CN117669349A

CN117669349A - Model training method, information generating method, device, equipment and medium

Info

Publication number: CN117669349A
Application number: CN202211007301.8A
Authority: CN
Inventors: 靳宏财
Original assignee: Beijing Jingdong Qianshi Technology Co Ltd
Current assignee: Beijing Jingdong Qianshi Technology Co Ltd
Priority date: 2022-08-22
Filing date: 2022-08-22
Publication date: 2024-03-08

Abstract

The embodiment of the disclosure discloses a model training method, an information generating method, a device, equipment and a medium. One embodiment of the method comprises the following steps: acquiring a historical crop scene data sequence and a crop actual growth data sequence; determining at least one crop growth experimental model; inputting the historical crop scene data sequence into at least one crop growth experimental model to obtain a historical crop growth simulation data sequence set; for each historical crop growth simulation data sequence in the historical crop growth simulation data sequence set, carrying out data corresponding combination on the historical crop growth simulation data sequence and the historical crop scene data sequence to obtain a historical crop combination data sequence; and performing model training on the initial crop growth information generation model to obtain a crop growth information generation model. This embodiment is related to artificial intelligence, and the predictive ability of crop generation information using the crop growth information generation model is more accurate.

Description

Model training method, information generating method, device, equipment and medium

Technical Field

The embodiment of the disclosure relates to the technical field of computers, in particular to a model training method, an information generating method, a device, equipment and a medium.

Background

With the development of modern agriculture, prediction of the daily growth conditions of crops is becoming an increasing focus of attention. For the prediction of the daily growth conditions of crops, the following methods are generally adopted: the pre-trained deep learning neural network is utilized to predict the daily growth condition of crops. Wherein the deep learning neural network is trained based on an actual growth condition dataset of the crop.

However, the inventors have found that when the above-described manner is used to predict the daily growth of crops, there are often the following technical problems:

the growth cycle of crops is longer, and the actual growth condition data set collection of the crops is difficult to cause, so that training samples are deficient, and further, the deep learning neural network obtained through training is insufficient in prediction capability for the growth information of the crops.

The above information disclosed in this background section is only for enhancement of understanding of the background of the inventive concept and, therefore, may contain information that does not form the prior art that is already known to those of ordinary skill in the art in this country.

Disclosure of Invention

The disclosure is in part intended to introduce concepts in a simplified form that are further described below in the detailed description. The disclosure is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.

Some embodiments of the present disclosure propose a model training method, an information generating method, an apparatus, a device, and a medium to solve the technical problems mentioned in the background section above.

In a first aspect, some embodiments of the present disclosure provide a model training method, comprising: acquiring a historical crop scene data sequence and a crop actual growth data sequence aiming at target crop information, wherein the historical crop scene data in the historical crop scene data sequence and the crop actual growth data in the crop actual growth data sequence have a time corresponding relation; determining at least one crop growth experimental model associated with the target crop information; inputting the historical crop scene data sequence into each crop growth experimental model in the at least one crop growth experimental model to generate a historical crop growth simulation data sequence, so as to obtain a historical crop growth simulation data sequence set; for each historical crop growth simulation data sequence in the historical crop growth simulation data sequence set, correspondingly combining the historical crop growth simulation data sequence with the historical crop scene data sequence to obtain a historical crop combination data sequence; and performing model training on the initial crop growth information generation model according to the obtained historical crop combination data sequence set and the crop actual growth data sequence to obtain a crop growth information generation model.

Optionally, the training of the initial crop growth information generation model according to the obtained historical crop combination data sequence set and the actual crop growth data sequence to obtain a crop growth information generation model includes: for each crop actual growth data in the crop actual growth data sequence, screening target historical crop growth simulation data meeting preset conditions in relation to the crop actual growth data from a target historical crop growth simulation data set according to the crop growth time corresponding to the crop actual growth data, and taking the target historical crop growth simulation data as candidate historical crop growth simulation data, wherein the target historical crop growth simulation data set is historical crop growth simulation data of which the corresponding time is related to the crop growth time; generating a historical crop combination data sequence corresponding to the obtained candidate historical crop growth simulation data sequence as a target historical crop combination data sequence; and performing model training on the initial crop growth information generation model according to the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model.

Optionally, the model training is performed on the initial crop growth information generation model according to the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model, which includes: taking the target historical crop combined data sequence as a primary training data sequence, and carrying out primary training on the initial crop growth information model to obtain a crop growth information generation model after primary training; and taking the crop actual growth data sequence as a secondary training data sequence, and performing secondary training on the crop growth information generation model after primary training to obtain a crop growth information generation model after secondary training, and taking the crop growth information generation model after secondary training as the crop growth information generation model, wherein model parameters of at least one sub-model at a preset position included in the crop growth information generation model after primary training are not changed in the secondary training process.

Optionally, the method further comprises: acquiring verification data and crop growth verification information aiming at the target crop information, wherein the verification data is farm scene data of a farm where the target crop information corresponding to the crop is located aiming at a first time, and the crop growth verification information is crop actual growth data aiming at the first time; inputting the verification data into the crop growth information generation model to obtain crop growth information; and generating information representing successful verification of the crop growth information generation model in response to determining that the information difference between the crop growth information and the crop growth verification information is less than or equal to a preset threshold.

Optionally, the method further comprises: in response to determining that the information difference between the crop growth information and the crop growth verification information is greater than a preset threshold, adjusting a model structure of the crop growth information generation model to obtain an adjusted crop growth information generation model; and performing model training on the adjusted crop growth information generation model according to the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model.

Optionally, the crop actual growth data in the crop actual growth data sequence includes: crop scene data; and screening target historical crop growth simulation data which has a relation meeting a preset condition with the actual crop growth data from the target historical crop growth simulation data set, wherein the target historical crop growth simulation data is used as candidate historical crop growth simulation data, and the method comprises the following steps: encoding the crop scene data to obtain an actual scene data vector; encoding each historical crop growth simulation data in the target historical crop growth simulation data set to generate a simulation data vector, and obtaining a simulation data vector set; screening simulation data vectors, the distance between which and the actual scene data vectors meet the preset distance condition, from the simulation data vector set to serve as candidate simulation data vectors; and determining the target historical crop growth simulation data corresponding to the candidate simulation data vector as candidate historical crop growth simulation data.

