Rice aboveground biomass inversion method and device based on component analysis, medium and equipment

Abstract

The application discloses a rice above-ground biomass inversion method and device based on component analysis, a medium and equipment. The method comprises the following steps: acquiring a multispectral image and a rice RGB image; extracting multiple types of spectral features, texture features and canopy height data; constructing a leaf biomass inversion model, inputting multiple types of spectral features and multiple types of texture features, and screening out first key type spectral features and first key type texture features; establishing a spike biomass inversion model to determine multiple spectral feature correlations, multiple texture feature correlations respectively, and screen out second key type spectral features and second key type texture features; fitting the canopy height data and the stem biomass, and determining the use of the canopy height data to predict the rice above-ground biomass according to the fitting result; and inputting the above-mentioned features and the canopy height data into a random forest model to obtain the rice above-ground biomass. The application can improve the prediction accuracy of the above-ground biomass.

Description

Technical Field

[0001] This application relates to the field of aboveground biomass prediction technology, and in particular to a method, apparatus, storage medium and electronic equipment for rice aboveground biomass inversion based on component analysis. Background Technology

[0002] Aboveground biomass (AGB) is a crucial indicator for monitoring crop growth and field management. Defined as the total dry weight of all living biomass above the soil surface, it reflects crop growth status and primary GDP. AGB is a vital basis for crop yield estimation and is directly related to grain production. Rapid and accurate assessment of crop AGB is essential for predicting grain yield and improving field management strategies.

[0003] Currently, the relevant methods for predicting crop aboveground biomass generally rely on calculating vegetation indices. However, due to the complexity of biomass information in various crop components and the fact that crop organs can overlap, the accuracy of methods using spectral information and vegetation indices to predict crop aboveground biomass is relatively low. Summary of the Invention

[0004] This application provides a method, apparatus, storage medium, and electronic device for inverting rice aboveground biomass based on component analysis, which can improve the accuracy of predicting crop aboveground biomass.

[0005] This application provides a method for inverting aboveground biomass of rice based on component analysis, including:

[0006] Images were acquired from rice-growing areas to obtain multispectral images and RGB images of rice.

[0007] Multiple types of spectral features and multiple types of texture features are extracted from the multispectral image, and canopy height data is extracted from the rice RGB image.

[0008] A leaf biomass inversion model was constructed to screen features for predicting leaf biomass; a spike biomass inversion model was constructed to screen features for predicting spike biomass; and a stem biomass inversion model was constructed to determine features for predicting stem biomass.

[0009] The process of constructing a leaf biomass inversion model for feature selection to predict leaf biomass includes: constructing a leaf biomass inversion model, inputting the multiple types of spectral features and the multiple types of texture features into the leaf biomass inversion model to invert leaf biomass, and selecting a first key type of spectral feature and a first key type of texture feature based on the leaf biomass inversion results;

[0010] The construction of a panicle biomass inversion model for feature screening to predict panicle biomass includes: establishing a panicle biomass inversion model to determine the correlation between multiple spectral features of the multiple types and panicle biomass, determining the correlation between multiple texture features of the multiple types and panicle biomass, and screening out a second key type of spectral features and a second key type of texture features based on the correlation between the multiple spectral features and the correlation between the multiple texture features.

[0011] The construction of a stem biomass inversion model for predicting stem biomass includes: fitting the canopy height data with the stem biomass, and determining, based on the fitting results, to use the canopy height data to predict the aboveground biomass of rice.

[0012] Determining rice aboveground biomass based on the selected features and the canopy height data includes: inputting the first key type spectrum, the first key type texture feature, the second key type spectral feature, the second key type texture feature, and the canopy height data into a random forest model to obtain the rice aboveground biomass.

[0013] Furthermore, the above-mentioned method for retrieving aboveground rice biomass based on component analysis includes, prior to the step of acquiring multispectral images and RGB images of the rice planting area, the following steps:

[0014] Before the rice is transplanted, obtain RGB images of the ground soil in the rice planting area;

[0015] A first digital surface model is constructed based on the RGB image of the soil.

[0016] Furthermore, the above-mentioned method for retrieving aboveground biomass of rice based on component analysis, wherein extracting canopy height data based on the rice RGB image includes:

[0017] A second digital surface model of the rice is constructed based on the RGB image of the rice.

[0018] The canopy height data is determined based on the first digital surface model and the second digital surface model.

[0019] Furthermore, in the above-mentioned method for retrieving aboveground biomass of rice based on component analysis, the construction of the leaf biomass inversion model involves inputting the multiple types of spectral features and the multiple types of texture features into the leaf biomass inversion model to invert leaf biomass, and selecting the first key type of spectral feature and the first key type of texture feature based on the leaf biomass inversion results.

[0020] Multiple types of spectral features were input into the leaf biomass inversion model to obtain the feature importance of each type of spectral feature for leaf biomass inversion;

[0021] Multiple types of texture features are input into the leaf biomass inversion model to obtain the feature importance of each type of texture feature for leaf biomass inversion;

[0022] Based on feature importance, the multiple types of spectral features are sorted from largest to smallest, and several first key type spectral features with higher feature importance are selected.

[0023] Based on feature importance, the multiple types of texture features are sorted from largest to smallest, and several first key type texture features with higher feature importance are selected.

