Method and apparatus for radiation modeling of microwave rainforest radiation field automatically mined by artificial intelligence

Abstract

The present invention discloses a method and device for microwave rainforest radiation field radiation modeling by artificial intelligence automatic mining, belonging to the technical field of electronic information. The method includes: selecting a series of time points, obtaining the measured values of the internal environment temperature field and humidity field of the rainforest; and obtaining the observation geometric angles when the external microwave radiation reference source observes the rainforest and the observation results of the external microwave radiation reference source; using the measured values of the internal environment temperature field and humidity field of the rainforest and the observation geometric angles when the microwave radiation reference source observes the rainforest as the input of the neural network, and using the observation results of the external microwave radiation reference source as the output of the neural network to train the neural network; inputting the measured values of the rainforest temperature field, the rainforest humidity field, the observation zenith angle and the azimuth angle into the trained model, and obtaining the calibration results of the remote sensing instrument at any time through the network output. The present invention can provide the theoretical radiation values of the remote sensing instrument under different observation geometric conditions under any working conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic information, and particularly to a method and device for radiation modeling of a microwave rainforest radiation field for artificial intelligence automatic mining. Background Art

[0002] As a surface radiation calibration target for microwave band space remote sensing instruments, the rainforest has the advantages of stable brightness temperature and relatively high degree of isotropy in all directions. For example, the Amazon rainforest is a commonly used radiation calibration field for microwave remote sensing instruments. Calibration based on a radiation calibration field often uses a co-pupil observation method. When a remote sensing satellite observes a radiation calibration field, through near-simultaneous observation by a reference satellite and in cooperation with ground observation equipment, radiation calibration field observation data with a radiation reference is formed to perform radiation calibration on the remote sensing instrument to be calibrated.

[0003] However, in fact, for microwave remote sensing, when a specific microwave remote sensing instrument observes the rainforest, the microwave penetration determines that the observed radiation amount (or brightness temperature) is a complex mixture of various factors such as surface emissivity, rainforest temperature environment, humidity environment, and canopy emissivity according to the weights formed under different observation conditions. This complex process leads to relatively large uncertainty in the observation of the rainforest calibration field by a single microwave remote sensing instrument. Although the rainforest is a relatively good ground-based radiation calibration field for microwave remote sensing instruments, there is still a certain gap for high-precision calibration requirements.

[0004] The information disclosed in this background art section is only intended to increase the understanding of the overall background of the present invention and should not be regarded as an admission or any form of suggestion that this information constitutes prior art already known to those of ordinary skill in the art. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and device for radiation modeling of a microwave rainforest radiation field for artificial intelligence automatic mining to solve some or all of the above problems.

[0006] To achieve the above purpose, the present invention provides a method for radiation modeling of a microwave rainforest radiation field for artificial intelligence automatic mining, including the following steps:

[0007] S1: Select a series of time points, obtain the measured values of the internal environment temperature field and humidity field of the rainforest at this series of time points; and obtain the observation geometric angle when an external microwave radiation reference source observes the rainforest and the observation result of the external microwave radiation reference source.

[0008] S2: Take the measured values of the internal environment temperature field and humidity field of the rainforest at a certain time point and the observation geometric angle when the microwave radiation reference source observes the rainforest as the input of the neural network, and take the observation result of the external microwave radiation reference source corresponding to this time point as the output of the neural network to train the neural network.

[0009] S3: Input the measured values of the rainforest temperature field, the measured values of the rainforest humidity field, the observed zenith angle, and the azimuth angle into the trained model, and obtain the calibration result of the remote sensing instrument at any time through the network output.

[0010] In an embodiment of the present invention, the observation geometric angles of the external microwave radiation reference source when observing the rainforest include the zenith angle and the azimuth angle.