In a second aspect, some embodiments of the present disclosure provide a model training apparatus comprising: a first acquisition unit configured to acquire a historical crop scene data sequence and a crop actual growth data sequence for target crop information, wherein the historical crop scene data in the historical crop scene data sequence has a time correspondence relationship with the crop actual growth data in the crop actual growth data sequence; a determining unit configured to determine at least one crop growth experimental model associated with the target crop information; a first input unit configured to input the historical crop scene data sequence into each of the at least one crop growth experimental model to generate a historical crop growth simulation data sequence, resulting in a historical crop growth simulation data sequence set; a combination unit configured to perform data corresponding combination on the historical crop growth simulation data sequence and the historical crop scene data sequence for each historical crop growth simulation data sequence in the historical crop growth simulation data sequence set to obtain a historical crop combination data sequence; and the training unit is configured to perform model training on the initial crop growth information generation model according to the obtained historical crop combination data sequence set and the crop actual growth data sequence to obtain a crop growth information generation model.

Alternatively, the training unit may be configured to: for each crop actual growth data in the crop actual growth data sequence, screening target historical crop growth simulation data meeting preset conditions in relation to the crop actual growth data from a target historical crop growth simulation data set according to the crop growth time corresponding to the crop actual growth data, and taking the target historical crop growth simulation data as candidate historical crop growth simulation data, wherein the target historical crop growth simulation data set is historical crop growth simulation data of which the corresponding time is related to the crop growth time; generating a historical crop combination data sequence corresponding to the obtained candidate historical crop growth simulation data sequence as a target historical crop combination data sequence; and performing model training on the initial crop growth information generation model according to the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model.

Alternatively, the training unit may be configured to: taking the target historical crop combined data sequence as a primary training data sequence, and carrying out primary training on the initial crop growth information model to obtain a crop growth information generation model after primary training; and taking the crop actual growth data sequence as a secondary training data sequence, and performing secondary training on the crop growth information generation model after primary training to obtain a crop growth information generation model after secondary training, and taking the crop growth information generation model after secondary training as the crop growth information generation model, wherein model parameters of at least one sub-model at a preset position included in the crop growth information generation model after primary training are not changed in the secondary training process.

Alternatively, the training unit may be configured to: acquiring verification data and crop growth verification information aiming at the target crop information, wherein the verification data is farm scene data of a farm where the target crop information corresponding to the crop is located aiming at a first time, and the crop growth verification information is crop actual growth data aiming at the first time; inputting the verification data into the crop growth information generation model to obtain crop growth information; and generating information representing successful verification of the crop growth information generation model in response to determining that the information difference between the crop growth information and the crop growth verification information is less than or equal to a preset threshold.

Alternatively, the training unit may be configured to: in response to determining that the information difference between the crop growth information and the crop growth verification information is greater than a preset threshold, adjusting a model structure of the crop growth information generation model to obtain an adjusted crop growth information generation model; and performing model training on the adjusted crop growth information generation model according to the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model.

Optionally, the crop actual growth data in the crop actual growth data sequence includes: crop scene data; and the training unit may be configured to: encoding the crop scene data to obtain an actual scene data vector; encoding each historical crop growth simulation data in the target historical crop growth simulation data set to generate a simulation data vector, and obtaining a simulation data vector set; screening simulation data vectors, the distance between which and the actual scene data vectors meet the preset distance condition, from the simulation data vector set to serve as candidate simulation data vectors; and determining the target historical crop growth simulation data corresponding to the candidate simulation data vector as candidate historical crop growth simulation data.

In a third aspect, some embodiments of the present disclosure provide an information generating method, including: acquiring crop scene data corresponding to the target farm and aiming at a second time by using the target crop information; the crop scene data is input to a crop growth information generation model to generate crop growth prediction information, wherein the crop growth information generation model is generated by a method as described in any of the implementations of the first aspect.

In a fourth aspect, some embodiments of the present disclosure provide an information generating apparatus including: a second acquisition unit configured to acquire crop scene data for a second time corresponding to the target farm with respect to the target crop information; and a second input unit configured to input the crop scene data to a crop growth information generation model to generate crop growth prediction information, wherein the crop growth information generation model is generated by a method as described in any implementation manner of the first aspect.

In a fifth aspect, some embodiments of the present disclosure provide an electronic device comprising: one or more processors; a storage device having one or more programs stored thereon, which when executed by one or more processors, cause the one or more processors to implement the method as described in any of the implementations of the first and third aspects.

In a sixth aspect, some embodiments of the present disclosure provide a computer readable medium having a computer program stored thereon, wherein the program when executed by a processor implements a method as described in any of the implementations of the first and third aspects.

In a seventh aspect, some embodiments of the present disclosure provide a computer program product comprising a computer program which, when executed by a processor, implements the method described in any one of the implementations of the first and third aspects above.

The above embodiments of the present disclosure have the following advantageous effects: the model training method of some embodiments of the present disclosure improves the prediction accuracy of the crop growth information generated model obtained by training. Specifically, the reason for the insufficient accuracy of the crop growth information generation model prediction is that: the growth period of the crops is longer, so that the actual growth condition data set of the crops is difficult to collect, and the training samples are deficient. Furthermore, the deep learning neural network obtained through training is not accurate enough in prediction capability of crop generation information. Based on this, the model training method of some embodiments of the present disclosure first acquires a historical crop scene data sequence and a crop actual growth data sequence for target crop information. Here, the historical crop scene data sequence is used for the generation of a subsequent historical crop growth simulation data sequence set. The actual crop growth data sequence is used for training a subsequent initial crop growth information generation model. At least one crop growth experimental model associated with the target crop information is then determined. By determining at least one crop growth experimental model for the target crop information, the simulation data is subsequently used to generate more accurate, fit target crop characteristics. And then, inputting the historical crop scene data sequence into each crop growth experimental model in the at least one crop growth experimental model to generate a historical crop growth simulation data sequence, so as to obtain a historical crop growth simulation data sequence set. The historical crop growth simulation data sequence set obtained here is used for subsequent generation of training data sets for the initial crop growth information generation model. Therefore, when the actual growth condition data set of the crops is difficult to collect and the data is less due to the longer growth period of the crops, the problem of lack of training samples can be solved by utilizing the training data set generated by the historical crop growth simulation data sequence set. Further, for each historical crop growth simulation data sequence in the set of historical crop growth simulation data sequences, the historical crop growth simulation data sequence and the historical crop scene data sequence are correspondingly combined in data so as to obtain a historical crop combination data sequence for subsequent training of the initial crop growth information generation model. And finally, performing model training on the initial crop growth information generation model through the historical crop combination data sequence set and the crop actual growth data sequence. Through the training mode, the problem that the prediction capability of the deep learning neural network obtained through training is not accurate enough for the crop growth information due to the fact that the data sets are fewer can be effectively solved.

