CN117114917B - AI prediction processing method and system applied to digital agriculture - Google Patents
AI prediction processing method and system applied to digital agriculture Download PDFInfo
- Publication number
- CN117114917B CN117114917B CN202311387219.7A CN202311387219A CN117114917B CN 117114917 B CN117114917 B CN 117114917B CN 202311387219 A CN202311387219 A CN 202311387219A CN 117114917 B CN117114917 B CN 117114917B
- Authority
- CN
- China
- Prior art keywords
- agricultural product
- sample
- data
- prediction
- network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 33
- 238000012545 processing Methods 0.000 claims abstract description 49
- 238000000034 method Methods 0.000 claims abstract description 26
- 230000008569 process Effects 0.000 claims abstract description 22
- 239000012634 fragment Substances 0.000 claims description 38
- 239000002689 soil Substances 0.000 claims description 24
- 241000238631 Hexapoda Species 0.000 claims description 21
- 238000005286 illumination Methods 0.000 claims description 18
- 230000004720 fertilization Effects 0.000 claims description 13
- 238000013507 mapping Methods 0.000 claims description 9
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 6
- 230000011218 segmentation Effects 0.000 claims description 5
- 230000000694 effects Effects 0.000 abstract description 6
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 241000607479 Yersinia pestis Species 0.000 description 19
- 238000012795 verification Methods 0.000 description 15
- 230000002159 abnormal effect Effects 0.000 description 6
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 4
- 235000015097 nutrients Nutrition 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000005065 mining Methods 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- 125000004122 cyclic group Chemical group 0.000 description 2
- 230000002950 deficient Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000003337 fertilizer Substances 0.000 description 2
- 238000003973 irrigation Methods 0.000 description 2
- 230000002262 irrigation Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 235000009434 Actinidia chinensis Nutrition 0.000 description 1
- 244000298697 Actinidia deliciosa Species 0.000 description 1
- 235000009436 Actinidia deliciosa Nutrition 0.000 description 1
- 241000196324 Embryophyta Species 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000008713 feedback mechanism Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000003306 harvesting Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000000575 pesticide Substances 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Mining
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
- Y02A40/10—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in agriculture
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Mining & Mineral Resources (AREA)
- Life Sciences & Earth Sciences (AREA)
- Marine Sciences & Fisheries (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Agronomy & Crop Science (AREA)
- Animal Husbandry (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides an AI prediction processing method and system applied to digital agriculture, and relates to the technical field of artificial intelligence. According to the invention, network updating is carried out on candidate agricultural product prediction networks according to sample agricultural product data to form a target agricultural product prediction network; acquiring a target agricultural product growth record and predicting target matters; determining a corresponding target agricultural product data combination based on the target agricultural product growth record and the predicted target item; and (3) through a target agricultural product prediction network, performing prediction processing on the target agricultural product growth record based on target agricultural product data combination, and outputting the reference agricultural product quality and the corresponding quality determination cause of the target agricultural product. Based on the above, the problem of relatively poor effect of the prediction process in digital agriculture can be improved to some extent.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to an AI prediction processing method and system applied to digital agriculture.
Background
AI prediction processing has wide application in digital agriculture, with the following being some common fields of application:
Weather forecast: the AI algorithm is utilized to analyze a large amount of meteorological data, forecast weather changes and extreme weather events, and help farmers to decide planting time, irrigation management and the like; predicting plant diseases and insect pests: by monitoring the crop growth environment and sensor data, the AI model is applied to predict the risk of pest outbreak, corresponding measures are adopted in advance, and crop loss and pesticide use are reduced; crop growth prediction: the crop growth condition is monitored and predicted by combining an image recognition technology and a machine learning algorithm, so that farmers are helped to determine the optimal harvesting time and optimize fertilization and management strategies; and (3) water resource management: the AI model is utilized to analyze data such as soil humidity, rainfall and the like, so as to predict farmland water demand, and intelligent irrigation and reasonable water resource utilization are realized; market demand prediction: through analysis of market data and consumer behaviors, an AI algorithm is applied to predict the demand trend of agricultural products, so that farmers are helped to adjust planting scales and select varieties suitable for marketing and road matching; and (3) detecting the quality of agricultural products: and the quality detection and classification are carried out on the agricultural products by utilizing the image recognition and machine learning technology, so that the market competitiveness is improved.
Among them, the prediction of the quality of agricultural products based on the growth records of agricultural products can also be realized based on AI prediction, but there is a problem in that the effect of the prediction process is relatively poor.
Disclosure of Invention
In view of the above, an object of the present invention is to provide an AI prediction processing method and system for digital agriculture, which can improve the problem of relatively poor effect of the AI prediction processing in digital agriculture.
In order to achieve the above purpose, the embodiment of the present invention adopts the following technical scheme:
an AI prediction processing method applied to digital agriculture, comprising:
according to sample agricultural product data, carrying out network updating on the candidate agricultural product prediction network to form a target agricultural product prediction network, wherein the sample agricultural product data is provided with growth process description data of sample agricultural products;
obtaining a target agricultural product growth record describing a growth process of the target agricultural product and a predicted target item guiding: predicting the quality of the agricultural product based on a plurality of predetermined reference agricultural product quality, and outputting the reference agricultural product quality of the loaded agricultural product growth record and a corresponding quality determination cause;
determining a corresponding target agricultural product data combination based on the target agricultural product growth record and the predicted target item, wherein the target agricultural product growth record in the target agricultural product data combination is used as the loaded agricultural product growth record;
And through the target agricultural product prediction network, performing prediction processing on the target agricultural product growth record based on the target agricultural product data combination, and outputting a reference agricultural product quality of the target agricultural product and a corresponding quality determination reason, wherein the reference agricultural product quality of the target agricultural product is used for reflecting a quality prediction result of the target agricultural product, and the quality determination reason is used for reflecting the reason for predicting the quality prediction result.
In some preferred embodiments, in the above AI prediction processing method applied to digital agriculture, the step of performing network update on the candidate agricultural product prediction network according to the sample agricultural product data to form the target agricultural product prediction network includes:
extracting sample agricultural product data corresponding to a candidate agricultural product prediction network, wherein the sample agricultural product data comprises a sample agricultural product growth record, sample agricultural product quality corresponding to the sample agricultural product growth record and a corresponding quality determination cause, the sample agricultural product quality of the sample agricultural product growth record belongs to one of a plurality of predetermined reference agricultural product qualities, and the sample agricultural product growth record is used for describing the growth process of the sample agricultural product;
Determining a prediction target item of the candidate agricultural product prediction network;
determining a sample agricultural product data combination for updating the candidate agricultural product prediction network according to the prediction target item and the sample agricultural product data, wherein a sample agricultural product growth record in the sample agricultural product data is used as a loaded agricultural product growth record;
and carrying out network updating processing on the candidate agricultural product prediction network according to the sample agricultural product data combination, and obtaining a target agricultural product prediction network based on the updated candidate agricultural product prediction network.
In some preferred embodiments, in the above AI prediction processing method applied to digital agriculture, the step of determining a sample agricultural product data combination for updating the candidate agricultural product prediction network according to the prediction target item and the sample agricultural product data includes:
combining the predicted target item and sample agricultural product growth records in the sample agricultural product data to form corresponding first agricultural product combined data, wherein the sample agricultural product growth records are at least used for describing soil humidity, air temperature, fertilization amount, growth time, insect condition, illumination, rainfall, air component content and agricultural product appearance characteristics in the growth process of the sample agricultural products;
Combining the quality of the sample agricultural products in the sample agricultural product data and the quality determination cause in the sample agricultural product data to form corresponding second agricultural product combination data, wherein the quality of the sample agricultural products at least comprises high quality, general and inferior products, and the quality determination cause at least comprises a local cause corresponding to each factor of soil humidity, air temperature, fertilization amount, insect condition, illumination, rainfall and air component content;
determining a sample agricultural product data combination for updating the candidate agricultural product prediction network based on the first agricultural product merging data and the second agricultural product merging data;
and the step of performing network update processing on the candidate agricultural product prediction network according to the sample agricultural product data combination, and obtaining a target agricultural product prediction network based on the updated candidate agricultural product prediction network, includes:
carrying out quality prediction on a sample agricultural product growth record in the sample agricultural product data based on first agricultural product merging data in the sample agricultural product data combination through the candidate agricultural product prediction network, and outputting corresponding agricultural product quality prediction data, wherein the agricultural product quality prediction data comprises reference agricultural product quality and a corresponding quality determination cause of the sample agricultural product growth record;
And updating network parameters of the candidate agricultural product prediction network based on the difference between the agricultural product quality prediction data and the second agricultural product merging data in the sample agricultural product data combination to obtain a target agricultural product prediction network.
