross-cutting manner to flexibly construct new embodiments, as long as such combination does not depart from the inventive spirit of this application and can meet the needs of the prior art or solve a certain deficiency in the prior art. Those skilled in the art should be aware of such modifications.

[0069] Please see Figure 1 In one embodiment of the rice grain radiative transfer model construction method of this application, the method includes:

[0070] Step S10: Obtain the ellipsoidal space equation corresponding to the physical model of the target rice grain, and the ray parameter equation of the incident light of the target rice grain. Substitute the ray parameter equation into the ellipsoidal space equation to solve for the intersection point of the incident light and the physical model of the rice grain. Calculate the reflection direction and refraction direction of the light at the intersection point based on the laws of reflection and refraction, respectively, to complete the physical constraint of the propagation path of the incident light during the interaction between the incident light and the physical model of the rice grain.

[0071] The rice grain radiative transfer model construction system in the terminal device can obtain the ellipsoidal space equation corresponding to the physical model of the target rice grain, as well as the ray parameter equation of the incident light of the target rice grain. The ray parameter equation is substituted into the ellipsoidal space equation to solve for the intersection point of the incident light and the physical model of the rice grain. Based on the laws of reflection and refraction, the reflection direction and refraction direction of the light at the intersection point are calculated respectively to complete the physical constraint of the propagation path of the incident light during the interaction between the incident light and the physical model of the rice grain.

[0072] Please see Figure 3 Using real rice grains with their outer skin as the model prototype, and after morphological optimization and regularization (ignoring minor geometric features to reduce computational complexity), the target rice grain is assumed to be an ideal ellipsoidal structure. The physical relationship between the rice grain physical model and light rays is used to constrain the light transmission path. Furthermore, the energy distribution of the light rays is clarified by combining the interaction process between the light rays and the rice grain physical model, thus deriving analytical expressions for the overall emissivity and overall absorptivity when a single beam of light interacts with the rice grain physical model. A two-stage optimization algorithm is then used to calibrate the absorption coefficient of the internal medium of the rice grain physical model. and refractive index Finally, the overall spectral reflectance and transmittance were statistically analyzed using a ray tracing algorithm, enabling spectral simulation of the forward modeling process of rice grain radiative transfer model (RGRT-MODEL) and completing the construction of the rice grain radiative transfer model.

[0073] To calculate the optical properties (reflectivity, transmittance, and absorptivity) of light rays after they interact with the ellipsoidal rice grain physical model, it is necessary to establish a spatial coordinate system, introduce the structural spatial equations of the rice grain physical model, the light ray vector equations, and the laws of reflection and refraction to describe the physical path of the interaction between light rays and the rice grain physical model.

[0074] In some embodiments, the step of substituting the ray parameter equation into the ellipsoidal space equation to solve for the intersection point of the incident ray and the physical model of the rice grain includes:

[0075] Step S101: Substitute the incident light ray parameter equation into the ellipsoidal space equation corresponding to the rice grain physical model to derive the quadratic equation about the light propagation distance, and calculate the coefficients of the quadratic equation about the light propagation distance.

[0076] Step S102: Calculate the discriminant of the quadratic equation for the propagation distance of the light ray. Determine whether there is an intersection point between the incident light ray and the physical model of the rice grain based on the discriminant. If the discriminant is greater than zero, then there is an intersection point. Take the positive root of the quadratic equation and calculate the coordinates of the near-side intersection point corresponding to the near-side intersection point of the incident light ray and the physical model of the rice grain, so as to complete the solution of the intersection point between the incident light ray and the physical model of the rice grain.

[0077] In a specific embodiment, firstly, the target rice grain is abstracted as an ellipsoid to construct a physical model of the rice grain, and the spatial geometric structure of the physical model of the rice grain is defined. To reduce computational complexity, this application makes idealized assumptions about the morphological structure of the rice grain, ignores the interference of secondary geometric features of the rice grain, and abstracts it as an ellipsoid. The calculation formula of the ellipsoidal spatial equation corresponding to the physical model of the rice grain is expressed as follows:

[0078]

[0079] Among them, represent the three semi-axial lengths of the target rice grain. Indicates the length of the target rice grain. Indicates the width of the target rice grain. Indicates the thickness of the target rice grain; Represents the three-dimensional coordinates of the ellipsoidal coordinate system.

[0080] Next, the interaction between the incident light ray and the ellipsoidal rice grain was simulated. It was assumed that a beam of incident light ray entered from the upper surface of the target rice grain in the external space, and the parametric equation of the light ray was expressed as:

[0081]

[0082] in, Indicates the distance light travels The spatial position vector after; The spatial coordinate vector of the ray's origin; This is the unit direction vector for the propagation of light. This represents the distance light travels.

[0083] Furthermore, after propagating a certain distance, the incident light ray may intersect with the ellipsoidal surface of the rice grain. Substituting the light ray parameter equation into the ellipsoidal space equation corresponding to the physical model of the rice grain, the relationship between the light ray propagation distance and the ellipsoidal space equation is derived. The quadratic equation is expressed as:

[0084]

[0085] in, Indicates the distance the light travels. The coefficients of the quadratic equation; This represents the distance light travels.

[0086] Furthermore, the coefficients of the quadratic equation for the propagation distance of the light are calculated, and are expressed as follows:

[0087]

[0088] The discriminant of the quadratic equation for calculating the propagation distance of the light ray is used to determine whether there is an intersection point between the incident light ray and the physical model of the rice grain. If the discriminant is greater than zero, an intersection point exists. The positive root of the quadratic equation is then taken to calculate the coordinates of the near-side intersection point corresponding to the near-side intersection point of the incident light ray and the physical model of the rice grain. Specifically, the discriminant of the quadratic equation is used to... Confirmed: If The incident light ray has no intersection with the physical model of the rice grain, and the light ray does not contact the surface of the rice grain; if Take the positive root Calculate the coordinates of the nearest intersection point. .

[0089] In a further embodiment, the reflection direction and refraction direction of the light ray at the intersection point are calculated based on the laws of reflection and refraction, respectively, to complete the physical constraint of the propagation path of the incident light ray during the interaction with the physical model of the rice grain. This includes:

[0090] Step S1001: Calculate the normal vector of the surface of the rice grain physical model at the near-side intersection and complete the normalization process. Based on the law of reflection, calculate the reflection direction of the light at the near-side intersection by the incident light direction vector and the normalized normal vector. The normal vector of the surface of the rice grain physical model represents the vector at any near-side intersection of the rice grain physical model constructed based on the idealized ellipsoidal structure of the rice grain, which is perpendicular to the tangent plane of the intersection point and points to the outside of the rice grain physical model.