Drawings

The above and other features, advantages, and aspects of embodiments of the present disclosure will become more apparent by reference to the following detailed description when taken in conjunction with the accompanying drawings. The same or similar reference numbers will be used throughout the drawings to refer to the same or like elements. It should be understood that the figures are schematic and that elements and components are not necessarily drawn to scale.

FIG. 1 is a schematic illustration of one application scenario of a model training method according to some embodiments of the present disclosure;

FIG. 2 is a flow chart of some embodiments of a model training method according to the present disclosure;

FIG. 3 is a flow chart of further embodiments of a model training method according to the present disclosure;

FIG. 4 is a schematic illustration of one application scenario of an information generation method according to some embodiments of the present disclosure;

FIG. 5 is a flow chart of some embodiments of an information generation method according to the present disclosure;

FIG. 6 is a schematic structural view of some embodiments of a model training apparatus according to the present disclosure;

FIG. 7 is a schematic diagram of the structure of some embodiments of an information generating apparatus according to the present disclosure;

fig. 8 is a schematic structural diagram of an electronic device suitable for use in implementing some embodiments of the present disclosure.

Detailed Description

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete. It should be understood that the drawings and embodiments of the present disclosure are for illustration purposes only and are not intended to limit the scope of the present disclosure.

It should be noted that, for convenience of description, only the portions related to the present invention are shown in the drawings. Embodiments of the present disclosure and features of embodiments may be combined with each other without conflict.

It should be noted that the terms "first," "second," and the like in this disclosure are merely used to distinguish between different devices, modules, or units and are not used to define an order or interdependence of functions performed by the devices, modules, or units.

It should be noted that references to "one", "a plurality" and "a plurality" in this disclosure are intended to be illustrative rather than limiting, and those of ordinary skill in the art will appreciate that "one or more" is intended to be understood as "one or more" unless the context clearly indicates otherwise.

The names of messages or information interacted between the various devices in the embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

The present disclosure will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.

Fig. 1 is a schematic diagram of one application scenario of a model training method according to some embodiments of the present disclosure.

In the application scenario of fig. 1, first, the electronic device 101 may acquire a historical crop scenario data sequence 103 and a crop actual growth data sequence 104 for the target crop information 102. Wherein, the historical crop scene data in the historical crop scene data sequence 103 and the crop actual growth data in the crop actual growth data sequence 104 have a time corresponding relation. In this application scenario, the target crop information 102 may be a "×fruit tree". The electronic device 101 may then determine at least one crop growth experimental model 105 associated with the target crop information 102. In this application scenario, the at least one crop growth experimental model 105 may include: crop growth experimental model 1051 and crop growth experimental model 1052. Next, the electronic device 101 may input the historical crop scene data sequence 103 into each of the at least one crop growth experimental model 105 to generate a historical crop growth simulation data sequence, resulting in a historical crop growth simulation data sequence set 106. In this application scenario, the historical crop growth simulation data sequence set 106 may include: historical crop growth simulation data sequence 1061, historical crop growth simulation data sequence 1062, historical crop growth simulation data sequence 1063, and historical crop growth simulation data sequence 1064. Next, for each historical crop growth simulation data sequence in the set of historical crop growth simulation data sequences 106, the electronic device 101 may perform data corresponding combination on the historical crop growth simulation data sequence and the historical crop scene data sequence 103 to obtain a historical crop combination data sequence. Finally, based on the obtained historical crop combination data sequence set 107 and the actual crop growth data sequence 104, model training is performed on the initial crop growth information generation model 108 to obtain a crop growth information generation model 109. In this application scenario, the historical crop combination data sequence set 107 may include: a historical crop combination data sequence 1071, a historical crop combination data sequence 1072, a historical crop combination data sequence 1073, and a historical crop combination data sequence 1074.

The electronic device 101 may be hardware or software. When the electronic device is hardware, the electronic device may be implemented as a distributed cluster formed by a plurality of servers or terminal devices, or may be implemented as a single server or a single terminal device. When the electronic device is embodied as software, it may be installed in the above-listed hardware device. It may be implemented as a plurality of software or software modules, for example, for providing distributed services, or as a single software or software module. The present invention is not particularly limited herein.

It should be understood that the number of electronic devices in fig. 1 is merely illustrative. There may be any number of electronic devices as desired for an implementation.

With continued reference to fig. 2, a flow 200 of some embodiments of a model training method according to the present disclosure is shown. The model training method comprises the following steps:

step 201, a historical crop scene data sequence and a crop actual growth data sequence for target crop information are acquired.

In some embodiments, the execution subject of the model training method (e.g., the electronic device 101 shown in fig. 1) may acquire the historical crop scene data sequence and the crop actual growth data sequence for the target crop information through a wired connection or a wireless connection. Wherein, the historical crop scene data in the historical crop scene data sequence and the crop actual growth data in the crop actual growth data sequence have a time corresponding relation. The target crop information may be crop information of the target crop. The target crop may be a crop for which growth information is to be generated. Specifically, the crop information may be the name of the crop, or may be the crop number of the crop. The growth information may be information of the growth of the crop over a life cycle. In practice, the growth information may include, but is not limited to, at least one of: the climatic period of the crops and the yield of the crops. For example, for crops that are fruit trees, the climatic period may include: bud picking, flowering, full-bloom stage, flower falling and fruiting. Each historical crop scene data in the historical crop scene data sequence has a one-to-one corresponding data acquisition time. The respective data acquisition times may be preset times. The order of the historical crop scene data in the historical crop scene data sequence is generated according to the data acquisition time. The historical crop scene data may be scene data for a scene at which the target crop is located at a historical time. In practice, the scene data may include, but is not limited to, at least one of: meteorological data, soil data, leaf area index, crop growth diameter. Wherein the crop growth diameter may comprise: diameter of trunk of crops and diameter of fruits of crops. The meteorological data may include, but is not limited to, at least one of: air temperature, air humidity, rainfall, illumination intensity and carbon dioxide content. Soil data may include, but is not limited to, at least one of: soil organic matter content, soil thickness, soil water content and soil temperature and humidity. The order of the crop actual growth data in the sequence of crop actual growth data is also generated in accordance with the data acquisition time. The crop actual growth data may be actual growth data of the target crop. Here, the crop actual growth data may be acquired based on various sensors. For example, various sensors may include, but are not limited to: satellite sensor, unmanned aerial vehicle sensor, radar sensor.