In some preferred embodiments, in the above AI prediction processing method applied to digital agriculture, the sample agricultural product data combination is a data set formed by combining the prediction target item and the sample agricultural product data, and before the step of outputting corresponding agricultural product quality prediction data by the candidate agricultural product prediction network, based on first agricultural product combination data in the sample agricultural product data combination, performing quality prediction on a sample agricultural product growth record in the sample agricultural product data; the step of performing network update processing on the candidate agricultural product prediction network according to the sample agricultural product data combination, and obtaining a target agricultural product prediction network based on the updated candidate agricultural product prediction network, further comprises:
carrying out data segmentation on the sample agricultural product data combination to form a first number of agricultural product data fragments;
Determining network loading data by using a first second number of agricultural product data fragments in the first number of agricultural product data fragments based on the precedence relation of the first number of agricultural product data fragments, and determining network expected data corresponding to the network loading data by using a second number of agricultural product data fragments except for the first number of agricultural product data fragments in the first number of agricultural product data fragments, wherein the difference value between the first number and the second number is equal to 1;
sequentially carrying out prediction output on the network loading data according to the granularity of the agricultural product data fragments through the candidate agricultural product prediction network to form a corresponding prediction output result, wherein the prediction output result comprises a second number of agricultural product data fragments which are predicted to be output, and the a-th agricultural product data fragment in the prediction output result is predicted to be output based on the first a agricultural product data fragments in the network loading data, wherein a is smaller than or equal to the second number;
and updating the network parameters of the candidate agricultural product prediction network based on the difference between the network expected data and the prediction output result to obtain the candidate agricultural product prediction network after preliminary updating.
In some preferred embodiments, in the above AI prediction processing method applied to digital agriculture, the step of updating the network parameters of the candidate agricultural product prediction network based on the difference between the network expected data and the prediction output result to obtain a candidate agricultural product prediction network after preliminary updating includes:
based on the network expected data and the prediction output result, determining error parameters corresponding to all agricultural product data segments in the prediction output result, wherein the error parameters corresponding to the a-th agricultural product data segment in the prediction output result are used for reflecting: a distinction between the a-th agricultural product data segment in the predicted output result and the a-th agricultural product data segment in the network expected data;
combining error parameters corresponding to each agricultural product data segment in the prediction output result, and outputting network error parameters of the candidate agricultural product prediction network;
and updating the network parameters of the candidate agricultural product prediction network along the direction of reducing the network error parameters to obtain the candidate agricultural product prediction network after preliminary updating.
In some preferred embodiments, in the above AI-prediction processing method for digital agriculture, the sample agricultural product data belongs to one candidate sample agricultural product event in a sample agricultural product event set, the AI-prediction processing method for digital agriculture further includes a step of forming the sample agricultural product event set, the step of forming including:
collecting a plurality of candidate sample agricultural product events of the candidate agricultural product prediction network, wherein one candidate sample agricultural product event comprises a sample agricultural product growth record and an agricultural product quality related index of the sample agricultural product growth record, the agricultural product quality related index of any sample agricultural product growth record comprises sample agricultural product quality and a corresponding quality determination cause of the sample agricultural product growth record, and the sample agricultural product quality in any candidate sample agricultural product event belongs to one reference agricultural product quality in the plurality of reference agricultural product qualities;
determining at least one sample agricultural product growth record from among sample agricultural product growth records included in the plurality of candidate sample agricultural product events;
respectively adjusting the determined sample agricultural product growth records to form a third number of adjustment sample agricultural product growth records, wherein one adjustment sample agricultural product growth record corresponds to one sample agricultural product growth record, and any one adjustment sample agricultural product growth record and the corresponding sample agricultural product growth record have the same semantic characteristics;
Marking the agricultural product quality related indexes of the sample agricultural product growth record corresponding to each adjustment sample agricultural product growth record as the agricultural product quality related indexes of the adjustment sample agricultural product growth record, respectively taking each adjustment sample agricultural product growth record as a new sample agricultural product growth record, and forming a third number of new candidate sample agricultural product events according to the third number of adjustment sample agricultural product growth records and the corresponding agricultural product quality related indexes;
forming the sample agricultural product event set based on the plurality of candidate sample agricultural product events and the third number of new candidate sample agricultural product events.
In some preferred embodiments, in the above AI prediction processing method applied to digital agriculture, the step of determining at least one sample agricultural product growth record from among sample agricultural product growth records included in the plurality of candidate sample agricultural product events includes:
determining a statistical quantity of each of the plurality of reference agricultural product qualities according to the plurality of candidate sample agricultural product events, wherein the statistical quantity of any one of the reference agricultural product qualities is used for reflecting: the number of candidate sample agricultural product events of the plurality of candidate sample agricultural product events having the arbitrary reference agricultural product quality;
Selecting a reference agricultural product quality satisfying a target condition from the plurality of reference agricultural product qualities based on the statistical quantity of each reference agricultural product quality, the reference agricultural product quality satisfying the target condition being: the corresponding statistical quantity is smaller than or equal to the reference agricultural product quality of the preset reference statistical quantity;
at least one sample agricultural product growth record is determined from sample agricultural product growth records included in candidate sample agricultural product events having the reference agricultural product quality that satisfies the target condition.
In some preferred embodiments, in the above AI prediction processing method applied to digital agriculture, the step of performing adjustment processing on each determined sample agricultural product growth record to form a third number of adjusted sample agricultural product growth records includes:
extracting an adjustment target item to the growth record adjustment network, the adjustment target item to guide: outputting an adjusted sample agricultural product growth record having the same semantic features for the loaded sample agricultural product growth record based on the loaded sample agricultural product growth record and the sample agricultural product quality;
extracting at least one agricultural product growth record combination, wherein the loaded data in any agricultural product growth record combination comprises the following components: one sample agricultural product growth record and corresponding sample agricultural product quality, the expected output data in any one of said agricultural product growth record combinations comprising: a sample agricultural product growth record having the same semantic characteristics as the sample agricultural product growth record in the loaded data;
And analyzing, by the growth record adjustment network, adjusted sample agricultural product growth records having the same semantic features for each determined sample agricultural product growth record based on the adjustment target item and the at least one agricultural product growth record combination, thereby forming a third number of adjusted sample agricultural product growth records.
In some preferred embodiments, in the above AI prediction processing method applied to digital agriculture, the step of analyzing, by the growth record adjustment network, adjustment sample agricultural product growth records having the same semantic features for each determined sample agricultural product growth record based on the adjustment target item and the at least one agricultural product growth record combination, thereby forming a third number of adjustment sample agricultural product growth records, includes:
combining the adjustment target item and the at least one agricultural product growth record combination to form a network update basis of the growth record adjustment network;
based on the determined growth records of the sample agricultural products and the corresponding quality of the sample agricultural products, respectively forming loaded agricultural product information combinations corresponding to the determined growth records of the sample agricultural products;
Updating the growth record adjustment network according to the network updating basis to form an updated growth record adjustment network;
and through the updated growth record adjustment network, adjusting sample agricultural product growth records with the same semantic characteristics according to analysis of corresponding sample agricultural product growth records based on the loaded agricultural product information combination corresponding to the determined sample agricultural product growth records, so as to form a third number of adjustment sample agricultural product growth records.
The embodiment of the invention also provides an AI prediction processing system applied to digital agriculture, which comprises a processor and a memory, wherein the memory is used for storing a computer program, and the processor is used for executing the computer program so as to realize the AI prediction processing method applied to digital agriculture.
According to the AI prediction processing method and system applied to digital agriculture, provided by the embodiment of the invention, the candidate agricultural product prediction network is updated according to sample agricultural product data to form a target agricultural product prediction network; acquiring a target agricultural product growth record and predicting target matters; determining a corresponding target agricultural product data combination based on the target agricultural product growth record and the predicted target item; and (3) through a target agricultural product prediction network, performing prediction processing on the target agricultural product growth record based on target agricultural product data combination, and outputting the reference agricultural product quality and the corresponding quality determination cause of the target agricultural product. Based on the foregoing, since the corresponding quality determination factors are output together in the process of predicting the quality of the agricultural product, it can be better understood that the predicted quality of the agricultural product is further evaluated, for example, when further evaluation is required (such as manual or other modes), the further verification is required when the estimated quality of the agricultural product is poor, if the verification is pertinent, the verification cost can be greatly reduced and the verification efficiency can be improved, and the problem that the effect of the prediction processing in the digital agriculture is relatively poor in the prior art can be improved to a certain extent based on the fact that the purposeful manual verification is performed according to the analyzed quality determination factors.