[0091] Step S1002: Based on the law of refraction and the relationship between the refractive index of the internal medium, the incident angle and the refraction angle, calculate the refraction direction of the light at the near intersection point to complete the physical constraint of the propagation path of the incident light during the interaction between the incident light and the physical model of rice grains.

[0092] In a specific embodiment, the normal vector at the near-side intersection point of the incident ray and the physical model of the rice grain, and the near-side intersection point of the ellipsoidal surface corresponding to the target rice grain. The normal vector (pointing outwards) at that location is represented as:

[0093]

[0094] in, Indicates the intersection point of the ellipsoidal surface on the near side. The unit normal vector at that location, The coordinates of the near-side intersection point corresponding to the near-side intersection point of the incident ray and the physical model of the rice grain are indicated.

[0095] The normalized normal vector is represented as:

[0096]

[0097] in, This represents the normalized unit normal vector; Indicates the intersection point of the ellipsoidal surface on the near side. The unit normal vector at that location; Indicates the intersection point of the ellipsoidal surface on the near side. The magnitude of the normal vector at that point.

[0098] At the point where the incident ray intersects the near side of the physical model of the rice grain, the ray undergoes reflection and refraction. According to the law of reflection, the direction of specular reflection satisfies the following conditions: the reflected ray, the incident ray, and the normal are coplanar, the angle of incidence equals the angle of reflection, and the vector formula for the direction of reflection is expressed as:

[0099]

[0100] Where is the unit direction vector of the incident ray. The intersection of the ellipsoidal surface on the near side The unit normal vector at that location, This is the unit direction vector of the reflected ray.

[0101] Furthermore, according to the law of refraction, the direction of the refracted ray is from... It is confirmed that, among them, Represents the refractive index of the incident medium. Represents the refractive index of the medium on the refracting side; It is expressed as the angle of incidence, which represents the angle between the incident ray and the normal vector; Let be the angle of refraction, which represents the angle between the refracted ray and the normal vector. Based on the law of refraction and the relationship between the refractive index of the internal medium, the angle of incidence, and the angle of refraction, the direction of refraction of the ray at the near-side intersection point is calculated using the following formula:

[0102]

[0103] in, , (when hour).

[0104] When light enters the physical model of a rice grain, it travels from an optically denser medium to an optically less dense medium. ),and When total internal reflection occurs, then, Indicates the critical angle.

[0105] Step S20: Analyze the radiative transmission process of a single incident light beam in the physical model of rice grains, and combine the Lambert-Beer law, Fresnel law and the law of conservation of energy to derive analytical expressions for the total emissivity and total absorptivity when a single incident light beam interacts with the physical model of rice grains.

[0106] The ellipsoidal space equation corresponding to the physical model of the target rice grain and the ray parameter equation of the incident light of the target rice grain are obtained. The ray parameter equation is substituted into the ellipsoidal space equation to solve for the intersection point of the incident light and the physical model of the rice grain. Based on the laws of reflection and refraction, the reflection direction and refraction direction of the light at the intersection point are calculated respectively to complete the physical constraint of the propagation path of the incident light during the interaction with the physical model of the rice grain. Then, the radiative transmission process of a single incident light beam in the physical model of the rice grain is analyzed. Combining the Lambert-Beer law, Fresnel law and the law of conservation of energy, analytical expressions for the total emissivity and total absorptivity when a single incident light beam interacts with the physical model of the rice grain are derived.

[0107] Please see Figure 4 The interaction process between light rays and ellipsoidal rice grains based on the physical constraints of the propagation path is as follows: Figure 4 As shown. Beam (Energy is 1) The reflectance when it reaches the surface of a rice grain from air (medium 1) is _____. Transmittance Energy is absorbed by two parts of the medium (transmittance during internal transmission). When the energy reaches the lower inner surface (medium 2), the remaining energy is The reflectivity inside the medium is Transmittance The energy after reflection is The energy after transmission is Light continues to travel within the medium until its intensity gradually decreases to zero.

[0108] In some embodiments, the step of analyzing the radiative transport process of a single incident light beam within a rice grain physical model, and deriving analytical expressions for the total emissivity and total absorptivity when the single incident light beam interacts with the rice grain physical model, by combining the Lambert-Beer law, Fresnel's law, and the law of conservation of energy, includes:

[0109] Step S201: Analyze the radiative transmission process of a single incident light beam in the physical model of rice grains, and determine the energy attenuation law and single transmittance of the single incident light beam propagating in the internal medium of the target rice grains by combining the Lambert-Beer law.

[0110] Step S202: Based on Fresnel's law, determine the reflectivity and transmittance of a single incident ray at different medium interfaces. By using the law of conservation of energy, clarify the energy distribution relationship of reflected energy, transmitted energy, and absorbed energy of the ray within the model, so as to derive analytical expressions for the total emissivity and total absorptivity when a single incident ray interacts with the physical model of the rice grain.

[0111] In a specific embodiment, based on the interaction process between the light beam and the physical model of rice grains described above, analytical expressions for the overall emissivity and overall absorptivity of a single beam of light propagating within the physical model of rice grains were derived:

[0112]

[0113] Here, represents the overall emissivity, which is the percentage of total energy of light that is reflected or penetrates to the outside of the rice grain. This represents the reflectivity of light reflected from medium 1 (air) back to medium 1; This represents the reflectivity of light reflected from medium 2 back to medium 2. This represents the transmittance of light when it travels from medium 1 (air) to medium 2 (rice grains); This represents the transmittance of light when it travels from medium 2 (rice grains) to medium 1 (air). Indicates the first ray of light Transmittance through the interior of rice grains.

[0114] Simplified and organized:

[0115]

[0116] Where t represents the transmittance of light as it passes through the internal medium of a rice grain in one pass:

[0117]

[0118] in, For incident energy, The remaining energy after propagating a certain distance.

[0119] Regarding the absorption of light by the internal medium of rice grains, assuming the internal medium of rice grains is uniform and there is no scattering interference in the direction of light propagation, according to the Lambert-Beer law, the energy of light decreases due to absorption when it propagates inside the rice grain:

[0120]

[0121] in, For incident energy, For the distance of transmission The remaining energy after is the absorption coefficient.