It should be noted that, there is a correspondence between the historical data acquisition time sequences corresponding to the historical crop scene data sequences and the data acquisition time sequences corresponding to the actual crop growth data sequences. That is, the time intervals between the respective historical data collection times in the historical data collection time series are the same as the time intervals between the respective data collection times in the data collection time series. For example, the historical data collection time series is { 1/2012/3/1/2012/4/1/2012 }. The data collection time sequence is {2018, 3, 1, 4, 1, 5, 1, 6, 1 }. Thus, the time interval between each of the historical data collection times in the historical data collection time series is one month. Likewise, the time interval between each data acquisition time in the data acquisition time series is also one month.

Step 202, determining at least one crop growth experimental model associated with the target crop information.

In some embodiments, the executing entity may determine at least one crop growth experimental model associated with the target crop information. The crop growth experimental model is an idealized model which is obtained based on a laboratory cultivation mode and is used for crop growth prediction. For example, the crop growth experimental model may be, but is not limited to, one of the following: CO ₂ A WOFOST (World Food Studies) model driven, an EPIC (Environmental Policy-Integrated Climate) model, a SAFY (Simple Algorithm For Yield estimates) model driven by light energy, and an aquaCrop model driven by water.

As an example, first, the above-described execution subject may determine the crop attribute and the crop type of the crop corresponding to the target crop information. And then, selecting at least one crop growth experimental model according to the crop attribute and the crop type.

Step 203, inputting the historical crop scene data sequence into each crop growth experimental model in the at least one crop growth experimental model to generate a historical crop growth simulation data sequence, so as to obtain a historical crop growth simulation data sequence set.

In some embodiments, the execution body may input the historical crop scene data sequence into each of the at least one crop growth experimental model to generate a historical crop growth simulation data sequence, resulting in a historical crop growth simulation data sequence set. For each crop growth experimental model, the corresponding historical crop growth simulation data in the historical crop growth simulation data sequence has a one-to-one correspondence simulation relationship with the historical crops in the historical crop scene data sequence. The historical crop growth simulation data may be crop growth data simulated using a crop growth experimental model for corresponding historical crop scene data.

It should be noted that, for the target crop under the historical crop scene data, there may be a certain error between the corresponding historical crop growth simulation data and the actual growth data.

And 204, for each historical crop growth simulation data sequence in the historical crop growth simulation data sequence set, correspondingly combining the historical crop growth simulation data sequence with the historical crop scene data sequence to obtain a historical crop combination data sequence.

In some embodiments, the execution body may perform data corresponding combination on the historical crop growth simulation data sequence and the historical crop scene data sequence for each historical crop growth simulation data sequence in the historical crop growth simulation data sequence set to obtain a historical crop combination data sequence.

As an example, for each historical crop growth simulation data in the sequence of historical crop growth simulation data, first, the above-described execution subject may determine historical crop scene data corresponding to the above-described historical crop growth simulation data as target historical crop scene data. Then, the target historical crop scene data is used as training data, the historical crop growth simulation data is used as a label, and the target historical crop scene data and the corresponding historical crop scene data are combined to generate combined data which is used as historical crop combined data.

And 205, performing model training on the initial crop growth information generation model according to the obtained historical crop combination data sequence set and the crop actual growth data sequence to obtain a crop growth information generation model.

In some embodiments, the execution subject may perform model training on the initial crop growth information generation model according to the obtained historical crop combination data sequence set and the crop actual growth data sequence, so as to obtain a crop growth information generation model. Wherein the initial crop growth information generation model may be a model after model parameter initialization (i.e., a model for which model parameters have not been updated). The crop growth information generation model may be a model that generates crop growth information. In practice, the crop growth information generation model may be a time-series neural network model. For example, the temporal neural network model may be one of: a Long Short-Term Memory network (LSTM) model, a gated loop unit (Gated Recurrent Unit, GRU) model. In addition, the crop growth information generation model may be a combination model based on a time-series neural network model. For example, the crop growth information generation model is a combined model composed of a long-short-term memory network model and a convolutional neural network.

As an example, the execution subject may perform model training on the initial crop growth information generation model by using the historical crop combination data sequence set and the crop actual growth data sequence as the training sample set of the initial crop growth information, to obtain the crop growth information generation model.

With further reference to fig. 3, a flow 300 of further embodiments of the model training method according to the present disclosure is shown. The model training method comprises the following steps:

step 301, a historical crop scene data sequence and a crop actual growth data sequence for target crop information are acquired.

At step 302, at least one crop growth experimental model associated with the target crop information is determined.

Step 303, inputting the historical crop scene data sequence into each crop growth experimental model in the at least one crop growth experimental model to generate a historical crop growth simulation data sequence, so as to obtain a historical crop growth simulation data sequence set.

And 304, for each historical crop growth simulation data sequence in the historical crop growth simulation data sequence set, correspondingly combining the historical crop growth simulation data sequence with the historical crop scene data sequence to obtain a historical crop combination data sequence.

In some embodiments, the specific implementation of steps 301 to 304 and the technical effects thereof may refer to steps 201 to 204 in the corresponding embodiment of fig. 2, which are not described herein.

Step 305, for each crop actual growth data in the crop actual growth data sequence, selecting target historical crop growth simulation data meeting a preset condition with relation between the target historical crop growth simulation data set and the crop actual growth data according to the crop growth time corresponding to the crop actual growth data, and using the target historical crop growth simulation data as candidate historical crop growth simulation data.

In some embodiments, the executing entity (e.g., the electronic device 101 shown in fig. 1) may, for each crop actual growth data in the sequence of crop actual growth data, screen, as candidate historical crop growth simulation data, target historical crop growth simulation data satisfying a preset condition with respect to the relationship between the crop actual growth data from the target historical crop growth simulation data set according to the crop growth time corresponding to the crop actual growth data.