In order to make the above objects, features and advantages of the present invention more comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
Fig. 1 is a block diagram of an AI prediction processing system applied to digital agriculture according to an embodiment of the present invention.
Fig. 2 is a flowchart illustrating steps included in an AI prediction processing method applied to digital agriculture according to an embodiment of the present invention.
Fig. 3 is a schematic diagram of each module included in the AI prediction processing device applied to digital agriculture according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, but not all embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
Thus, the following detailed description of the embodiments of the invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
As shown in FIG. 1, an embodiment of the present invention provides an AI prediction processing system for digital agriculture. Wherein, the AI prediction processing system applied to digital agriculture can comprise a memory and a processor. And the memory and the processor are directly or indirectly electrically connected to realize the transmission or interaction of data. For example, electrical connection may be made to each other via one or more communication buses or signal lines. The memory may store at least one software functional module (computer program) that may exist in the form of software or firmware. The processor may be configured to execute the executable computer program stored in the memory, so as to implement the AI prediction processing method applied to digital agriculture provided by the embodiment of the present invention.
Alternatively, in some specific applications, the Memory may be, but is not limited to, random access Memory (Random Access Memory, RAM), read Only Memory (ROM), programmable Read Only Memory (Programmable Read-Only Memory, PROM), erasable Read Only Memory (Erasable Programmable Read-Only Memory, EPROM), electrically erasable Read Only Memory (Electric Erasable Programmable Read-Only Memory, EEPROM), and the like. The processor may be a general purpose processor including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), a System on Chip (SoC), etc.; but also Digital Signal Processors (DSPs), application Specific Integrated Circuits (ASICs), field Programmable Gate Arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
Alternatively, in some specific applications, the AI-prediction processing system for digital agriculture may be a server with data processing capabilities.
With reference to fig. 2, the embodiment of the invention further provides an AI prediction processing method applied to digital agriculture, which can be applied to the AI prediction processing system applied to digital agriculture. The method steps defined by the flow related to the AI prediction processing method applied to the digital agriculture can be realized by the AI prediction processing system applied to the digital agriculture.
The specific flow shown in fig. 2 will be described in detail.
And step S110, carrying out network updating on the candidate agricultural product prediction network according to the sample agricultural product data to form a target agricultural product prediction network.
In the embodiment of the invention, the AI prediction processing system applied to digital agriculture can update the candidate agricultural product prediction network according to sample agricultural product data to form a target agricultural product prediction network, wherein the sample agricultural product data has growth process description data of sample agricultural products. That is, the candidate agricultural product prediction network learns the information in the sample agricultural product data, such as learning the mapping relation between the agricultural product growth record and the agricultural product quality and the corresponding quality determination cause, so that the target agricultural product prediction network formed after learning can predict based on the mapping relation.
Step S120, obtaining a target agricultural product growth record and predicting target items.
In the embodiment of the invention, the AI prediction processing system applied to digital agriculture can acquire the growth record of the target agricultural product and predict the target item. The target agricultural product growth record is used for describing the growth process of the target agricultural product, and can be text data, image data and the like, such as soil humidity, air temperature, fertilization amount, growth time, insect condition, illumination, rainfall, air component content and agricultural product appearance characteristics (color, size, shape and the like) in the growth process of the agricultural product are recorded, and the prediction target items are used for guiding: predicting the quality of the agricultural product based on a plurality of predetermined reference agricultural product quality, and outputting the reference agricultural product quality of the loaded agricultural product growth record and a corresponding quality determination cause; illustratively, the prediction target item may specifically be one of the following:
the following loaded agricultural product growth records can be matched to which agricultural product quality, and account for (i.e., give a quality determination cause), the agricultural product quality includes a reference agricultural product quality 1, a reference agricultural product quality 2, a reference agricultural product quality 3, and a reference agricultural product quality 4; the loaded agricultural product growth record may be determined as to which agricultural product quality matches: reference agricultural product quality 1, reference agricultural product quality 2, reference agricultural product quality 3, and reference agricultural product quality 4, and account for (i.e., give a quality determination cause); four reference agricultural product masses are given, respectively: reference agricultural product quality 1, reference agricultural product quality 2, reference agricultural product quality 3, and reference agricultural product quality 4, please determine that the loaded agricultural product growth record matches one of the four reference agricultural product qualities, and account for reasons such as soil moisture being too low for a long period of time, light being insufficient for a long period of time, and the like.
And step S130, determining corresponding target agricultural product data combinations based on the target agricultural product growth records and the predicted target matters.
In the embodiment of the invention, the AI prediction processing system applied to digital agriculture can determine the corresponding target agricultural product data combination based on the target agricultural product growth record and the predicted target item. A target agricultural product growth record in the target agricultural product data combination as the loaded agricultural product growth record. For example, the target agricultural product data combination may be:
the following loaded agricultural product growth records can be matched to which agricultural product quality, and account for (i.e., give a quality determination cause), the agricultural product quality includes a reference agricultural product quality 1, a reference agricultural product quality 2, a reference agricultural product quality 3, and a reference agricultural product quality 4; record of loaded agricultural product growth: and recording the growth of the target agricultural product.
And step S140, through the target agricultural product prediction network, the target agricultural product growth record is subjected to prediction processing based on the target agricultural product data combination, and the reference agricultural product quality and the corresponding quality determination cause of the target agricultural product are output.
In the embodiment of the invention, the AI prediction processing system applied to digital agriculture can predict the target agricultural product growth record based on the target agricultural product data combination through the target agricultural product prediction network, and output the reference agricultural product quality and the corresponding quality determination cause of the target agricultural product. The target agricultural product has a reference agricultural product quality for reflecting quality predictions of the target agricultural product, such as quality, general, and inferior (other grades may also be included), the quality determination being due to reasons for reflecting the predictions of the quality predictions.
The target agricultural product prediction network may be a neural network, and include a feature mining unit, a quality prediction unit and a cause generation unit, where the feature mining unit may be a convolutional neural network, and is configured to perform convolution operation on the target agricultural product growth record (corresponding word embedded vector sequence) to form a corresponding agricultural product growth semantic feature, and the quality prediction unit is configured to, after performing full-connection processing on the agricultural product growth semantic feature, further process a result of the full-connection processing based on a classification function such as softmax, to obtain a corresponding probability distribution, where the probability distribution has probability parameters of each reference agricultural product quality (i.e. probabilities matched with each reference agricultural product quality), so that one reference agricultural product quality with the largest probability parameter may be used as the reference agricultural product quality that the target agricultural product has. In addition, the cause generation unit may be an RNN model (a recurrent neural network, which is a neural network model commonly used in sequence data processing, and has a feedback mechanism capable of capturing time sequence information in sequence data, gradually generating descriptive text, predicting one word at a time, and selecting the next most likely word using greedy Search, beam Search (Beam Search), or the like, until the complete descriptive text is generated), so that agricultural product growth semantic features may be loaded into the RNN model to generate the quality determination cause corresponding text.
Based on the foregoing, since the corresponding quality determination factors are output together in the process of predicting the quality of the agricultural product, it can be better understood that the predicted quality of the agricultural product is further evaluated, for example, when further evaluation is required (such as manual or other modes), the further verification is required when the estimated quality of the agricultural product is poor, if the verification is pertinent, the verification cost can be greatly reduced and the verification efficiency can be improved, and the problem that the effect of the prediction processing in the digital agriculture is relatively poor in the prior art can be improved to a certain extent based on the fact that the purposeful manual verification is performed according to the analyzed quality determination factors.
For example, the first agricultural product growth record (for kiwi, quality agricultural product is good quality agricultural product): soil humidity: is kept in a proper range of 40 to 60 percent; air temperature: maintaining the optimum temperature range of 20-30 ℃; fertilizing amount: fertilizing on time, providing proper amount of nutrients, and fertilizing 100 g each month; insect pest condition: the occurrence rate of insect pests is low, the influence of the insect pests is small, and a small amount of insect pests occasionally appear; illumination: the illumination intensity is moderate, the sunlight time is sufficient and stable, and the sunlight is sufficient for 10 hours; the corresponding quality determination cause may be: each growth index is normal;
Second agricultural product growth record (quality of agricultural product is general agricultural product): soil humidity: the fluctuation of soil humidity is large, and the situations of excessive drought and excessive humidity occur occasionally, namely 30% and 70% respectively; air temperature: the temperature fluctuation is large, and the temperature is high and low, namely 35 ℃ and 15 ℃ respectively; fertilizing amount: uneven fertilization, sometimes too much and too little fertilization, such as 200 grams and 50 grams fertilization every two months; insect pest condition: occasionally insect pests occur, but have a minor impact on agricultural products, such as a small number of pests, without affecting yield and quality; illumination: the light fluctuation is large, the light is strong and insufficient, and the light is insufficient in cloudy days or short days; the corresponding quality determination cause may be: each growth index is slightly abnormal (in other agricultural product growth records, partial growth index abnormality can be generated, and based on the abnormal growth index, quality determination can be caused by abnormal illumination, abnormal soil humidity, abnormal air temperature and the like, so that targeted analysis and the like can be performed);
third agricultural product growth record (agricultural product quality is defective agricultural product): soil humidity: soil humidity is kept unstable, and drought and over-wet phenomena are frequently generated, namely 20% and 80% respectively; air temperature: the temperature change is severe and is frequently influenced by high temperature and low temperature, namely 40 ℃ and 10 ℃; fertilizing amount: improper fertilization, lack of necessary nutrients or excessive fertilization causes problems, 50 grams or 500 grams per week; insect pest condition: frequently affected by insect attack, obvious damage is caused to agricultural products, a large number of insect pests are caused, and fruits are injured or yield is reduced; illumination: the illumination is insufficient, the normal growth requirement of agricultural products cannot be met, and the illumination is insufficient due to long-term cloudy days; the corresponding quality determination cause may be: various growth indexes are seriously abnormal.