[0122] According to the Lambert-Beer law, the transmittance of light passing through the internal medium of a rice grain is related to the propagation path d, and the calculation formula includes:

[0123]

[0124] in, Indicates the first ray of light Transmittance that penetrates the interior of a rice grain; Indicates the first ray of light The distance traveled through the interior of each rice grain; This indicates the number of times light passes through the interior of a rice grain; This represents the absorption coefficient of the medium inside the rice grain.

[0125]

[0126] in, This represents the relative refractive index of medium 2 (rice grains) and medium 1 (air). .

[0127] Assuming that light has no absorption loss at the interface between two media, and only reflection and refraction occur, the relationship between reflectivity and transmittance is satisfied by the law of energy conservation:

[0128]

[0129] in, This is the reflectance of light incident from medium 1 (air) onto the interface of medium 2 (rice grains), which corresponds exactly to the "natural light reflectance R" in Fresnel's formula. According to Fresnel's law, It can be directly determined by the angle of incidence. and relative refractive index The calculation formula includes:

[0130] For s-polarized light:

[0131]

[0132] For p-polarized light:

[0133]

[0134] If the incident light is natural light. It can be represented as:

[0135]

[0136] in, The reflectivity of an interface represents natural light; This represents the reflectivity of s-polarized light (perpendicular to the plane of incidence); This represents the reflectivity of p-polarized light (parallel to the incident plane).

[0137] Combining the above equations, we can obtain the analytical expression for the overall emissivity of a single beam of light:

[0138]

[0139] in, It represents the overall emissivity of a single beam of light; This represents the reflectivity of s-polarized light (perpendicular to the plane of incidence); This represents the reflectivity of p-polarized light (parallel to the incident plane). This indicates the transmittance of light when it first passes through the internal medium of a rice grain. This represents the reflectivity of light reflected from medium 2 back to medium 2. This represents the transmittance of light when it travels from medium 1 (air) to medium 2 (rice grains); It represents the transmittance of light when it travels from medium 2 (rice grains) to medium 1 (air).

[0140] According to the law of conservation of energy, the overall absorption rate of an odd number of rays propagating inside a medium can be derived as follows:

[0141]

[0142] in, It represents the overall absorption rate, which is the percentage of total energy absorbed by the internal medium of rice grains during the transmission of light within the grains.

[0143] Based on the above process, the main parameters of the rice grain radiative transfer model (RGRT-MODEL) are shown in Table 1.

[0144] Table 1. Main parameters of the rice grain radiative transfer model (RGRT-MODEL)

[0145]

[0146] Note: In represents input parameters; Int represents intermediate parameters; Out represents output parameters.

[0147] Step S30: Using the geometric dimensions, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters, a two-stage optimization strategy combining global search and local optimization is adopted. With the measured reflectance spectrum and measured transmittance spectrum as constraints, the optimal combination of model parameters matching the measured reflectance spectrum and measured transmittance spectrum is solved in reverse to complete the calibration of the key input parameters of the model.

[0148] The radiative transfer process of a single incident light beam within a rice grain physical model was analyzed. Combining the Lambert-Beer law, Fresnel's law, and the law of conservation of energy, analytical expressions for the total emissivity and total absorptivity of the single incident light beam interacting with the rice grain physical model were derived. Using the geometric dimensions, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters, a two-stage optimization strategy combining global search and local optimization was employed. With measured reflectance and transmittance spectra as constraints, the optimal combination of model parameters matching the measured reflectance and transmittance spectra was solved in reverse to calibrate the key input parameters of the model.

[0149] In some embodiments, the steps of using the geometric dimensions, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters, employing a two-stage optimization strategy combining global search and local optimization, and using measured reflectance and transmittance spectra as constraints, to inversely solve for the optimal combination of model parameters matching the measured reflectance and transmittance spectra, include:

[0150] Step S301: Using the geometric size parameters, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters, construct objective functions with the normalized squared error between the simulated reflectance spectrum and the measured reflectance spectrum, and between the simulated transmittance spectrum and the measured transmittance spectrum as error terms, and introduce regularization terms to constrain the range of parameter values.

[0151] Step S302: The objective function is solved by a two-stage optimization strategy that combines global search with local optimization. The differential evolution algorithm is used to traverse the global range of parameters to initially locate the optimal parameter interval.

[0152] Step S303: Using the optimal parameter range as the initial value, perform local refinement and adjustment using the limited memory Broyden-Fletcher-Goldfarb-Shannon band boundary constraint algorithm. Independently invert each spectral band and evaluate the goodness of fit with the coefficient of determination. Select the parameter result with the highest goodness of fit as the parameter value of the corresponding band. Solve in reverse to obtain the optimal model parameter combination that matches the measured reflectance spectrum and measured transmittance spectrum of the target rice grain.

[0153] Step S40: Based on the preset ray tracing algorithm, initialize the geometric size parameters of the rice grain physical model and the optical parameters of the incident ray, uniformly sample the incident points on the surface of the rice grain physical model, iteratively trace the propagation path of the internal ray of the rice grain physical model segment by segment, and calculate the energy attenuation value of each segment of the propagation path of the internal ray.

[0154] Using the geometric dimensions, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters, a two-stage optimization strategy combining global search and local optimization is adopted. With the measured reflectance spectrum and measured transmittance spectrum as constraints, the optimal combination of model parameters matching the measured reflectance spectrum and measured transmittance spectrum is solved in reverse to complete the calibration of the key input parameters of the model. Based on a preset ray tracing algorithm, the geometric dimensions of the rice grain physical model and the optical parameters of the incident light are initialized. The incident points are uniformly sampled on the surface of the rice grain physical model, and the propagation path of the internal light rays of the rice grain physical model is iteratively traced segment by segment, and the energy attenuation value of each segment of the internal light ray propagation path is calculated.

[0155] In some embodiments, the steps of initializing the geometric parameters of the rice grain physical model and the optical parameters of the incident light rays based on a preset ray tracing algorithm, uniformly sampling the incident points on the surface of the rice grain physical model, iteratively tracing the propagation path of the internal light rays of the rice grain physical model segment by segment, and calculating the energy attenuation value of each segment of the propagation path of the internal light rays include:

[0156] Step S401: Call the preset ray tracing algorithm to initialize the geometric size parameters and incident light optical parameters of the rice grain physical model and form a parameter dataset. On the ellipsoidal surface of the rice grain physical model, uniformly sample the incident points at a preset density and generate a set of incident point coordinates.