[0024] Furthermore, the above-mentioned method for inverting aboveground biomass of rice based on component analysis, wherein determining the correlations between multiple spectral features of the multiple types and panicle biomass, determining the correlations between multiple texture features of the multiple types and panicle biomass, and screening out second key type spectral features and second key type texture features based on the correlations between the spectral features and the texture features, includes:

[0025] The correlations between multiple types of spectral features and ear biomass are calculated separately. Several spectral features with high correlation are selected from the multiple spectral feature correlations and the spectral features are determined as the second key type of spectral features.

[0026] The correlations between multiple texture features and ear biomass are calculated separately. Several texture features with high correlations are selected from the multiple texture feature correlations and the texture features are determined as the second key type texture features.

[0027] Furthermore, in the above-mentioned method for retrieving aboveground biomass of rice based on component analysis, the central bands of the multispectral image include: 490nm, 520nm, 550nm, 570nm, 670nm, 680nm, 700nm, 720nm, 800nm, 850nm, 900nm, and 950nm. The multiple types of spectral features include: normalized vegetation index, normalized red edge index, red edge index, enhanced vegetation index, terrestrial chlorophyll index, and wide dynamic range vegetation index. The types of texture features include: contrast, dissimilarity, homogeneity, correlation, second moment, entropy, inverse differential moment, and mean, wherein the correlation includes horizontal correlation value and vertical correlation value.

[0028] The first key type texture features selected include: horizontal correlation value of 670nm band, horizontal correlation value of 680nm band, horizontal correlation value of 700nm band, vertical correlation value of 550nm band, and entropy of 700nm band. The first key type spectral features include: enhanced vegetation index, wide dynamic range vegetation index, and normalized red edge index.

[0029] Furthermore, in the above-mentioned rice aboveground biomass inversion method based on component analysis, the second key type texture features selected include: horizontal correlation value of 950nm band, vertical correlation of 900nm band and horizontal correlation value of 570nm band, and the second key type spectral features include: normalized red edge index, red edge index and terrestrial chlorophyll index.

[0030] This application also provides a rice aboveground biomass inversion device based on component analysis, including:

[0031] The acquisition module is used to acquire images of rice planting areas to obtain multispectral images and RGB images of rice.

[0032] The extraction module is used to extract multiple types of spectral features and multiple types of texture features based on the multispectral image, and to extract canopy height data based on the rice RGB image;

[0033] The component prediction module is used to construct a leaf biomass inversion model to select features for predicting leaf biomass, construct a spike biomass inversion model to select features for predicting spike biomass, and construct a stem biomass inversion model to determine features for predicting stem biomass.

[0034] The process of constructing a leaf biomass inversion model for feature selection to predict leaf biomass includes: constructing a leaf biomass inversion model, inputting the multiple types of spectral features and the multiple types of texture features into the leaf biomass inversion model to invert leaf biomass, and selecting a first key type of spectral feature and a first key type of texture feature based on the leaf biomass inversion results;

[0035] The construction of a panicle biomass inversion model for feature screening to predict panicle biomass includes: establishing a panicle biomass inversion model to determine the correlation between multiple spectral features of the multiple types and panicle biomass, determining the correlation between multiple texture features of the multiple types and panicle biomass, and screening out a second key type of spectral features and a second key type of texture features based on the correlation between the multiple spectral features and the correlation between the multiple texture features.

[0036] The construction of a stem biomass inversion model for predicting stem biomass includes: fitting the canopy height data with the stem biomass, and determining, based on the fitting results, to use the canopy height data to predict the aboveground biomass of rice.

[0037] The determination module is used to determine the aboveground biomass of rice based on the selected features and the canopy height data, including: inputting the first key type spectrum, the first key type texture feature, the second key type spectral feature, the second key type texture feature and the canopy height data into a random forest model to obtain the aboveground biomass of rice.

[0038] This application also provides a computer-readable storage medium storing a plurality of instructions adapted for loading by a processor to execute any of the above-described component analysis-based rice aboveground biomass inversion methods.

[0039] This application also provides an electronic device, including a processor and a memory, wherein the processor is electrically connected to the memory, the memory is used to store instructions and data, and the processor is used in the steps of the rice above-described component analysis-based above-described above-described method for retrieving aboveground biomass of rice.

[0040] This application provides a method, apparatus, storage medium, and electronic equipment for rice aboveground biomass inversion based on component analysis. This application conducts targeted analysis on the biomass of different components (leaf, panicle, and stem) of rice, and screens features to predict leaf biomass, panicle biomass, and stem biomass. It identifies key features for inverting rice aboveground biomass: first key type spectral, first key type texture features, second key type spectral, second key type texture features, and canopy height data. Based on these key features, the prediction accuracy of rice aboveground biomass can be improved. Attached Figure Description

[0041] The technical solution and other beneficial effects of this application will become apparent from the following detailed description of specific embodiments in conjunction with the accompanying drawings.

[0042] Figure 1 A schematic flowchart of the rice aboveground biomass inversion method based on component analysis provided in this application embodiment.

[0043] Figure 2 Another flowchart of the rice aboveground biomass inversion method based on component analysis provided in the embodiments of this application.

[0044] Figure 3 A graph showing the relationship between stem weight and canopy height provided in an embodiment of this application.

[0045] Figure 4 The measured AGB values ​​for various parts of rice provided in the embodiments of this application are as follows.

[0046] Figure 5 This is a schematic diagram comparing an RGB image with a texture feature (670nm band COR) image provided in an embodiment of this application.

[0047] Figure 6 A schematic diagram comparing the conventional method and the method of this application in predicting AGB, as provided in the embodiments of this application.