The target historical crop growth simulation data set is historical crop growth simulation data of which the corresponding time is related to the crop growth time. Specifically, the crop actual growth data has a one-to-one time relationship with the simulation time corresponding to the target historical crop growth simulation data. The above-described time relationship may be a relationship in which the time difference between the crop growth time and the simulation time is a predetermined threshold. And aiming at each crop actual growth data, taking the historical crop growth simulation data with the time difference between the corresponding time and the crop actual growth data corresponding to the crop growth time being a preset threshold value as target historical crop growth simulation data to obtain a target historical crop growth simulation data set. The preset relationship may be that the similarity between the data content corresponding to the target historical crop growth simulation data and the data content corresponding to the crop actual growth data is the largest.

Step 306, generating a historical crop combination data sequence corresponding to the obtained candidate historical crop growth simulation data sequence as a target historical crop combination data sequence.

In some embodiments, the execution subject may generate a historical crop combination data sequence corresponding to the obtained candidate historical crop growth simulation data sequence as the target historical crop combination data sequence. The candidate historical crop growth simulation data in the candidate historical crop growth simulation data sequence and the historical crop combination data in the historical crop combination data sequence have a one-to-one correspondence.

As an example, for each candidate history generation simulation data in the candidate history crop growth simulation data sequence, first, the execution body may determine history crop scene data corresponding to the candidate history generation simulation data as a candidate history crop scene data sequence. The execution body may then perform corresponding data combining of the candidate historical crop growth simulation data sequence and the candidate historical crop scene data sequence to generate the target historical crop combination data sequence.

Step 307, performing model training on the initial crop growth information generation model according to the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model.

In some embodiments, the execution subject may perform model training on the initial crop growth information generation model according to the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model.

As an example, the execution subject may perform model training on the initial crop growth information generation model using the target historical crop combination data sequence and the crop actual growth data sequence as training sample sequences, to obtain the crop growth information generation model.

In some optional implementations of some embodiments, the training the initial crop growth information generation model according to the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model may include the following steps:

the first step, the execution subject may train the initial crop growth information model once using the target historical crop combination data sequence as a training data sequence, and obtain a crop growth information generation model after training once.

And secondly, the execution main body can take the actual crop growth data sequence as a secondary training data sequence, and perform secondary training on the crop growth information generation model after primary training to obtain a crop growth information generation model after secondary training, and the crop growth information generation model is taken as the crop growth information generation model.

The model parameters of the sub-model at least one preset position included in the crop growth information generation model after the primary training are not changed in the secondary training process. The sub-model at the at least one preset position is a model positioned at the rear position in a sub-model sequence included in the crop growth information generation model.

Here, the model located in front in the sub-model sequence is often a feature extraction model. The subsequent models are key models generated for individualization of crop growth information. Dynamic adjustment of the latter model helps to generate more accurate crop growth information.

Optionally, after the step of taking the actual crop growth data sequence as a secondary training data sequence and performing secondary training on the crop growth information generation model after primary training to obtain a crop growth information generation model after secondary training, the step further includes:

first, verification data and crop growth verification information for the target crop information are acquired.

The verification data is farm scene data of a farm where the target crop information corresponds to the crop at the first time, and the crop growth verification information is crop actual growth data for the first time. The first time may be a preset time.

And secondly, inputting the verification data into the crop growth information generation model to obtain crop growth information.

And thirdly, generating information representing successful verification of the crop growth information generation model in response to determining that the information difference between the crop growth information and the crop growth verification information is smaller than or equal to a preset threshold value.

Wherein the crop growth information comprises: a plurality of crop growth characteristic information. The crop growth verification information includes: a plurality of crop growth characteristic information. The crop growth characteristic information may be growth characteristic information of the crop. In practice, crop growth characteristics may include: crop yield, crop leaf area. The preset threshold may include: subthreshold for multiple crop growth characteristics. For example, for crop yield, the corresponding sub-threshold may be 200. The information difference between the crop growth information and the crop growth verification information described above may be a crop growth characteristic difference between individual crop growth characteristics.

Optionally, in response to determining that each crop growth characteristic difference between the crop growth information and the crop growth verification information is less than a corresponding sub-threshold, generating information indicative of success of verification of the crop growth information generation model.

Optionally, in response to determining that the average value corresponding to each crop growth characteristic difference between the crop growth information and the crop growth verification information is smaller than the average value corresponding to each sub-threshold, information representing that verification of the crop growth information generation model is successful is generated.

Optionally, after generating the information representing success of verification of the crop growth information generation model in response to determining that the information difference between the crop growth information and the crop growth verification information is equal to or less than a preset threshold, the steps further include:

and in the first step, in response to determining that the information difference between the crop growth information and the crop growth verification information is greater than a preset threshold, adjusting a model structure of the crop growth information generation model to obtain an adjusted crop growth information generation model.

As an example, the model structure of the crop growth information generation model is dynamically adjusted according to the received model structure adjustment information of the crop growth information generation model, so as to obtain an adjusted crop growth information generation model.

And secondly, performing model training on the adjusted crop growth information generation model according to the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model.

As an example, the execution subject may perform model training on the adjusted crop growth information generation model by using the target historical crop combination data sequence and the crop actual growth data sequence as a training sample set, to obtain the crop growth information generation model.

Optionally, the crop actual growth data in the crop actual growth data sequence includes: crop scene data; and

the step of screening target historical crop growth simulation data, which satisfy a preset condition with respect to the actual crop growth data, from the target historical crop growth simulation data set as candidate historical crop growth simulation data may include the steps of:

the first step is to encode the crop scene data to obtain the actual scene data vector.

As an example, the execution subject may encode the crop scene data using a word embedding model to obtain an actual scene data vector. For example, the word embedding model may be, but is not limited to, at least one of: the coding model in the BERT model and the coding model in the transducer model.

And secondly, encoding each historical crop growth simulation data in the target historical crop growth simulation data set to generate simulation data vectors, and obtaining a simulation data vector set.

As an example, the execution body may encode each historical crop growth simulation data in the target historical crop growth simulation data set using a word embedding model to generate a simulation data vector, resulting in a set of simulation data vectors.

And thirdly, screening the simulation data vectors, the distance between which and the actual scene data vector meets the preset distance condition, from the simulation data vector set to serve as candidate simulation data vectors.