[0157] Step S402: Based on the parameter dataset and the incident point coordinate set, the propagation path of the internal light rays formed after the incident light rays enter the rice grain physical model is iteratively tracked segment by segment. At the same time, the energy attenuation value corresponding to each segment of the propagation path of the internal light rays is calculated in combination with the Lambert-Beer law to construct an energy attenuation dataset.

[0158] Step S50: Based on the energy attenuation value, determine the remaining energy after energy attenuation through each propagation path. When total internal reflection does not occur during the emission stage of the light rays inside the rice grain physical model, allocate the transmitted energy according to the emission direction of the internal light rays, accumulate the reflected energy and transmitted energy corresponding to all incident light rays, calculate the absorbed energy of the light rays by the rice grain physical model in combination with the law of conservation of energy, correct the numerical error, and output the overall reflectivity, transmittance and absorptivity of the rice grain physical model to complete the construction of the rice grain radiative transfer model.

[0159] Based on a preset ray tracing algorithm, the geometric parameters of the rice grain physical model and the optical parameters of the incident rays are initialized. Incident points are uniformly sampled on the surface of the rice grain physical model. The propagation path of the internal rays of the rice grain physical model is iteratively traced segment by segment. After calculating the energy attenuation value of each segment of the propagation path, the remaining energy after energy attenuation through each segment of the propagation path is determined based on the energy attenuation value. When total internal reflection does not occur in the outgoing stage of the internal rays of the rice grain physical model, the transmitted energy is allocated according to the outgoing direction of the internal rays. The reflected energy and transmitted energy corresponding to all incident rays are accumulated. The energy absorption energy of the rice grain physical model is calculated by combining the law of conservation of energy and the numerical error is corrected. The overall reflectivity, transmittance and absorptivity of the rice grain physical model are output to complete the construction of the rice grain radiative transfer model.

[0160] In some embodiments, the remaining energy after energy attenuation through each propagation path is determined based on the energy attenuation value. When total internal reflection does not occur during the emission stage of the light rays inside the rice grain physical model, the transmitted energy is allocated according to the emission direction of the internal light rays. The reflected and transmitted energy corresponding to all incident light rays are accumulated. The energy absorbed by the rice grain physical model for light is calculated using the law of conservation of energy, and numerical errors are corrected. The overall reflectivity, transmittance, and absorptivity of the rice grain physical model are then output.

[0161] Step S501: Call the energy attenuation value of the energy attenuation dataset, and determine the remaining energy dataset of the internal light rays of the rice grain physical model after energy attenuation through each propagation path by iterative calculation;

[0162] Step S502: Based on the remaining energy dataset, determine the total internal reflection state of the internal light during the emission stage. If it is determined that no total internal reflection has occurred, then complete the quantitative allocation of the transmitted energy according to the emission direction of the internal light and generate a transmitted energy dataset.

[0163] Step S503: Synchronously collect the reflected energy dataset corresponding to all incident rays and accumulate the transmitted energy dataset. Calculate the absorbed energy of the rice grain physical model based on the law of conservation of energy and correct numerical errors for various energy data. Finally, calculate and output the reflectivity, transmittance, and absorptivity of the rice grain physical model based on the corrected energy data.

[0164] Specifically, the rice grain radiative transfer model (RGRT-MODEL) has several key input parameters, namely... , , , and. Among them, , , The geometric dimensions of the rice grain physical model can be directly measured. Regarding the spectral absorption coefficient of the internal medium of the rice grain... and the refractive index of the internal medium of rice grains The approach employs a "joint inversion" strategy. Based on a two-stage optimization strategy combining global search and local optimization, and using measured reflectance and transmittance spectra as constraints, the optimal parameter combination that matches the measured spectra is solved in reverse.

[0165] First, an objective function is constructed, using the "normalized squared error between simulated and measured spectra" as the error term (to eliminate magnitude bias and balance the fit between reflectance and transmittance). Simultaneously, a regularization term is introduced to constrain the parameter range (balancing fitting accuracy and the physical rationality of the inversion results). Second, a two-stage optimization strategy combining global search and local optimization is employed. First, a differential evolution algorithm is used to traverse the global parameter range, refining iterations and population size to initially locate a relatively optimal interval. Then, using this result as the initial value, a limited-memory Broyden-Fletcher-Goldfarb-Shannon band boundary constraint algorithm (L-BFGS-B algorithm) is used for refinement and adjustment to improve inversion accuracy. Finally, each spectral band is inverted independently, and the coefficient of determination (…) is used to determine the optimal parameters. To assess the goodness of fit, select... The highest result is used as the final parameter for that band to ensure the reliability of the inversion.

[0166] Furthermore, given that the overall expression for the optical properties of a rice grain model is difficult to solve directly using analytical methods, and that ray tracing algorithms are well-suited for solving the optical parameters of such complex geometric models, this application combines ray tracing algorithms to track the interaction between light rays and the rice grain model throughout the entire process. By determining the final emission direction of the light rays, it achieves statistical calculations of the total reflectivity and total transmittance.

[0167] First, the geometric and optical parameters of the rice grains are initialized, and incident points are uniformly sampled on the surface of the rice grains. For each ray of light incident from the air onto the rice grain, the normal vector and incident angle of the incident point are calculated, and the initial reflectivity and transmittance are obtained. Then, the propagation path of the light rays inside the rice grain is iteratively tracked, and the energy attenuation of each path segment is calculated. During the stage where the light rays exit from inside the rice grain towards the air, it is determined whether total internal reflection occurs: if total internal reflection occurs, the light rays are 100% reflected back into the rice grain; if not, the transmitted energy is allocated according to the exit point position (if the exit direction of the light rays is upward, reflection is included; if the exit direction of the light rays is downward, transmission is included). Finally, the reflected energy and transmitted energy of all light rays are accumulated, the absorbed energy is calculated, and numerical errors are corrected to output the overall reflectivity, transmittance, and absorptivity of the rice grain.

