CN110687619A - Meteorological data verification method and system for farmland meteorological station - Google Patents

Meteorological data verification method and system for farmland meteorological station Download PDF

Info

Publication number
CN110687619A
CN110687619A CN201910882810.7A CN201910882810A CN110687619A CN 110687619 A CN110687619 A CN 110687619A CN 201910882810 A CN201910882810 A CN 201910882810A CN 110687619 A CN110687619 A CN 110687619A
Authority
CN
China
Prior art keywords
data
verification
meteorological data
farmland
meteorological
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.)
Granted
Application number
CN201910882810.7A
Other languages
Chinese (zh)
Other versions
CN110687619B (en
Inventor
于景鑫
张钟莉莉
孟范玉
岳焕芳
吴文彪
李文龙
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Research Center for Information Technology in Agriculture
Original Assignee
Beijing Research Center for Information Technology in Agriculture
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing Research Center for Information Technology in Agriculture filed Critical Beijing Research Center for Information Technology in Agriculture
Priority to CN201910882810.7A priority Critical patent/CN110687619B/en
Publication of CN110687619A publication Critical patent/CN110687619A/en
Application granted granted Critical
Publication of CN110687619B publication Critical patent/CN110687619B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01W—METEOROLOGY
    • G01W1/00—Meteorology
    • G01W1/18—Testing or calibrating meteorological apparatus
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01W—METEOROLOGY
    • G01W1/00—Meteorology
    • G01W1/02—Instruments for indicating weather conditions by measuring two or more variables, e.g. humidity, pressure, temperature, cloud cover or wind speed
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Environmental & Geological Engineering (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • Biophysics (AREA)
  • General Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Biomedical Technology (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Atmospheric Sciences (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Ecology (AREA)
  • Environmental Sciences (AREA)
  • Image Processing (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The embodiment of the invention provides a meteorological data verification method and system for a farmland meteorological station. The method comprises the steps that an equipment edge node acquires environment information and farmland image information collected by a farmland weather station; wherein the environmental information comprises at least one meteorological data; the equipment edge nodes obtain a verification result of each meteorological data through a preset quality verification model; the device edge node sends the environment information, the farmland image information and the verification results of all meteorological data to a cloud server, so that the cloud server performs correction interpolation operation on the meteorological data with abnormal verification results through a preset correction interpolation model.