[0011] In an embodiment of the present invention, step S2 specifically includes the following steps:

[0012] S201: Take the measured values of the temperature field of the internal environment of the rainforest , the measured values of the rainforest humidity field , the observed zenith angle and the azimuth angle as inputs together, and take the observation result of the external microwave radiation reference source corresponding to this time point

[0013] as the output of the neural network as the label;

[0014] S202: Perform normalization processing on all inputs and labels;

[0015] S203: Create a fully connected type containing an input layer, at least one hidden layer, and an output layer as the neural network model to be trained;

[0016] In an embodiment of the present invention, step S304 specifically includes the following steps: During the training process, calculate the predicted value output by the neural network through forward propagation, calculate the loss value of the current model according to the predicted value and the true label y ; Calculate and record the gradient according to the loss value in each round of training, then clear the gradient and execute the backpropagation algorithm according to the loss value to calculate the new gradient value of the model parameters, and adjust the parameters of the model according to the calculated new gradient value; Repeat the above operations until the loss value is less than the set threshold, and stop the training process.

[0017] In an embodiment of the present invention, in step S304:

[0018] The loss function adopts form, where represents the predicted value output by the neural network, y represents the cross-calibration result value obtained in step S2.

[0019] The present invention also provides a microwave rainforest radiation field radiation modeling device for artificial intelligence automatic mining, including: a data processing module, a training module, and a data acquisition module;

[0020] The data processing module is used to select a series of time points, obtain the measured values of the internal environment temperature field and humidity field of the rainforest at this series of time points; and obtain the observation geometric angles when the external microwave radiation reference source observes the rainforest and the observation results of the external microwave radiation reference source;

[0021] The training module is used to use the measured values of the internal environment temperature field and humidity field of the rainforest at a certain time point, and the observation geometric angles when the microwave radiation reference source observes the rainforest as the input of the neural network, and use the observation results of the external microwave radiation reference source corresponding to this time point as the output of the neural network to train the neural network;

[0022] The data acquisition module is used to input the measured values of the rainforest temperature field, the measurement of the rainforest humidity field, the observation zenith angle and the azimuth angle into the trained model, and obtain the calibration result of the remote sensing instrument at any time through the network output.

[0023] In an embodiment of the present invention, the observation geometric angles when the external microwave radiation reference source observes the rainforest include the zenith angle and the azimuth angle.

[0024] In an embodiment of the present invention, the training module includes: an input sub-module, a processing sub-module, a model creation sub-module, and a training sub-module

[0025] The input sub-module is used to use the measured value of the temperature field of the internal environment of the rainforest , the measured value of the rainforest humidity field , the observation zenith angle and the azimuth angle together as the input, and use the observation results of the external microwave radiation reference source corresponding to this time point as the output of the neural network as the label;

[0026] The processing sub-module is used to perform normalization processing on all inputs and labels;

[0027] The model creation sub-module is used to create a fully connected type containing an input layer, at least one hidden layer, and an output layer as the neural network model to be trained;

[0028] The training sub-module is used to train the neural network model with the normalized inputs and labels.

[0029] In an embodiment of the present invention, the training sub-module is specifically configured to: during the training process, forward-propagate to calculate the predicted value y_pred output by the neural network, calculate the loss value of the current model according to the predicted value y_pred and the true label y; calculate and record the gradient according to the loss value in each round of training, then clear the gradient and execute the backpropagation algorithm according to the loss value to calculate the new gradient value of the model parameters, and adjust the parameters of the model according to the calculated new gradient value; repeat the above operations until the loss value is less than the set threshold, and stop the training process.

[0030] In an embodiment of the present invention, the loss function adopts form, where represents the predicted value output by the neural network, y represents the cross-calibration result value obtained in step S2.

[0031] Compared with the prior art, the microwave rainforest radiation field radiation modeling method and device automatically mined by artificial intelligence according to the present invention utilize the powerful learning ability of the neural network to automatically mine the complex relationship between the external radiation performance of the rainforest and the indication parameters through artificial intelligence, greatly improving the efficiency and accuracy of modeling. Artificial intelligence can process a large amount of data and quickly find the rules and relationships therein, reducing manual intervention and errors, and ensuring the theoretical radiation values of remote sensing instruments under different observation geometric conditions in any working condition (i.e., the state represented by the rainforest temperature field and humidity field). BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a flowchart of a microwave rainforest radiation field radiation modeling method automatically mined by artificial intelligence according to an embodiment of the present invention;

[0033] Figure 2 is a schematic structural diagram of a microwave rainforest radiation field radiation modeling device automatically mined by artificial intelligence according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0034] The following will describe in detail the specific embodiments of the present invention with reference to the accompanying drawings, but it should be understood that the protection scope of the present invention is not limited by the specific embodiments.