The preset distance condition may be that a distance between the two vectors is less than or equal to a target value.

As an example, the execution body may determine a cosine distance between each simulation data vector in the set of simulation data vectors and the actual scene data vector. And then, screening out simulation data vectors with the corresponding cosine distances smaller than the target value from the simulation data vector set, and taking the simulation data vectors as candidate simulation data vectors.

And fourthly, determining the target historical crop growth simulation data corresponding to the candidate simulation data vector as candidate historical crop growth simulation data.

As can be seen in fig. 3, the flow 300 of the model training method in some embodiments corresponding to fig. 3 results in a more accurate training sample set for subsequent initial crop growth information generation model training by generating historical crop combination data sequences corresponding to candidate historical crop growth simulation data sequences, as compared to the description of some embodiments corresponding to fig. 2. Therefore, based on the target historical crop combination data sequence and the crop actual growth data sequence, model training is carried out on the initial crop growth information generation model, and a crop growth information generation model with more accurate predicted crop growth information can be obtained.

Fig. 4 is a schematic diagram of an application scenario of an information generation method according to some embodiments of the present disclosure.

In the application scenario of fig. 4, the electronic device 401 may acquire crop scenario data 402 for the second time corresponding to the target farm for the target crop information 401. Then, the electronic device 401 inputs the above-described crop scene data 402 to the crop growth information generation model 403 to generate crop growth prediction information 404. Wherein the crop growth information generation model 403 is generated by the model training method of some embodiments of the present disclosure.

The electronic device 401 may be hardware or software. When the electronic device is hardware, the electronic device may be implemented as a distributed cluster formed by a plurality of servers or terminal devices, or may be implemented as a single server or a single terminal device. When the electronic device is embodied as software, it may be installed in the above-listed hardware device. It may be implemented as a plurality of software or software modules, for example, for providing distributed services, or as a single software or software module. The present invention is not particularly limited herein.

It should be understood that the number of electronic devices in fig. 4 is merely illustrative. There may be any number of electronic devices as desired for an implementation.

With continued reference to fig. 5, a flow 500 of some embodiments of an information generation method according to the present disclosure is shown. The information generating method comprises the following steps:

step 501, crop scene data for a second time corresponding to the target crop information is obtained.

In some embodiments, the execution subject of the above-described information generation method (e.g., the electronic device 401 shown in fig. 4) may acquire crop scene data for the second time corresponding to the target crop information. The target crop information may be crop information of the target crop. The target crop may be a crop for which growth information is to be generated. Specifically, the crop information may be the name of the crop, or may be the crop number of the crop. The growth information may be information of the growth of the crop over a life cycle. In practice, the growth information may include, but is not limited to, at least one of: the climatic period of the crops and the yield of the crops. For example, for crops that are fruit trees, the climatic period may include: bud picking, flowering, full-bloom stage, flower falling and fruiting. The target farm may be a farm on which the target crop is planted.

Step 502, inputting the crop scene data into a crop growth information generation model to generate crop growth prediction information.

In some embodiments, the execution subject may input the crop scene data into a crop growth information generation model to generate crop growth prediction information. Wherein, the crop growth information generation model is generated by a model training method of some embodiments of the present disclosure.

The above embodiments of the present disclosure have the following advantageous effects: more accurate crop growth information may be obtained by the information generation method of some embodiments of the present disclosure, i.e., using the crop growth information generation model.

With further reference to fig. 6, as an implementation of the method illustrated in the above figures, the present disclosure provides some embodiments of a model training apparatus, which apparatus embodiments correspond to those illustrated in fig. 2, which apparatus is particularly applicable in a variety of electronic devices.

As shown in fig. 6, a model training apparatus 600 includes: a first acquisition unit 601, a determination unit 602, a first input unit 603, a combination unit 604, and a training unit 605. Wherein, the first obtaining unit 601 is configured to obtain a historical crop scene data sequence and a crop actual growth data sequence for target crop information, wherein, the historical crop scene data in the historical crop scene data sequence has a time corresponding relation with the crop actual growth data in the crop actual growth data sequence; a determining unit 602 configured to determine at least one crop growth experimental model associated with the target crop information; a first input unit 603 configured to input the historical crop scene data sequence into each of the at least one crop growth experimental model to generate a historical crop growth simulation data sequence, resulting in a historical crop growth simulation data sequence set; a combination unit 604 configured to, for each historical crop growth simulation data sequence in the set of historical crop growth simulation data sequences, perform data corresponding combination on the historical crop growth simulation data sequence and the historical crop scene data sequence to obtain a historical crop combination data sequence; the training unit 605 is configured to perform model training on the initial crop growth information generation model according to the obtained historical crop combination data sequence set and the actual crop growth data sequence, so as to obtain a crop growth information generation model.

In some alternative implementations of some embodiments, the training unit 605 may be further configured to: for each crop actual growth data in the crop actual growth data sequence, screening target historical crop growth simulation data meeting preset conditions in relation to the crop actual growth data from a target historical crop growth simulation data set according to the crop growth time corresponding to the crop actual growth data, and taking the target historical crop growth simulation data as candidate historical crop growth simulation data, wherein the target historical crop growth simulation data set is historical crop growth simulation data of which the corresponding time is related to the crop growth time; generating a historical crop combination data sequence corresponding to the obtained candidate historical crop growth simulation data sequence as a target historical crop combination data sequence; and performing model training on the initial crop growth information generation model according to the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model.

In some alternative implementations of some embodiments, the training unit 605 may be further configured to: taking the target historical crop combined data sequence as a primary training data sequence, and carrying out primary training on the initial crop growth information model to obtain a crop growth information generation model after primary training; and taking the crop actual growth data sequence as a secondary training data sequence, and performing secondary training on the crop growth information generation model after primary training to obtain a crop growth information generation model after secondary training, and taking the crop growth information generation model after secondary training as the crop growth information generation model, wherein model parameters of at least one sub-model at a preset position included in the crop growth information generation model after primary training are not changed in the secondary training process.

In some alternative implementations of some embodiments, the training unit 605 may be further configured to: acquiring verification data and crop growth verification information aiming at the target crop information, wherein the verification data is farm scene data of a farm where the target crop information corresponding to the crop is located aiming at a first time, and the crop growth verification information is crop actual growth data aiming at the first time; inputting the verification data into the crop growth information generation model to obtain crop growth information; and generating information representing successful verification of the crop growth information generation model in response to determining that the information difference between the crop growth information and the crop growth verification information is less than or equal to a preset threshold.

