Inversion method, system and equipment of crop growth vigor and medium

Abstract

The invention discloses an inversion method, system and device for crop growth vigor and a medium, and relates to the technical field of remote sensing, and the method comprises the steps: obtaining a red edge index REIP and a red edge index IRECI according to the reflectivity of a spectral image of a crop planting region, and combining the red edge index REIP and the red edge index IRECI according to different proportions; training a first deep neural network DNN through the combined index of the first proportion to predict a leaf area index, and optimizing a learning rate and regularization parameters of the first deep neural network DNN through a particle swarm method in the training process to construct a leaf area index inversion model; training a second deep neural network DNN through the combination index of the second proportion to predict plant height data, and obtaining a plant height inversion model; and obtaining leaf area index and plant height data according to the leaf area index inversion model and the plant height inversion model, and fitting through an entropy weight method to obtain an index weight result for evaluating the growth vigor of the crops. According to the invention, the crop growth vigor can be accurately predicted.

Description

Technical Field

[0001] The invention relates to the field of remote sensing technology, and in particular to an inversion method, system, equipment and medium for crop growth. Background Art

[0002] Remote sensing technology is characterized by periodic observation and large-area coverage in acquiring ground information. It plays an important role in agricultural resource monitoring. It can capture the distribution information of key biophysical parameters of terrestrial vegetation in space and time, and is the basis for corn growth research.

[0003] In the existing technology, corn growth prediction is mostly analyzed from a single factor, such as soil factors, meteorological factors or factors of the crop itself, or analyzed from a small scale direction of collecting data from drones.

[0004] However, a single factor can hardly reflect the overall growth of crops, making it impossible to accurately predict the growth of crops. Summary of the invention

[0005] The embodiments of the present invention provide a method, system, device and medium for inverting the growth of crops, which can solve the problem in the prior art that the growth of crops cannot be accurately predicted.

[0006] The embodiment of the present invention provides an inversion method for crop growth, comprising the following steps: obtaining a red edge index REIP and a red edge index IRECI according to the reflectivity of a spectral image of a crop planting area; combining the red edge index REIP and the red edge index IRECI according to different proportions to obtain a combination index of a first proportion and a combination index of a second proportion; training a first deep neural network DNN through the combination index of the first proportion to predict a leaf area index, and optimizing the learning rate and regularization parameter of the first deep neural network DNN through a particle swarm method during the training process to construct a leaf area index inversion model; training a second deep neural network DNN through the combination index of the second proportion to predict plant height data to obtain a plant height inversion model; combining the red edge index REIP and the red edge index IRECI of the spectral image of the crop planting area to be tested according to the first proportion and the second proportion respectively; substituting the combination index of the first proportion to be tested into the leaf area index inversion model to obtain the leaf area index, and substituting the combination index of the second proportion to be tested into the plant height inversion model to obtain the plant height data; fitting the leaf area index and the plant height data through an entropy weight method to obtain an index weight result for evaluating the growth of crops.

[0007] Furthermore, the leaf area index inversion model is obtained by specifically obtaining the following steps: obtaining a red edge index RE IP and a red edge index I REC I combined according to a first ratio; setting initial hyperparameters in a first deep neural network DNN including 5 linear layers, optimizing the hyperparameters using a particle swarm optimization algorithm to obtain a learning rate and an L2 regularization strength, and predicting the leaf area index by training the optimized first deep neural network DNN with the combined index of the first ratio to construct a leaf area index inversion model.

[0008] Furthermore, the plant height inversion model is obtained by specifically obtaining the following steps: obtaining a red edge index RE IP and a red edge index I REC I combined according to a second ratio; setting hyperparameters in a second deep neural network DNN comprising 5 linear layers, setting the learning rate to 0.0001, adjusting the regularization strength to 0.2, and training the second deep neural network DNN through the combination index of the second ratio to predict plant height data, thereby obtaining a plant height inversion model.