[0035] Unless otherwise clearly stated, throughout the specification and claims, the term "comprising" or its variations such as "comprises" or "including" etc. will be understood to include the stated elements or components, without excluding other elements or other components.

[0036] Such as Figure 1As shown in the figure, the microwave rainforest radiation field radiation modeling method automatically mined by artificial intelligence according to the preferred embodiment of the present invention is applied to the calibration device for the radiation modeling of the microwave rainforest radiation correction field automatically mined by artificial intelligence, that is, the device that executes the microwave rainforest radiation correction field modeling method automatically mined by artificial intelligence.

[0037] The present invention is based on the basic principle of microwave remote sensing to establish a microwave rainforest radiation correction field radiation modeling method automatically mined by artificial intelligence. In the present invention, the internal environmental variables of the rainforest (including the temperature field and humidity field) and the time variable (characterizing the growth characteristics of the rainforest at different times) are used as the indication parameters of the external radiation performance of the rainforest. Then, the observed radiation amount (or brightness temperature) of the rainforest by the external microwave radiation reference source (including the microwave remote sensing satellite with high-precision calibration, the manned / unmanned aerial vehicle-borne microwave radiometer, or the vehicle-borne microwave radiometer) is used as the benchmark of the external radiation performance of the rainforest. By automatically mining the complex relationship between the external radiation performance of the rainforest and the indication parameters by artificial intelligence, considering the radiation difference in all directions of the rainforest target, the observation of the external microwave radiation reference source is described by the observation geometric angle and characterized by the zenith angle and azimuth angle.

[0038] Since the process of expressing the relationship between the external radiation performance and the indication parameters is relatively complex, it is difficult to find an accurate formula to describe this objective physical process, especially to find the accurate formula parameter values. Therefore, the present invention selects artificial intelligence technology to automatically mine this relationship, and obtains the neural network through training with a large number of sample data to obtain this relationship.

[0039] Specifically, the present invention includes the following steps:

[0040] S1: Select a series of time points, and obtain the measured values of the internal environmental temperature field and humidity field of the rainforest at this series of time points; and obtain the observation geometric angle when the external microwave radiation reference source observes the rainforest and the observation result of the external microwave radiation reference source.

[0041] In this step, a series of time points are selected as the anchor points for description, which are represented by , where represents the number of time points when the external microwave radiation reference source is used to observe the rainforest radiation correction field.

[0042] The internal environmental temperature field of the rainforest at these time points is represented by , and the humidity field is represented by , where represents the number of rainforest temperature fields that can be measured and selected, and represents the number of rainforest humidity fields that can be measured and selected.

[0043] The observation geometric angle when the external microwave radiation reference source observes the rainforest includes the zenith angle and azimuth angle , the observation results of the external microwave radiation reference source are represented by .

[0044] S2: Take the measured values of the internal environment temperature field and humidity field of the rainforest at a certain time point, and the observation geometric angles during the observation of the rainforest by the microwave radiation reference source as the input of the neural network, and take the observation results of the external microwave radiation reference source corresponding to this time point as the output of the neural network, and train the neural network.

[0045] In this step, the measured value of the temperature field of the internal environment of the rainforest at a specific time point , the measured value of the rainforest humidity field , the observed zenith angle and azimuth angle are combined to jointly form a dimensional vector as the input of the neural network, and then the observation results of the external microwave radiation reference source corresponding to this time point are used as the output of the neural network. At all time points that meet the method requirements , a dimensional input and dimensional output are formed, and then the neural network is trained. The present invention names it the Rainforest Radiometric Neural Network (RRNN). Regarding the number of samples, generally, the work of this application can be carried out and the sample size will be sufficient.

[0046] The neural network can be selected (but not limited to) such as a fully connected neural network, and an optimizer (Adam) can be used (but not limited to).

[0047] When training the neural network, it is specifically carried out through the following steps:

[0048] S201: Take the measured value of the temperature field of the internal environment of the rainforest , the measured value of the rainforest humidity field , the observed zenith angle and azimuth angle together as the input, and take the observation results of the external microwave radiation reference source corresponding to this time point as the output label of the neural network.