[0048] Figure 7 A schematic diagram of the structure of the rice aboveground biomass inversion device based on component analysis provided in this application embodiment.

[0049] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0050] Figure 9 Another structural schematic diagram of the electronic device provided in the embodiments of this application. Detailed Implementation

[0051] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0052] Currently, most methods for predicting aboveground biomass (AGB) of crops rely on vegetation indices (VIs). While VIS is relatively accurate for estimating AGB from tillering to jointing stages, it easily becomes saturated in later stages of crop growth. Due to the mixed shading of leaves and panicles, the spectral information after heading is highly complex, increasing the difficulty of estimating AGB using spectral data. Therefore, the accuracy of methods using vegetation indices to predict aboveground biomass is limited. Furthermore, as the rice canopy gradually densifies during growth, the stem portion is obscured by the dense canopy, making it difficult to capture effective information about this part of the spectral data. Therefore, this shading must be considered when improving AGB estimation models.

[0053] Because the biomass information of various rice organs is quite complex, and the characteristics exhibited by each organ at different growth stages vary significantly, it is necessary to explore a method that considers the influencing factors of aboveground organ biomass separately, based on the characteristics of different rice organs, in order to improve the overall accuracy of rice AGB estimation.

[0054] Based on the above, embodiments of this application provide a method, apparatus, storage medium, and electronic device for inverting aboveground biomass of rice based on component analysis. The aboveground biomass inversion apparatus for rice based on component analysis provided in this application can be integrated into an electronic device, which can be a terminal, server, or other device. The terminal can include a tablet computer, laptop computer, personal computer (PC), microprocessor box, or other devices.

[0055] Please see Figure 1 and Figure 2 , Figure 1 This is a flowchart illustrating the rice aboveground biomass inversion method based on component analysis provided in an embodiment of this application. Figure 2 Another schematic flowchart of the rice aboveground biomass inversion method based on component analysis provided in this application embodiment, which is applied in electronic devices, includes the following steps:

[0056] S1, acquires images of the rice planting area to obtain multispectral images and RGB images of the rice.

[0057] Specifically, the multispectral images and RGB images of rice were acquired by drones and needed to be taken on a clear, cloudless day to ensure image quality.

[0058] In one embodiment, the center bands of the acquired multispectral images are: 490nm, 520nm, 550nm, 570nm, 670nm, 680nm, 700nm, 720nm, 800nm, 850nm, 900nm, and 950nm. Furthermore, multiple calibration mats need to be laid in advance along the edge of the rice planting area. These mats can be eight mats with standard reflectances of 0.03, 0.06, 0.12, 0.24, 0.36, 0.48, 0.56, and 0.80, used for radiometric calibration of the images. The vertical overlap of the RGB images is set to 90%, the horizontal overlap to 70%, the photo interval to 2 seconds, and the image size to 5472*3648 pixels.

[0059] Before acquiring the multispectral image, calibration is performed using a calibration blanket to obtain a multispectral image with standard reflectance.

[0060] Prior to step S1, the rice aboveground biomass inversion method based on component analysis includes the following steps:

[0061] SA11 is used to acquire RGB images of the ground soil in the rice planting area before the rice is transplanted.

[0062] SA12, the first digital surface model is constructed based on soil RGB images.

[0063] Specifically, multiple images from different perspectives are acquired using a drone's camera. Based on a structure-of-motion (SOG) algorithm, feature point detection and matching are performed on these images to calculate the camera's trajectory and attitude information. Then, based on a multi-view stereo algorithm, real-time stereo matching is performed on the images from different perspectives according to the camera's attitude information to obtain the depth information of the corresponding points. Surface reconstruction is then performed based on the depth information to generate a first digital surface model.

[0064] S2 extracts multiple types of spectral features and multiple types of texture features based on multispectral images, and extracts canopy height data based on rice RGB images.

[0065] In one embodiment, extracting canopy height data based on a rice RGB image includes:

[0066] S21, Construct a second digital surface model of rice based on RGB images of rice.

[0067] The specific construction methods and steps involved in constructing the first digital surface model in SA12 are not described in detail here.

[0068] S22, determine the canopy height data based on the first digital surface model and the second digital surface model.

[0069] Specifically, it can be calculated using the following formula:

[0070] H = DSM - DSMsoil

[0071] Where H represents the canopy height data, DSM represents the second digital surface model, DSMsoil represents the first digital surface model, and DSM-DSMsoil refers to subtracting the heights of data points in two digital surface models at the same projection point.

[0072] It should be noted that in multiple experiments, stem biomass was calculated using texture features, spectral features, and canopy height data respectively. The calculation results showed that the stem biomass calculated using canopy height data was closest to the measured stem biomass, indicating that estimating stem biomass using canopy height data is the most accurate method. Figure 3 The stem weight and canopy height relationship curve provided in the embodiments of this application shows that the two have very similar trends.

[0073] Figure 4 The measured AGB values ​​for various parts of rice provided in the embodiments of this application are as follows: Figure 4As shown, the measured AGB values ​​of different parts of rice exhibit different trends at different growth stages. The AGB of the stem gradually increases with the growth stage, while the growth rate gradually slows down in the later stages. The AGB of the leaves first gradually increases at maturity and then slightly decreases. The AGB of the panicle starts from 0 at the heading stage and reaches its maximum value at maturity, at which point it exceeds the AGB of the stem and leaves. Based on the characteristics of different parts of rice, different methods should be used to explore the key factors for AGB estimation.