[0168] In some embodiments, to verify the spectral simulation performance of the RGRT-MODEL model, this application conducts evaluation experiments based on a self-built dataset. The experiments selected grains from six different rice varieties, with one sample randomly drawn from each variety for spectral simulation. Please refer to [link to relevant documentation]. Figure 5 , Figure 5 Images (a) and (b) in the figure present the measured reflectance curve and the simulated reflectance curve of the grain sample, respectively. Figure 5 (c) and (d) in the figure show the measured transmittance curve and the simulated transmittance curve of the grain sample, respectively.

[0169] Comparative analysis results show that the simulated spectral curves and measured spectral curves of different rice varieties have a high degree of consistency in their overall trends, and the peak and valley shapes and band fluctuation patterns are highly similar. Although there is a slight jitter in the simulated curves, the RGRT-MODEL model can accurately capture the core characteristics of the changes in reflectance and transmittance of rice grains with different bands.

[0170] Furthermore, combining the quantitative analysis results in Table 2, it can be seen that the reflectance and transmittance of different rice grain samples all exhibit relatively small root mean square error (RMSE) values ​​in the three spectral ranges of 400-700 nm, 700-1000 nm, and 400-1000 nm. Taking sample 401 as an example, its RMSE for reflectance in the 400-700 nm band is 0.014, and its RMSE for transmittance in the same band is even lower at 0.002. The RMSE of the remaining samples in each band also remains at a low level. This quantitative result corroborates the qualitative comparison conclusion of the spectral curves above, intuitively reflecting that the deviation between the simulated spectrum and the measured spectrum is small, further verifying the spectral simulation accuracy of the RGRT-MODEL model from the perspective of error quantification.

[0171] Table 2. Root mean square error (RMSE) of measured and simulated spectra in different bands.

[0172]

[0173] Furthermore, to evaluate the accuracy of the RGRT-MODEL model in simulating the optical parameters of rice grains, this experiment focused on analyzing the root mean square error (RMSE) distribution characteristics of three core optical parameters—reflectivity, transmittance, and absorptivity—within three key wavelength ranges: 400–700 nm, 700–1000 nm, and 400–1000 nm. The relevant results are presented in histogram form. Please refer to [link / reference]. Figure 6 ,in, Figure 6 This is a histogram showing the error distribution of different optical properties of rice grains. Figure 6 In the table, (a) to (c) correspond to the percentage of root mean square error (RMSE) for reflectivity in the 400-1000nm, 400-700nm, and 700-1000nm wavelength bands, respectively. Figure 6 In the table, (d) to (f) and (g) to (i) represent the corresponding relationships between transmittance and absorbance in the same spectral range, respectively. Figure 6 The horizontal axis represents the error value of each optical parameter, and the vertical axis represents the sample proportion of the corresponding error value.

[0174] From the error distribution pattern, the root mean square error (RMSE) of reflectance and transmittance is mainly concentrated near 0 in all spectral bands, indicating not only a small overall error but also a concentrated distribution. The RMSE of transmittance shows a significant peak near 0, indicating strong error concentration. Although the simulation error distribution of reflectance and absorptivity is slightly more dispersed than that of transmittance, statistical results show that the simulation errors of optical characteristics for the vast majority of samples are controlled within 0.05, without significant large error deviations. In summary, the simulation errors of the RGRT-MODEL model are generally concentrated, accurately replicating the reflection, transmission, and absorption optical characteristics of rice grains, thus verifying the model's reliability in optical parameter simulation tasks.

[0175] Furthermore, Table 3 presents the statistical results of the root mean square error (RMSE) of the measured and simulated spectra of optical properties of rice grains from different varieties. It is evident that there are significant differences in RMSE between varieties and between different wavelength bands. Looking at the performance by wavelength band, the simulated RMSE for reflectance is within 0.275 in all bands, the RMSE for transmittance in all bands does not exceed 0.338, and the maximum RMSE for absorptivity in all bands is 0.414. Overall, the simulation errors for these three parameters are at a low level. In particular, the mean RMSE for transmittance in each band is relatively lower, indicating that the model has better simulation accuracy for transmittance. From a variety-specific analysis, varieties 401 and 2512 generally have higher maximum and mean RMSE values ​​for reflectance, transmittance, and absorptivity in all bands than other varieties, indicating relatively weaker simulation performance. Varieties 2022 and 2509 showed even better simulation performance, with their mean root mean square error (RMSE) values ​​remaining at a low level across all wavelength bands. Overall statistical results indicate that, despite differences in variety and wavelength band, the RMSE values ​​for reflectance, transmittance, and absorptivity of rice grains from all varieties are generally small in the three wavelength bands of 400–700 nm, 700–1000 nm, and 400–1000 nm. This further verifies the reliability of the RGRT-MODEL model in simulating the optical properties of rice grains from different varieties.

[0176] Table 3. Statistical analysis of the root mean square error (RMSE) between measured and simulated values ​​of rice grain spectra for different varieties.

[0177]

[0178] For further details, please refer to Figure 7 To evaluate the stability of the simulated optical parameters, this application analyzed the root mean square error (RMSE) of reflectance, transmittance, and absorptivity in three different wavelength ranges (400–700 nm, 700–1000 nm, and 400–1000 nm) as a function of sample size. The results are presented as a scatter plot (e.g., Figure 7The scatter plots show that the model's overall simulation of rice grain characteristics is quite good. In all subplots, the root mean square error (RMSE) values ​​are mostly concentrated in the range of 0 to 0.1, with only a very few points having an RMSE close to 0.2, indicating that the model's predicted values ​​deviate little from the actual values, and the simulation error for most samples is at a low level. Within a sample size range of 0 to 180, the RMSE does not show a significant upward or downward trend with increasing sample size, but rather is relatively evenly distributed in the low error range, demonstrating good model stability. It does not experience large fluctuations in error due to changes in sample size, maintaining relatively stable simulation accuracy under different sample sizes. While there is some overlap in the RMSE distribution of the points corresponding to different rice grain varieties represented by different colors, there are also subtle differences. Points for some varieties are more concentrated in the low RMSE region, while a few varieties have slightly higher RMSE values ​​for individual data points. This may be due to differences in the structure and composition of rice grains among different varieties, but overall, the errors for all varieties are at an acceptable low level.