[0049] S202: Normalize all inputs and labels:

[0050]

[0051] Among them, represents the th input, and are the maximum and minimum values of the input vector, and are the values after normalization processing. The label normalization processing adopts the same method as above.

[0052] S203: Create a fully connected neural network containing an input layer, at least one hidden layer, and an output layer as the model to be trained.

[0053] In the present invention, the fully connected neural network processes the input data (the measured values of the temperature field, humidity field, observation zenith angle, and azimuth angle of the internal environment of the rainforest) through a series of linear transformations and activation functions, and finally obtains the output (calibration result).

[0054] In the training stage, known inputs and corresponding labels (the results after normalization) are used to adjust the parameters of the model, so that the model can predict the calibration results of new input data as accurately as possible.

[0055] In the present invention, the activation function used is the relu function. After experimentation, other activation functions can also be used if they meet the accuracy requirements.

[0056] The radiation modeling network (RRNN) of the microwave rainforest radiation automatically mined by artificial intelligence adopts the following code (taking Python code as an example, but not limited to Python code):

[0057] model = Sequential(

[0058] Linear(input_feature, hidden_feature),

[0059] ActiveFunction,

[0060] Linear(hidden_feature, hidden_feature),

[0061] ActiveFunction,

[0062] Linear(hidden_feature, output_feature),

[0063] ActiveFunction )

[0065] S204: Train the model with the input and labels after normalization processing.

[0066] During the training process, use the forward propagation of RRNN to calculate the predicted values of the neural network output , according to the predicted value and the true label y calculate the loss value of the current model ; according to the loss values in each round of training , calculate and record the gradients, then clear the gradients and according to the loss values perform the backpropagation algorithm to calculate the new gradient values of the model parameters, and adjust the parameters of the model according to the calculated new gradient values; repeat the above operations until the loss value is less than the set threshold, and stop the training process. In the present invention, the set threshold is 0.1, and this threshold is used to adjust the required radiometric calibration accuracy. The smaller its value, the higher the required radiometric calibration accuracy. When the required radiometric calibration accuracy decreases, its value can be increased.

[0067] During the training process of the model, an optimizer is also set, which is used to update the parameters of the model according to the calculated gradients, so that the model is adjusted in the direction of reducing the loss. Among them, the parameters of the model are the network and training parameters during the training process.

[0068] Among them, the loss function adopts form, where represents the predicted value output by the neural network, y represents the label, that is, the observation result of the external microwave radiation reference source obtained in step S1. This formula only shows the form of the loss function.

[0069] During the training process, through multiple loops, the parameters of the model are continuously adjusted to minimize the loss function, so that the model can better fit the given training data.

[0070] The logical process of training is in the form of the following code (taking Python code as an example, but not limited to Python code):

[0071] for epoch in tqdm.tqdm(range(1, epochs+1)):

[0072] # Forward propagation

[0073] y_pred = model(x)

[0074] loss = loss_f(y_pred, y)

[0075] Epoch.append(epoch)

[0076] Loss.append(loss.data)

[0077] if epoch % 100 == 0:

[0078] print("Epoch:{},loss:{}".format(epoch, loss))

[0079] optim.zero_grad()

[0080] # Backward propagation

[0081] loss.backward()

[0082] # Parameter fine-tuning

[0083] optim.step()

[0084] By continuously repeating forward propagation, loss calculation, backward propagation, and parameter update, the parameters of the model are gradually adjusted so that the model can better fit the training data. During the training process, the training effect of the model can be monitored by observing the change of the loss value, using the validation set for evaluation, etc., and hyperparameters (such as learning rate, number of layers, number of neurons, etc.) can be adjusted as needed to improve the performance of the model.

[0085] S3: Input the measured values of the rainforest temperature field, the measurement of the rainforest humidity field, the observed zenith angle and azimuth angle into the trained model, and obtain the calibration result of the remote sensing instrument at any time through the network output.

[0086] The calibration result is:

[0087] .

[0088] When the external microwave radiation reference source is a microwave remote sensing satellite, the calibration result in the above formula R is the observation result of the atmosphere appearance of the rainforest radiation correction field. When the external radiation reference source is airborne / vehicle-mounted, the calibration result in the above formula R is the radiation value of the upper layer of the rainforest, not the observation result of the atmosphere appearance. It is necessary to further calculate its theoretical observation value outside the atmosphere through radiative transfer.