In some alternative implementations of some embodiments, the training unit 605 may be further configured to: in response to determining that the information difference between the crop growth information and the crop growth verification information is greater than a preset threshold, adjusting a model structure of the crop growth information generation model to obtain an adjusted crop growth information generation model; and performing model training on the adjusted crop growth information generation model according to the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model.

In some alternative implementations of some embodiments, the crop actual growth data in the sequence of crop actual growth data includes: crop scene data; and training unit 605 may be further configured to: encoding the crop scene data to obtain an actual scene data vector; encoding each historical crop growth simulation data in the target historical crop growth simulation data set to generate a simulation data vector, and obtaining a simulation data vector set; screening simulation data vectors, the distance between which and the actual scene data vectors meet the preset distance condition, from the simulation data vector set to serve as candidate simulation data vectors; and determining the target historical crop growth simulation data corresponding to the candidate simulation data vector as candidate historical crop growth simulation data.

It will be appreciated that the elements described in the apparatus 600 correspond to the various steps in the method described with reference to fig. 2. Thus, the operations, features and resulting benefits described above with respect to the method are equally applicable to the apparatus 600 and the units contained therein, and are not described in detail herein.

With further reference to fig. 7, as an implementation of the method shown in the above figures, the present disclosure provides some embodiments of an information generating apparatus, which correspond to those method embodiments shown in fig. 5, and which are particularly applicable in various electronic devices.

As shown in fig. 7, a model training apparatus 700 includes: a second acquisition unit 701 and a second input unit 702. Wherein, the second obtaining unit 701 is configured to obtain crop scene data for a second time corresponding to the target farm by the target crop information; the second input unit 702 is configured to input the crop scene data to a crop growth information generation model to generate crop growth prediction information, wherein the crop growth information generation model is generated by a model training method of some embodiments of the present disclosure.

It will be appreciated that the elements described in the apparatus 700 correspond to the various steps in the method described with reference to fig. 5. Thus, the operations, features and resulting benefits described above for the method are equally applicable to the apparatus 700 and the units contained therein, and are not described in detail herein.

Referring now to fig. 8, a schematic diagram of an electronic device 800 (e.g., electronic device 101 of fig. 1) suitable for use in implementing some embodiments of the present disclosure is shown. The electronic device shown in fig. 8 is merely an example, and should not impose any limitations on the functionality and scope of use of embodiments of the present disclosure.

As shown in fig. 8, the electronic device 800 may include a processing means (e.g., a central processor, a graphics processor, etc.) 801, which may perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) 802 or a program loaded from a storage means 808 into a Random Access Memory (RAM) 803. In the RAM 803, various programs and data required for the operation of the electronic device 800 are also stored. The processing device 801, the ROM 802, and the RAM 803 are connected to each other by a bus 804. An input/output (I/O) interface 805 is also connected to the bus 804.

In general, the following devices may be connected to the I/O interface 805: input devices 806 including, for example, a touch screen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, and the like; an output device 807 including, for example, a Liquid Crystal Display (LCD), speakers, vibrators, etc.; storage 808 including, for example, magnetic tape, hard disk, etc.; communication means 809. The communication means 809 may allow the electronic device 800 to communicate wirelessly or by wire with other devices to exchange data. While fig. 8 shows an electronic device 800 having various means, it is to be understood that not all of the illustrated means are required to be implemented or provided. More or fewer devices may be implemented or provided instead. Each block shown in fig. 8 may represent one device or a plurality of devices as needed.

In particular, according to some embodiments of the present disclosure, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, some embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method shown in the flow chart. In such embodiments, the computer program may be downloaded and installed from a network via communication device 809, or from storage device 808, or from ROM 802. The above-described functions defined in the methods of some embodiments of the present disclosure are performed when the computer program is executed by the processing device 801.

It should be noted that, in some embodiments of the present disclosure, the computer readable medium may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In some embodiments of the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In some embodiments of the present disclosure, however, the computer-readable signal medium may comprise a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, fiber optic cables, RF (radio frequency), and the like, or any suitable combination of the foregoing.

In some implementations, the clients, servers may communicate using any currently known or future developed network protocol, such as HTTP (HyperText Transfer Protocol ), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), the internet (e.g., the internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed networks.

The computer readable medium may be contained in the electronic device; or may exist alone without being incorporated into the electronic device. The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: acquiring a historical crop scene data sequence and a crop actual growth data sequence aiming at target crop information, wherein the historical crop scene data in the historical crop scene data sequence and the crop actual growth data in the crop actual growth data sequence have a time corresponding relation; determining at least one crop growth experimental model associated with the target crop information; inputting the historical crop scene data sequence into each crop growth experimental model in the at least one crop growth experimental model to generate a historical crop growth simulation data sequence, so as to obtain a historical crop growth simulation data sequence set; for each historical crop growth simulation data sequence in the historical crop growth simulation data sequence set, correspondingly combining the historical crop growth simulation data sequence with the historical crop scene data sequence to obtain a historical crop combination data sequence; and performing model training on the initial crop growth information generation model according to the obtained historical crop combination data sequence set and the crop actual growth data sequence to obtain a crop growth information generation model. Acquiring crop scene data corresponding to the target farm and aiming at a second time by using the target crop information; the crop scene data is input into a crop growth information generation model to generate crop growth prediction information, wherein the crop growth information generation model is generated by a model training method of some embodiments of the present disclosure.

Computer program code for carrying out operations for some embodiments of the present disclosure may be written in one or more programming languages, including an object oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).

The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The units described in some embodiments of the present disclosure may be implemented by means of software, or may be implemented by means of hardware. The described units may also be provided in a processor, for example, described as: a processor includes a first acquisition unit, a determination unit, a first input unit, a combining unit, and a training unit. Wherein the names of the units do not constitute a limitation of the unit itself in certain cases, for example, the first input unit may also be described as "a unit for inputting the above-mentioned historical crop scene data sequence into each of the above-mentioned at least one crop growth experimental model to generate a historical crop growth simulation data sequence, resulting in a set of historical crop growth simulation data sequences".

The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: a Field Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC), an Application Specific Standard Product (ASSP), a system on a chip (SOC), a Complex Programmable Logic Device (CPLD), and the like.