[0009] Furthermore, the red edge index RE IP and the red edge index I REC I are combined in different proportions, and the specific steps include: combining the red edge index RE IP and the red edge index I REC I in a first ratio of 0.3:0.7; and combining the red edge index RE IP and the red edge index I REC I in a second ratio of 0.2:0.8.

[0010] Furthermore, the obtaining of the red edge index RE IP and the red edge index I RECI specifically includes: obtaining multiple vegetation indices according to the reflectance contained in the spectral image of the crop planting area, performing feature selection on the multiple vegetation indices using a feature recursive elimination algorithm, and selecting the red edge index RE IP and the red edge index I RECI.

[0011] Furthermore, the specific steps of obtaining the weight results of indicators for evaluating the growth of crops include: standardizing the leaf area index and plant height data to eliminate the influence of different dimensions; obtaining the information entropy of each indicator from the standardized data, and using information redundancy to obtain the weight results.

[0012] An embodiment of the present invention provides an inversion system for crop growth, comprising:

[0013] A data acquisition module is used to obtain a red edge index RE IP and a red edge index I RECI according to the reflectivity of the spectral image of the crop planting area; and combine the red edge index RE IP and the red edge index I RECI in different proportions to obtain a combination index of a first proportion and a combination index of a second proportion;

[0014] A model training module is used to train a first deep neural network DNN through a combination index of a first ratio to predict a leaf area index, and optimize a learning rate and a regularization parameter of the first deep neural network DNN through a particle swarm method during the training process to construct a leaf area index inversion model; train a second deep neural network DNN through a combination index of a second ratio to predict plant height data, and obtain a plant height inversion model;

[0015] The growth inversion module is used to combine the red edge index REIP and the red edge index IRECI of the spectral image of the crop planting area to be tested according to the first ratio and the second ratio respectively; substitute the combined index of the first ratio to be tested into the leaf area index inversion model to obtain the leaf area index, and substitute the combined index of the second ratio to be tested into the plant height inversion model to obtain the plant height data; fit the leaf area index and the plant height data through the entropy weight method to obtain the indicator weight result for evaluating the growth of crops.

[0016] An embodiment of the present invention provides a computer device, comprising: a memory and a processor; the memory stores a computer program, and the processor implements the above-mentioned crop growth inversion method when executing the computer program.

[0017] An embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned method for inverting the growth potential of crops is implemented.

[0018] The embodiments of the present invention provide a method, system, device and medium for inversion of crop growth. Compared with the prior art, the beneficial effects thereof are as follows:

[0019] The red edge index RE IP and the red edge index I RECI are obtained according to the reflectivity of the spectral image of the crop planting area; the red edge index RE IP and the red edge index I RECI are combined in different proportions to obtain a combination index of a first proportion and a combination index of a second proportion; a first deep neural network DNN is trained by the combination index of the first proportion to predict the leaf area index, and the learning rate and regularization parameter of the first deep neural network DNN are optimized by the particle swarm method during the training process to construct a leaf area index inversion model; a second deep neural network DNN is trained by the combination index of the second proportion to predict plant height data to obtain a plant height inversion model; the red edge index RE IP and the red edge index I RECI of the spectral image of the crop planting area to be tested are combined according to the first proportion and the second proportion respectively; the combination index of the first proportion to be tested is substituted into the leaf area index inversion model to obtain the leaf area index, and the combination index of the second proportion to be tested is substituted into the plant height inversion model to obtain the plant height data; the leaf area index and the plant height data are fitted by the entropy weight method to obtain the index weight result for evaluating the growth of crops.

[0020] Among them, the red edge index RE IP and the red edge index IRECI reflecting the growth of crops are combined in different proportions and input into the leaf area index inversion model to achieve accurate prediction of the leaf area index. The entropy weight method is then used to fit the predicted plant height data. The leaf area index and plant height data are used as indicators for evaluating the growth of crops, and the indicator weight results for evaluating the growth of crops can be accurately obtained, ultimately achieving accurate prediction of the growth of crops. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A flow chart of a crop growth inversion method provided by an embodiment of the present invention;

[0022] Figure 2 Flowchart of the main stages of the DNN algorithm application provided by the embodiment of the present invention;

[0023] Figure 3 Flowchart of particle swarm optimization algorithm provided by an embodiment of the present invention;

[0024] Figure 4 The deep neural network DNN optimized by the particle swarm optimization algorithm provided in the embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention, so the present invention is not limited by the specific embodiments disclosed below.

[0026] See also Figure 1 to Figure 4 The embodiment of the present invention provides a method for inverting the growth of crops, comprising the following steps:

[0027] Step 1: Obtain multiple vegetation indices based on the reflectance contained in the spectral image of the crop planting area, use the feature recursive elimination algorithm to perform feature selection on the multiple vegetation indices, and select the red edge index REIP and the red edge index IRECI, where REIP is the red edge position index and IRECI is the inverted red edge chlorophyll index. Combine the red edge index REIP and the red edge index IRECI in different proportions, specifically: combine the red edge index REIP and the red edge index IRECI in a first ratio of 0.3:0.7. Combine the red edge index REIP and the red edge index IRECI in a second ratio of 0.2:0.8.

[0028] Step 2: Train the first deep neural network DNN through the combination index of the first ratio to predict the leaf area index, and optimize the learning rate and regularization parameter of the first deep neural network DNN through the particle swarm method during the training process to construct a leaf area index inversion model; train the second deep neural network DNN through the combination index of the second ratio to predict the plant height data and obtain the plant height inversion model.

[0029] Step 3: Combine the red edge index RE IP and the red edge index I RECI of the spectral image of the crop planting area to be tested according to the first ratio and the second ratio respectively; substitute the combined index of the first ratio to be tested into the leaf area index inversion model to obtain the leaf area index, and substitute the combined index of the second ratio to be tested into the plant height inversion model to obtain the plant height data. Fit the leaf area index and plant height data by the entropy weight method to obtain the index weight result for evaluating the growth of crops.

[0030] Descriptive statistical analysis and visualization were performed on the field measurement data set, and 9 vegetation indices were selected to analyze their relationship with LAI and plant height. The NDVI index can quantify the amount of vegetation. High values ​​indicate dense tree crowns, and low or negative values ​​indicate urban and waterscape areas. The PVI index is the vertical distance from the vegetation pixel to the soil brightness line in the two-dimensional coordinate system of RN IR. The SAVI index is a hybrid between a ratio-based index and a vertical index. The REIP index was developed for use in the measurement / management of biomass and nitrogen (N) absorption in heterogeneous farmland. The inflection point of strong red light absorption of near-infrared reflectance contains information about growth conditions, and the reflectance near the red edge is very sensitive to large-scale changes in crop LAI. The IRECI index was developed for use in the measurement / management of biomass and nitrogen (N) absorption in heterogeneous farmland. The inflection point of strong red light absorption of near-infrared reflectance contains information about growth conditions, and the reflectance near the red edge is very sensitive to large-scale changes in crop LAI. The PSSRa index is used to estimate chlorophyll content from MERIS (Medium Resolution Imaging Spectroradiometer) data, and the goal is to estimate REIP. The goal of the MTCI index is to estimate the red edge position (REP), which is the maximum tilt point in the red and near-infrared regions of the plant spectral reflectance. It is important for observing chlorophyll content, vegetation aging, and stress of water and nutrient deficiency. The MCARI index responds to chlorophyll changes. The index algorithm responds to leaf chlorophyll concentration and ground reflectance. The above index calculation methods are shown in Table 1 Index calculation formula.

[0031] Table 1 Index calculation formula

[0032]

[0033]

[0034] To ensure the accuracy of LAI estimation, it is crucial to establish a temporal correspondence between the Sentinel-2 satellite reflectivity data and the LAI measured in situ.

[0035] Feature selection (FS) is a widely used technique in pattern recognition applications. By removing irrelevant, noisy, and redundant features from the original feature space, FS alleviates the problem of overfitting and improves the performance of the model. It can also reduce the time and space cost of the learning algorithm. More importantly, we can gain a deeper understanding of the data by analyzing the importance of features. The Recursive Feature Elimination (RFE) algorithm is a wrapper method for feature selection that gradually optimizes feature subsets by repeatedly building models and eliminating the least important features, thereby improving the predictive performance of the model. The following is a detailed introduction to the principle of the RFE algorithm. The purpose of the RFE algorithm is to eliminate redundancy between features, select the optimal feature combination, and reduce the feature dimension. It is committed to finding the best-performing feature subset by repeatedly creating models. Its basic principles can be summarized in the following steps:

[0036] 1. Initial feature subset selection: First, select K features from the original feature set as the initial feature subset. This K value can be set according to actual conditions or experience, or it can be gradually increased from 1 to explore the impact of different feature numbers on model performance. In this experiment, the original features are increased one by one from 1 to 8. Each case is traversed.

[0037] 2. Model training and evaluation: Use the initial feature subset to train a machine learning model and evaluate the performance of the model. Evaluation methods usually include cross-validation to ensure the reliability of model performance. This experiment adopts the cross-validation method.

[0038] 3. Feature importance evaluation: After the model training is completed, the importance of each feature is evaluated. The importance of features can be determined by different methods, such as feature weight, coefficient, information gain, etc., depending on the machine learning algorithm used.

[0039] 4. Feature elimination and subset update: Sort the features by importance, eliminate the least important features (or keep the most important features, depending on the algorithm implementation), and use the remaining features to build a new feature subset. This realizes the functions of dimensionality reduction and feature subset optimization.

[0040] 5. Iterative process: Repeat the above steps, i.e. train the model with a new feature subset, evaluate the model performance, evaluate the feature importance, and remove (or retain) features until the predetermined number of features is reached or other stopping conditions are met (such as model performance is no longer significantly improved).

[0041] 6. Select the optimal feature subset: During all iterations, multiple feature subsets with different numbers of features will be generated. Finally, the feature subset with the highest accuracy is selected. The RFE analysis results show that the optimal number of features is 2, and the corresponding features are REIP and IRECI.

[0042] Establish plant height inversion model based on DNN:

[0043] A deep neural network (DNN) is a multi-layer neural network that forms a deep structure by connecting multiple neurons together. DNN developed from ANN has a deeper architecture and is able to build complex mapping functions from input data to output data. The basic structure of a feedforward DNN usually consists of an input layer, multiple hidden layers, and an output layer. There are one or more neurons in each layer, and each neuron is connected to other neurons in the upper and lower layers. The neuron receives the output of the neurons in the previous layer and transfers it through the motivation function, and then passes it to other neurons in the next layer during the forward pass. The motivation function of the neuron is usually a nonlinear function, which enables DNN to handle the highly nonlinear relationship between input and output. During the training process, the weights and biases of each neuron are adjusted through iterations until the optimal value is reached. It is the backward propagation of the error (loss) that passes the error between the predicted value and the true value to the layer above. The weights and biases can be optimized by many algorithms, such as the gradient descent method. DNN can be used to process various types of data, including images, text, and speech. In a deep neural network, the number of hidden layers and the number of neurons in each layer can depend on the specific task and data type. The training process of DNN usually adopts the back-propagation algorithm and gradient descent optimization method to minimize the prediction error and loss function by continuously adjusting the parameters of the neural network.

[0044] Deep neural networks have multiple nonlinear mapping feature transformations and can fit highly complex functions. Compared with shallow modeling methods, deep modeling can represent actual complex nonlinear problems more carefully and efficiently.

[0045]

[0046] where x i represents the i-th input of the neuron (the output of the previous layer of neurons), w i is the weight between two neurons, b is the bias of the neuron, and f represents the activation function.

[0047] The vegetation index was used as input and plant height as output. Five linear layers were included. Hyperparameters were set, the learning rate was set to 0.0001, and the regularization strength was adjusted to 0.2. The corn crop was trained in a small area to obtain a corn plant height model, which was then verified in a large area.

[0048] DNN Optimization In theory, neural networks can approximate arbitrary functions through appropriate architectures. However, as the number of layers increases, the architecture of the neural network will become more complex and the parameters of the neural network will be more. It makes the learning process much slower. In neural networks with multiple hidden layers, it is difficult to train using classical algorithms because the gradient of the error disappears during backpropagation. Similar to traditional neural networks, the increase in the number of DNN layers will make the network structure more complex. At this time, the hyperparameters of DNN (such as the number of network layers, the number of hidden units, the activation function, and the optimization method, etc.) are more difficult to adjust, and it takes considerable time and effort to set and adjust them. Because the setting and adjustment of DNN hyperparameters are usually done manually based on a lot of experience and expertise. To solve this problem, this paper proposes a method for automatically optimizing DNN hyperparameters based on particle swarms.

[0049] Particle swarm algorithm, also known as bird flock algorithm, is inspired by the predation behavior of bird flocks. It belongs to genetic algorithm and swarm intelligence algorithm. Particle swarm algorithm focuses on two properties of particles: position and speed. Each particle searches separately in space. They remember the optimal solution they have found and the optimal solution currently found by the entire particle swarm. Where to go next depends on the current direction of the particle, the direction of the optimal solution they have found, and the direction of the current optimal solution of the entire particle swarm. The main process is as follows: First, PSO initializes a position Speed A population of individuals, each of which corresponds to a random candidate solution of the object function. Then, according to the fitness value, the equation is used to update the single extreme value pbest, i.e. the local optimal solution, and the global extreme value gbest, the global optimal solution. The particle swarm optimization is used for the dnn model hyperparameters. The parameter optimization process based on the particle swarm optimization algorithm can be summarized as the process of solving a mathematical optimization problem:

[0050]

[0051]

[0052] Among them, v id and x id Represents the kth generation and the ith generation respectively t The velocity component and position component of the particle, c1 and c2 are the learning rates, which control the evolution amplitude to the single best particle and the global best particle respectively. The above process will be repeated until the expected error value is reached or the maximum number of iterations is reached. Finally, PSO outputs the best position of the particle, which corresponds to the optimal solution to the problem. PSO can be a good choice for solving optimization problems in DNNs due to its easy implementation and minimal adjustable parameters.

[0053] This study introduces the particle swarm method, whose goal is to optimize the best learning rate and weight decay of DNN. The core idea of ​​this work is: through the particle swarm optimization algorithm, automatically search for the optimal learning rate and L2 regularization strength to optimize the performance of the deep neural network, thereby improving the accuracy and generalization ability of the model. In the iterative evolution process, it gradually converges to the optimal value as the population evolves. Then the evaluation function connects the particle swarm optimization algorithm and model training, first receives the parameters, then uses them to train the model, and returns the loss on the validation set of the model as an evaluation indicator. This loss value is used for the search process of the particle swarm optimization algorithm. According to the optimal position of the individual particle swarm, it iteratively updates the position of each particle swarm, and then finds the optimal hyperparameter combination, and then trains the final model. Then find the optimal hyperparameter combination and train the final optimal model. Specifically, it includes the following 4 steps:

[0054] 1) Initialize parameters and population. In the constructor of the PSO class, some parameters of the PSO algorithm are initialized, including the inertia weight w, the individual and social learning factors c1 and c2, the population size pN, the dimension dim, and the maximum number of iterations max_iter. Then, the init_Population method is called to initialize the position and velocity of the population, initialize the individual optimal solution and the global optimal solution, and then randomly initialize the values ​​of these hyperparameters within the given value range of the hyperparameters.

[0055] 2) DNN model construction. The DNN model is constructed layer by layer using the current values ​​of the hyperparameters.

[0056] 3) Obtain optimization results. Obtain the global optimal solution, that is, the hyperparameter combination that minimizes the validation set loss found during the entire optimization process, and record the fitness value of the global optimal solution after each iteration (that is, the validation set loss.

[0057] 4) Train the final model using the best hyperparameters. Train the deep regression model and return the model's performance, create a new deep learning model instance, such as the value of the loss function or the accuracy on the validation set, and use the best hyperparameters found by PSO to configure the optimizer (i.e., learning rate and L2 regularization coefficient).

[0058] The vegetation index was used as input and LAI as output. Five linear layers were included. The initial hyperparameters were set and the particle swarm optimization algorithm was used to optimize the hyperparameters to obtain the optimal hyperparameters: learning rate and L2 regularization intensity. The corn crop model was trained in a small area and then verified in a large area.

[0059] Combining plant height and LAI to fit corn growth potential and grade evaluation:

[0060] LAI inversion model based on particle swarm optimization DNN:

[0061] Use python to optimally fit the extracted different indexes with LAI and plant height. Specifically, first use some commonly used fitting methods, including linear and nonlinear (binomial, exponential, power and logarithmic) fitting functions, to construct the correlation between different vegetation indices and LAI and plant height. Then, for each index, by comparing and analyzing the correlation between the index and LAI and plant height fitted by different fitting methods, the fitting method with the best correlation with LAI and plant height is selected as the optimal fitting model between the two. There are 8 commonly used weight calculation methods, namely AHP hierarchical analysis method, priority diagram method, entropy method, principal component analysis method, factor analysis method, CRI TIC weight, independence weight, and information weight. The comparison of its analysis principles is shown in Table 2 Weight calculation principle:

[0062] Table 2 Weight calculation principle

[0063]

[0064] The entropy method uses data entropy information, i.e., the amount of information, to calculate weights. Information is a measure of the order of a system, and entropy is a measure of the disorder of a system; according to the definition of information entropy, for a certain indicator, the entropy value can be used to judge the degree of dispersion of a certain indicator. The smaller the information entropy value, the greater the dispersion of the indicator, and the greater the impact of the indicator on the comprehensive evaluation (i.e., weight). If the values ​​of a certain indicator are all equal, the indicator will not work in the comprehensive evaluation. Therefore, the information entropy tool can be used to calculate the weights of each indicator to provide a basis for the comprehensive evaluation of multiple indicators. This type of method is suitable for data with fluctuations, and at the same time, it will use data fluctuations as a method of information. The entropy method is used frequently in papers. In the present invention, leaf area index and plant height data are used as indicators.

[0065] Entropy weight method steps:

[0066] 1. Since the measurement units of various indicators are not uniform, they must be standardized before they are used to calculate the comprehensive indicators, that is, the absolute values ​​of the indicators are converted into relative values, so as to solve the homogeneity problem of the values ​​of various indicators of different qualities. Data standardization first de-dimensionalizes each indicator. Assuming that m indicators are given:

[0067] X1,X2,...X m

[0068] in:

[0069] X i ={x1,x2,……x n}

[0070] Assume that the standardized values ​​of each indicator data are:

[0071] Y1,Y2,...Y m

[0072] The positive and negative indicators have different meanings (the higher the positive indicator value, the better, and the lower the negative indicator value, the better). Therefore, different algorithms are needed to normalize the data for positive and negative indicators:

[0073]

[0074] 2. Find the ratio of each indicator under each plan.

[0075] Assume that a certain primary indicator has m secondary indicators and has obtained data for n years, recorded as a matrix. Under the same indicator, calculate the proportion of each year's value to the total value, the formula is as follows:

[0076]

[0077] 3. Calculate the information entropy of each indicator.

[0078] According to the definition of information entropy in information theory, the information entropy of a set of data is:

[0079] (If p ij =0, define E j =0)

[0080] 4. Determine the weight of each indicator.

[0081] According to the calculation formula of information entropy, the information entropy of each indicator is calculated as E1, E2, …, E m

[0082] 4.1 Calculate the weight of each indicator through information entropy.

[0083]

[0084] Here k refers to the number of indicators, that is, k=m.

[0085] 4.2 Calculate the weight by calculating the information redundancy.

[0086] D j =1-E j

[0087] Then calculate the indicator weights:

[0088]

[0089] 5. Finally, calculate the comprehensive score of each plan.

[0090]

[0091] Principal Component Analysis and Factor Analysis:

[0092] The principal component analysis and factor analysis methods utilize the principle of data dimensionality reduction when determining weights, and mainly use characteristic roots, variance explanation rate and loading coefficient to calculate weights.

[0093] An embodiment of the present invention provides an inversion system for crop growth, comprising:

[0094] The data acquisition module is used to obtain the red edge index REIP and the red edge index IRECI according to the reflectivity of the spectral image of the crop planting area; and the red edge index REIP and the red edge index IRECI are combined in different proportions to obtain the combination index of the first proportion and the combination index of the second proportion. The model training module is used to train the first deep neural network DNN to predict the leaf area index through the combination index of the first proportion, and optimize the learning rate and regularization parameter of the first deep neural network DNN through the particle swarm method during the training process to construct the leaf area index inversion model; train the second deep neural network DNN through the combination index of the second proportion to predict the plant height data and obtain the plant height inversion model. The growth inversion module is used to combine the red edge index REIP and the red edge index IRECI of the spectral image of the crop planting area to be tested according to the first proportion and the second proportion respectively; substitute the combination index of the first proportion to be tested into the leaf area index inversion model to obtain the leaf area index, and substitute the combination index of the second proportion to be tested into the plant height inversion model to obtain the plant height data; fit the leaf area index and the plant height data through the entropy weight method to obtain the index weight result for evaluating the growth of crops.

[0095] An embodiment of the present invention provides a computer device, comprising: a memory and a processor; the memory stores a computer program, and the processor implements the steps of an inversion method for crop growth when executing the computer program.

[0096] An embodiment of the present invention provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of a method for inverting the growth of crops are implemented.

[0097] A specific embodiment is as follows:

[0098] This study used Sentinel-2 images as the basis of remote sensing data, combined with field-measured plant height and LAI data.

[0099] Firstly, the corn-planting area was extracted from the preprocessed sentinel remote sensing images, and then the eight vegetation indices, including NDVI, PVI, SAVI, IRECI, PSSRa, MTCI, REIP and MCARI, were calculated for the corn-planting areas. Multiple vegetation indices were obtained according to the reflectance contained in the spectral image of the crop-planting area. The feature recursive elimination algorithm was used to select the multiple vegetation indices, and the red edge index REIP and the red edge index I RECI were selected. REIP and IRECI were determined as the preferred red edge vegetation indices, and the red edge index REIP and the red edge index I RECI were combined in different proportions.

[0100] Then, the plant height inversion model and LAI inversion model were established based on the REIP and IRECI red edge indexes, respectively. The plant height inversion model was based on the DNN model, and 80% of the measured data was used as training data. Because the LAI inversion model directly established using DNN was not effective, the particle swarm method was introduced to optimize the learning rate and regularization parameters in DNN, and a DNN inversion LAI model based on particle swarm optimization was established.

[0101] Finally, the plant height and LAI obtained after remote sensing inversion were fitted using the entropy weight method to obtain the remote sensing inversion results of corn growth combined with key indicators, and an artificial grading evaluation of its growth was performed.

[0102] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for inverting the growth potential of crops, characterized in that: The following steps are involved: Obtain a red edge index REIP and a red edge index IRECI according to the reflectance of the spectral image of the crop planting area; and combine the red edge index REIP and the red edge index IRECI in different proportions to obtain a combination index of a first proportion and a combination index of a second proportion; The first deep neural network DNN is trained by the combination index of the first ratio to predict the leaf area index, and the learning rate and regularization parameter of the first deep neural network DNN are optimized by the particle swarm method during the training process to construct a leaf area index inversion model; the second deep neural network DNN is trained by the combination index of the second ratio to predict the plant height data and obtain the plant height inversion model; The red edge index REIP and the red edge index IRECI of the spectral image of the crop planting area to be tested are respectively combined according to a first ratio and a second ratio; the combined index of the first ratio to be tested is substituted into the leaf area index inversion model to obtain the leaf area index, and the combined index of the second ratio to be tested is substituted into the plant height inversion model to obtain the plant height data; The leaf area index and plant height data were fitted by the entropy weight method to obtain the index weight results for evaluating the growth of crops.

2. The inversion method for crop growth as claimed in claim 1, characterized in that: The obtaining of the leaf area index inversion model comprises the following specific steps: Obtaining a red edge index REIP and a red edge index IRECI combined according to a first ratio; Initial hyperparameters are set in the first deep neural network DNN containing 5 linear layers, and the hyperparameters are optimized using the particle swarm optimization algorithm to obtain the learning rate and L2 regularization strength. The optimized first deep neural network DNN is trained through the first proportional combined index to predict the leaf area index, so as to construct a leaf area index inversion model.

3. The inversion method for crop growth as claimed in claim 1, characterized in that: The specific steps of obtaining the plant height inversion model include: Obtaining a red edge index REIP and a red edge index IRECI combined according to a second ratio; Hyperparameters were set in the second deep neural network DNN containing 5 linear layers, the learning rate was set to 0.0001, the regularization strength was adjusted to 0.2, and the second deep neural network DNN was trained through the combination index of the second ratio to predict plant height data, and a plant height inversion model was obtained.

4. The inversion method for crop growth as claimed in claim 1, characterized in that: The red edge index REIP and the red edge index IRECI are combined in different proportions, and the specific steps include: The red edge index REIP and the red edge index IRECI are combined in a first ratio of 0.3:0.7; The red edge index REIP and the red edge index IRECI are combined in a second ratio of 0.2:0.

8.

5. The inversion method for crop growth as claimed in claim 1, characterized in that: The obtaining of the red edge index REIP and the red edge index IRECI specifically includes: According to the reflectance contained in the spectral image of the crop planting area, multiple vegetation indices are obtained, and the feature recursive elimination algorithm is used to perform feature selection on the multiple vegetation indices to select the red edge index REIP and the red edge index IRECI.

6. The inversion method for crop growth as claimed in claim 1, characterized in that: The specific steps of obtaining the index weight results for evaluating the growth of crops include: The leaf area index and plant height data were standardized to eliminate the effects of different dimensions; The information entropy of each indicator is obtained from the standardized data, and the weight result is obtained using information redundancy.

7. A crop growth inversion system, characterized in that: include: A data acquisition module, used for acquiring a red edge index REIP and a red edge index IRECI according to the reflectance of the spectral image of the crop planting area; and combining the red edge index REIP and the red edge index IRECI in different proportions to obtain a combination index of a first proportion and a combination index of a second proportion; A model training module is used to train a first deep neural network DNN through a combination index of a first ratio to predict a leaf area index, and optimize a learning rate and a regularization parameter of the first deep neural network DNN through a particle swarm method during the training process to construct a leaf area index inversion model; train a second deep neural network DNN through a combination index of a second ratio to predict plant height data, and obtain a plant height inversion model; The growth inversion module is used to combine the red edge index REIP and the red edge index IRECI of the spectral image of the crop planting area to be tested according to the first ratio and the second ratio respectively; substitute the combined index of the first ratio to be tested into the leaf area index inversion model to obtain the leaf area index, and substitute the combined index of the second ratio to be tested into the plant height inversion model to obtain the plant height data; fit the leaf area index and the plant height data through the entropy weight method to obtain the indicator weight result for evaluating the growth of crops.

8. A computer device comprising: Memory and processor; The memory stores a computer program, wherein the processor implements a crop growth inversion method according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for inverting the growth potential of crops as claimed in any one of claims 1 to 6 is implemented.

Images