[0089] Based on the high-precision fitting ability of the neural network for functions, the present invention takes the physical factors affecting the calibration accuracy (i.e., the rainforest temperature field and humidity field) as part of the input, ensuring that the method can obtain the theoretical radiation values of the remote sensing instrument under different observation geometric conditions under any working conditions (i.e., the states represented by the rainforest temperature field and humidity field).

[0090] As Figure 2 shown, the microwave rainforest radiation field radiation modeling device automatically mined by artificial intelligence according to the preferred embodiment of the present invention includes: a data processing module 1, a training module 2, and a data acquisition module 3.

[0091] The data processing module 1 is used to select a series of time points, obtain the measured values of the internal environmental temperature field and humidity field of the rainforest at this series of time points; and obtain the observation geometric angles when the external microwave radiation reference source observes the rainforest and the observation results of the external microwave radiation reference source.

[0092] Specifically, a series of time points are selected as the described anchor points, denoted here by , where represents the number of time points when the external microwave radiation reference source is used to observe the rainforest radiation correction field.

[0093] The internal environmental temperature field of the rainforest at these time points is denoted by , and the humidity field is denoted by , where represents the number of rainforest temperature fields that can be measured and selected, represents the number of rainforest humidity fields that can be measured and selected.

[0094] The observation geometric angles when the external microwave radiation reference source observes the rainforest include the zenith angle and the azimuth angle , and the observation results of the external microwave radiation reference source are denoted by .

[0095] The training module 2 is used to take the measured values of the internal environmental temperature field and humidity field of the rainforest at a certain time point, and the observation geometric angles when the microwave radiation reference source observes the rainforest as the input of the neural network, and take the observation results of the external microwave radiation reference source corresponding to this time point as the output of the neural network to train the neural network.

[0096] In the training module 2, the measured value of the temperature field of the internal environment of the rainforest at a specific time point , the measured value of the humidity field of the rainforest , the observed zenith angle and the azimuth angle are combined to jointly form -dimensional vector as the input of the neural network, and then the observation result of the external microwave radiation reference source corresponding to this time point is used as the output of the neural network. At all time points that meet the method requirements, -dimensional input and -dimensional output are formed, and then the neural network is trained. The present invention names it the Rainforest Radiometric Neural Network (RRNN). Regarding the number of samples, generally, the work of this application can be carried out and the sample size will be sufficient.

[0097] The neural network can be selected (but not limited to) as a fully connected neural network, and an optimizer (Adam) can be adopted (but not limited to).

[0098] When training the neural network, the training module 2 specifically includes:[[]]

[0099] An input sub-module 201, which is used to use the temperature field measurement value of the internal environment of the rainforest , the rainforest humidity field measurement value , the observed zenith angle and the azimuth angle as inputs together, and use the observation result of the external microwave radiation reference source corresponding to this time point as the output of the neural network as the label;

[0100] A processing sub-module 202, which is used to perform normalization processing on all inputs and labels:

[0101]

[0102] Among them, represents the th input, and are the maximum and minimum values of this input vector, is the value after normalization processing. The label normalization processing adopts the same method as above.

[0103] A model creation sub-module 203, which is used to create a fully connected type neural network containing an input layer, at least one hidden layer and an output layer as the model to be trained.

[0104] In the present invention, the fully connected type neural network processes the input data (the output value of the undetermined calibration remote sensing instrument and the optical path temperature measurement value) through a series of linear transformations and activation functions, and finally obtains the output (calibration result). In the training stage, the known inputs and the corresponding labels (the calibrated results after normalization) are used to adjust the parameters of the model, so that the model can predict the calibrated results of new input data as accurately as possible.

[0105] In the present invention, the activation function used is the relu function. After experiments, other activation functions can also be used if the accuracy requirements are met.

[0106] The radiation modeling network (RRNN) of the microwave rainforest radiation field automatically mined by artificial intelligence adopts the following code (taking Python code as an example, but not limited to Python code) form:

[0107] model = Sequential(

[0108] Linear(input_feature, hidden_feature),

[0109] ActiveFunction,

[0110] Linear(hidden_feature, hidden_feature),

[0111] ActiveFunction,

[0112] Linear(hidden_feature, output_feature),

[0113] ActiveFunction )

[0115] The training sub-module 204 is used to train the model with the normalized input and labels.