Some embodiments of the present disclosure also provide a computer program product comprising a computer program which, when executed by a processor, implements any of the model training methods described above and an information generation method.

The foregoing description is only of the preferred embodiments of the present disclosure and description of the principles of the technology being employed. It will be appreciated by those skilled in the art that the scope of the invention in the embodiments of the present disclosure is not limited to the specific combination of the above technical features, but encompasses other technical features formed by any combination of the above technical features or their equivalents without departing from the spirit of the invention. Such as the above-described features, are mutually substituted with (but not limited to) the features having similar functions disclosed in the embodiments of the present disclosure.

Claims

1. A model training method, comprising:

acquiring a historical crop scene data sequence and a crop actual growth data sequence aiming at target crop information, wherein a time corresponding relation exists between historical crop scene data in the historical crop scene data sequence and crop actual growth data in the crop actual growth data sequence;

determining at least one crop growth experimental model associated with the target crop information;

inputting the historical crop scene data sequence to each crop growth experimental model in the at least one crop growth experimental model to generate a historical crop growth simulation data sequence, so as to obtain a historical crop growth simulation data sequence set;

For each historical crop growth simulation data sequence in the historical crop growth simulation data sequence set, carrying out data corresponding combination on the historical crop growth simulation data sequence and the historical crop scene data sequence to obtain a historical crop combination data sequence;

and performing model training on the initial crop growth information generation model according to the obtained historical crop combination data sequence set and the crop actual growth data sequence to obtain a crop growth information generation model.

2. The method of claim 1, wherein the model training the initial crop growth information generation model based on the obtained historical crop combination data sequence set and the crop actual growth data sequence to obtain a crop growth information generation model comprises:

for each crop actual growth data in the crop actual growth data sequence, screening target historical crop growth simulation data meeting a preset condition in relation with the crop actual growth data from a target historical crop growth simulation data set according to the crop growth time corresponding to the crop actual growth data, and taking the target historical crop growth simulation data as candidate historical crop growth simulation data, wherein the target historical crop growth simulation data set is historical crop growth simulation data of which the corresponding time is related to the crop growth time;

Generating a historical crop combination data sequence corresponding to the obtained candidate historical crop growth simulation data sequence as a target historical crop combination data sequence;

and performing model training on an initial crop growth information generation model according to the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model.

3. The method of claim 2, wherein the model training an initial crop growth information generation model based on the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model comprises:

taking the target historical crop combined data sequence as a primary training data sequence, and carrying out primary training on the initial crop growth information model to obtain a crop growth information generation model after primary training;

and taking the crop actual growth data sequence as a secondary training data sequence, and performing secondary training on the crop growth information generation model after primary training to obtain a crop growth information generation model after secondary training, wherein the model parameters of the sub-model at least one preset position included in the crop growth information generation model after primary training are not changed in the secondary training process.

4. A method according to claim 3, wherein the method further comprises:

acquiring verification data and crop growth verification information aiming at the target crop information, wherein the verification data is farm scene data of a farm where the target crop information corresponding to the first time is located, and the crop growth verification information is crop actual growth data aiming at the first time;

inputting the verification data into the crop growth information generation model to obtain crop growth information;

and generating information representing successful verification of the crop growth information generation model in response to determining that the information difference between the crop growth information and the crop growth verification information is less than or equal to a preset threshold.

5. The method of claim 4, wherein the method further comprises:

in response to determining that the information difference between the crop growth information and the crop growth verification information is greater than a preset threshold, adjusting a model structure of the crop growth information generation model to obtain an adjusted crop growth information generation model;

and performing model training on the adjusted crop growth information generation model according to the target historical crop combination data sequence and the crop actual growth data sequence to obtain the crop growth information generation model.

6. The method of claim 2, wherein the crop actual growth data in the sequence of crop actual growth data comprises: crop scene data; and

the step of screening target historical crop growth simulation data meeting preset conditions with relation between the actual crop growth simulation data and the target historical crop growth simulation data set as candidate historical crop growth simulation data comprises the following steps:

encoding the crop scene data to obtain an actual scene data vector;

encoding each historical crop growth simulation data in the target historical crop growth simulation data set to generate a simulation data vector, and obtaining a simulation data vector set;

screening simulation data vectors, the distance between which and the actual scene data vector meets the preset distance condition, from the simulation data vector set, and taking the simulation data vectors as candidate simulation data vectors;

and determining the target historical crop growth simulation data corresponding to the candidate simulation data vector as candidate historical crop growth simulation data.

7. An information generation method, comprising:

acquiring crop scene data corresponding to the target farm and aiming at a second time by using the target crop information;

Inputting the crop scene data into a crop growth information generation model to generate crop growth prediction information, wherein the crop growth information generation model is generated based on the method of claims 1-6.

8. A model training apparatus comprising:

a first acquisition unit configured to acquire a historical crop scene data sequence and a crop actual growth data sequence for target crop information, wherein historical crop scene data in the historical crop scene data sequence has a time correspondence relationship with crop actual growth data in the crop actual growth data sequence;

a determining unit configured to determine at least one crop growth experimental model associated with the target crop information;

a first input unit configured to input the historical crop scene data sequence to each of the at least one crop growth experimental model to generate a historical crop growth simulation data sequence, resulting in a historical crop growth simulation data sequence set;

the combination unit is configured to correspondingly combine the historical crop growth simulation data sequences with the historical crop scene data sequences to obtain a historical crop combination data sequence for each historical crop growth simulation data sequence in the historical crop growth simulation data sequence set;

And the training unit is configured to perform model training on the initial crop growth information generation model according to the obtained historical crop combination data sequence set and the crop actual growth data sequence to obtain a crop growth information generation model.

9. An information generating apparatus comprising:

a second acquisition unit configured to acquire crop scene data for a second time corresponding to the target farm with respect to the target crop information;

a second input unit configured to input the crop scene data to a crop growth information generation model to generate crop growth prediction information, wherein the crop growth information generation model is generated based on the method of claims 1-6.

10. An electronic device, comprising:

one or more processors;

a storage device having one or more programs stored thereon,

when executed by the one or more processors, causes the one or more processors to implement the method of any of claims 1-7.

11. A computer readable medium having stored thereon a computer program, wherein the program when executed by a processor implements the method of any of claims 1-7.

12. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any of claims 1-7.