[0074] S3, construct a leaf biomass inversion model to screen features for predicting leaf biomass, construct a spike biomass inversion model to screen features for predicting spike biomass, and construct a stem biomass inversion model to determine features for predicting stem biomass.

[0075] Specifically, the aboveground biomass of various organs in the rice-growing area is calculated in advance to prepare for subsequent feature selection.

[0076] The construction of a leaf biomass inversion model for predicting leaf biomass includes feature selection, which includes:

[0077] S31, construct a leaf biomass inversion model, input the multiple types of spectral features and the multiple types of texture features into the leaf biomass inversion model to invert leaf biomass, and select the first key type of spectral features and the first key type of texture features based on the leaf biomass inversion results.

[0078] In the leaf biomass inversion model, the random forest algorithm can be used to screen the first key type spectral features and the first key type texture features.

[0079] In one embodiment, step S31 includes the following steps:

[0080] S311, input multiple types of spectral features into the leaf biomass inversion model to obtain the feature importance of each type of spectral feature for leaf biomass inversion;

[0081] S312, input multiple types of texture features into the leaf biomass inversion model respectively, and obtain the feature importance of each type of texture feature for leaf biomass inversion;

[0082] S313, based on feature importance, sort multiple types of spectral features from largest to smallest, and select several first key type spectral features with higher feature importance;

[0083] S314. Based on feature importance, sort multiple types of texture features from largest to smallest, and select several first key type texture features with higher feature importance.

[0084] The types of spectral features can include: Normalized Difference Vegetation Index (NDVI), Normalized Red Edge Index (NDRE), Red Mark Index (CIre), Enhanced Vegetation Index (EVI2), Terrestrial Chlorophyll Index (MTCI), and Wide Dynamic Range Vegetation Index (WDRVI). The types of texture features can include: Contrast (CON), Dissimilarity (DIS), Homogeneity (HOM), Correlation (COR), Second Moment (ASM), Entropy (ENT), Inverse Differential Moment (IDM), and Mean (MEA), where correlation includes horizontal and vertical correlation values.

[0085] The aforementioned texture features are extracted from regions of interest within a 3x3 windowed multispectral image. The horizontal correlation value is obtained by sliding the window horizontally, and the vertical correlation value is obtained by sliding the window vertically. For example... Figure 5 As shown, Figure 5 This is a schematic diagram comparing an RGB image with a texture feature (670nm band COR) image provided in an embodiment of this application.

[0086] The calculation methods for spectral characteristics are shown in Table 1 below.

[0087] Table 1. Formulas for calculating spectral characteristics

[0088]

[0089] Where ρ refers to the reflectivity in a certain wavelength band, such as ρ 800 The reflectance is at 800 nm.

[0090] Texture features are obtained by calculating the gray-level co-occurrence matrix of rice images from seedling stage to maturity stage. The specific calculation method of texture features is shown in Table 2 below.

[0091] Table 2 Formulas for Calculating Texture Features

[0092]

[0093]

[0094] in, V i,j Let N be the values ​​in the i-th and j-th cells (row i and column j) of the moving window, where N is the number of rows or columns. P(i,j) is the (i,j)-th value in the normalized gray-level co-occurrence matrix.

[0095] As an example, when the center bands of the acquired multispectral image include 490nm, 520nm, 550nm, 570nm, 670nm, 680nm, 700nm, 720nm, 800nm, 850nm, 900nm, and 950nm, texture features of 8 types under 12 bands and 2 directions (0′ and 90′) are calculated respectively, resulting in 192 texture feature data.

[0096] 192 texture features and 6 vegetation indices were input into the leaf biomass inversion model, and their importance was ranked to obtain the first key type of texture features, including: horizontal correlation values ​​in the 670nm and 680nm bands, horizontal correlation values ​​in the 700nm band, vertical correlation values ​​in the 550nm band, and entropy in the 700nm band. The first key type of spectral features included: enhanced vegetation index, wide dynamic range vegetation index, and normalized red edge index. Leaf biomass was calculated based on the first key type of texture features, and the results were compared with the actual measured leaf biomass. It was found that the leaf biomass calculated using the first key type of texture features (spectral features) had higher accuracy.

[0097] The construction of a panicle biomass inversion model and the feature selection for predicting panicle biomass include:

[0098] S32, establish a panicle biomass inversion model to determine the correlation between multiple types of spectral features and panicle biomass, and to determine the correlation between multiple types of texture features and panicle biomass. Based on the correlation between multiple spectral features and the correlation between multiple texture features, select the second key type of spectral features and the second key type of texture features.

[0099] In one embodiment, step S32 includes the following steps:

[0100] S321, calculate the correlation between multiple types of spectral features and ear biomass, select several spectral features with high correlation from the multiple spectral feature correlations, and determine the spectral features as the second key type of spectral features;

[0101] S322, calculate the correlation between multiple texture features and ear biomass, select several texture features with high correlation from the multiple texture feature correlations, and determine the texture features as the second key type texture features.

[0102] Specifically, the correlation between multiple types of texture features (spectral features) and the actual measured ear biomass is calculated. The correlation can be calculated by linear fitting, and several texture features (spectral features) with better linear fitting results are selected as the filtered features.

[0103] As an example, the second key type of texture features selected include: horizontal correlation value in the 950nm band, vertical correlation value in the 900nm band, and horizontal correlation value in the 570nm band. The second key type of spectral features include: normalized red edge index, red edge index, and terrestrial chlorophyll index.