It will be appreciated that the agricultural product growth record described above is only one example, and in practice, more indicators and more details may be included, such as each day, each time period.
Optionally, in some specific applications, the step S110 may include:
extracting sample agricultural product data corresponding to a candidate agricultural product prediction network, wherein the sample agricultural product data comprises a sample agricultural product growth record, sample agricultural product quality corresponding to the sample agricultural product growth record and a corresponding quality determination cause, the sample agricultural product quality of the sample agricultural product growth record belongs to one of a plurality of predetermined reference agricultural product qualities (such as high quality, general and defective products), and the sample agricultural product growth record is used for describing the growth process of the sample agricultural product;
determining the prediction target matters of the candidate agricultural product prediction network, as described in the related description;
determining a sample agricultural product data combination for updating the candidate agricultural product prediction network according to the prediction target item and the sample agricultural product data, wherein a sample agricultural product growth record in the sample agricultural product data is used as a loaded agricultural product growth record, as described in the related description;
According to the sample agricultural product data combination, carrying out network updating processing on the candidate agricultural product prediction network, and obtaining a target agricultural product prediction network based on the updated candidate agricultural product prediction network; for example, the candidate agricultural product prediction network may form the target agricultural product prediction network by learning the agricultural product data combination, i.e., learning a mapping relationship between sample agricultural product growth records and corresponding sample agricultural product quality and corresponding quality determination cause.
Optionally, in some specific applications, the sample agricultural product data belongs to a candidate sample agricultural product event in a sample agricultural product event set, and the AI prediction processing method applied to digital agriculture further includes a step of forming the sample agricultural product event set, which may include:
collecting a plurality of candidate sample agricultural product events of the candidate agricultural product prediction network, wherein one candidate sample agricultural product event comprises a sample agricultural product growth record and an agricultural product quality related index of the sample agricultural product growth record, the agricultural product quality related index of any sample agricultural product growth record comprises a sample agricultural product quality (namely actual agricultural product quality) of the sample agricultural product growth record and a corresponding quality determination cause, and the sample agricultural product quality in any candidate sample agricultural product event belongs to one reference agricultural product quality in the plurality of reference agricultural product qualities;
Determining at least one sample agricultural product growth record, such as randomly or according to a certain rule, from sample agricultural product growth records included in the plurality of candidate sample agricultural product events;
respectively adjusting the determined sample agricultural product growth records to form a third number of adjustment sample agricultural product growth records, wherein one adjustment sample agricultural product growth record corresponds to one sample agricultural product growth record, and any one adjustment sample agricultural product growth record and the corresponding sample agricultural product growth record have the same semantic characteristics; illustratively, the sample agricultural product growth record 1 may be subjected to adjustment processing to obtain a corresponding adjusted sample agricultural product growth record 1, the sample agricultural product growth record 2 may be subjected to adjustment processing to obtain a corresponding adjusted sample agricultural product growth record 2, and the sample agricultural product growth record 3 may be subjected to adjustment processing to obtain a corresponding adjusted sample agricultural product growth record 3, wherein the adjusted sample agricultural product growth record 1 and the sample agricultural product growth record 1 have the same semantic features, the adjusted sample agricultural product growth record 2 and the sample agricultural product growth record 2 have the same semantic features, and the adjusted sample agricultural product growth record 3 and the sample agricultural product growth record 3 have the same semantic features, i.e., the semantics are not changed when adjusted;
Marking the agricultural product quality related indexes of the sample agricultural product growth record corresponding to each adjustment sample agricultural product growth record as the agricultural product quality related indexes of the adjustment sample agricultural product growth record, respectively taking each adjustment sample agricultural product growth record as a new sample agricultural product growth record, and forming a third number of new candidate sample agricultural product events according to the third number of adjustment sample agricultural product growth records and the corresponding agricultural product quality related indexes; illustratively, as described above, the agricultural product quality related index of the sample agricultural product growth record 1 may be used as the agricultural product quality related index of the whole sample agricultural product growth record 1, the agricultural product quality related index of the sample agricultural product growth record 2 may be used as the agricultural product quality related index of the whole sample agricultural product growth record 2, the agricultural product quality related index of the sample agricultural product growth record 3 may be used as the agricultural product quality related index of the whole sample agricultural product growth record 3; the generated adjusted sample agricultural product growth record has the same semantic characteristics as the original sample agricultural product growth record, so that when a new candidate sample agricultural product event is formed by adopting the adjusted sample agricultural product growth record, the reference agricultural product quality and the quality determination cause of the original sample agricultural product growth record can be directly used, the workload of manually collecting and marking the sample agricultural product growth record can be effectively reduced on the basis of enlarging the scale and richness of a sample agricultural product event set, and the content and the data volume of the agricultural product growth record are very large, so that the actual application value is higher;
Forming the sample agricultural product event set according to the plurality of candidate sample agricultural product events and the third number of new candidate sample agricultural product events; based on this, sample agricultural product data can be reliably expanded based on the plurality of candidate sample agricultural product events.
Optionally, in some specific applications, the step of determining at least one sample agricultural product growth record from among sample agricultural product growth records included in the plurality of candidate sample agricultural product events may include:
determining a statistical quantity of each of the plurality of reference agricultural product qualities according to the plurality of candidate sample agricultural product events, wherein the statistical quantity of any one of the reference agricultural product qualities is used for reflecting: the number of candidate sample agricultural product events with any one of the reference agricultural product quality among the plurality of candidate sample agricultural product events is X, Y, Z, and W respectively corresponding to the reference agricultural product quality 1, 2, and 4;
Selecting a reference agricultural product quality satisfying a target condition from the plurality of reference agricultural product qualities based on the statistical quantity of each reference agricultural product quality, the reference agricultural product quality satisfying the target condition being: the corresponding reference agricultural product quality with the statistical quantity smaller than or equal to the pre-configured reference statistical quantity is not limited, and the specific numerical value of the reference statistical quantity can be configured according to actual demands, such as 10, 20, 50 and the like;
determining at least one sample agricultural product growth record from sample agricultural product growth records included in the candidate sample agricultural product events having the reference agricultural product quality satisfying the target condition; based on the method, the corresponding sample agricultural product growth records with smaller statistical quantity are determined so as to carry out adjustment and expansion, so that the sample quantity of the smaller reference agricultural product quality can be increased, and the samples can be balanced on the basis of enlarging the scale and the richness of the sample agricultural product event set.
Optionally, in some specific applications, the step of performing adjustment processing on the determined growth records of each sample agricultural product to form a third number of adjustment sample agricultural product growth records may include:
Extracting an adjustment target item to the growth record adjustment network, the adjustment target item to guide: outputting an adjusted sample agricultural product growth record having the same semantic features for the loaded sample agricultural product growth record based on the loaded sample agricultural product growth record and the sample agricultural product quality;
extracting at least one agricultural product growth record combination, wherein the loaded data in any agricultural product growth record combination comprises the following components: one sample agricultural product growth record and corresponding sample agricultural product quality, the expected output data (which can be used for comparison with the actual output data) in any one of said agricultural product growth record combinations comprises: a sample agricultural product growth record having the same semantic characteristics as the sample agricultural product growth record in the loaded data;
and analyzing, by the growth record adjustment network, adjusted sample agricultural product growth records having the same semantic features for each determined sample agricultural product growth record based on the adjustment target item and the at least one agricultural product growth record combination, thereby forming a third number of adjusted sample agricultural product growth records, that is, the growth record adjustment network may learn a mapping relationship between the sample agricultural product growth records and the corresponding sample agricultural product quality in the agricultural product growth record combination and the sample agricultural product growth records having the same semantic features, so that text generation processing may be performed on each determined sample agricultural product growth record based on the mapping relationship, that is, the adjusted sample agricultural product growth records having the same semantic features may be generated. That is, the growth record adjustment network may include a Convolutional Neural Network (CNN) and a cyclic neural network (RNN), where the convolutional neural network is configured to perform a convolution operation on the sample agricultural product growth record and the quality of the corresponding sample agricultural product to obtain a corresponding convolution feature, and the cyclic neural network may process the convolution feature to generate an adjusted sample agricultural product growth record with the same semantic feature.