Description

Meteorological data verification method and system for farmland meteorological station
Technical Field
The invention relates to the technical field of agriculture, in particular to a meteorological data verification method and system for a farmland meteorological station.
Background
The farmland meteorological station realizes the automatic acquisition of data such as farmland air temperature, air humidity, rainfall, wind speed, wind direction, illumination, agricultural condition pictures and the like by integrating sensor probes such as an air hygrothermograph, a rain gauge, a two-dimensional ultrasonic anemoscope, an illuminometer, a camera and the like. The meteorological data of the farmland environment is closely related to the growth and development, water transpiration, nutrient absorption and the like of crops, and is an important decision basis for guiding farmland irrigation, agricultural condition estimation, agricultural operation and the like. The farmland weather station is different from the ordinary weather station in that the arrangement height of the farmland weather station is 2-2.5m, and the arrangement height of the ordinary weather station can reach 10m, so that the parameters acquired by the ordinary weather station cannot be directly used for agricultural application. The farmland weather station is generally arranged in the farmland or in the vicinity of the farmland, is very complicated to be influenced by the farmland environment, and influences the weather data of the farmland such as farmland irrigation and crop growth. In addition, the requirement of agricultural operation on data monitoring density in a critical growth period is high, so that the monitoring data volume is large, even if the average daily monitoring data of a single device exceeds 1440 items according to the hourly monitoring frequency, the manual data quality verification and discrimination are very difficult. The meteorological data of the farmland show a complex nonlinear relationship among various parameters, and the equipment has personalized differences of regions, farmland types, soil types, different planted crops and the like, and the monitored interruption and loss of the meteorological data can greatly weaken the usability of a data set, thereby causing data resource waste.
The existing quality check of meteorological data mainly comprises the steps of performing data quality control through climate extreme value range check, time consistency check and space consistency check, or establishing a discrimination model according to historical data, performing quality judgment through comparison of predicted data and actually measured data of the meteorological data, or removing confirmed abnormal data, and finally establishing a linear correlation relationship with data of an adjacent unit to perform missing time interval data interpolation and storage.
The prior art is mainly directed at the traditional meteorological station at present, is not suitable for the meteorological data that obtains to the farmland meteorological station, and is not accurate enough to the quality testing of meteorological data.
Disclosure of Invention
Because the existing method has the problems, the embodiment of the invention provides a meteorological data verification method and a meteorological data verification system for a farmland meteorological station.
In a first aspect, an embodiment of the present invention provides a meteorological data verification method for a meteorological station of a farmland, including:
the method comprises the steps that an equipment edge node acquires environment information and farmland image information collected by a farmland weather station; wherein the environmental information comprises at least one meteorological data;
the equipment edge node obtains a verification result of each meteorological data in the environmental information through a preset quality verification model; the quality verification model is obtained by training a sample by using training environment information marked on meteorological data in advance;
the equipment edge node sends the environment information, the farmland image information and the verification results of all meteorological data to a cloud server, so that the cloud server performs correction interpolation operation on the meteorological data with abnormal verification results through a preset correction interpolation model according to the environment information and the farmland image information sent by the equipment edge node at the moment earlier than the current moment; the correction interpolation model is obtained by training meteorological data with training environment information and training farmland image information labeled in advance as samples.
Further, the equipment edge node obtains a verification result of each meteorological data in the environmental information through a preset quality verification model; the method specifically comprises the following steps:
and the equipment edge node inputs the environmental information with the collection time closest to the current moment and with the preset first quantity into the quality verification model corresponding to each meteorological data to obtain the verification result of each meteorological data.
Further, the equipment edge node sequentially inputs the environmental information with the collection time closest to the current time and in the preset first quantity into the quality verification model corresponding to each meteorological data to obtain the verification result of each meteorological data; the method specifically comprises the following steps:
the equipment edge node inputs the environmental information with the collection time closest to the current moment and with the preset first quantity into a quality verification model corresponding to each meteorological data, and obtains a verification identifier of each meteorological data according to a verification result; wherein, the check mark specifically comprises: the first identification is used for representing that the verification result of the meteorological data is normal, the second identification is used for representing that the verification result of the meteorological data is abnormal, or the third identification is used for representing that the verification result of the meteorological data is missing.
Further, the meteorological data verification method for the farmland meteorological station further comprises the following steps:
if the verification result of the meteorological data is abnormal or missing, retesting the meteorological data through the farmland meteorological station to obtain new meteorological data, and obtaining the verification result of the new meteorological data through the quality verification model again;
and if the verification result of the new meteorological data is normal, updating the environmental information according to the new meteorological data.
Further, the equipment edge node sends the environment information, the current farmland image information and the verification result of each meteorological data to a cloud server; the method specifically comprises the following steps:
the equipment edge node sends first data information and second data information containing the same preset data identification to the cloud server according to the environment information, the farmland image information and the verification identification of each meteorological data; the first data information comprises the environment information and the verification identification of each meteorological data, and the second data information comprises the farmland image information.
Further, the equipment edge node obtains a verification result of each meteorological data in the environmental information through a preset quality verification model; the method specifically comprises the following steps:
the equipment edge node obtains a verification result of each meteorological data in the environmental information through a preset quality verification model constructed based on a depth confidence network; correspondingly, the cloud server performs correction interpolation operation on the meteorological data with abnormal verification result according to the environmental information and the farmland image information which are sent by the equipment edge node at the moment earlier than the current moment through a preset correction interpolation model; the method specifically comprises the following steps:
and the cloud server performs correction interpolation operation on the meteorological data with abnormal verification results according to the environmental information and the farmland image information which are sent by the equipment edge nodes at the moment earlier than the current moment through a preset correction interpolation model which is constructed based on an integrated learning strategy and combined with a convolutional neural network and a cyclic neural network.
In a second aspect, an embodiment of the present invention further provides a meteorological data verification method for a meteorological station of a farmland, including:
the cloud server receives environment information, farmland image information and verification results of all meteorological data sent by the equipment edge nodes; the environment information and the farmland image information are acquired by the equipment edge node through a farmland meteorological station, the environment information comprises at least one meteorological data, the verification result of each meteorological data is obtained through a quality verification model preset by the equipment edge node, and the quality verification model is obtained by training a training environment information marked on the meteorological data as a sample in advance;
the cloud server performs correction interpolation operation on the meteorological data with abnormal verification result according to the environmental information and the farmland image information which are sent by the equipment edge node at the moment earlier than the current moment through a preset correction interpolation model; the correction interpolation model is obtained by training meteorological data with training environment information and training farmland image information labeled in advance as samples.
Further, the cloud server performs correction interpolation operation on meteorological data with the verification result abnormal according to environmental information and farmland image information which are sent by the equipment edge node at the moment earlier than the current moment through a preset correction interpolation model; the method specifically comprises the following steps:
and if the verification result of the meteorological data is not normal, the cloud server inputs a preset second amount of environmental information and farmland image information which are sent by the equipment edge node at the moment earlier than the current moment into a correction interpolation model corresponding to the meteorological data to obtain a prediction result corresponding to the meteorological data so as to replace the meteorological data in the environmental information.
Further, the meteorological data verification method for the farmland meteorological station further comprises the following steps:
and if the verification result of the meteorological data is judged to be wrong, marking the meteorological data, and updating the quality verification model and the correction interpolation model by taking the corresponding environmental information and farmland image information as new samples.
In a third aspect, an embodiment of the present invention further provides a weather data verification system for a meteorological station of a farmland, including:
the system comprises a cloud server, at least one equipment edge node and farmland weather stations in one-to-one correspondence with the equipment edge nodes, wherein the cloud server is connected with each equipment edge node, and each equipment edge node is connected with the corresponding farmland weather station; the cloud server is used for executing the meteorological data verification method for the farmland meteorological station, and the equipment edge node is used for executing the meteorological data verification method for the farmland meteorological station.
According to the meteorological data verification method and system for the farmland meteorological station, provided by the embodiment of the invention, each meteorological data verification result is judged through the quality verification model arranged at the edge node corresponding to the farmland meteorological station, and when the verification result is abnormal, the abnormal meteorological data is corrected by the cloud server according to the preset correction interpolation model, so that the accuracy of the acquired meteorological data is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a flow chart of a method for verifying meteorological data at a meteorological station in a farmland according to an embodiment of the invention;
FIG. 2 is a flow chart of another method for verifying weather data at a meteorological station in a farm according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a weather data verification system for a weather station of a farm according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a weather data verification system for a weather station of a farm according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 is a flowchart of a weather data verification method for a meteorological station in a farmland, according to an embodiment of the present invention, as shown in fig. 1, the method includes:
step S01, the equipment edge node acquires environmental information and farmland image information collected by a farmland weather station; wherein the environmental information comprises at least one meteorological data.
The embodiment of the invention sets corresponding equipment edge nodes for each farmland weather station, and the equipment edge nodes acquire the environment information and the farmland image information of the farmland weather stations through the farmland weather stations according to the needs. The farmland weather station is used for acquiring environmental information of the farmland weather station at a certain moment according to a plurality of preset data acquisition devices, such as sensors, and sending the environmental information to the edge nodes of the devices, wherein the environmental information comprises acquired data obtained by the data acquisition devices. The collected data in the environmental information can be specifically divided into meteorological data, soil data, time and space data, equipment information data and the like according to different data types. The meteorological data may include: air temperature (T), air humidity (H), precipitation (R), sunlight (S), atmospheric pressure (P), wind speed (U) and the like. The soil data can comprise soil temperature distribution, entropy data and the like of the soil. The temporal and spatial data may include: temporal dimensional data, e.g., year, month, day, hour, minute, second, etc., and spatial dimensional data, e.g., longitude, latitude, elevation, etc. The device information data may include: the battery power of the farmland weather station and the like.
The data acquisition process of the environmental information can be started by the equipment edge node running a preset information acquisition program regularly, the farmland meteorological station acquires current instantaneous numerical values from each data acquisition equipment, and acquired data corresponding to each data acquisition equipment is obtained through conversion of correlation coefficients and formulas.
In addition, the farmland weather station is also provided with a plurality of image acquisition devices for acquiring the farmland image information of the farmland weather station and sending the farmland image information to the equipment edge node, and the specific image acquisition process can be synchronized with the data acquisition of the data acquisition devices or preset time in advance, and can be preset in an information acquisition program by the equipment edge node. For example: after an information acquisition program of each equipment edge node is started, 4 pictures can be respectively shot by image acquisition equipment carried by a farmland weather station in 4 directions, and the sky, the canopy, the leaves and the ground soil are respectively shot, wherein the resolution of the obtained pictures is 1920 multiplied by 1080, the horizontal and vertical resolution is 96dpi, and the bit depth is 24. The farmland image information of the last 10 times can be stored in the edge node of the equipment.
S02, the equipment edge node obtains a verification result of each meteorological data in the environmental information through a preset quality verification model; the quality verification model is obtained by training a sample by using training environment information labeled on meteorological data in advance.
And the equipment edge node is preset with a quality check model used for checking the meteorological data in the environmental information so as to obtain the check result of each meteorological data in the environmental information acquired each time. The quality verification model is obtained by training pre-acquired training environment information, and the corresponding verification result of each meteorological data in the training environment information is marked. Due to the fact that the computing capacity of the equipment edge nodes is limited, the training process of the quality check model is carried out by the cloud server, and the model parameters of the quality check model are sent to the equipment edge nodes after the training is finished.
The verification result may be classified into a plurality of categories according to actual needs, or may be simply classified into a normal category and a non-normal category, which is not specifically limited herein.
Step S03, the equipment edge node sends the environment information, the farmland image information and the verification results of the meteorological data to a cloud server, so that the cloud server performs correction interpolation operation on the meteorological data with abnormal verification results through a preset correction interpolation model according to the environment information and the farmland image information sent by the equipment edge node earlier than the current moment; the correction interpolation model is obtained by training meteorological data with training environment information and training farmland image information labeled in advance as samples.
And the equipment edge node sends the acquired environmental information, the farmland image information and the verification results of all meteorological data in the environmental information to a cloud server. And the cloud server identifies the verification result of each meteorological data, if the meteorological data with the verification result not normal is determined to exist, the meteorological data is corrected and interpolated according to a preset correction interpolation model, so that the meteorological data which are not normal are corrected or interpolated, and the finally corrected or interpolated meteorological data are sent back to the equipment edge node for updating the environmental information stored by the equipment edge node, so that each meteorological data in the environmental information obtained by the equipment edge node and the cloud server can meet the normal requirement. The correction interpolation model is obtained by performing pre-training on training environment information and training farmland image information of the meteorological data which are labeled.
According to the embodiment of the invention, the quality verification model arranged at the edge node corresponding to the farmland meteorological station is used for judging the verification result of each meteorological data, and the cloud server corrects the abnormal meteorological data according to the preset correction interpolation model when the verification result is abnormal, so that the accuracy of the acquired meteorological data is improved.
Based on the foregoing embodiment, further, the step S02 specifically includes:
and S021, the equipment edge node inputs the environmental information with the collection time closest to the current time and the preset first quantity into a quality verification model corresponding to each meteorological data to obtain a verification result of each meteorological data.
In order to verify each meteorological data in the environmental information, the equipment edge nodes can directly construct a uniform quality verification model to verify each input meteorological data simultaneously, and can also construct a corresponding quality verification model for each meteorological data respectively. However, the uniform quality check model is too complex, a large number of training processes are required, and it is not easy to flexibly select the required meteorological data to be added into the environmental information according to the actual situation, so the embodiment of the invention adopts a mode of respectively constructing the quality check model for each meteorological data.
And after the equipment edge node acquires the environmental information, respectively calling a quality verification model corresponding to each meteorological data in the environmental information to respectively obtain a verification result of each meteorological data.
The input of the quality check model is specifically the environment information of a preset first quantity collected recently, including the environment information collected this time, for example: environmental information collected the last 5 or 6 times. Therefore, not only the mutual influence relationship between each meteorological data and other data in the environmental information acquired at this time is considered, but also the correlation between the continuously acquired environmental information is considered, and whether the meteorological data are normal or not can be better judged. And the output of the quality verification model is the verification result of the meteorological data.
Further, the step S021 specifically includes:
s022, inputting a preset first amount of environmental information with the collection time closest to the current time into a quality verification model corresponding to each meteorological data by an equipment edge node, and obtaining a verification identifier of each meteorological data according to a verification result; wherein, the check mark specifically comprises: the first identification is used for representing that the verification result of the meteorological data is normal, the second identification is used for representing that the verification result of the meteorological data is abnormal, or the third identification is used for representing that the verification result of the meteorological data is missing.
The verification result obtained by the quality verification model is divided into three types, namely normal, abnormal and missing, and respectively corresponds to a first identifier, a second identifier and a third identifier, for example: respectively 0, 1 and 2. At this time, the output of the quality verification model may be directly displayed with a verification identifier corresponding to the verification result.