[0116] During the training process, the training sub-module 204 is used to calculate the predicted value of the neural network output using the RRNN forward propagation , according to the predicted value and the true label y to calculate the loss value of the current model ; according to the loss values in each round of training , calculate and record the gradients, then clear the gradients and execute the backpropagation algorithm according to the loss value to calculate the new gradient values of the model parameters, and adjust the model parameters according to the calculated new gradient values; repeat the above operations until the loss value is less than the set threshold, and stop the training process. In the present invention, the set threshold is 0.1, and this threshold is used to adjust the required radiometric calibration accuracy. The smaller the value, the higher the required radiometric calibration accuracy. When the required radiometric calibration accuracy decreases, its value can be increased.

[0117] During the training process of the model, an optimizer is also set, which is used to update the model parameters according to the calculated gradients, so that the model is adjusted in the direction of reducing the loss. Among them, the model parameters are the network and training parameters during the training process.

[0118] Among them, the loss function adopts form, where represents the predicted value of the neural network output, y represents the label, that is, the observation result of the external microwave radiation reference source. This formula only shows the form of the loss function.

[0119] During the training process, through multiple loops, the parameters of the model are continuously adjusted to minimize the loss function, enabling the model to better fit the given training data.

[0120] The logical process of training is in the form of the following code (taking Python code as an example, but not limited to Python code):

[0121] for epoch in tqdm.tqdm(range(1, epochs+1)):

[0122] # Forward propagation

[0123] y_pred = model(x)

[0124] loss = loss_f(y_pred, y)

[0125] Epoch.append(epoch)

[0126] Loss.append(loss.data)

[0127] if epoch % 100 == 0:

[0128] print("Epoch:{},loss:{}".format(epoch, loss))

[0129] optim.zero_grad()

[0130] # Backward propagation

[0131] loss.backward()

[0132] # Parameter fine-tuning

[0133] optim.step()

[0134] By continuously repeating forward propagation, calculating the loss, backward propagation, and parameter update, the parameters of the model are gradually adjusted to enable the model to better fit the training data. During the training process, the training effect of the model can be monitored by observing the change of the loss value, using the validation set for evaluation, etc., and hyperparameters (such as learning rate, number of layers, number of neurons, etc.) can be adjusted as needed to improve the performance of the model.

[0135] The data acquisition module 3 is used to input the measured values of the rainforest temperature field, the measurement of the rainforest humidity field, the observed zenith angle and azimuth angle into the trained model, and obtain the calibration result of the remote sensing instrument at any time through the network output.

[0136] The calibration result is:

[0137] 。

[0138] When the external microwave radiation reference source is a microwave remote sensing satellite, the calibration result in the above formula R is the observation result of the outer atmosphere of the rainforest radiation calibration field. When the external radiation reference source is airborne / vehicle-mounted, the calibration result in the above formula R is the radiation value of the upper layer of the rainforest and is not the observation result of the outer atmosphere. It is necessary to further calculate its theoretical observation value outside the atmosphere through radiative transfer.

[0139] The foregoing description of the specific exemplary embodiments of the present invention is for purposes of illustration and exemplification. These descriptions are not intended to limit the invention to the precise forms disclosed, and it is apparent that many changes and variations are possible in light of the above teaching. The purpose of selecting and describing the exemplary embodiments is to explain the specific principles of the present invention and its practical applications, so that those skilled in the art can implement and utilize the various different exemplary embodiments of the present invention, as well as various different selections and changes. The scope of the present invention is intended to be defined by the claims and their equivalents.