For example, one adjustment target item may be: based on the loaded sample agricultural product growth record and the corresponding reference agricultural product quality of the sample agricultural product growth record, the sample agricultural product growth record is adjusted so as to have the same semantic characteristics and the reference agricultural product quality, and the description modes are different. Illustratively, one sample agricultural product growth is recorded as "soil moisture: is kept in a proper range of 40 to 60 percent; air temperature: maintaining the optimum temperature range of 20-30 ℃; fertilizing amount: fertilizing on time, providing proper amount of nutrients, and fertilizing 100 g each month; insect pest condition: the occurrence rate of insect pests is low, the influence of the insect pests is small, and a small amount of insect pests occasionally appear; illumination: the illumination intensity is moderate, the sunlight time is sufficient and stable, the sunlight is sufficient for 10 hours, and the corresponding adjustment sample agricultural product growth record is' the soil humidity which is stabilized in a proper range of 40% -60%; having an air temperature stabilized in an optimum temperature range of 20-30 ℃; the fertilizer is applied on time, and a fertilizer application state of 100 g is provided for providing proper amount of nutrients for each month; the pest situation has low pest incidence rate, is less affected by pests and occasionally causes a small amount of pests; the sunlight irradiation device has the illumination state of moderate illumination intensity, sufficient and stable sunlight irradiation time and sufficient sunlight irradiation time of 10 hours.
Optionally, in some specific applications, the step of analyzing, by the growth record adjustment network, the adjusted sample agricultural product growth records having the same semantic features for each determined sample agricultural product growth record based on the adjustment target item and the at least one agricultural product growth record combination, thereby forming a third number of adjusted sample agricultural product growth records may include:
combining the adjustment target item and the at least one agricultural product growth record combination to form a network update basis of the growth record adjustment network, namely a learned object;
based on the determined growth records of the sample agricultural products and the corresponding quality of the sample agricultural products, respectively forming loaded agricultural product information combinations corresponding to the determined growth records of the sample agricultural products;
updating the growth record adjustment network according to the network updating basis to form an updated growth record adjustment network; illustratively, generation of an adjusted sample agricultural product growth record having the same semantic features may be performed in accordance with the included sample agricultural product growth record and the corresponding sample agricultural product quality based on the network update, and then a difference between the generated adjusted sample agricultural product growth record and the network update in accordance with the included sample agricultural product growth record having the same semantic features may be calculated and the growth record adjustment network updated based on the difference; for example, the network update is based on a sample agricultural product growth record 2 comprising a sample agricultural product growth record 1, a corresponding sample agricultural product mass, and a sample agricultural product growth record 1 having the same semantic characteristics as the sample agricultural product growth record 1, such that the sample agricultural product growth record 1 and the corresponding sample agricultural product mass can be processed by the growth record adjustment network to generate an adjusted sample agricultural product growth record 1, then a difference between the adjusted sample agricultural product growth record 1 and the sample agricultural product growth record 2 can be calculated, and based on the difference, network parameters of the growth record adjustment network are updated;
Based on the determined agricultural product information combinations loaded corresponding to the sample agricultural product growth records, adjusting sample agricultural product growth records with the same semantic features according to analysis of the corresponding sample agricultural product growth records through the updated growth record adjustment network, so as to form a third number of adjustment sample agricultural product growth records; that is, according to the sample agricultural product growth record and the corresponding sample agricultural product quality included in the loaded agricultural product information combination, according to the learned mapping relation, an adjusted sample agricultural product growth record with the same semantic characteristics is generated.
Wherein, optionally, in some specific applications, the step of analyzing, by the growth record adjustment network, the adjusted sample agricultural product growth records having the same semantic features for each determined sample agricultural product growth record based on the adjustment target item and the at least one agricultural product growth record combination, thereby forming a third number of adjusted sample agricultural product growth records may include:
polling the determined growth records of each sample agricultural product, and marking the currently polled growth records of the sample agricultural product as the growth records of the agricultural product to be processed; illustratively, the determined sample agricultural product growth record may include a sample agricultural product growth record a, a sample agricultural product growth record B, and a sample agricultural product growth record C, so that the present agricultural product growth record a, the sample agricultural product growth record B, and the sample agricultural product growth record C may be sequentially taken as the agricultural product growth record to be processed;
Forming a corresponding one loaded agricultural product information combination based on the agricultural product growth record to be processed and the corresponding sample agricultural product quality, and cascading the adjustment target item, the at least one agricultural product growth record combination and the formed loaded agricultural product information combination to form a growth record adjustment target item corresponding to the agricultural product growth record to be processed;
after learning the growth record adjustment target items corresponding to the agricultural product growth record to be processed through the growth record adjustment network, analyzing an adjustment sample agricultural product growth record with the same semantic characteristics as the agricultural product growth record to be processed, namely, carrying out corresponding sequential network update on each determined sample agricultural product growth record, and then generating the adjustment sample agricultural product growth record;
and after all the determined sample agricultural product growth records are polled, forming a third number of adjusted sample agricultural product growth records, such as an adjusted sample agricultural product growth record corresponding to the agricultural product growth record A, an adjusted sample agricultural product growth record corresponding to the sample agricultural product growth record B and an adjusted sample agricultural product growth record corresponding to the sample agricultural product growth record C.
Optionally, in some specific applications, the step of determining a sample agricultural product data combination for updating the candidate agricultural product prediction network according to the prediction target item and the sample agricultural product data may include:
combining the predicted target item and sample agricultural product growth records in the sample agricultural product data, such as splicing two text data, to form corresponding first agricultural product combined data, wherein the sample agricultural product growth records are at least used for describing soil humidity, air temperature, fertilization amount, growth time, insect condition, illumination, rainfall, air component content and agricultural product appearance characteristics in the growth process of the sample agricultural product;
combining the quality of the sample agricultural products in the sample agricultural product data and the quality determination cause in the sample agricultural product data, such as splicing two text data, to form corresponding second agricultural product combined data, wherein the quality of the sample agricultural products at least comprises high quality, general and inferior products, and the quality determination cause at least comprises local causes corresponding to various factors including soil humidity, air temperature, fertilization amount, insect condition, illumination, rainfall and air component content;
Determining a sample agricultural product data combination for updating the candidate agricultural product prediction network based on the first agricultural product merging data and the second agricultural product merging data; that is, the sample agricultural product data combination may include first agricultural product mix data and second agricultural product mix data.
Optionally, in some specific applications, the step of performing network update processing on the candidate agricultural product prediction network according to the sample agricultural product data combination, and obtaining the target agricultural product prediction network based on the updated candidate agricultural product prediction network may include:
carrying out quality prediction on a sample agricultural product growth record in the sample agricultural product data based on first agricultural product merging data in the sample agricultural product data combination through the candidate agricultural product prediction network, and outputting corresponding agricultural product quality prediction data, wherein the agricultural product quality prediction data comprises reference agricultural product quality and a corresponding quality determination cause of the sample agricultural product growth record; that is, analysis and prediction can be performed on the sample agricultural product growth records in the first agricultural product merging data to obtain a reference agricultural product quality and a corresponding quality determination cause;
Updating network parameters of the candidate agricultural product prediction network based on the difference between the agricultural product quality prediction data and the second agricultural product merging data in the sample agricultural product data combination to obtain a target agricultural product prediction network; illustratively, an error (which may be cross entropy between two probability distributions, the predicted reference agricultural product quality may be a reference agricultural product quality 1-probability 11, a reference agricultural product quality 2-probability 12, a reference agricultural product quality 3-probability 13, and a reference agricultural product quality 4-probability 14) between a reference agricultural product quality in the agricultural product quality prediction data and a sample agricultural product quality in the second agricultural product merging data, which may be a reference agricultural product quality 1-probability 21, a reference agricultural product quality 2-probability 22, a reference agricultural product quality 3-probability 23, and a reference agricultural product quality 4-probability 24, may be calculated, and an error (such as a text edit distance or the like or a distance mapped into a feature space or the like) between a quality determination cause in the agricultural product quality prediction data and a quality determination cause in the second agricultural product merging data may be calculated, and then a sum of the two errors may be calculated, and a network parameter of a candidate agricultural product prediction network may be updated based on the sum value, to obtain a target prediction network.