[0104] The construction of a stem biomass inversion model for predicting stem biomass involves determining the characteristics of the model, including fitting canopy height data with stem biomass, and determining the use of canopy height data to predict aboveground biomass of rice based on the fitting results.

[0105] Specifically, after multiple experiments, stem biomass was fitted using spectral information, texture information, and canopy height data respectively. It was found that canopy height data had the best fitting effect, and stem biomass was significantly correlated with upper layer height data. Therefore, canopy height data was directly used as the key type feature.

[0106] S4. Determine the aboveground biomass of rice based on the selected features and canopy height data, including: inputting the first key type spectrum, the first key type texture features, the second key type spectral features, the second key type texture features and the canopy height data into the random forest model to obtain the aboveground biomass of rice.

[0107] Specifically, the selected features and canopy height data are input into a trained random forest model to predict the aboveground biomass of rice.

[0108] It should be noted that after the random forest model is trained, a real-world dataset should be provided to validate the model. Specifically, this can be achieved using the coefficient of determination R0. 2 The performance of the model was verified using RMSE.

[0109]

[0110] in, y represents the predicted value of aboveground biomass. i This is the measured aboveground biomass. R represents the mean of the measured aboveground biomass. 2 The closer the value is to 1, the better the model performance.

[0111] Figure 6This diagram illustrates a comparison between the conventional method and the method of this application for predicting AGB (Above Ground Biomass). By comparing the conventional method of predicting aboveground biomass solely through vegetation indices with the method provided in this application, the verification results shown in Table 3 are obtained. Figure 6 A comparison diagram.

[0112] Table 3. Comparison of verification results between traditional methods and the method of this application.

[0113]

[0114] As can be seen from the table above, the method provided in this application yields more accurate predictions of aboveground biomass compared to traditional methods.

[0115] This application identifies key features for retrieving aboveground biomass of rice by grouping and predicting leaf biomass, panicle biomass, and stem biomass, and by screening the features. Based on these key features and canopy height data, the prediction of aboveground biomass of rice can be improved, and the destructive impact and high labor costs of field sampling can be avoided.

[0116] Based on the method described in the above embodiments, this embodiment will further describe the rice aboveground biomass inversion device based on component analysis. The rice aboveground biomass inversion device based on component analysis can be implemented as an independent entity or integrated into an electronic device. The electronic device can be a terminal, server, or other devices. The terminal can include a tablet computer, a laptop computer, a personal computer (PC), a microprocessor box, or other devices.

[0117] Please see Figure 7 , Figure 7 This application provides a specific description of a rice aboveground biomass inversion device based on component analysis, which is applied in an electronic device. The device may include:

[0118] The acquisition module is used to acquire images of rice planting areas to obtain multispectral images and RGB images of rice.

[0119] The extraction module is used to extract multiple types of spectral features and multiple types of texture features based on multispectral images, and to extract canopy height data based on rice RGB images;

[0120] The component prediction module is used to construct a leaf biomass inversion model to select features for predicting leaf biomass, construct a spike biomass inversion model to select features for predicting spike biomass, and construct a stem biomass inversion model to determine features for predicting stem biomass.

[0121] Among them, constructing a leaf biomass inversion model for feature selection to predict leaf biomass includes: constructing a leaf biomass inversion model, inputting multiple types of spectral features and multiple types of texture features into the leaf biomass inversion model to invert leaf biomass, and selecting the first key type of spectral features and the first key type of texture features based on the leaf biomass inversion results;

[0122] The construction of a panicle biomass inversion model for feature screening to predict panicle biomass includes: establishing a panicle biomass inversion model to determine the correlation between multiple types of spectral features and multiple spectral features between panicle biomass, determining the correlation between multiple types of texture features and multiple texture features between panicle biomass, and screening out second key type spectral features and second key type texture features based on the correlation between multiple spectral features and multiple texture features.

[0123] The construction of a stem biomass inversion model for predicting stem biomass involves determining the characteristics of the model, including: fitting canopy height data with stem biomass, and determining the use of canopy height data to predict aboveground biomass of rice based on the fitting results.

[0124] The determination module is used to determine the aboveground biomass of rice based on the selected features and canopy height data. This includes: inputting the first key type spectrum, the first key type texture feature, the second key type spectral feature, the second key type texture feature, and the canopy height data into the random forest model to obtain the aboveground biomass of rice.

[0125] In specific implementation, the above modules and / or units can be implemented as independent entities, or they can be arbitrarily combined and implemented as the same or several entities. For the specific implementation of the above modules and / or units, please refer to the previous method embodiments. For the specific beneficial effects that can be achieved, please also refer to the beneficial effects in the previous method embodiments, which will not be repeated here.

[0126] In addition, embodiments of this application also provide an electronic device, which may be a computer, tablet computer, or other similar device. Figure 8 As shown, the electronic device 400 includes a processor 401 and a memory 402. The processor 401 and the memory 402 are electrically connected.

[0127] The processor 401 is the control center of the electronic device 400. It connects various parts of the electronic device through various interfaces and lines. By running or loading the application program stored in the memory 402 and calling the data stored in the memory 402, it performs various functions of the electronic device and processes data, thereby monitoring the electronic device as a whole.

[0128] In this embodiment, the processor 401 in the electronic device 400 loads the instructions corresponding to the processes of one or more application programs into the memory 402 according to the following steps, and the processor 401 runs the application programs stored in the memory 402 to realize various functions:

[0129] Images were acquired from rice-growing areas to obtain multispectral images and RGB images of rice.