[0179] As can be seen from the above embodiments, compared with the prior art, this application addresses the problems of traditional radiative transfer models based on leaves and canopies, whose design logic is mostly adapted to the structural characteristics of "flat and thin" leaves or "uniformly layered" canopies, making it difficult to directly meet the modeling requirements of grain optical properties. This application has, but is not limited to, the following beneficial effects:

[0180] Firstly, existing crop radiative transfer models (RTMs) all rely on the core assumptions of "flat, thin leaf" and "uniformly layered canopy," which are only suitable for optical simulations at the leaf and canopy scales. However, rice grains have an ellipsoidal, three-dimensional, convex structure with surface curvature varying with location, leading to multi-angle scattering and refraction of incident light. The heterogeneity of internal component distribution also complicates the radiative transfer path, making traditional model frameworks unsuitable for direct application. This application abstracts the rice grain into an ideal ellipsoidal structure. By constructing ellipsoidal spatial equations and ray parameter equations, it solves for the intersection points of light rays and grains and constrains the propagation path based on the laws of reflection and refraction. This specifically adapts to the three-dimensional structural characteristics of the grain, achieving for the first time a dedicated modeling of radiative transfer at the organ scale of rice grains. This fills the technical gap in grain-specific modeling and eliminates spectral simulation biases caused by structural mismatches in traditional models.

[0181] Secondly, this application analyzes the radiative transmission process of a single beam of light within the grain, integrating the Lambert-Beer law (energy decay), Fresnel's law (interface transmission and reflection), and the law of conservation of energy (energy distribution) to derive analytical expressions for the total emissivity and total absorptivity of the interaction between the single beam of light and the grain. This reveals the formation principle of the spectral signal of rice grains from a physical perspective, upgrading grain spectral simulation from "statistical correlation" to "physical mechanism driven," significantly improving the model's cross-variety and cross-scenario portability, and avoiding excessive reliance on samples by data-driven models.

[0182] Thirdly, this application proposes a two-stage optimization strategy combining global search and local optimization, using grain geometry, internal absorption coefficient, and medium refractive index as key input parameters. First, a differential evolution algorithm is used to traverse the global range of parameters to locate the optimal interval. Then, a limited-memory Broyden-Fletcher-Goldfarb-Shannon band boundary constraint algorithm (L-BFGS-B algorithm) is used for local refinement. Simultaneously, measured reflectance and transmittance spectra are used as constraints to independently invert each spectral band, and the goodness of fit is evaluated using the coefficient of determination. This strategy avoids the problem of single local optimization algorithms easily getting trapped in local optima, and ensures the matching between parameters and actual grains through measured spectral constraints. It achieves accurate joint inversion of the core optical parameters of the grain, laying a high-precision parameter foundation for subsequent spectral simulations.

[0183] Fourth, this application relies on a ray tracing algorithm. By initializing the grain and ray parameters, uniformly sampling the incident point, iteratively tracing the internal ray path segment by segment and calculating energy attenuation, combining the remaining energy to determine the total internal reflection state and distributing the transmitted energy according to the emission direction, and finally accumulating the transmitted and reflected energy of all rays, calculating the absorbed energy, and correcting numerical errors, it achieves accurate statistical calculation of the overall optical properties of the grain. This method avoids the limitations of analytical methods. Through refined simulation on a ray-by-ray and path-by-path basis, it significantly reduces spectral simulation errors. The simulated spectra of different grain varieties are highly consistent with the peak and valley shapes and band fluctuations of the measured spectra. The RMSE values ​​are generally at a low level, and the simulated RMSE of transmittance is even as low as 0.001, which fully demonstrates the accuracy and reliability of the spectral simulation of the model in this application.

[0184] Fifth, this application achieves precise simulation of rice grain spectra, effectively analyzing the contribution ratio of grains to the canopy spectrum, eliminating monitoring interference caused by this spectral component, and making canopy spectrum analysis more accurate in the later stages of rice growth. At the same time, the optical properties of grains are closely related to quality parameters such as grain maturity, starch, and protein content. The model of this application can serve as the core physical basis for non-destructive inversion of grain quality parameters, providing key technical support for precise monitoring of rice throughout its entire growth period and efficient evaluation of yield and quality, and contributing to precise water and fertilizer regulation in rice cultivation and sustainable agricultural development.

[0185] Please see Figure 8 This application provides a rice grain radiative transfer model construction device, comprising a physical interaction path constraint module 1100, an optical characteristic analysis module 1200, a key parameter calibration module 1300, an energy attenuation calculation module 1400, and a radiative transfer model construction module 1500. The physical interaction path constraint module 1100 is configured to obtain the ellipsoidal space equation corresponding to the physical model of the target rice grain, and the ray parameter equation of the incident light ray of the target rice grain. It substitutes the ray parameter equation into the ellipsoidal space equation to solve for the intersection point of the incident light ray and the rice grain physical model, and calculates the reflection direction and refraction direction of the light ray at the intersection point based on the laws of reflection and refraction, respectively, to complete the physical constraint of the propagation path of the incident light ray during the interaction with the rice grain physical model; the optical characteristic analysis module 1200... This system is designed to analyze the radiative transfer process of a single incident light beam within a rice grain physical model. By combining the Lambert-Beer law, Fresnel's law, and the law of conservation of energy, analytical expressions for the total emissivity and total absorptivity of the single incident light beam interacting with the rice grain physical model are derived. The key parameter calibration module 1300 is configured to use the geometric dimensions, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters. It employs a two-stage optimization strategy combining global search and local optimization, using measured reflectance and measured transmittance as the calibration parameters. Using the spectrum as a constraint, the optimal combination of model parameters matching the measured reflectance spectrum and the measured transmittance spectrum is solved in reverse to complete the calibration of the key input parameters of the model. The energy attenuation calculation module 1400 is set to initialize the geometric size parameters of the rice grain physical model and the optical parameters of the incident light based on a preset ray tracing algorithm. The incident points are uniformly sampled on the surface of the rice grain physical model, and the propagation path of the internal light of the rice grain physical model is iteratively traced segment by segment, and the energy attenuation value of each segment of the propagation path of the internal light is calculated. The radiation transfer model construction module 1500 is set to determine the remaining energy after energy attenuation through each segment of the propagation path based on the energy attenuation value. When total internal reflection does not occur in the emission stage of the internal light of the rice grain physical model, the transmitted energy is allocated according to the emission direction of the internal light. The reflected energy and transmitted energy corresponding to all incident light are accumulated. The energy absorption energy of the rice grain physical model is calculated by combining the law of conservation of energy and the numerical error is corrected. The overall reflectance, transmittance and absorptivity of the rice grain physical model are output to complete the construction of the rice grain radiation transfer model.

[0186] Based on any embodiment of this application, please refer to Figure 9 Another embodiment of this application also provides an electronic device, which can be implemented by a computer device, such as... Figure 9The diagram shows the internal structure of a computer device. The computer device includes a processor, a computer-readable storage medium, a memory, and a network interface connected via a system bus. The computer-readable storage medium stores an operating system, a database, and computer-readable instructions. The database may store control information sequences. When executed by the processor, the computer-readable instructions enable the processor to implement a method for constructing a rice grain radiative transfer model. The processor provides computational and control capabilities, supporting the operation of the entire computer device. The memory stores computer-readable instructions, which, when executed by the processor, enable the processor to execute the rice grain radiative transfer model construction method of this application. The network interface of the computer device is used for communication with a terminal. Those skilled in the art will understand that… Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0187] In this embodiment, the processor is used to execute... Figure 8 The specific functions of each module are described, and the memory stores the program code and various data required to execute these modules. The network interface is used for data transmission between the user terminal and the server. In this embodiment, the memory stores the program code and data required to execute all modules in the rice grain radiative transfer model construction device of this application, and the server can call the server's program code and data to execute the functions of all modules.

[0188] This application also provides a storage medium storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the steps of the rice grain radiative transfer model construction method described in any embodiment of this application.

[0189] This application also provides a computer program product, including a computer program / instructions that, when executed by one or more processors, implement the steps of the rice grain radiative transfer model construction method described in any embodiment of this application.

[0190] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. This computer program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The aforementioned storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0191] The above description is only a partial embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for constructing a rice grain radiative transport model, characterized in that, include: Obtain the ellipsoidal space equation corresponding to the physical model of the target rice grain, and the ray parameter equation of the incident light of the target rice grain. Substitute the ray parameter equation into the ellipsoidal space equation to solve for the intersection point of the incident light and the physical model of the rice grain. Calculate the reflection direction and refraction direction of the light at the intersection point based on the laws of reflection and refraction, respectively, to complete the physical constraint of the propagation path of the incident light during the interaction between the incident light and the physical model of the rice grain. The radiative transmission process of a single incident light beam within a rice grain physical model is analyzed. By combining the Lambert-Beer law, Fresnel's law, and the law of conservation of energy, analytical expressions for the total emissivity and total absorptivity when a single incident light beam interacts with the rice grain physical model are derived. Using the geometric dimensions, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters, a two-stage optimization strategy combining global search and local optimization is adopted. With the measured reflectance spectrum and measured transmittance spectrum as constraints, the optimal combination of model parameters matching the measured reflectance spectrum and measured transmittance spectrum is solved in reverse to complete the calibration of the key input parameters of the model. Based on a preset ray tracing algorithm, the geometric dimensions of the rice grain physical model and the optical parameters of the incident ray are initialized. The incident points are uniformly sampled on the surface of the rice grain physical model. The propagation path of the internal ray of the rice grain physical model is iteratively traced segment by segment, and the energy attenuation value of each segment of the propagation path of the internal ray is calculated. Based on the energy attenuation value, the remaining energy after energy attenuation through each propagation path is determined. When total internal reflection does not occur during the emission stage of the light rays inside the rice grain physical model, the transmitted energy is allocated according to the emission direction of the light rays inside the model. The reflected energy and transmitted energy corresponding to all incident light rays are accumulated. The energy absorbed by the rice grain physical model is calculated by combining the law of conservation of energy and the numerical error is corrected. The overall reflectivity, transmittance and absorptivity of the rice grain physical model are output to complete the construction of the rice grain radiative transfer model.

2. The method for constructing a rice grain radiative transfer model according to claim 1, characterized in that, The steps of substituting the ray parameter equation into the ellipsoidal space equation to solve for the intersection point of the incident ray and the physical model of the rice grain include: Substituting the incident ray parameter equation into the ellipsoidal space equation corresponding to the rice grain physical model, a quadratic equation for the ray propagation distance is derived, and the coefficients of the quadratic equation for the ray propagation distance are calculated. The discriminant of the quadratic equation for the propagation distance of the light ray is calculated. Based on the discriminant, it is determined whether there is an intersection point between the incident light ray and the physical model of the rice grain. If the discriminant is greater than zero, then there is an intersection point. The positive root of the quadratic equation is taken, and the coordinates of the near-side intersection point corresponding to the near-side intersection point between the incident light ray and the physical model of the rice grain are calculated to complete the solution of the intersection point between the incident light ray and the physical model of the rice grain.

3. The method for constructing a rice grain radiative transfer model according to claim 1, characterized in that, The steps of calculating the reflection and refraction directions of the light rays at the intersection point based on the laws of reflection and refraction, to complete the physical constraints on the propagation path of the incident light rays during the interaction with the rice grain physical model, include: The normal vector of the surface of the rice grain physical model at the near-side intersection is calculated and normalized. Based on the law of reflection, the reflection direction of the light at the near-side intersection is calculated by the incident light direction vector and the normalized normal vector. The normal vector of the surface of the rice grain physical model represents the vector at any near-side intersection of the surface of the rice grain physical model constructed based on the idealized ellipsoidal structure of the rice grain, which is perpendicular to the tangent plane of the intersection point and points to the outside of the rice grain physical model. Based on the law of refraction and the relationship between the refractive index of the internal medium, the angle of incidence and the angle of refraction, the direction of refraction of the light at the near intersection is calculated to complete the physical constraint of the propagation path of the incident light during the interaction between the incident light and the physical model of rice grains.

4. The method for constructing a rice grain radiative transfer model according to claim 1, characterized in that, The steps for analyzing the radiative transport process of a single incident light beam within a rice grain physical model, and deriving analytical expressions for the total emissivity and total absorptivity when a single incident light beam interacts with the rice grain physical model, by combining the Lambert-Beer law, Fresnel's law, and the law of conservation of energy, include: The radiative transmission process of a single incident light beam in a physical model of a rice grain was analyzed, and the energy attenuation law and single-pass transmittance of the single incident light beam in the internal medium of the target rice grain were determined by combining the Lambert-Beer law. Based on Fresnel's law, the reflectivity and transmittance of a single incident ray at different medium interfaces are determined. The energy distribution relationship of reflected energy, transmitted energy, and absorbed energy of the ray within the model is clarified by the law of conservation of energy, so as to derive analytical expressions for the total emissivity and total absorptivity when a single incident ray interacts with the physical model of rice grains.

5. The method for constructing a rice grain radiative transfer model according to claim 1, characterized in that, Using the geometric dimensions, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters, a two-stage optimization strategy combining global search and local optimization is employed. This involves using measured reflectance and transmittance spectra as constraints to inversely solve for the optimal combination of model parameters matching the measured reflectance and transmittance spectra. The steps include: Using the geometric dimensions, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters, objective functions are constructed with the normalized squared error between the simulated reflectance spectrum and the measured reflectance spectrum, and between the simulated transmittance spectrum and the measured transmittance spectrum as error terms, and regularization terms are introduced to constrain the range of parameter values. The objective function is solved by a two-stage optimization strategy that combines global search with local optimization. The differential evolution algorithm is used to traverse the global range of parameters to initially locate the optimal parameter interval. Using the aforementioned optimal parameter range as initial values, local refinement adjustments are made using the limited-memory Broyden-Fletcher-Goldfarb-Shannon band boundary constraint algorithm. Each spectral band is independently inverted, and the goodness of fit is evaluated using the coefficient of determination. The parameter result with the highest goodness of fit is selected as the parameter value for the corresponding band. The optimal model parameter combination that matches the measured reflectance spectrum and measured transmittance spectrum of the target rice grain is obtained by reverse solving.

6. The method for constructing a rice grain radiative transfer model according to claim 1, characterized in that, Based on a preset ray tracing algorithm, the steps of initializing the geometric parameters of the rice grain physical model and the optical parameters of the incident ray, uniformly sampling the incident points on the surface of the rice grain physical model, iteratively tracing the propagation path of the internal ray of the rice grain physical model segment by segment, and calculating the energy attenuation value of each segment of the internal ray propagation path include: The preset ray tracing algorithm is invoked to initialize the geometric dimension parameters and incident ray optical parameters of the rice grain physical model and form a parameter dataset. Incident points are uniformly sampled on the ellipsoidal surface of the rice grain physical model at a preset density and an incident point coordinate set is generated. Based on the parameter dataset and the incident point coordinate set, the propagation path of the internal light rays formed after the incident light rays enter the rice grain physical model is iteratively tracked segment by segment. At the same time, the energy attenuation value corresponding to each segment of the internal light ray propagation path is calculated in combination with the Lambert-Beer law to construct an energy attenuation dataset.

7. The method for constructing a rice grain radiative transfer model according to claim 6, characterized in that, Based on the energy attenuation value, the remaining energy after energy attenuation through each propagation path is determined. When total internal reflection does not occur during the emission stage of the light rays inside the rice grain physical model, the transmitted energy is allocated according to the emission direction of the internal light rays. The reflected and transmitted energy corresponding to all incident light rays are accumulated. The energy absorbed by the rice grain physical model is calculated using the law of conservation of energy, and numerical errors are corrected. The overall reflectivity, transmittance, and absorptivity of the rice grain physical model are then output. The steps include: The energy decay values ​​of the energy decay dataset are called, and the remaining energy dataset after the internal light rays of the rice grain physical model have decayed through each propagation path is determined by iterative calculation. Based on the remaining energy dataset, the total internal reflection state of the internal light during the emission stage is determined. If it is determined that no total internal reflection has occurred, the transmission energy is quantified and allocated according to the emission direction of the internal light and a transmission energy dataset is generated. The reflected energy dataset corresponding to all incident rays is collected synchronously and accumulated with the transmitted energy dataset. Based on the law of conservation of energy, the absorbed energy of the rice grain physical model is calculated and numerical error correction is performed on various energy data. Finally, the reflectivity, transmittance and absorptivity of the rice grain physical model are calculated and output based on the corrected energy data.

8. A device for constructing a rice grain radiative transport model, characterized in that, include: The physical interaction path constraint module is configured to obtain the ellipsoidal space equation corresponding to the physical model of the target rice grain, and the ray parameter equation of the incident light of the target rice grain. The ray parameter equation is substituted into the ellipsoidal space equation to solve for the intersection point of the incident light and the physical model of the rice grain. Based on the laws of reflection and refraction, the reflection direction and refraction direction of the light at the intersection point are calculated respectively to complete the physical constraint of the propagation path of the incident light during the interaction between the incident light and the physical model of the rice grain. The optical properties analysis module is set to analyze the radiative transmission process of a single incident light beam in the physical model of rice grains. Combining the Lambert-Beer law, Fresnel law and the law of conservation of energy, analytical expressions for the total emissivity and total absorptivity when a single incident light beam interacts with the physical model of rice grains are derived. The key parameter calibration module is configured to use the geometric size parameters, internal medium spectral absorption coefficient, and internal medium refractive index of the rice grain physical model as key input parameters. It adopts a two-stage optimization strategy combining global search and local optimization, and uses the measured reflectance spectrum and measured transmittance spectrum as constraints to solve in reverse the optimal combination of model parameters that matches the measured reflectance spectrum and measured transmittance spectrum, so as to complete the calibration of the key input parameters of the model. The energy attenuation calculation module is configured to initialize the geometric parameters of the rice grain physical model and the optical parameters of the incident light rays based on a preset ray tracing algorithm, uniformly sample the incident points on the surface of the rice grain physical model, iteratively trace the propagation path of the internal light rays of the rice grain physical model segment by segment, and calculate the energy attenuation value of each segment of the propagation path of the internal light rays. The radiative transfer model construction module is configured to determine the remaining energy after energy attenuation through each propagation path based on the energy attenuation value. When total internal reflection does not occur during the emission stage of light rays inside the rice grain physical model, the transmitted energy is allocated according to the emission direction of the internal light rays. The reflected and transmitted energy corresponding to all incident light rays are accumulated. The energy absorption energy of the rice grain physical model is calculated by combining the law of conservation of energy and the numerical error is corrected. The overall reflectivity, transmittance and absorptivity of the rice grain physical model are output to complete the construction of the rice grain radiative transfer model.

9. An electronic device comprising a central processing unit and a memory, characterized in that, The central processing unit is used to invoke and run a computer program stored in the memory to perform the steps of the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores, in the form of computer-readable instructions, a computer program implemented according to any one of claims 1 to 7, which, when invoked by a computer, executes the steps included in the corresponding method.

Images