Optionally, in some specific applications, the sample agricultural product data combination is a data set formed by combining the prediction target item and the sample agricultural product data, and before the step of outputting corresponding agricultural product quality prediction data by the candidate agricultural product prediction network, performing quality prediction on a sample agricultural product growth record in the sample agricultural product data based on the first agricultural product combination data in the sample agricultural product data combination; the step of performing network update processing on the candidate agricultural product prediction network according to the sample agricultural product data combination, and obtaining a target agricultural product prediction network based on the updated candidate agricultural product prediction network may further include:
the sample agricultural product data combination is subjected to data segmentation to form a first number of agricultural product data fragments, and as mentioned above, the sample agricultural product data combination can be text data, so that the text data can be segmented, for example, according to granularity levels such as words or sentences, and the obtained agricultural product data fragments can be text words or text sentences;
Determining network loading data by utilizing a first second number of agricultural product data fragments in the first number of agricultural product data fragments based on the precedence relation of the first number of agricultural product data fragments (namely, the precedence of the first number of agricultural product data fragments in the sample agricultural product data combination), and determining network expected data corresponding to the network loading data by utilizing a second number of agricultural product data fragments except the first number of agricultural product data fragments in the first number of agricultural product data fragments, wherein the difference value between the first number and the second number is equal to 1;
sequentially carrying out prediction output on the network loading data according to the granularity of the agricultural product data fragments through the candidate agricultural product prediction network to form a corresponding prediction output result, wherein the prediction output result comprises a second number of agricultural product data fragments which are predicted to be output, and the a-th agricultural product data fragment in the prediction output result is predicted to be output based on the first a agricultural product data fragments in the network loading data, wherein a is smaller than or equal to the second number; that is, the 1 st agricultural product data segment in the prediction output result is predicted to be output based on the first 1 agricultural product data segment in the network loading data, the 2 nd agricultural product data segment in the prediction output result is predicted to be output based on the first 2 agricultural product data segments in the network loading data, the 3 rd agricultural product data segment in the prediction output result is predicted to be output based on the first 3 agricultural product data segments in the network loading data, and the 4 th agricultural product data segment in the prediction output result is predicted to be output based on the first 4 agricultural product data segments in the network loading data;
And updating network parameters of the candidate agricultural product prediction network based on the difference between the network expected data and the prediction output result (namely, the error between the actual data and the prediction data), so as to obtain the candidate agricultural product prediction network after preliminary updating.
For example, the sample agricultural product data is combined as "soil moisture: the text word segmentation processing can be carried out to obtain each word segmentation word while keeping the proper range of 40% -60% ": "soil moisture", "holding", "40%", "60%", and "suitable range". The corresponding network loading data may be ("soil moisture", "holding", "40%", "60%"), and the corresponding network desired data may be ("holding", "40%", "60%", "suitable range"), based on which the 1 st agricultural product data segment in the predicted output result may be predicted to correspond to the "holding" in the network desired data based on the first 1 agricultural product data segment in the network loading data, i.e., "soil moisture"; the 2 nd agricultural product data segment in the predicted output result can be predicted based on the first 2 agricultural product data segments in the network loading data, namely 'soil humidity', 'holding', and corresponds to '40%' in network expected data; the 3 rd agricultural product data segment in the prediction output result can be predicted based on the first 3 agricultural product data segments in the network loading data, namely ' soil humidity ', ' holding ', 40% ', and corresponds to ' 60% ' in network expected data; the 4 th agricultural product data segment in the prediction output result can be predicted to correspond to the "suitable range" in the network expected data based on the first 4 agricultural product data segments in the network loading data, namely, "soil humidity", "holding", 40% ", and" 60% ". Based on the method, the preliminarily updated candidate agricultural product prediction network can learn some text-generated mapping relations in the agricultural product quality prediction field, so that the efficiency is higher when the quality prediction is learned later, namely, the preliminary training is performed firstly, such as updating the parameters of the feature mining unit; based on this, sufficient learning and updating can be performed with limited sample agricultural product data, so that a more reliable network can be obtained with less sample agricultural product data.
Optionally, in some specific applications, the step of updating the network parameters of the candidate agricultural product prediction network based on the difference between the network expected data and the prediction output result to obtain the candidate agricultural product prediction network after preliminary updating may include:
based on the network expected data and the prediction output result, determining error parameters corresponding to each agricultural product data segment in the prediction output result (namely calculating edit distances and the like corresponding to each agricultural product data segment), wherein the error parameters corresponding to the a-th agricultural product data segment in the prediction output result are used for reflecting: a distinction between the a-th agricultural product data segment in the predicted output result and the a-th agricultural product data segment in the network expected data;
combining error parameters corresponding to each agricultural product data segment in the prediction output result, such as summation calculation, and outputting network error parameters of the candidate agricultural product prediction network;
and updating the network parameters of the candidate agricultural product prediction network along the direction of reducing the network error parameters to obtain a preliminarily updated candidate agricultural product prediction network, namely, performing quality prediction on the sample agricultural product growth record in the sample agricultural product data based on the first agricultural product merging data in the sample agricultural product data combination by using the candidate agricultural product prediction network, and outputting corresponding agricultural product quality prediction data.
Referring to fig. 3, the embodiment of the invention further provides an AI prediction processing device applied to digital agriculture, which can be applied to the AI prediction processing system applied to digital agriculture. Wherein, the AI prediction processing apparatus applied to digital agriculture may include:
the prediction network updating module is used for carrying out network updating on the candidate agricultural product prediction network according to sample agricultural product data to form a target agricultural product prediction network, wherein the sample agricultural product data is provided with growth process description data of sample agricultural products;
the to-be-predicted information acquisition module is used for acquiring a target agricultural product growth record and predicting target matters, wherein the target agricultural product growth record is used for describing the growth process of the target agricultural product, and the predicting target matters are used for guiding: predicting the quality of the agricultural product based on a plurality of predetermined reference agricultural product quality, and outputting the reference agricultural product quality of the loaded agricultural product growth record and a corresponding quality determination cause;
the information to be predicted combination module is used for determining a corresponding target agricultural product data combination based on the target agricultural product growth record and the predicted target item, wherein the target agricultural product growth record in the target agricultural product data combination is used as the loaded agricultural product growth record;
The agricultural product quality prediction module is used for predicting the target agricultural product growth record based on the target agricultural product data combination through the target agricultural product prediction network, outputting the reference agricultural product quality of the target agricultural product and a corresponding quality determination reason, wherein the reference agricultural product quality of the target agricultural product is used for reflecting the quality prediction result of the target agricultural product, and the quality determination reason is used for reflecting the reason for predicting the quality prediction result.
In summary, according to the AI prediction processing method and system for digital agriculture provided by the invention, the candidate agricultural product prediction network is updated according to the sample agricultural product data to form the target agricultural product prediction network; acquiring a target agricultural product growth record and predicting target matters; determining a corresponding target agricultural product data combination based on the target agricultural product growth record and the predicted target item; and (3) through a target agricultural product prediction network, performing prediction processing on the target agricultural product growth record based on target agricultural product data combination, and outputting the reference agricultural product quality and the corresponding quality determination cause of the target agricultural product. Based on the foregoing, since the corresponding quality determination factors are output together in the process of predicting the quality of the agricultural product, it can be better understood that the predicted quality of the agricultural product is further evaluated, for example, when further evaluation is required (such as manual or other modes), the further verification is required when the estimated quality of the agricultural product is poor, if the verification is pertinent, the verification cost can be greatly reduced and the verification efficiency can be improved, and the problem that the effect of the prediction processing in the digital agriculture is relatively poor in the prior art can be improved to a certain extent based on the fact that the purposeful manual verification is performed according to the analyzed quality determination factors.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. An AI prediction processing method applied to digital agriculture, characterized by comprising the following steps:
according to sample agricultural product data, carrying out network updating on a candidate agricultural product prediction network to form a target agricultural product prediction network, wherein the sample agricultural product data is provided with growth process description data of sample agricultural products, and the candidate agricultural product prediction network learns a mapping relation between an agricultural product growth record and agricultural product quality and a corresponding quality determination cause by learning information in the sample agricultural product data, so that the target agricultural product prediction network formed after learning predicts based on the mapping relation;
obtaining a target agricultural product growth record describing a growth process of the target agricultural product and a predicted target item guiding: predicting the quality of the agricultural product based on a plurality of predetermined reference agricultural product quality, and outputting the reference agricultural product quality of the loaded agricultural product growth record and a corresponding quality determination cause; the loaded agricultural product growth record refers to the agricultural product growth record loaded to the target agricultural product prediction network;
Determining a corresponding target agricultural product data combination based on the target agricultural product growth record and the predicted target item, wherein the target agricultural product growth record in the target agricultural product data combination is used as the loaded agricultural product growth record;
and through the target agricultural product prediction network, performing prediction processing on the target agricultural product growth record based on the target agricultural product data combination, and outputting a reference agricultural product quality of the target agricultural product and a corresponding quality determination reason, wherein the reference agricultural product quality of the target agricultural product is used for reflecting a quality prediction result of the target agricultural product, and the quality determination reason is used for reflecting the reason for predicting the quality prediction result.