According to the embodiment of the invention, the corresponding quality verification model is set for each meteorological data, the input of the quality verification model is the first amount of environmental information which is latest at the acquisition time, the output is the verification result of the meteorological data, and the verification result is divided into three types of normal, abnormal and missing, so that the meteorological data in the environmental information can be more accurately verified.
Based on the above embodiment, further, after S22, the weather data verification method for a meteorological station of a farm further includes:
and S023, if the verification result of the meteorological data is abnormal or missing, retesting the meteorological data through the farmland meteorological station to obtain new meteorological data, and obtaining the verification result of the new meteorological data through the quality verification model again.
After the equipment edge node obtains the verification result of the meteorological data through the quality verification model, in order to eliminate the abnormal or missing condition of the meteorological data caused by accidental factors, if the verification result of any meteorological data is abnormal or missing, namely the verification result of any meteorological data is not normal, the meteorological data can be retested again, and the retesting instruction of the meteorological data is sent to the farmland meteorological station, so that the farmland meteorological station collects new meteorological data through the data acquisition equipment corresponding to the meteorological data and sends the new meteorological data to the equipment edge node. And the equipment edge nodes obtain a new meteorological data verification result by utilizing the corresponding quality verification model again.
And S024, if the verification result of the new meteorological data is normal, updating the environmental information according to the new meteorological data.
If the verification result of the new meteorological data becomes normal, determining the new meteorological data as the meteorological data in the environmental information; and if the verification result of the new meteorological data is still abnormal, retesting can be carried out again until the verification result is normal or the number of continuous retesting reaches a preset retesting threshold value. In order to be able to give the data acquisition device time to self-repair, the successive retests need to be separated by a preset separation time, for example: 1 minute, etc.
According to the embodiment of the invention, the meteorological data with the abnormal verification result is retested, so that abnormal data caused by the fault of the data acquisition equipment due to accidental factors is eliminated, and the reliability of the acquired meteorological data is improved.
Based on the foregoing embodiment, further, the step S03 specifically includes:
step 031, the device edge node sends first data information and second data information containing the same preset data identifier to the cloud server according to the environment information, the farmland image information and the check identifier of each meteorological data; the first data information comprises the environment information and the verification identification of each meteorological data, and the second data information comprises the farmland image information.
The specific process of sending the acquired environmental information, the acquired farmland image information and the check marks of the meteorological data to the cloud server by the edge node of the equipment can be set to be a corresponding transmission mode according to a transmission protocol applied in practice. The embodiment of the invention only gives an illustration of one of the embodiments.
The device edge node randomly generates a unique corresponding data identifier for each acquisition process, for example: the cloud server comprises a Global Unique Identifier (GUID), and information to be sent is divided into first data information and second data information, and the first data information and the second data information both comprise the data identification, so that the cloud server can associate the first data information and the second data information according to the data identification after receiving the first data information and the second data information. Due to the particularity of the image information, the environment information and the verification identification of each meteorological data are divided into first data information, and the farmland image information is divided into second data information.
The first data information may be sent in the form of a data sequence after splicing environmental information and a check identifier of weather data, and a data protocol format of the data sequence includes:
data identification, transmission protocol number, farmland weather station equipment number, date and time, air temperature, air humidity, daily rainfall, maximum wind speed, minimum wind speed, average wind speed, wind direction, radiation, ultraviolet rays, hour ET, daily accumulation ET, effective rainfall, accumulated effective rainfall, battery voltage, NC, address information 1, address information 2, quality verification model version, air temperature verification identification, air humidity verification identification, daily rainfall verification identification, maximum wind speed verification identification, minimum wind speed verification identification, average wind speed verification identification, wind direction verification identification, radiation verification identification, ultraviolet verification identification, hour ET verification identification, daily accumulation ET verification identification, effective rainfall verification identification, and accumulated effective rainfall verification identification.
And the second data information is sent after all the farmland image information collected this time is packaged.
The equipment edge node and the cloud server can adopt a 4G/5G network to carry out data transmission by a TCP/IP protocol. The cloud server analyzes the received first data information and the second data information according to a preset transmission protocol, so that environment information, the check marks of the meteorological data and farmland image information are obtained.
According to the embodiment of the invention, the environment information, the farmland image information and the check marks of the meteorological data are divided into the first data information and the second data information, the first data information and the second data information are sent to the cloud server, and the correlation is carried out according to the data marks which are distributed in advance and uniquely correspond to the current acquisition process, so that the reliability of data transmission from the equipment edge node to the cloud server is improved.
Based on the above embodiment, further, the equipment edge node obtains a verification result of each meteorological data in the environmental information through a preset quality verification model; the method specifically comprises the following steps:
and the equipment edge node obtains the verification result of each meteorological data in the environmental information through a preset quality verification model constructed based on the deep belief network.
There are many methods that can be used to construct the quality check model of the device edge node, and the embodiment of the present invention is exemplified by constructing based on a deep belief network DBN. The basic unit is composed of a double-layer network unit structure which comprises a visible unit v ═ {0, 1}dAnd hidden unit h ═ {0, 1}LThe mathematical expression of the combination cooperative operation of the units is as follows:
Figure BDA0002206397480000111
wherein the content of the first and second substances,θ＝{bi，aj，wij}，wijas a weight between visible cell i and hidden cell j, biAnd ajThe bias values for the visible layer and the hidden layer, respectively. In the base unit, the hidden unit is conditionally independent, and unbiased samples can be obtained from the posterior distribution given the visible unit data vector. And sequentially overlapping all the basic units to construct a DBN network, and adding a Logistic regression layer at the tail of the network as a classifier.
Correspondingly, the cloud server performs correction interpolation operation on the meteorological data with abnormal verification result according to the environmental information and the farmland image information which are sent by the equipment edge node at the moment earlier than the current moment through a preset correction interpolation model; the method specifically comprises the following steps:
and the cloud server performs correction interpolation operation on the meteorological data with abnormal verification results according to the environmental information and the farmland image information which are sent by the equipment edge nodes at the moment earlier than the current moment through a preset correction interpolation model which is constructed based on an integrated learning strategy and combined with a convolutional neural network and a cyclic neural network.
The cloud server can be used for constructing a correction interpolation model, and the embodiment of the invention is exemplified by constructing the correction interpolation model based on an ensemble learning strategy and combining a convolutional neural network and a cyclic neural network.
The correction interpolation model adopts a deep learning model, comprises a convolutional neural network CNN for processing images and a cyclic neural network RNN for processing time sequence data, and realizes data correction and interpolation by integrating and learning high-dimensional feature vectors acquired by the images and the time sequence to realize model building.
The scheme provides that a convolutional neural network CNN is adopted to process image data, each image comprises R, G, B three channels, and high-latitude characteristic information implied by different spectrums in the image is provided through convolution and pooling operations. The convolutional neural network utilizes local connections to extract 2-D spatial features of image context, and is composed of a group of alternate convolutional pooling operations, and network parameters are reduced through a weight sharing mechanism. The pooling layer compresses the feature map size created by the convolutional layer to obtain more general and abstract features. And finally, outputting the feature map by further converting the feature map into a feature vector. The convolutional neural network comprises three network structure layers of a convolutional layer, a pooling layer and a full-connection layer.
The convolutional layer is passed through a convolution kernel to generate a multi-dimensional feature map. Specifically, let X be an input cube, whose size is m × n × d, where m × n represents the spatial dimension of X, d is the number of channels, and Xi is the ith feature map of X. Assuming there are k filters, the convolutional layer and the jth filter can be represented by a weight wiAnd bias bjAnd (4) defining. The jth output of a convolutional layer can be expressed as follows:
Figure BDA0002206397480000121
where is the convolution operator and f (·) is the activation function.
The pooling layer is used to reduce redundant information after multiple convolution operations. Specifically, for a p × p window size pooling kernel defined as S, the mathematical expression for average pooling is:
Figure BDA0002206397480000122
wherein F is the number of elements in S and xijIs the activation value for the corresponding location (i, j).
The full-connection layer obtains depth and abstract features by transforming the feature map into an n-dimensional feature map, and inputs the feature map output by the pooling layer into the full-connection layer after flattening operation, wherein the mathematical expression of the full-connection layer is as follows:
Figure BDA0002206397480000123
where X ', Y', W, and b represent the input, output, weight, and offset, respectively, of the fully connected layer.
The embodiment of the invention provides a method for processing time sequence data by using a Recurrent Neural Network (RNN), wherein data continuously acquired by an edge node of equipment can be regarded as a series of time sequence data with fixed intervals, a sequence data mode and dynamic time sequence characteristics are identified by the RNN, and the method is characterized in that each step is in a relation with the previous step through a recursive hidden state, and the acquired characteristics are used for correcting and interpolating missing data.
Let x be x1，x2…xnIs time series data, where xiIs the data of the ith time step. Recursive hidden state h of RNN at time series t<t>Can be updated, and the mathematical expression of the update rule is as follows:
h<t>＝f1(wx<t>+uh<t-1>+bh)
where w and u represent the input of the current step and the activation coefficient matrix of the preceding recursive hidden unit, respectively, bhRepresenting the corresponding offset vector.
h<t>Will be used to predict y in time step t<t>The mathematical expression is as follows:
y<t>＝f2(ph<t>+by)
in the formula f2Is a non-linear function, p is the coefficient matrix after the activation of the recursive hidden unit in the current step, byAre the corresponding offset vectors.
The correction interpolation model constructed in the embodiment of the invention realizes the functions of abnormal data correction and missing data interpolation by fusing the image characteristic information extracted by the CNN and the time series data characteristic information extracted by the RNN. The adopted integrated learning fusion strategy is a superposition mode, the constructed CNN and RNN networks are flattened through a Flatten layer and then spliced and input to a fully-connected neural network for learning, and the output of the fully-connected neural network is the network output prediction result.
And if the cloud server receives the environment information, the farmland image information and the check marks of the meteorological data sent by any equipment edge node, performing correction interpolation operation on the meteorological data with the check marks 1 and 2 through the correction interpolation model, recording a correction interpolation result value, correction interpolation time and a model version number in a database after operation, and simultaneously recording an operation log.
According to the embodiment of the invention, through the quality verification model established based on the deep belief network and the correction interpolation model established based on the integrated learning strategy and the convolution neural network and the cyclic neural network, the verification and correction interpolation operation can be more accurately carried out on the meteorological data, and the accuracy and the reliability of the meteorological data are improved.
Fig. 2 is a flowchart of another weather data verification method for a meteorological station in a farm, according to an embodiment of the present invention, as shown in fig. 2, the method includes:
step S11, the cloud server receives the environment information, the farmland image information and the verification results of the meteorological data sent by the equipment edge nodes; the environment information and the farmland image information are acquired by the equipment edge nodes through a farmland meteorological station, the environment information comprises at least one meteorological data, the verification result of each meteorological data is acquired through a quality verification model preset by the equipment edge nodes, and the quality verification model is acquired by training the meteorological data with the marked training environment information as a sample in advance.
The cloud server periodically receives the environment information and the farmland image information sent by the edge nodes of the equipment and the verification results of the meteorological data contained in the environment information.
The environment information and the farmland image information are acquired by the equipment edge node through a plurality of data acquisition equipment and image acquisition equipment which are preset in the farmland weather station, and a specific acquisition process can be started by the equipment edge node periodically running a preset information acquisition program and then sends an acquisition instruction to the farmland weather station.
The environmental information includes collected data obtained by each data collection device, and the collected data may be specifically classified into meteorological data, soil data, time and space data, device information data, and the like, according to the type of the data.
The image collected by the farmland image information can be specifically set according to actual needs, such as: when image acquisition is carried out every time, 4 pictures can be respectively shot by image acquisition equipment carried by a farmland weather station in 4 directions, and sky, canopy, leaves and ground soil are respectively shot, wherein the resolution of the obtained pictures is 1920 multiplied by 1080, the horizontal and vertical resolution is 96dpi, and the bit depth is 24. The farmland image information of the last 10 times can be stored in the image acquisition equipment.
And the verification result of the meteorological data is obtained by presetting a quality verification model by the equipment edge node and verifying the meteorological data in the acquired environmental information. The quality verification model is obtained by training pre-acquired training environment information, and the corresponding verification result of each meteorological data in the training environment information is marked. Due to the fact that the computing capacity of the equipment edge nodes is limited, the training process of the quality check model is carried out by the cloud server, and the model parameters of the quality check model are sent to the equipment edge nodes after the training is finished. The verification result may be divided into a plurality of categories according to actual needs, or simply scored as normal or abnormal, and is not specifically limited herein.
And the equipment edge node sends the acquired environmental information, the farmland image information and the verification results of all meteorological data in the environmental information to a cloud server.
Step S12, the cloud server performs correction interpolation operation on meteorological data with the verification result abnormal according to environmental information and farmland image information sent by the equipment edge node at the moment earlier than the current moment through a preset correction interpolation model; the correction interpolation model is obtained by training meteorological data with training environment information and training farmland image information labeled in advance as samples.
The cloud server identifies the verification result of each meteorological data, if the meteorological data with the verification result not normal is determined to exist, the meteorological data are corrected and interpolated according to a preset correction interpolation model, so that the meteorological data which are not normal are corrected or interpolated, and the finally corrected or interpolated meteorological data are sent back to the equipment edge node to update the environmental information stored by the equipment edge node, so that the meteorological data in the environmental information obtained by the equipment edge node and the cloud server meet the normal requirement. The correction interpolation model is obtained by performing pre-training on training environment information and training farmland image information of the meteorological data which are labeled.
According to the embodiment of the invention, the quality verification model arranged at the edge node corresponding to the farmland meteorological station is used for judging the verification result of each meteorological data, and the cloud server corrects the abnormal meteorological data according to the preset correction interpolation model when the verification result is abnormal, so that the accuracy of the acquired meteorological data is improved.
Based on the foregoing embodiment, further, the step S12 specifically includes:
step S121, if the verification result of the meteorological data is not normal, the cloud server inputs a preset second amount of environmental information and farmland image information, which are sent by the device edge node earlier than the current time, into a correction interpolation model corresponding to the meteorological data, and obtains a prediction result corresponding to the meteorological data to replace the meteorological data in the environmental information.
When the cloud server receives environment information, meteorological data verification results and farmland image information sent by the edge nodes of the equipment, if the verification result of any meteorological data is determined to be abnormal or missing, namely the corresponding verification identifier is 1 or 2, the fact that the meteorological data needs to be corrected and interpolated is judged. And calling a correction interpolation model corresponding to the meteorological data, and extracting a preset second amount of environment information and farmland image information which are sent by the equipment edge node earlier than the current moment from a database, such as: and 7 pieces of environment information and farmland image information sent earlier than the environment information and farmland image information collected this time. And inputting the extracted second amount of environmental information and farmland image information into the correction interpolation model, wherein the correction interpolation model outputs a prediction result of the meteorological data, and the prediction result is used as the meteorological data to update the acquired environmental information.
According to the embodiment of the invention, the correction interpolation model which is arranged on the cloud server and corresponds to each meteorological data is used for performing correction interpolation operation on the meteorological data of which the verification result is abnormal, so that the reliability and the accuracy of the meteorological data are ensured.
Based on the above embodiment, further, the weather data verification method for the farmland weather station further includes:
and if the verification result of the meteorological data is judged to be wrong, marking the meteorological data, and updating the quality verification model and the correction interpolation model by taking the corresponding environmental information and farmland image information as new samples.
In order to ensure that the quality check model and the correction interpolation model are suitable for the current application environment in real time and have better accuracy and generalization capability, the cloud server can periodically optimize and update the quality check model and the correction interpolation model by adding samples for training.
The added samples can be derived from the judgment of the verification result of the meteorological data, if the verification result obtained by the quality verification model is judged to be inconsistent with the actual verification result, the verification result of the meteorological data is marked again, and the corresponding environmental information and the farmland image information are used as new samples. In addition, the newly added samples can also be derived from external training samples.
After the quality verification model and the correction interpolation model are updated regularly, the cloud server needs to record the model version after each update, and timely sends the model parameters of the updated quality verification model to the equipment edge nodes, so that the quality verification model of the equipment edge nodes is synchronously optimized.
The generalization ability of model training is a core index considering the model ability, and when the model training is stopped in the model training determines the generalization performance of the model. The method for terminating the model training in the model training of the scheme comprises the following steps: model training is terminated when the validation set error grows for a number of consecutive cycles.
When the model is trained, 80% of a data set is used as a training set, 20% of the data set is used as a verification set, and the updated model is measured by adopting a root mean square error MAE, a mean square error MSE, a mean absolute error MRSE and the like:
Figure BDA0002206397480000161
Figure BDA0002206397480000162
Figure BDA0002206397480000171
Figure BDA0002206397480000172
wherein the content of the first and second substances,to predict value, yiIn order to be the true value of the value,
Figure BDA0002206397480000174
are averages.
When the accuracy rate of the verification result of the quality verification model training set is more than 90%, the verification result of the correction interpolation model meets the requirements of MAE, MSE and RMSE which are all less than 0.2 and R2>The model takes effect at 0.90.
According to the embodiment of the invention, the timeliness, the accuracy and the migration generalization performance of the model are enhanced by updating and optimizing the quality check model and the correction interpolation model.
Fig. 3 is a schematic structural diagram of a weather data verification system for a meteorological station in an embodiment of the present invention, and as shown in fig. 3, the weather data verification system for the meteorological station in the farmland includes: the system comprises a cloud server 10, at least one equipment edge node 11 and farmland weather stations 12 corresponding to the equipment edge nodes 11 one by one; wherein the content of the first and second substances,
the cloud server 10 is connected with each equipment edge node 11, and each equipment edge node 11 is connected with a corresponding farmland weather station 12; the cloud server 10 and the device edge node 11 are configured to execute the weather data verification method for the meteorological station in the farmland according to the embodiment. Specifically, the method comprises the following steps:
the equipment edge node 11 acquires the environment information and farmland image information of the farmland weather station 12 through the connected farmland weather station 12 according to actual needs. The farmland weather station 12 is preset with a plurality of data acquisition devices for acquiring environmental information of the farmland weather station 12 at a certain moment and sending the environmental information to the equipment edge node 11, wherein the environmental information comprises acquired data obtained by each data acquisition device. The collected data in the environmental information can be specifically divided into meteorological data, soil data, time and space data, equipment information data and the like according to different data types. The data acquisition process of the environmental information may be specifically started by the device edge node 11 periodically running a preset information acquisition program, acquiring a current instantaneous value from each data acquisition device, and obtaining acquisition data corresponding to each data acquisition device through conversion of a correlation coefficient and a formula.
In addition, the farmland weather station 12 is further provided with a plurality of image acquisition devices for acquiring the farmland image information of the farmland weather station and sending the farmland image information to the device edge node 11, and a specific image acquisition process can be synchronized with the data acquisition of the data acquisition devices or preset in advance for a preset time, and the device edge node 11 can be preset in the information acquisition program.
The device edge node 11 is preset with a quality check model for checking the meteorological data in the environmental information to obtain a check result of each meteorological data in the environmental information acquired each time. The quality verification model is obtained by training pre-acquired training environment information, and the corresponding verification result of each meteorological data in the training environment information is marked. Due to the limited computing capability of the device edge nodes 11, the training process of the quality check model is performed by the cloud server 10, and the model parameters of the quality check model are sent to the device edge nodes 11 after the training is completed.
The verification result may be divided into a plurality of categories according to actual needs, or simply scored as normal or abnormal, and is not specifically limited herein.
The device edge node 11 sends the collected environmental information, the farmland image information, and the verification results of the meteorological data in the environmental information to the cloud server 10. The cloud server 10 identifies the verification result of each meteorological data, and if it is determined that there is meteorological data whose verification result is not normal, performs correction interpolation operation on the meteorological data according to a preset correction interpolation model, so as to correct the meteorological data which is not normal, and sends the finally corrected meteorological data back to the equipment edge node 11 for updating the environmental information stored by the equipment edge node 11, so as to ensure that each meteorological data in the environmental information obtained by the equipment edge node 11 and the cloud server 10 at last meets the normal requirement. The correction interpolation model is obtained by performing pre-training on training environment information and training farmland image information of the meteorological data which are labeled.
The system provided in the embodiment of the present invention is configured to execute the method, and the functions of the system are specifically referred to the method embodiment, and the specific method flow is not described herein again.
According to the embodiment of the invention, the quality verification model arranged at the edge node corresponding to the farmland meteorological station is used for judging the verification result of each meteorological data, and the cloud server corrects the abnormal meteorological data according to the preset correction interpolation model when the verification result is abnormal, so that the accuracy of the acquired meteorological data is improved.
FIG. 4 is a schematic diagram of a weather data verification system for a weather station of a farm according to an embodiment of the present invention; as shown in fig. 4, an equipment edge node connected to each of the meteorological stations of the farmland is provided, and the equipment edge node performs a verification process on the meteorological data, and is divided into an equipment data acquisition unit, a data quality verification unit, and a data transmission unit.
The equipment data acquisition unit executes a preset information acquisition program periodically to instruct the farmland meteorological station to send acquired environmental information and farmland image information to the equipment edge node and store the environmental information and the farmland image information in the database. The environment information includes at least: weather data and equipment information data, the weather data may include: air temperature (T), air humidity (H), precipitation (R), sunlight (S), atmospheric pressure (P), wind speed (U) and the like. The device information data includes: battery voltage, latitude and longitude, time, address information, equipment number, protocol number and the like. The farmland image information comprises four images of sky, canopy, leaf and soil.
The data quality checking unit is provided with a quality checking model, and when the data quality checking unit judges that newly-added environmental information occurs in the process of scanning the database, the checking result of the meteorological data in the newly-added environmental information is obtained through the quality checking model. And respectively marking the result as 0, 1 and 2 according to the check result, and recording the log. And if the verification result of the meteorological data is not normal, sending a retest instruction to the equipment data acquisition unit, acquiring the meteorological data again, and verifying the meteorological data again.
And after the data quality verification unit finishes the verification process, triggering a data sending unit, performing data fusion on the environment information, the verification identification of each meteorological data, the farmland image information and the distributed GUID to obtain multidimensional data, and transmitting the multidimensional data to a data receiving unit of the cloud server through a 4G/5G network.
The cloud server comprises a data receiving unit, a data correction interpolation unit and a model updating unit.
The data receiving unit carries out protocol analysis verification on the received multidimensional data, if the received multidimensional data passes the protocol analysis verification, the received multidimensional data is stored in a database, if the received multidimensional data does not pass the protocol analysis verification, the received multidimensional data is archived and backed up, and meanwhile, logs are recorded.
The data correction interpolation unit presets a correction interpolation model and determines whether new multidimensional data exist in the timing scanning database. If the abnormal meteorological data exists, the data of the meteorological data with abnormal verification results are corrected through the correction interpolation model, the data of the meteorological data with missing verification results are complemented, and therefore correct meteorological data are obtained, and meanwhile the correction interpolation results are recorded in a log.
The model updating unit is used for updating the quality verification model and the correction interpolation model, performing manual marking on newly added data to serve as new samples for performing model training on the quality verification model and the correction interpolation model, evaluating the models through a preset model precision evaluation method in the training process, performing model deployment on the passed models, performing archive backup on the failed models, not performing model deployment, and recording logs.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims (10)

1. A meteorological data verification method for a farmland meteorological station is characterized by comprising the following steps:
the method comprises the steps that an equipment edge node acquires environment information and farmland image information collected by a farmland weather station; wherein the environmental information comprises at least one meteorological data;
the equipment edge node obtains a verification result of each meteorological data in the environmental information through a preset quality verification model; the quality verification model is obtained by training a sample by using training environment information marked on meteorological data in advance;
the equipment edge node sends the environment information, the farmland image information and the verification results of all meteorological data to a cloud server, so that the cloud server performs correction interpolation operation on the meteorological data with abnormal verification results through a preset correction interpolation model according to the environment information and the farmland image information sent by the equipment edge node at the moment earlier than the current moment; the correction interpolation model is obtained by training meteorological data with training environment information and training farmland image information labeled in advance as samples.
2. The weather data verification method for the farmland weather station as claimed in claim 1, wherein the equipment edge node obtains the verification result of each weather data in the environmental information through a preset quality verification model; the method specifically comprises the following steps:
and the equipment edge node inputs the environmental information with the collection time closest to the current moment and with the preset first quantity into the quality verification model corresponding to each meteorological data to obtain the verification result of each meteorological data.
3. The weather data verification method for the farmland weather station as claimed in claim 2, wherein the equipment edge node sequentially inputs the environmental information of the preset first quantity whose collection time is closest to the current time into the quality verification model corresponding to each weather data to obtain the verification result of each weather data; the method specifically comprises the following steps:
the equipment edge node inputs the environmental information with the collection time closest to the current moment and with the preset first quantity into a quality verification model corresponding to each meteorological data, and obtains a verification identifier of each meteorological data according to a verification result; wherein, the check mark specifically comprises: the first identification is used for representing that the verification result of the meteorological data is normal, the second identification is used for representing that the verification result of the meteorological data is abnormal, or the third identification is used for representing that the verification result of the meteorological data is missing.
4. The weather data verification method for the meteorological station of a farmland, further comprising:
if the verification result of the meteorological data is abnormal or missing, retesting the meteorological data through the farmland meteorological station to obtain new meteorological data, and obtaining the verification result of the new meteorological data through the quality verification model again;
and if the verification result of the new meteorological data is normal, updating the environmental information according to the new meteorological data.
5. The weather data verification method for the farmland weather station as claimed in claim 3, wherein the equipment edge node transmits the environment information, the current farmland image information and the verification results of the weather data to a cloud server; the method specifically comprises the following steps:
the equipment edge node sends first data information and second data information containing the same preset data identification to the cloud server according to the environment information, the farmland image information and the verification identification of each meteorological data; the first data information comprises the environment information and the verification identification of each meteorological data, and the second data information comprises the farmland image information.
6. The weather data verification method for the farmland weather station as claimed in claim 1, wherein the equipment edge node obtains the verification result of each weather data in the environmental information through a preset quality verification model; the method specifically comprises the following steps:
the equipment edge node obtains a verification result of each meteorological data in the environmental information through a preset quality verification model constructed based on a depth confidence network; correspondingly, the cloud server performs correction interpolation operation on the meteorological data with abnormal verification result according to the environmental information and the farmland image information which are sent by the equipment edge node at the moment earlier than the current moment through a preset correction interpolation model; the method specifically comprises the following steps:
and the cloud server performs correction interpolation operation on the meteorological data with abnormal verification results according to the environmental information and the farmland image information which are sent by the equipment edge nodes at the moment earlier than the current moment through a preset correction interpolation model which is constructed based on an integrated learning strategy and combined with a convolutional neural network and a cyclic neural network.
7. A meteorological data verification method for a farmland meteorological station is characterized by comprising the following steps:
the cloud server receives environment information, farmland image information and verification results of all meteorological data sent by the equipment edge nodes; the environment information and the farmland image information are acquired by the equipment edge node through a farmland meteorological station, the environment information comprises at least one meteorological data, the verification result of each meteorological data is obtained through a quality verification model preset by the equipment edge node, and the quality verification model is obtained by training a training environment information marked on the meteorological data as a sample in advance;
the cloud server performs correction interpolation operation on the meteorological data with abnormal verification result according to the environmental information and the farmland image information which are sent by the equipment edge node at the moment earlier than the current moment through a preset correction interpolation model; the correction interpolation model is obtained by training meteorological data with training environment information and training farmland image information labeled in advance as samples.
8. The weather data verification method for the farmland weather station as claimed in claim 7, wherein the cloud server performs a correction interpolation operation on weather data, the verification result of which is not normal, according to the environmental information and the farmland image information sent by the equipment edge node earlier than the current time through a preset correction interpolation model; the method specifically comprises the following steps:
and if the verification result of the meteorological data is not normal, the cloud server inputs a preset second amount of environmental information and farmland image information which are sent by the equipment edge node earlier than the current moment into a correction interpolation model corresponding to the meteorological data to obtain a prediction result corresponding to the meteorological data so as to replace the meteorological data in the environmental information.
9. The weather data verification method for the farmland weather station as claimed in claim 7 or 8, wherein the weather data verification method for the farmland weather station further comprises:
and if the verification result of the meteorological data is judged to be wrong, marking the meteorological data, and updating the quality verification model and the correction interpolation model by taking the corresponding environmental information and farmland image information as new samples.
10. A weather data verification system for a weather station in a farm, comprising:
the system comprises a cloud server, at least one equipment edge node and farmland weather stations in one-to-one correspondence with the equipment edge nodes, wherein the cloud server is connected with each equipment edge node, and each equipment edge node is connected with the corresponding farmland weather station; the cloud server is used for executing the weather data verification method for the meteorological station of the farmland according to any one of claims 7 to 9, and the device edge node is used for executing the weather data verification method for the meteorological station of the farmland according to any one of claims 1 to 6.
CN201910882810.7A 2019-09-18 2019-09-18 Meteorological data verification method and system for farmland meteorological station Active CN110687619B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910882810.7A CN110687619B (en) 2019-09-18 2019-09-18 Meteorological data verification method and system for farmland meteorological station

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910882810.7A CN110687619B (en) 2019-09-18 2019-09-18 Meteorological data verification method and system for farmland meteorological station

Publications (2)

Publication Number Publication Date
CN110687619A true CN110687619A (en) 2020-01-14
CN110687619B CN110687619B (en) 2021-09-03

Family

ID=69109459

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910882810.7A Active CN110687619B (en) 2019-09-18 2019-09-18 Meteorological data verification method and system for farmland meteorological station

Country Status (1)

Country Link
CN (1) CN110687619B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111830595A (en) * 2020-06-09 2020-10-27 上海眼控科技股份有限公司 Meteorological element prediction method and equipment
WO2021208326A1 (en) * 2020-04-17 2021-10-21 中国科学院上海微系统与信息技术研究所 Fog computing and blockchain platform-based distributed weather station system and operating method
CN114325875A (en) * 2021-12-24 2022-04-12 中国气象局气象探测中心 Weather station equipment control method and equipment
WO2022139683A1 (en) * 2020-12-23 2022-06-30 National University Of Singapore Edge computing based face and gesture recognition

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104280791A (en) * 2014-09-26 2015-01-14 宁波绮耘软件有限公司 Meteorological data quality control and processing method and system
US20170223509A1 (en) * 2016-06-17 2017-08-03 Nextnav, Llc Characterizing and using building weather profile parameters
JP2018170971A (en) * 2017-03-31 2018-11-08 沖電気工業株式会社 Temperature prediction device, temperature prediction method, temperature prediction program and temperature prediction system
CN108802856A (en) * 2018-03-30 2018-11-13 南京大学 A kind of source data dynamic corrections forecast system and its working method based on AI
CN109596179A (en) * 2019-01-14 2019-04-09 沈阳农业大学 Wisdom agricultural system based on wireless sensor network

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104280791A (en) * 2014-09-26 2015-01-14 宁波绮耘软件有限公司 Meteorological data quality control and processing method and system
US20170223509A1 (en) * 2016-06-17 2017-08-03 Nextnav, Llc Characterizing and using building weather profile parameters
JP2018170971A (en) * 2017-03-31 2018-11-08 沖電気工業株式会社 Temperature prediction device, temperature prediction method, temperature prediction program and temperature prediction system
CN108802856A (en) * 2018-03-30 2018-11-13 南京大学 A kind of source data dynamic corrections forecast system and its working method based on AI
CN109596179A (en) * 2019-01-14 2019-04-09 沈阳农业大学 Wisdom agricultural system based on wireless sensor network

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021208326A1 (en) * 2020-04-17 2021-10-21 中国科学院上海微系统与信息技术研究所 Fog computing and blockchain platform-based distributed weather station system and operating method
CN111830595A (en) * 2020-06-09 2020-10-27 上海眼控科技股份有限公司 Meteorological element prediction method and equipment
WO2022139683A1 (en) * 2020-12-23 2022-06-30 National University Of Singapore Edge computing based face and gesture recognition
CN114325875A (en) * 2021-12-24 2022-04-12 中国气象局气象探测中心 Weather station equipment control method and equipment

Also Published As

Publication number Publication date
CN110687619B (en) 2021-09-03

Similar Documents

Publication Publication Date Title
CN110687619B (en) Meteorological data verification method and system for farmland meteorological station
CN109214592B (en) Multi-model-fused deep learning air quality prediction method
CN110456026B (en) Soil moisture content monitoring method and device
CN112232554A (en) Construction method of local short-term rainfall forecast model based on BP neural network
CN113919231A (en) PM2.5 concentration space-time change prediction method and system based on space-time diagram neural network
CN112287294A (en) Time-space bidirectional soil water content interpolation method based on deep learning
CN110896761B (en) Irrigation decision-making method and system for greenhouse
CN115267945A (en) Thunder and lightning early warning method and system based on graph neural network
CN115345076A (en) Wind speed correction processing method and device
CN113984969A (en) Air quality prediction method and system based on multi-source space-time data fusion
CN116415730A (en) Fusion self-attention mechanism time-space deep learning model for predicting water level
CN113610286A (en) PM accounting for spatio-temporal correlations and meteorological factors2.5Concentration prediction method and device
CN109543911B (en) Sunlight radiation prediction method and system
CN116720156A (en) Weather element forecasting method based on graph neural network multi-mode weather data fusion
CN117008217A (en) Short-time strong precipitation prediction method integrating self-attention module and Unet model
CN114118634B (en) Soil moisture prediction method and device
CN116245008A (en) Dynamic estimation method of drainage basin hydrologic model parameters based on digital twin
CN117079130B (en) Intelligent information management method and system based on mangrove habitat
CN116527135B (en) Site availability determination method, system, device, electronic equipment and storage medium
CN110826810B (en) Regional rainfall prediction method combining spatial reasoning and machine learning
CN116680651A (en) Aquaculture monitoring method and system based on multi-source data fusion
CN113988351A (en) Numerical mode air temperature forecast product release method based on space-time local model
CN115080629B (en) Meteorological element forecasting system and method based on remote sensing technology
CN115238967A (en) Photovoltaic power prediction method and device combining cloud picture and adjacent power station cluster
CN114926744A (en) Local cloud and mist prediction method and system

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