[0130] Multiple types of spectral features and multiple types of texture features are extracted from the multispectral image, and canopy height data is extracted from the rice RGB image.

[0131] A leaf biomass inversion model was constructed to screen features for predicting leaf biomass; a spike biomass inversion model was constructed to screen features for predicting spike biomass; and a stem biomass inversion model was constructed to determine features for predicting stem biomass.

[0132] The process of constructing a leaf biomass inversion model for feature selection to predict leaf biomass includes: constructing a leaf biomass inversion model, inputting the multiple types of spectral features and the multiple types of texture features into the leaf biomass inversion model to invert leaf biomass, and selecting a first key type of spectral feature and a first key type of texture feature based on the leaf biomass inversion results;

[0133] The construction of a panicle biomass inversion model for feature screening to predict panicle biomass includes: establishing a panicle biomass inversion model to determine the correlation between multiple spectral features of the multiple types and panicle biomass, determining the correlation between multiple texture features of the multiple types and panicle biomass, and screening out a second key type of spectral features and a second key type of texture features based on the correlation between the multiple spectral features and the correlation between the multiple texture features.

[0134] The construction of a stem biomass inversion model for predicting stem biomass includes: fitting the canopy height data with the stem biomass, and determining, based on the fitting results, to use the canopy height data to predict the aboveground biomass of rice.

[0135] Determining rice aboveground biomass based on the selected features and the canopy height data includes: inputting the first key type spectrum, the first key type texture feature, the second key type spectral feature, the second key type texture feature, and the canopy height data into a random forest model to obtain the rice aboveground biomass.

[0136] This electronic device can implement the steps in any embodiment of the rice aboveground biomass inversion method based on component analysis provided in the embodiments of this application. Therefore, it can achieve the beneficial effects that any rice aboveground biomass inversion method based on component analysis provided in the embodiments of this invention can achieve, as detailed in the preceding embodiments, which will not be repeated here.

[0137] Figure 9 A specific structural block diagram of an electronic device provided in an embodiment of the present invention is shown. This electronic device can be used to implement the rice aboveground biomass inversion method based on component analysis provided in the above embodiments. The electronic device 500 can be a terminal, server, or other device. The terminal can include a tablet computer, laptop computer, personal computer (PC), microprocessor box, or other devices.

[0138] RF circuit 510 is used to receive and transmit electromagnetic waves, converting electromagnetic waves into electrical signals and vice versa, thereby enabling communication with communication networks or other devices. RF circuit 510 may include various existing circuit elements used to perform these functions, such as antennas, radio frequency transceivers, digital signal processors, encryption / decryption chips, Subscriber Identity Module (SIM) cards, memory, etc. RF circuit 510 can communicate with various networks such as the Internet, corporate intranets, and wireless networks, or communicate with other devices via wireless networks. The aforementioned wireless networks may include cellular telephone networks, wireless local area networks (WLANs), or metropolitan area networks (MANs). The aforementioned wireless networks may use various communication standards, protocols, and technologies, including but not limited to Global System for Mobile Communication (GSM), Enhanced Data GSM Environment (EDGE), Wideband Code Division Multiple Access (WCDMA), Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Wireless Fidelity (Wi-Fi) (such as IEEE 802.11a, IEEE 802.11b, IEEE 802.11g, and / or IEEE 802.11n), Voice over Internet Protocol (VoIP), Worldwide Interoperability for Microwave Access (Wi-Max), other protocols for email, instant messaging, and short messages, and any other suitable communication protocols, including those that have not yet been developed.

[0139] The memory 520 can be used to store software programs and modules, such as the program instructions / modules corresponding to those in the above embodiments. The processor 580 executes various functional applications and data processing by running the software programs and modules stored in the memory 520, such as taking pictures with the front-facing camera, processing the captured images, and switching the display colors of the content displayed on the screen. The memory 520 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 520 may further include memory remotely located relative to the processor 580, and these remote memories can be connected to the electronic device 500 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0140] The input unit 530 can be used to receive input numeric or character information, and to generate a keyboard and mouse related to user settings and function control.

[0141] Display unit 540 can be used to display information input by the user or information provided to the user, as well as various graphical user interfaces, which can be composed of graphics, text, icons, video, and any combination thereof. Display unit 540 may include display panel 541, which may optionally be configured in the form of LCD (Liquid Crystal Display), OLED (Organic Light-Emitting Diode), or other similar forms.

[0142] Audio circuitry 560, speaker 561, and microphone 562 provide an audio interface between the user and electronic device 500. Audio circuitry 560 converts received audio data into electrical signals and transmits them to speaker 561, where speaker 561 converts them into sound signals for output. Conversely, microphone 562 converts collected sound signals into electrical signals, which are then received by audio circuitry 560, converted back into audio data, and processed by processor 580. The audio data is then transmitted via RF circuitry 510 to, for example, another terminal, or output to memory 520 for further processing. Audio circuitry 560 may also include an earphone jack to facilitate communication between external headphones and electronic device 500.

[0143] Electronic device 500, through transmission module 570 (e.g., Wi-Fi module), can help users receive requests, send information, etc., providing users with wireless broadband internet access. Although transmission module 570 is shown in the figure, it is understood that it is not an essential component of electronic device 500 and can be omitted as needed without changing the essence of the invention.

[0144] The processor 580 is the control center of the electronic device 500. It connects to various parts of the phone via various interfaces and lines, and performs various functions and processes data of the electronic device 500 by running or executing software programs and / or modules stored in the memory 520, and by calling data stored in the memory 520, thereby providing overall monitoring of the electronic device. Optionally, the processor 580 may include one or more processing cores; in some embodiments, the processor 580 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, and the modem processor mainly handles wireless communication. It is understood that the modem processor may also not be integrated into the processor 580.

[0145] Electronic device 500 also includes a power supply 590 (such as a battery) for supplying power to various components. In some embodiments, the power supply may be logically connected to processor 580 through a power management system, thereby enabling functions such as managing charging, discharging, and power consumption through the power management system. The power supply 590 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0146] Although not shown, the electronic device 500 also includes cameras (such as front-facing cameras and rear-facing cameras), Bluetooth modules, etc., which will not be described in detail here. Specifically, in this embodiment, the display unit of the electronic device is a touch screen display, and the mobile terminal also includes a memory and one or more programs, wherein one or more programs are stored in the memory and configured to be executed by one or more processors. One or more programs contain instructions for performing the following operations:

[0147] Images were acquired from rice-growing areas to obtain multispectral images and RGB images of rice.

[0148] Multiple types of spectral features and multiple types of texture features are extracted from the multispectral image, and canopy height data is extracted from the rice RGB image.

[0149] A leaf biomass inversion model was constructed to screen features for predicting leaf biomass; a spike biomass inversion model was constructed to screen features for predicting spike biomass; and a stem biomass inversion model was constructed to determine features for predicting stem biomass.

[0150] The process of constructing a leaf biomass inversion model for feature selection to predict leaf biomass includes: constructing a leaf biomass inversion model, inputting the multiple types of spectral features and the multiple types of texture features into the leaf biomass inversion model to invert leaf biomass, and selecting a first key type of spectral feature and a first key type of texture feature based on the leaf biomass inversion results;

[0151] The construction of a panicle biomass inversion model for feature screening to predict panicle biomass includes: establishing a panicle biomass inversion model to determine the correlation between multiple spectral features of the multiple types and panicle biomass, determining the correlation between multiple texture features of the multiple types and panicle biomass, and screening out a second key type of spectral features and a second key type of texture features based on the correlation between the multiple spectral features and the correlation between the multiple texture features.

[0152] The construction of a stem biomass inversion model for predicting stem biomass includes: fitting the canopy height data with the stem biomass, and determining, based on the fitting results, to use the canopy height data to predict the aboveground biomass of rice.

[0153] Determining rice aboveground biomass based on the selected features and the canopy height data includes: inputting the first key type spectrum, the first key type texture feature, the second key type spectral feature, the second key type texture feature, and the canopy height data into a random forest model to obtain the rice aboveground biomass.

[0154] In practice, the above modules can be implemented as independent entities or combined in any way to be implemented as the same or several entities. For the specific implementation of the above modules, please refer to the previous method implementation examples, which will not be repeated here.

[0155] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor. Therefore, embodiments of the present invention provide a storage medium storing multiple instructions that can be loaded by a processor to execute the steps of any embodiment of the rice aboveground biomass inversion method based on component analysis provided by the present invention.

[0156] The computer-readable storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0157] Since the instructions stored in the storage medium can execute the steps in any embodiment of the rice aboveground biomass inversion method based on component analysis provided in the embodiments of the present invention, the beneficial effects that any rice aboveground biomass inversion method based on component analysis provided in the embodiments of the present invention can achieve can be realized, as detailed in the preceding embodiments, and will not be repeated here.

[0158] The foregoing provides a detailed description of a method, apparatus, storage medium, and electronic device for rice aboveground biomass inversion based on component analysis, as provided in the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and its core ideas. At the same time, those skilled in the art will recognize that there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for inverting aboveground biomass of rice based on component analysis, characterized in that, The method includes: Images were acquired from rice-growing areas to obtain multispectral images and RGB images of rice. Multiple types of spectral features and multiple types of texture features are extracted from the multispectral image, and canopy height data is extracted from the rice RGB image. A leaf biomass inversion model was constructed to screen features for predicting leaf biomass; a spike biomass inversion model was constructed to screen features for predicting spike biomass; and a stem biomass inversion model was constructed to determine features for predicting stem biomass. The process of constructing a leaf biomass inversion model for feature selection to predict leaf biomass includes: constructing a leaf biomass inversion model, inputting the multiple types of spectral features and the multiple types of texture features into the leaf biomass inversion model to invert leaf biomass, and selecting a first key type of spectral feature and a first key type of texture feature based on the leaf biomass inversion results; The construction of a panicle biomass inversion model for feature screening to predict panicle biomass includes: establishing a panicle biomass inversion model to determine the correlation between multiple spectral features of the multiple types and panicle biomass, and to determine the correlation between multiple texture features of the multiple types and panicle biomass, and to screen out a second key type of spectral feature and a second key type of texture feature based on the correlation between the multiple spectral features and the correlation between the multiple texture features. The construction of a stem biomass inversion model for predicting stem biomass includes: fitting the canopy height data with the stem biomass, and determining, based on the fitting results, to use the canopy height data to predict the aboveground biomass of rice. Determining rice aboveground biomass based on the selected features and the canopy height data includes: inputting the first key type spectrum, the first key type texture feature, the second key type spectral feature, the second key type texture feature, and the canopy height data into a random forest model to obtain the rice aboveground biomass.

2. The method for inverting aboveground biomass of rice based on component analysis according to claim 1, characterized in that, Prior to the step of acquiring multispectral images and RGB images of the rice planting area, the following steps are included: Before the rice is transplanted, obtain RGB images of the ground soil in the rice planting area; A first digital surface model is constructed based on the RGB image of the soil.

3. The method for inverting aboveground biomass of rice based on component analysis according to claim 2, characterized in that, Extracting canopy height data based on the rice RGB image includes: A second digital surface model of the rice is constructed based on the RGB image of the rice. The canopy height data is determined based on the first digital surface model and the second digital surface model.

4. The method for inverting rice aboveground biomass based on component analysis according to claim 1, characterized in that, The leaf biomass inversion model is constructed by inputting the multiple types of spectral features and the multiple types of texture features into the model to invert leaf biomass. Based on the leaf biomass inversion results, a first key type of spectral feature and a first key type of texture feature are selected. Multiple types of spectral features were input into the leaf biomass inversion model to obtain the feature importance of each type of spectral feature for leaf biomass inversion; Multiple types of texture features are input into the leaf biomass inversion model to obtain the feature importance of each type of texture feature for leaf biomass inversion; Based on feature importance, the multiple types of spectral features are sorted from largest to smallest, and several first key type spectral features with higher feature importance are selected. Based on feature importance, the multiple types of texture features are sorted from largest to smallest, and several first key type texture features with higher feature importance are selected.

5. The method for inverting aboveground biomass of rice based on component analysis according to claim 1, characterized in that, The leaf biomass inversion model is constructed by inputting the multiple types of spectral features and the multiple types of texture features into the model to invert leaf biomass. Based on the leaf biomass inversion results, a first key type of spectral feature and a first key type of texture feature are selected, including: The correlations between multiple types of spectral features and ear biomass are calculated separately. Several spectral features with high correlation are selected from the multiple spectral feature correlations and the spectral features are determined as the second key type of spectral features. The correlations between multiple texture features and ear biomass are calculated separately. Several texture features with high correlations are selected from the multiple texture feature correlations and the texture features are determined as the second key type texture features.

6. The method for inverting aboveground rice biomass based on component analysis according to claim 4, characterized in that, The center bands of the multispectral image include: 490nm, 520nm, 550nm, 570nm, 670nm, 680nm, 700nm, 720nm, 800nm, 850nm, 900nm, and 950nm. The multiple types of spectral features include: normalized vegetation index, normalized red edge index, red mark index, enhanced vegetation index, terrestrial chlorophyll index, and wide dynamic range vegetation index. The types of texture features include: contrast, dissimilarity, homogeneity, correlation, second moment, entropy, inverse differential moment, and mean, wherein the correlation includes horizontal correlation value and vertical correlation value. The first key type texture features selected include: horizontal correlation value of 670nm band, horizontal correlation value of 680nm band, horizontal correlation value of 700nm band, vertical correlation value of 550nm band, and entropy of 700nm band. The first key type spectral features selected include: enhanced vegetation index, wide dynamic range vegetation index, and normalized red edge index.

7. The method for inverting aboveground biomass of rice based on component analysis according to claim 5, characterized in that, The second key type texture features selected include: horizontal correlation value of 950nm band, vertical correlation of 900nm band and horizontal correlation value of 570nm band. The second key type spectral features include: normalized red edge index, red edge index and terrestrial chlorophyll index.

8. A rice aboveground biomass inversion device based on component analysis, characterized in that, include: The acquisition module is used to acquire images of rice planting areas to obtain multispectral images and RGB images of rice. The extraction module is used to extract multiple types of spectral features and multiple types of texture features based on the multispectral image, and to extract canopy height data based on the rice RGB image; The component prediction module is used to construct a leaf biomass inversion model to select features for predicting leaf biomass, construct a spike biomass inversion model to select features for predicting spike biomass, and construct a stem biomass inversion model to determine features for predicting stem biomass. The process of constructing a leaf biomass inversion model for feature selection to predict leaf biomass includes: constructing a leaf biomass inversion model, inputting the multiple types of spectral features and the multiple types of texture features into the leaf biomass inversion model to invert leaf biomass, and selecting a first key type of spectral feature and a first key type of texture feature based on the leaf biomass inversion results; The construction of a panicle biomass inversion model for feature screening to predict panicle biomass includes: establishing a panicle biomass inversion model to determine the correlation between multiple spectral features of the multiple types and panicle biomass, determining the correlation between multiple texture features of the multiple types and panicle biomass, and screening out a second key type of spectral features and a second key type of texture features based on the correlation between the multiple spectral features and the correlation between the multiple texture features. The construction of a stem biomass inversion model for predicting stem biomass includes feature determination: fitting the canopy height data with the stem biomass, and determining, based on the fitting result, to use the canopy height data to predict the aboveground biomass of rice; a determination module is used to determine the aboveground biomass of rice based on the selected features and the canopy height data, including: inputting the first key type spectrum, the first key type texture feature, the second key type spectral feature, the second key type texture feature, and the canopy height data into a random forest model to obtain the aboveground biomass of rice.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a plurality of instructions adapted to be loaded by a processor to execute the rice aboveground biomass inversion method based on component analysis as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, It includes a processor and a memory, the processor being electrically connected to the memory, the memory being used to store instructions and data, and the processor being used to execute the steps in the rice aboveground biomass inversion method based on component analysis as described in any one of claims 1 to 7.

Images