2. The AI-prediction processing method for digital agriculture according to claim 1, wherein the step of performing network update on the candidate agricultural product prediction network based on the sample agricultural product data to form the target agricultural product prediction network comprises:
extracting sample agricultural product data corresponding to a candidate agricultural product prediction network, wherein the sample agricultural product data comprises a sample agricultural product growth record, sample agricultural product quality corresponding to the sample agricultural product growth record and a corresponding quality determination cause, the sample agricultural product quality corresponding to the sample agricultural product growth record belongs to one of a plurality of predetermined reference agricultural product qualities, and the sample agricultural product growth record is used for describing the growth process of the sample agricultural product;
Determining a prediction target item of the candidate agricultural product prediction network;
determining a sample agricultural product data combination for updating the candidate agricultural product prediction network according to the prediction target item and the sample agricultural product data, wherein a sample agricultural product growth record in the sample agricultural product data is used as an agricultural product growth record loaded into the candidate agricultural product prediction network;
and carrying out network updating processing on the candidate agricultural product prediction network according to the sample agricultural product data combination, and obtaining a target agricultural product prediction network based on the updated candidate agricultural product prediction network.
3. The AI prediction processing method for digital agriculture according to claim 2, wherein the step of determining a sample agricultural product data combination for updating the candidate agricultural product prediction network based on the prediction target item and the sample agricultural product data comprises:
combining the predicted target item and sample agricultural product growth records in the sample agricultural product data to form corresponding first agricultural product combined data, wherein the sample agricultural product growth records are at least used for describing soil humidity, air temperature, fertilization amount, growth time, insect condition, illumination, rainfall, air component content and agricultural product appearance characteristics in the growth process of the sample agricultural products;
Combining the quality of the sample agricultural products in the sample agricultural product data and the quality determination cause in the sample agricultural product data to form corresponding second agricultural product combination data, wherein the quality of the sample agricultural products at least comprises high quality, general and inferior products, and the quality determination cause at least comprises a local cause corresponding to each factor of soil humidity, air temperature, fertilization amount, insect condition, illumination, rainfall and air component content;
determining a sample agricultural product data combination for updating the candidate agricultural product prediction network based on the first agricultural product merging data and the second agricultural product merging data;
and the step of performing network update processing on the candidate agricultural product prediction network according to the sample agricultural product data combination, and obtaining a target agricultural product prediction network based on the updated candidate agricultural product prediction network, includes:
carrying out quality prediction on a sample agricultural product growth record in the sample agricultural product data based on first agricultural product merging data in the sample agricultural product data combination through the candidate agricultural product prediction network, and outputting corresponding agricultural product quality prediction data, wherein the agricultural product quality prediction data comprises reference agricultural product quality and a corresponding quality determination cause of the sample agricultural product growth record;
And updating network parameters of the candidate agricultural product prediction network based on the difference between the agricultural product quality prediction data and the second agricultural product merging data in the sample agricultural product data combination to obtain a target agricultural product prediction network.
4. The AI prediction processing method for digital agriculture according to claim 3, wherein the sample agricultural product data combination is a data set formed by combining the prediction target item and the sample agricultural product data, and before the step of outputting the corresponding agricultural product quality prediction data by the candidate agricultural product prediction network, performing network update processing on the candidate agricultural product prediction network based on the updated candidate agricultural product prediction network, the step of obtaining a target agricultural product prediction network further includes:
the sample agricultural product data combination is subjected to data segmentation to form a first number of agricultural product data fragments, the sample agricultural product data combination is text data, and the agricultural product data fragments are text words or text sentences;
Determining network loading data by using a first second number of agricultural product data fragments in the first number of agricultural product data fragments based on the precedence relation of the first number of agricultural product data fragments, and determining network expected data corresponding to the network loading data by using a second number of agricultural product data fragments except for the first number of agricultural product data fragments in the first number of agricultural product data fragments, wherein the difference between the first number and the second number is equal to 1, and the precedence relation represents precedence in text data corresponding to sample agricultural product data combination;
sequentially carrying out prediction output on the network loading data according to the granularity of the agricultural product data fragments through the candidate agricultural product prediction network to form a corresponding prediction output result, wherein the prediction output result comprises a second number of agricultural product data fragments, an a-th agricultural product data fragment in the prediction output result is predicted and output based on the first a-th agricultural product data fragment in the network loading data, a is smaller than or equal to the second number, and the second number of agricultural product data fragments in the prediction output result corresponds to the agricultural product data fragments in the network expected data;
And updating the network parameters of the candidate agricultural product prediction network based on the difference between the network expected data and the prediction output result to obtain the candidate agricultural product prediction network after preliminary updating.
5. The AI-prediction processing method for digital agriculture as claimed in claim 4, wherein said step of updating network parameters of said candidate agricultural product prediction network based on a difference between said network expected data and said prediction output result to obtain a preliminarily updated candidate agricultural product prediction network comprises:
based on the network expected data and the prediction output result, determining error parameters corresponding to all agricultural product data segments in the prediction output result, wherein the error parameters corresponding to the a-th agricultural product data segment in the prediction output result are used for reflecting: a distinction between the a-th agricultural product data segment in the predicted output result and the a-th agricultural product data segment in the network expected data;
combining error parameters corresponding to each agricultural product data segment in the prediction output result, and outputting network error parameters of the candidate agricultural product prediction network;
And updating the network parameters of the candidate agricultural product prediction network along the direction of reducing the network error parameters to obtain the candidate agricultural product prediction network after preliminary updating.
6. The AI-prediction processing method for digital agriculture of any of claims 2 to 5, wherein the sample agricultural product data belongs to a candidate sample agricultural product event in a sample agricultural product event set, the AI-prediction processing method for digital agriculture further comprising a step of forming the sample agricultural product event set, the step of forming comprising:
collecting a plurality of candidate sample agricultural product events of the candidate agricultural product prediction network, wherein one candidate sample agricultural product event comprises a sample agricultural product growth record and an agricultural product quality related index of the sample agricultural product growth record, the agricultural product quality related index of any sample agricultural product growth record comprises a sample agricultural product quality and a corresponding quality determination cause of the sample agricultural product growth record, and the sample agricultural product quality in any candidate sample agricultural product event belongs to one reference agricultural product quality in the plurality of reference agricultural product qualities;
Determining at least one sample agricultural product growth record from among sample agricultural product growth records included in the plurality of candidate sample agricultural product events;
respectively adjusting the determined sample agricultural product growth records to form a third number of adjustment sample agricultural product growth records, wherein one adjustment sample agricultural product growth record corresponds to one sample agricultural product growth record, and any one adjustment sample agricultural product growth record and the corresponding sample agricultural product growth record have the same semantic characteristics;
marking the agricultural product quality related indexes of the sample agricultural product growth record corresponding to each adjustment sample agricultural product growth record as the agricultural product quality related indexes of the adjustment sample agricultural product growth record, respectively taking each adjustment sample agricultural product growth record as a new sample agricultural product growth record, and forming a third number of new candidate sample agricultural product events according to the third number of adjustment sample agricultural product growth records and the corresponding agricultural product quality related indexes;
forming the sample agricultural product event set based on the plurality of candidate sample agricultural product events and the third number of new candidate sample agricultural product events.
7. The AI prediction processing method for digital agriculture of claim 6, wherein the step of determining at least one sample agricultural product growth record from among sample agricultural product growth records included in the plurality of candidate sample agricultural product events comprises:
determining a statistical quantity of each of the plurality of reference agricultural product qualities according to the plurality of candidate sample agricultural product events, wherein the statistical quantity of any one of the reference agricultural product qualities is used for reflecting: the number of candidate sample agricultural product events of the plurality of candidate sample agricultural product events having the arbitrary reference agricultural product quality;
selecting a reference agricultural product quality satisfying a target condition from the plurality of reference agricultural product qualities based on the statistical quantity of each reference agricultural product quality, the reference agricultural product quality satisfying the target condition being: the corresponding statistical quantity is smaller than or equal to the reference agricultural product quality of the preset reference statistical quantity;
at least one sample agricultural product growth record is determined from sample agricultural product growth records included in candidate sample agricultural product events having the reference agricultural product quality that satisfies the target condition.