Claims

1. A microwave rainforest radiation field modeling method automatically mined by artificial intelligence, characterized in that: The following steps are involved: S1: Select a series of time points to obtain the measured values ​​of the temperature field and humidity field of the rainforest environment at the series of time points; and obtain the observation geometry angle when an external microwave radiation reference source observes the rainforest and the observation results of the external microwave radiation reference source; S2: Take the measured values ​​of the temperature field and humidity field of the rainforest environment at a certain time point and the observed geometric angle when the rainforest is observed by the microwave radiation reference source as the input of the neural network, and take the observation result of the external microwave radiation reference source corresponding to the time point as the output of the neural network to train the neural network; S3: Input the measured values ​​of the rainforest temperature field, the rainforest humidity field, the observed zenith angle and the azimuth angle into the trained model, and obtain the calibration results of the remote sensing instrument at any time through the network output; Wherein, the observation geometric angles of the external microwave radiation reference source when observing the rainforest include the zenith angle and the azimuth angle; Step S2 specifically includes the following steps: S201: The temperature field measurement value of the internal environment of the rainforest , Rainforest humidity field measurement values , Observation zenith angle and azimuth The observation results of the external microwave radiation reference source corresponding to this time point are taken as input. The output of the neural network is the label; S202: normalize all inputs and labels; S203: creating a fully connected type including an input layer, at least one hidden layer and an output layer as a neural network model to be trained; S204: Train the neural network model using the normalized input and labels.

2. The microwave rainforest radiation field radiation modeling method automatically mined by artificial intelligence according to claim 1, characterized in that: Step S304 specifically includes the following steps: During the training process, forward propagation calculates the predicted value output by the neural network, and based on the predicted value and the true label y Calculate the loss value of the current model; calculate and record the gradient according to the loss value in each round of training, then clear the gradient and execute the back propagation algorithm according to the loss value to calculate the new gradient value of the model parameters, and adjust the parameters of the model according to the calculated new gradient value; repeat the above operation until the loss value is less than the set threshold, and stop the training process.

3. The microwave rainforest radiation field radiation modeling method automatically mined by artificial intelligence as claimed in claim 2 is characterized in that: In step S304: The loss function is form, in which represents the predicted value of the neural network output, y represents the cross calibration result value obtained in step S2.

4. A microwave rainforest radiation field modeling device automatically mined by artificial intelligence, characterized in that: include: Data processing module, training module and data acquisition module; The data processing module is used to select a series of time points to obtain the measured values ​​of the temperature field and humidity field of the rainforest environment at the series of time points; and to obtain the observation geometric angle when the external microwave radiation reference source observes the rainforest and the observation results of the external microwave radiation reference source; The training module is used to use the measured values ​​of the temperature field and humidity field of the rainforest environment at a certain time point and the observed geometric angle when the rainforest is observed by the microwave radiation reference source as the input of the neural network, and the observation result of the external microwave radiation reference source corresponding to the time point as the output of the neural network to train the neural network; The data acquisition module is used to input the rainforest temperature field measurement value, rainforest humidity field measurement, observation zenith angle and azimuth angle into the trained model, and obtain the calibration result of the remote sensing instrument at any time through network output; Wherein, the observation geometric angles of the external microwave radiation reference source when observing the rainforest include the zenith angle and the azimuth angle; The training module includes: an input submodule, a processing submodule, a model creation submodule and a training submodule The input submodule is used to measure the temperature field of the rainforest internal environment. , Rainforest humidity field measurement values , Observation zenith angle and azimuth The observation results of the external microwave radiation reference source corresponding to this time point are taken as input. The output of the neural network is the label; The processing submodule is used to normalize all inputs and labels; The model creation submodule is used to create a fully connected type neural network model to be trained, which includes an input layer, at least one hidden layer and an output layer; The training submodule is used to train the neural network model using the normalized input and labels.

5. The microwave rainforest radiation field radiation modeling device automatically mined by artificial intelligence as claimed in claim 4 is characterized in that: The training submodule is specifically used for: in the training process, forward propagation is used to calculate the predicted value y_pred output by the neural network, and the loss value of the current model is calculated according to the predicted value y_pred and the true label y; according to the loss value in each round of training, the gradient is calculated and recorded, and then the gradient is cleared and the back propagation algorithm is executed according to the loss value to calculate the new gradient value of the model parameter, and the parameters of the model are adjusted according to the calculated new gradient value; the above operation is repeated until the loss value is less than the set threshold, and the training process is stopped.

6. The microwave rainforest radiation field radiation modeling device automatically mined by artificial intelligence as claimed in claim 5, characterized in that: The loss function is form, in which represents the predicted value of the neural network output, y represents the cross calibration result value obtained in step S2.

Images