8. The AI-prediction processing method for digital agriculture according to claim 6, wherein the step of adjusting each of the determined sample agricultural product growth records to form a third number of adjusted sample agricultural product growth records comprises:
extracting an adjustment target item to the growth record adjustment network, the adjustment target item to guide: outputting an adjusted sample agricultural product growth record having the same semantic features for the loaded sample agricultural product growth record based on the loaded sample agricultural product growth record and the sample agricultural product quality;
extracting at least one agricultural product growth record combination, wherein the loaded data in any agricultural product growth record combination comprises the following components: one sample agricultural product growth record and corresponding sample agricultural product quality, the expected output data in any one of said agricultural product growth record combinations comprising: a sample agricultural product growth record having the same semantic characteristics as the sample agricultural product growth record in the loaded data;
and analyzing, by the growth record adjustment network, adjusted sample agricultural product growth records having the same semantic features for each determined sample agricultural product growth record based on the adjustment target item and the at least one agricultural product growth record combination, thereby forming a third number of adjusted sample agricultural product growth records.
9. The AI prediction processing method for digital agriculture of claim 8, wherein the step of analyzing, by the growth record adjustment network, adjustment sample agricultural product growth records having the same semantic characteristics for each of the determined sample agricultural product growth records based on the adjustment target item and the at least one agricultural product growth record combination, thereby forming a third number of adjustment sample agricultural product growth records, comprises:
combining the adjustment target item and the at least one agricultural product growth record combination to form a network update basis of the growth record adjustment network;
based on the determined growth records of the sample agricultural products and the corresponding quality of the sample agricultural products, respectively forming loaded agricultural product information combinations corresponding to the determined growth records of the sample agricultural products;
updating the growth record adjustment network according to the network updating basis to form an updated growth record adjustment network;
and through the updated growth record adjustment network, adjusting sample agricultural product growth records with the same semantic characteristics according to analysis of corresponding sample agricultural product growth records based on the loaded agricultural product information combination corresponding to the determined sample agricultural product growth records, so as to form a third number of adjustment sample agricultural product growth records.
10. An AI prediction processing system for use in digital agriculture, comprising a processor and a memory, the memory for storing a computer program, the processor for executing the computer program to implement the method of any of claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311387219.7A CN117114917B (en) | 2023-10-25 | 2023-10-25 | AI prediction processing method and system applied to digital agriculture |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311387219.7A CN117114917B (en) | 2023-10-25 | 2023-10-25 | AI prediction processing method and system applied to digital agriculture |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117114917A CN117114917A (en) | 2023-11-24 |
| CN117114917B true CN117114917B (en) | 2024-01-26 |
Family
ID=88798846
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311387219.7A Active CN117114917B (en) | 2023-10-25 | 2023-10-25 | AI prediction processing method and system applied to digital agriculture |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117114917B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101162384A (en) * | 2006-10-12 | 2008-04-16 | 魏珉 | Artificial intelligence plant growth surroundings regulate and control expert decision-making system |
| CN110880118A (en) * | 2019-11-26 | 2020-03-13 | 徐州斯塬网络科技有限公司 | Block chain-based high-quality agricultural product guarantee method |
| CN111582324A (en) * | 2020-04-20 | 2020-08-25 | 广州海睿信息科技有限公司 | Agricultural big data analysis method and device |
| CN113177578A (en) * | 2021-04-06 | 2021-07-27 | 安徽农业大学 | Agricultural product quality classification method based on LSTM |
| CN113435649A (en) * | 2021-06-29 | 2021-09-24 | 布瑞克农业大数据科技集团有限公司 | Global agricultural data sorting method, system, device and medium |
| CN113469746A (en) * | 2021-07-09 | 2021-10-01 | 布瑞克农业大数据科技集团有限公司 | Agricultural product prediction method and system |
| CN113487376A (en) * | 2021-01-08 | 2021-10-08 | 苏州市吴中区新鹤数创信息科技有限公司 | AI intelligent agricultural batch cloud platform processing system and method |
| CN116668473A (en) * | 2023-05-05 | 2023-08-29 | 仲恺农业工程学院 | Edge recognition method, system, equipment and medium for agricultural abnormal data |
| CN116824362A (en) * | 2022-04-06 | 2023-09-29 | 布瑞克(苏州)农业互联网股份有限公司 | Agricultural product monitoring method and system |
-
2023
- 2023-10-25 CN CN202311387219.7A patent/CN117114917B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101162384A (en) * | 2006-10-12 | 2008-04-16 | 魏珉 | Artificial intelligence plant growth surroundings regulate and control expert decision-making system |
| CN110880118A (en) * | 2019-11-26 | 2020-03-13 | 徐州斯塬网络科技有限公司 | Block chain-based high-quality agricultural product guarantee method |
| CN111582324A (en) * | 2020-04-20 | 2020-08-25 | 广州海睿信息科技有限公司 | Agricultural big data analysis method and device |
| CN113487376A (en) * | 2021-01-08 | 2021-10-08 | 苏州市吴中区新鹤数创信息科技有限公司 | AI intelligent agricultural batch cloud platform processing system and method |
| CN113177578A (en) * | 2021-04-06 | 2021-07-27 | 安徽农业大学 | Agricultural product quality classification method based on LSTM |
| CN113435649A (en) * | 2021-06-29 | 2021-09-24 | 布瑞克农业大数据科技集团有限公司 | Global agricultural data sorting method, system, device and medium |
| CN113469746A (en) * | 2021-07-09 | 2021-10-01 | 布瑞克农业大数据科技集团有限公司 | Agricultural product prediction method and system |
| CN116824362A (en) * | 2022-04-06 | 2023-09-29 | 布瑞克(苏州)农业互联网股份有限公司 | Agricultural product monitoring method and system |
| CN116668473A (en) * | 2023-05-05 | 2023-08-29 | 仲恺农业工程学院 | Edge recognition method, system, equipment and medium for agricultural abnormal data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117114917A (en) | 2023-11-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11222385B2 (en) | Method and system for forecasting crop yield | |
| EP3482630A1 (en) | Method, system and computer program for performing a pest forecast | |
| EP3816880A1 (en) | A yield estimation method for arable crops and grasslands, coping with extreme weather conditions and with limited reference data requirements | |
| US20200111174A1 (en) | Probabilistic Load Forecasting via Point Forecast Feature Integration | |
| Chiu et al. | Modelling and forecasting of greenhouse whitefly incidence using time-series and ARIMAX analysis | |
| CN116012720B (en) | Crop growth condition monitoring method, device and medium based on high-resolution remote sensing image | |
| Fila et al. | Calibration and validation of grapevine budburst models using growth-room experiments as data source | |
| Li et al. | Comparison of parameter uncertainty analysis techniques for a TOPMODEL application | |
| CN112352523B (en) | Tea garden water and fertilizer irrigation control method and system based on intelligent decision | |
| CN114330934A (en) | Model parameter self-adaptive GRU new energy short-term power generation power prediction method | |
| CN117114917B (en) | AI prediction processing method and system applied to digital agriculture | |
| Rajesh et al. | Data assimilation of gross domestic product (gdp) in india using stochastic data mining approach | |
| CN117371727A (en) | Intelligent agricultural monitoring and planning method and system based on Internet of things | |
| Abdurahman et al. | IoT-Based Smart Farming Using Machine Learning For Red Spinach | |
| CN116629117A (en) | Crop disease occurrence area prediction method based on gridding monitoring | |
| CN113537645B (en) | Soybean yield prediction method based on machine learning fusion satellite and weather data | |
| KR20200070736A (en) | Method of predicting crop yield and apparatus for embodying the same | |
| CN114372611A (en) | Water resource optimal allocation method and device | |
| CN113723670A (en) | Photovoltaic power generation power short-term prediction method with variable time window | |
| Fabregas | Decision Tree Algorithm Applied in Suitability Assessment of Temporary Crops Based on Agrometeorological Forecasts | |
| CN117745148B (en) | Multi-source data-based rice stubble flue-cured tobacco planting quality evaluation method and system | |
| Islam et al. | Machine Learning Models to Predict Soil Moisture for Irrigation Schedule | |
| KR102471743B1 (en) | Method for forecasting future production of smart farms | |
| Nedoumaran et al. | Precision Farming Using Machine Learning and Data Analytics | |
| Dayalini et al. | Agro-Mate: A Virtual Assister to Maximize Crop Yield in Agriculture Sector |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |