CN111339237A - Farm risk prediction method, device, equipment and storage medium - Google Patents
Farm risk prediction method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN111339237A CN111339237A CN202010130767.1A CN202010130767A CN111339237A CN 111339237 A CN111339237 A CN 111339237A CN 202010130767 A CN202010130767 A CN 202010130767A CN 111339237 A CN111339237 A CN 111339237A
- Authority
- CN
- China
- Prior art keywords
- farm
- risk prediction
- information
- model
- map
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 63
- 238000003860 storage Methods 0.000 title claims abstract description 14
- 238000013058 risk prediction model Methods 0.000 claims abstract description 55
- 238000001514 detection method Methods 0.000 claims abstract description 48
- 238000009395 breeding Methods 0.000 claims abstract description 47
- 230000001488 breeding effect Effects 0.000 claims abstract description 47
- 238000012549 training Methods 0.000 claims description 125
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 27
- 238000013434 data augmentation Methods 0.000 claims description 14
- 238000004422 calculation algorithm Methods 0.000 claims description 11
- 230000003416 augmentation Effects 0.000 claims description 10
- 230000003190 augmentative effect Effects 0.000 claims description 6
- 238000007689 inspection Methods 0.000 claims 2
- 238000009360 aquaculture Methods 0.000 abstract 1
- 244000144974 aquaculture Species 0.000 abstract 1
- 230000002265 prevention Effects 0.000 description 17
- 230000008569 process Effects 0.000 description 15
- 241001465754 Metazoa Species 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 8
- 238000004891 communication Methods 0.000 description 6
- 208000007407 African swine fever Diseases 0.000 description 4
- 208000035473 Communicable disease Diseases 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 208000015181 infectious disease Diseases 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 3
- 238000003062 neural network model Methods 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 241000282887 Suidae Species 0.000 description 2
- 241000282898 Sus scrofa Species 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000002360 preparation method Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012502 risk assessment Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 241000701386 African swine fever virus Species 0.000 description 1
- 206010037660 Pyrexia Diseases 0.000 description 1
- 230000001154 acute effect Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000002458 infectious effect Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 231100000225 lethality Toxicity 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 230000003245 working effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
-
- 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
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- 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—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Resources & Organizations (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Software Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Databases & Information Systems (AREA)
- Entrepreneurship & Innovation (AREA)
- General Business, Economics & Management (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Agronomy & Crop Science (AREA)
- Primary Health Care (AREA)
- Mining & Mineral Resources (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Marine Sciences & Fisheries (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
Abstract
The invention discloses a risk prediction method, a risk prediction device, risk prediction equipment and a storage medium for an aquaculture farm, wherein the method comprises the following steps: inputting a map to be detected into a farm detection model to obtain farm information in the map to be detected; determining a farm to be predicted and culture information in a preset range around the farm to be predicted according to the farm information, and acquiring other influence factor information in the preset range; and inputting the breeding information and the other influence factor information into a preset risk prediction model to obtain a risk prediction result of the farm to be predicted. The method and the system realize rapid and accurate prediction of the risk condition of the farm.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a method, a device, equipment and a storage medium for predicting risk of a farm.
Background
African swine fever is an acute, febrile and highly contagious animal infectious disease of pigs caused by African swine fever virus, and the related lethality rate is up to 100%. Since the first African swine fever in China is confirmed to be diagnosed by the Chinese animal health and epidemiology center, the African swine fever epidemic situation occurs in part of China and the loss is serious. Once an infectious disease of an animal, which is similar to the African swine fever, is discovered, a prevention and control measure needs to be taken quickly and accurately to avoid the negative influence on all aspects of society caused by the continuous expansion of epidemic situations. In order to take rapid and accurate preventive and control measures, the risk condition of each farm needs to be known rapidly and accurately. However, the current situation is risk assessment for the farm, and the assessment needs to be performed after an expert inspects the situation of the farm on the spot, or depends on the reported information of the farmers. However, the scheme of field investigation by experts is obviously low in efficiency and difficult to deal with the outburst and rapidly spread epidemic situation; the situation that the breeding information is concealed by the breeding households can happen when the breeding households actively report, so that the risk assessment accuracy is low. The risk condition of the farm cannot be quickly and accurately obtained, and the prevention and control working effect is poor.
Disclosure of Invention
The invention mainly aims to provide a risk prediction method, a device, equipment and a storage medium for a farm, aiming at solving the problem that the prevention and control work effect is poor due to the fact that the risk condition of the farm cannot be rapidly and accurately obtained aiming at the epidemic situation of animals at present.
In order to achieve the above object, the present invention provides a farm risk prediction method, including the steps of:
inputting a map to be detected into a farm detection model to obtain farm information in the map to be detected;
determining a farm to be predicted according to the farm information, determining culture information in a preset range around the farm to be predicted, and acquiring other influence factor information except the farm in the preset range;
and inputting the breeding information and the other influence factor information into a preset risk prediction model to obtain a risk prediction result of the farm to be predicted.
Optionally, the step of inputting the map to be detected into a farm detection model to obtain farm information in the map to be detected includes:
inputting the map to be detected into the farm detection model to obtain the classification category of each pixel point in the map to be detected, wherein the classification category is used for indicating whether the corresponding pixel point belongs to the farm;
and determining the farm information in the map to be detected according to the classification type.
Optionally, the step of determining the farm information in the map to be detected according to the classification category includes:
determining target pixel points belonging to a farm in the map to be detected according to the classification category;
clustering the target pixel points by adopting a preset clustering algorithm to obtain each cluster group;
and determining the farm information in the map to be detected according to each cluster group.
Optionally, before the step of inputting the map to be detected into the farm detection model to obtain the farm information in the map to be detected, the method further includes:
performing primary training on a model to be trained by adopting the formal training data, wherein the formal training data comprises a plurality of pre-collected remote sensing satellite maps containing farms and farm marking data corresponding to each remote sensing satellite map;
adjusting the model to be trained after the initial training by adopting negative example training data or adopting the negative example training data and the positive example training data, wherein the negative example training data comprise a plurality of remote sensing satellite maps which are collected in advance and do not comprise a farm;
when the adjusted model to be trained is detected to meet the preset model condition, taking the adjusted model to be trained as the farm detection model, otherwise, executing the steps again based on the adjusted model to be trained: and (5) performing primary training on the model to be trained by adopting the normal training data.
Optionally, the step of performing preliminary training on the model to be trained by using the proper training data includes:
carrying out data augmentation operation on each remote sensing satellite map in the formal training data to obtain an augmented map, wherein the data augmentation operation at least comprises distortion operation, turning operation and noise adding operation;
and performing preliminary training on the model to be trained by adopting the sound training data and the augmentation map.
Optionally, the other influence factor information includes water network information, and the step of inputting the breeding information and the other influence factor information into a preset risk prediction model to obtain a risk prediction result of the farm to be predicted includes:
determining a waterway net density characteristic value of the preset range according to the waterway net information and the preset range;
determining a farm density characteristic value in the preset range according to the breeding information and the preset range;
and inputting the density characteristic value of the waterway network and the density characteristic value of the farm into the risk prediction model to obtain a risk prediction result of the farm to be predicted.
Optionally, the risk prediction model includes preset weight values corresponding to the waterway network density characteristic value and the farm density characteristic value, respectively, and the step of inputting the waterway network density characteristic value and the farm density characteristic value into the risk prediction model to obtain the risk prediction result of the farm to be predicted includes:
inputting the waterway network density characteristic value and the farm density characteristic value into the risk prediction model to call the risk prediction model, and calculating to obtain a risk coefficient of the farm to be predicted based on the weight value, the waterway network density characteristic value and the farm density characteristic value;
and taking the risk coefficient as a risk prediction result of the farm to be predicted, or taking the risk grade as a risk prediction result of the farm to be predicted after determining the risk grade according to the risk coefficient and a preset corresponding relation between the coefficient and the grade.
In order to achieve the above object, the present invention also provides a farm risk prediction device, including:
the input module is used for inputting the map to be detected into a farm detection model to obtain farm information in the map to be detected;
the determining module is used for determining a farm to be predicted according to the farm information, determining breeding information in a preset range around the farm to be predicted, and acquiring other influence factor information except the farm in the preset range;
and the prediction module is used for inputting the breeding information and the other influence factor information into a preset risk prediction model to obtain a risk prediction result of the farm to be predicted.
In order to achieve the above object, the present invention also provides a farm risk prediction apparatus, including: a memory, a processor, and a farm risk prediction program stored on the memory and executable on the processor, the farm risk prediction program when executed by the processor implementing the steps of the farm risk prediction method as described above.
Furthermore, to achieve the above object, the present invention also proposes a computer-readable storage medium having stored thereon a farm risk prediction program that, when executed by a processor, implements the steps of the farm risk prediction method as described above.
According to the method, the map to be detected is input into a farm detection model obtained through pre-training, so that farm information in the map to be detected is obtained; determining a farm to be predicted according to farm information, determining breeding information in a preset range around the farm to be predicted, and acquiring other influence factor information in the preset range; and inputting the breeding information and other influence factor information into a preset risk detection model to obtain a risk prediction result of the farm to be predicted. In the invention, each farm is detected from the map by training a farm detection model, and farm information is obtained and used as a data source for risk prediction, so that the dependence on information reported by farmers is reduced, the management efficiency of the farm is improved, and the risk prediction efficiency is also improved; the accurate plant information of in time statistics to for breeding owner provides schemes such as production, management, improve animal husbandry and breed the efficiency of managing. The risk prediction model is arranged to process the breeding information and other influence factor information within the preset range of the farm to be predicted to obtain the risk prediction result of the farm to be predicted, and a specialist is not required to examine the condition of each farm on the spot, so that the risk prediction efficiency of the farm is improved, and the condition that the epidemic situation spreads rapidly can be dealt with; compared with the scheme that the risk condition is actively reported by the farmers, the risk condition of each farm can be more accurately predicted by adopting a unified risk prediction model in the embodiment; based on a rapid and accurate risk prediction result, the epidemic situation prevention and control work can be rapidly and pertinently expanded, the situation is prevented from being out of control, the quality safety of agricultural products is improved, and the ecological environment is improved.
Drawings
FIG. 1 is a schematic diagram of a hardware operating environment according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart illustrating a first embodiment of a risk prediction method for a farm according to the present invention;
FIG. 3 is a schematic diagram of a risk prediction process according to an embodiment of the present invention;
FIG. 4 is a block diagram of a farm risk prediction system according to a preferred embodiment of the present invention.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
As shown in fig. 1, fig. 1 is a schematic device structure diagram of a hardware operating environment according to an embodiment of the present invention.
It should be noted that, the plant risk prediction device in the embodiment of the present invention may be a smart phone, a personal computer, a server, and the like, and is not limited herein.
As shown in fig. 1, the farm risk prediction apparatus may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, a communication bus 1002. Wherein a communication bus 1002 is used to enable connective communication between these components. The user interface 1003 may include a Display screen (Display), an input unit such as a Keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1005 may be a high-speed RAM memory or a non-volatile memory (e.g., a magnetic disk memory). The memory 1005 may alternatively be a storage device separate from the processor 1001.
Those skilled in the art will appreciate that the configuration of the apparatus shown in fig. 1 does not constitute a limitation of the plant risk prediction apparatus and may include more or fewer components than those shown, or some components in combination, or a different arrangement of components.
As shown in fig. 1, a memory 1005, which is a kind of computer storage medium, may include therein an operating system, a network communication module, a user interface module, and a farm risk prediction program. The operating system is a program for managing and controlling hardware and software resources of the equipment, and supports the operation of a farm risk prediction program and other software or programs.
In the device shown in fig. 1, the user interface 1003 is mainly used for data communication with a client; the network interface 1004 is mainly used for establishing communication connection with a server; and the processor 1001 may be configured to invoke the farm risk prediction program stored in the memory 1005 and perform the following operations:
inputting a map to be detected into a farm detection model to obtain farm information in the map to be detected;
determining a farm to be predicted according to the farm information, determining culture information in a preset range around the farm to be predicted, and acquiring other influence factor information except the farm in the preset range;
and inputting the breeding information and the other influence factor information into a preset risk prediction model to obtain a risk prediction result of the farm to be predicted.
Further, the step of inputting the map to be detected into a farm detection model to obtain farm information in the map to be detected includes:
inputting the map to be detected into the farm detection model to obtain the classification category of each pixel point in the map to be detected, wherein the classification category is used for indicating whether the corresponding pixel point belongs to the farm;
and determining the farm information in the map to be detected according to the classification type.
Further, the step of determining the farm information in the map to be detected according to the classification category includes:
determining target pixel points belonging to a farm in the map to be detected according to the classification category;
clustering the target pixel points by adopting a preset clustering algorithm to obtain each cluster group;
and determining the farm information in the map to be detected according to each cluster group.
Further, before the step of inputting the map to be detected into the farm detection model to obtain the farm information in the map to be detected, the processor 1001 may be further configured to invoke a farm risk prediction program stored in the memory 1005, and execute the following operations:
performing primary training on a model to be trained by adopting the formal training data, wherein the formal training data comprises a plurality of pre-collected remote sensing satellite maps containing farms and farm marking data corresponding to each remote sensing satellite map;
adjusting the model to be trained after the initial training by adopting negative example training data or adopting the negative example training data and the positive example training data, wherein the negative example training data comprise a plurality of remote sensing satellite maps which are collected in advance and do not comprise a farm;
when the adjusted model to be trained is detected to meet the preset model condition, taking the adjusted model to be trained as the farm detection model, otherwise, executing the steps again based on the adjusted model to be trained: and (5) performing primary training on the model to be trained by adopting the normal training data.
Further, the step of performing preliminary training on the model to be trained by using the normal training data includes:
carrying out data augmentation operation on each remote sensing satellite map in the formal training data to obtain an augmented map, wherein the data augmentation operation at least comprises distortion operation, turning operation and noise adding operation;
and performing preliminary training on the model to be trained by adopting the sound training data and the augmentation map.
Further, the information of the other influence factors includes information of a water network, and the step of inputting the breeding information and the information of the other influence factors into a preset risk prediction model to obtain a risk prediction result of the farm to be predicted includes:
determining a waterway net density characteristic value of the preset range according to the waterway net information and the preset range;
determining a farm density characteristic value in the preset range according to the breeding information and the preset range;
and inputting the density characteristic value of the waterway network and the density characteristic value of the farm into the risk prediction model to obtain a risk prediction result of the farm to be predicted.
Further, the risk prediction model includes preset weight values corresponding to the characteristic value of the density of the waterway network and the characteristic value of the density of the farm, and the step of inputting the characteristic value of the density of the waterway network and the characteristic value of the density of the farm into the risk prediction model to obtain the risk prediction result of the farm to be predicted includes:
inputting the waterway network density characteristic value and the farm density characteristic value into the risk prediction model to call the risk prediction model, and calculating to obtain a risk coefficient of the farm to be predicted based on the weight value, the waterway network density characteristic value and the farm density characteristic value;
and taking the risk coefficient as a risk prediction result of the farm to be predicted, or taking the risk grade as a risk prediction result of the farm to be predicted after determining the risk grade according to the risk coefficient and a preset corresponding relation between the coefficient and the grade.
Based on the above structure, various embodiments of a farm risk prediction method are provided.
Referring to fig. 2, fig. 2 is a schematic flow chart of a farm risk prediction method according to a first embodiment of the present invention.
While a logical order is shown in the flow chart, in some cases, the steps shown or described may be performed in an order different than that shown or described herein. The execution subject of each embodiment of the farm risk prediction method of the invention can be a smart phone, a personal computer, a server and other devices, and for convenience of description, the execution subject is omitted in the following embodiments for explanation. In this embodiment, the method for predicting risk of a farm includes:
step S10, inputting a map to be detected into a farm detection model to obtain farm information in the map to be detected, wherein the farm detection model is obtained by pre-training;
in this embodiment, for a certain area (area to be detected), the risk condition of each farm in the area needs to be predicted, and a map of the area can be obtained as the map to be detected. The risk may be the degree of influence that the farm may be affected when the infectious disease is transmitted by an animal such as swine fever. There are various ways to obtain the map, for example, the map may be downloaded from map software, or the remote sensing satellite map may be obtained by a remote sensing satellite. A farm detection model can be trained in advance and used for detecting the farm in the picture. The farm detection model can adopt a common image target detection model, such as a semantic image segmentation model DeepLab-v3 +. The training method of the farm detection model can adopt the existing training method of a machine learning model.
And inputting the map to be detected into a farm detection model, and obtaining farm information in the map to be detected according to an output result of the farm detection model. Specifically, the output result of the farm detection model may be the position of the farm in the map to be detected, and then the actual geographic position of the farm in the area to be detected may be calculated according to the scale of the map to be detected and the actual geographic position corresponding to the map to be detected. The acquired farm information may include at least the actual geographical location of the farm. Wherein, the farm can be seen as a point (the central point of the farm), and the actual geographic position of the farm is the position of the point; it is also possible to treat the farm as an area, the actual geographical location of the farm being the location of the area. The actual geographic location may be expressed in terms of latitude and longitude, or in terms of coordinates under a predetermined specific coordinate system, and the like.
Step S20, determining a farm to be predicted according to the farm information, determining breeding information in a preset range around the farm to be predicted, and acquiring other influence factor information except the farm in the preset range;
after the information of the farm in the map to be detected is obtained, the farm to be predicted can be determined according to the information of the farm. Specifically, how many farms are included in the area to be detected, the actual geographical position of each farm, and the like can be determined according to farm information. One of the detected multiple farms in the area to be detected can be selected as the farm to be predicted, specifically, each farm can be sequentially used as the farm to be predicted, or the farm to be predicted can be determined based on the selection of the user.
For a certain farm to be predicted, the risk of the farm is predicted, and the principle adopted in the embodiment is that the relevant data of the area around the farm can be used as the basis for judging the risk of the farm. Specifically, a predetermined range around the farm to be predicted may be determined. The determination method of the preset range can be various; for example, a radius value may be preset, a circular area may be divided by taking the farm as a circle center and the radius value as a radius, that is, the circular area may be used as a preset range of the farm, and according to the actual geographic position of the farm and the radius value, the actual geographic position of the preset range may be determined; the method can also be used for dividing a matrix area as a preset range of the farm by taking the farm as the center of a rectangle and presetting a matrix edge length value; there are other feasible methods, and different methods may be adopted to determine the preset range according to different actual application scenarios, which are not listed here.
After the preset range of the farm to be predicted is determined, the breeding information in the preset range can be determined according to the farm information in the map to be detected. Specifically, according to the actual geographical position of the preset range and the actual geographical position of each farm in the farm information, whether each farm falls into the preset range or not can be detected, the farm falling into the preset range is determined as the target farm, and then the information of the target farm can be extracted from the farm information as the breeding information in the preset range. For example, when the actual geographic location is expressed by longitude and latitude, the location of the preset range may be composed of a longitude interval and a latitude interval, and it is necessary to determine whether the longitude of each farm falls within the longitude interval and the latitude falls within the latitude interval, and if both of them fall, it is determined that the farm falls within the preset range.
Various risk influencing factor types, such as farms, water networks, road networks, villages, etc., which are different types of influencing factors, may be specified in advance. The water network may refer to rivers, lakes and the like with various sizes, and the road network may refer to various roads, such as national roads, high speed and the like. Hereinafter, the water network and the road network are collectively referred to as a "waterway network", and the water network and the road network may be viewed together or separately.
After the breeding information in the preset range is acquired, the information of other risk influence factors except the breeding farm in the preset range, namely other influence factor information, can be acquired. Specifically, other influencing factor information in the area to be detected may be collected in advance, where the other influencing factor information may be obtained by extracting from the electronic map data or by manual collection. The other influence factor information may specifically include the position, size, and the like of other risk influence factors, and the data may be different according to the type of the risk influence factor, for example, the collected village information may include the position information, area, and the like of the village, and the collected waterway network information may include the position information, length, and the like of the waterway network.
And determining the information of the other influence factors in the preset range according to the actual geographic position of each other influence factor in the information of the other influence factors in the area to be detected and the actual geographic position of the preset range.
And step S30, inputting the breeding information and the other influence factor information into a preset risk prediction model to obtain a risk prediction result of the farm to be predicted.
And inputting the breeding information and other influencing factor information within a preset range into a preset risk prediction model to obtain a risk prediction result of the farm to be predicted. Specifically, a risk prediction model may be preset, the risk prediction model may be a model structure such as a linear model or a neural network model, if the risk prediction model is the neural network model, the neural network model needs to be trained in advance, and the trained model is used for risk prediction. The setting principle of the risk prediction model is as follows: when the more and more intensive the input data represent the target influence factors in the preset range of the farm to be predicted, the higher the risk of the farm to be predicted represented by the output risk prediction result. The input data of the risk prediction model can be set according to the situation, for example, the input data can be set to be the total area of the farm and the total area of the waterway network within a preset range. The breeding information and other influence factor information can be processed into a form matched with the input data of the risk prediction model, the processed data is input into the risk prediction model, and the risk prediction result of the farm to be predicted is output after the data is processed by the risk prediction model. The risk prediction result may be in various forms, for example, a score or a grade. The risk levels corresponding to the various scores or grades may be predefined, for example, the score is 0 to 10, and the risk levels become larger from 0 to 10.
In the embodiment, the map to be detected is input into a farm detection model obtained by pre-training, so that farm information in the map to be detected is obtained; determining a farm to be predicted according to farm information, determining breeding information in a preset range around the farm to be predicted, and acquiring other influence factor information in the preset range; and inputting the breeding information and other influence factor information into a preset risk detection model to obtain a risk prediction result of the farm to be predicted. In the embodiment, each farm is detected from the map by training a farm detection model, and farm information is obtained and used as a data source for risk prediction, so that dependence on information reported by farmers is reduced, the management efficiency of the farm is improved, and the risk prediction efficiency is also improved; the accurate plant information of in time statistics to for breeding owner provides schemes such as production, management, improve animal husbandry and breed the efficiency of managing. The risk prediction model is arranged to process the breeding information and other influence factor information within the preset range of the farm to be predicted to obtain the risk prediction result of the farm to be predicted, and a specialist is not required to examine the condition of each farm on the spot, so that the risk prediction efficiency of the farm is improved, and the condition that the epidemic situation spreads rapidly can be dealt with; compared with the scheme that the risk condition is actively reported by the farmers, the risk condition of each farm can be more accurately predicted by adopting a unified risk prediction model in the embodiment; based on a rapid and accurate risk prediction result, the epidemic situation prevention and control work can be rapidly and pertinently expanded, the situation is prevented from being out of control, the quality safety of agricultural products is improved, and the ecological environment is improved.
Furthermore, different prevention and control schemes can be set in advance according to different risk prediction results, after the risk prediction result of the farm to be predicted is obtained, a proper prevention and control scheme is matched from the various prevention and control schemes according to the risk prediction result, the matched prevention and control scheme is determined as a target prevention and control scheme of the farm to be predicted, and the target prevention and control scheme can be output. The output mode can be output to a display screen for display or output in a voice mode, and the output mode can be set differently according to specific application scenes. The control scheme of the farm to be predicted is directly output and displayed, so that a user can visually know the control scheme, and control work is carried out according to the control scheme. Particularly, for users who need to take prevention and control measures but do not know which prevention and control measures exist or do not know the risk condition of the farm, such as farmers or residents around the farm, the prevention and control measures matched with the risk condition of the farm can be conveniently and directly obtained, so that the development of prevention and control work is facilitated.
Based on the risk prediction method in the embodiment, the risk prediction method can be practically applied in a mobile phone or computer application software mode, so that a user can input relevant information, such as position information, of the farm to be predicted based on an installed application program, and a risk prediction result of the farm to be predicted, even a prevention and control scheme, can be directly obtained.
Further, based on the first embodiment, a second embodiment of the farm risk prediction method according to the present invention is provided, and in this embodiment, the step S10 includes:
step S102, inputting a map to be detected into the farm detection model to obtain classification categories of all pixel points in the map to be detected, wherein the classification categories are used for indicating whether the corresponding pixel points belong to the farm or not;
the plant detection model can be a two-classification model, after the map to be detected is input into the plant detection model, the plant detection model can output classification results of all pixel points in the map to be detected, and the classification results are used for indicating whether the corresponding pixel points belong to the plant or not. For example, the classification result may be 0 and 1, where 1 indicates that the corresponding pixel belongs to the farm, and 0 indicates that the corresponding pixel does not belong to the farm.
And S103, determining the farm information in the map to be detected according to the classification type.
According to the classification category of each pixel point, the farm information in the map to be detected can be determined. Specifically, the target pixel point belonging to the farm can be determined from each pixel point according to the classification category of each pixel point; and then according to the scale of the map to be detected, determining the number of actual areas represented by each pixel point, and according to the longitude and latitude interval of the map to be detected, determining the longitude and latitude interval of each target pixel point.
Further, the step S103 includes:
step S1031, determining target pixel points belonging to the farm in the map to be detected according to the classification category;
and determining target pixel points belonging to the farm from the pixel points according to the classification categories of the pixel points, for example, when the classification categories are 0 and 1, taking all the pixel points with the classification category of 1 as the target pixel points.
Step S1032, clustering the target pixel points by adopting a preset clustering algorithm to obtain each cluster group;
a clustering algorithm can be preset, and the clustering algorithm is adopted to cluster the target pixel points to obtain each cluster group. It should be noted that when the area actually corresponding to one pixel point in the map to be detected is smaller, it is possible that one pixel point represents one breeding house (i.e., a unit smaller than the farm), it is possible that the area of one farm corresponds to a plurality of target pixel points, at this time, a clustering algorithm may be adopted to cluster the identified plurality of pixel points, or cluster the longitude and latitude sections corresponding to the plurality of pixel points, and cluster a plurality of breeding houses with closer distances into a cluster group. The clustering algorithm may employ a conventional clustering algorithm, such as a neighbor clustering method.
And step S1033, determining the farm information in the map to be detected according to each cluster group.
And (3) respectively using each clustered group obtained by clustering as a farm, wherein the longitude and latitude of the farm can be the center of the clustered group, the area of the farm can be the sum of the areas of all breeding houses in the cluster group range, and the farm information in the farm to be detected can at least comprise the position and the area of each farm.
Further, before the step S10, the method further includes:
step S40, performing preliminary training on a model to be trained by adopting the formal training data, wherein the formal training data comprise a plurality of pre-collected remote sensing satellite maps containing farms and farm marking data corresponding to the remote sensing satellite maps;
further, in this embodiment, the farm test model may be trained as follows:
the method comprises the steps of collecting a plurality of remote sensing satellite maps containing the farm in advance, and collecting marking data of the farm in the remote sensing satellite maps, namely the marking data of the farm. The marked data of the farm can be a mask graph corresponding to the remote sensing satellite picture, classification categories corresponding to each pixel point of the remote sensing satellite picture are arranged in the mask graph, if 0 indicates that the pixel point does not belong to the farm, and if 1 indicates that the pixel point belongs to the farm, different colors can be adopted in the mask graph to indicate different categories. A plurality of remote sensing satellite maps containing the farm and farm data are used as the regular training data.
The model to be trained is initially trained by adopting the normal training data. The structure of the model to be trained can adopt a common semantic segmentation model, such as a semantic image segmentation model deep lab-v3 +. The process of preliminary training may employ an existing machine learning model training process.
Step S50, adjusting the model to be trained after the initial training by adopting negative example training data or adopting the negative example training data and the positive example training data, wherein the negative example training data comprise a plurality of remote sensing satellite maps which are collected in advance and do not comprise a farm;
a plurality of remote sensing satellite maps which do not contain a farm are collected in advance and used as negative example training data, the proportion of the negative example training data to the positive example training data can be set according to practical experience, and for example, the negative example training data can be one tenth of the positive example training data. After the initial training, the semantic segmentation model after the initial training is adjusted by adopting the negative training data and the positive training data together, or the semantic segmentation model after the initial training is adjusted by adopting the negative training data alone, wherein the adjustment at this time can be fine adjustment. Specifically, the fine tuning is trained by using the positive training data and the negative training data together, and the training process can also be the training process of the existing machine learning model. It should be noted that, during fine tuning, the hyper-parameters of the model may be adjusted according to the ratio of the positive training data to the negative training data, and then fine tuning training is performed after the adjustment, for example, when the negative training data may be one tenth of the positive training data, the learning rate in the fine tuning stage may also be set to one tenth of the initial training stage.
Step S60, when the adjusted model to be trained is detected to meet the preset model condition, the adjusted model to be trained is used as the farm detection model, otherwise, the step is executed again based on the adjusted model to be trained: and (5) performing primary training on the model to be trained by adopting the normal training data.
And detecting whether the adjusted model to be trained accords with preset model conditions, and if so, taking the adjusted model to be trained as a farm detection model so as to detect the farm by adopting the farm detection model in the following process. Otherwise, if the preset model condition is not met, the adjusted model to be trained is subjected to preliminary training by adopting the normal training data, and then the adjustment training is carried out until the preset model condition is detected to be met. The preset model condition may be a condition set in advance according to a performance requirement of the model, for example, a loss function convergence of the model may be used as a condition, or a commonly used objective index for detecting the model performance may be used as a condition, for example, objective indexes such as accuracy, recall rate, and IOU (Intersection over unit, Intersection ratio). The calculation method of each objective index may refer to the existing index calculation method, which is not described in detail herein.
In the embodiment, a farm detection model is trained in advance, and a map to be detected is detected and identified to obtain farm information in the map to be detected, so that the farm information of each region can be intelligently and quickly obtained, and the risk prediction efficiency of the farm to be predicted is improved; compared with a mode of actively reporting by relying on farmers, the scheme in the embodiment can more comprehensively acquire farm information, so that data on which risk prediction is based is more accurate and comprehensive, and a risk prediction result is more accurate.
Further, the step S40 includes:
step S401, carrying out data augmentation operation on each remote sensing satellite map in the training data of the prime example to obtain an augmented map, wherein the data augmentation operation at least comprises distortion operation, turning operation and noise adding operation;
further, for each remote sensing satellite map in the training data of the prime example, data augmentation operation may be performed on each remote sensing satellite map, where the data augmentation operation may be operations such as distortion operation, flip operation, and noise addition operation performed on the remote sensing satellite image, and these operations are similar to the existing data augmentation operation, and are not described in detail herein. Through data augmentation operation, on one hand, the number of the remote sensing satellite images can be increased when the number of the remote sensing satellite images is small, so that training data are increased, the model is sufficiently trained, on the other hand, the model obtained through training can identify various different farms, and the universality of the model is also improved.
And S402, performing preliminary training on the model to be trained by adopting the normal training data and the augmentation map.
And an augmentation map is obtained by augmenting data by a remote sensing satellite map, and the annotation data of the augmentation map is the same as the annotation data of the remote sensing satellite. And (3) taking the remote sensing satellite map and the augmented map in the normal training data as input data of the model to be trained, and correcting the output of the model to be trained by adopting the marking data so as to perform primary training on the model to be trained.
Further, the augmentation operation may be applied to negative example training data as well.
As shown in fig. 3, a possible risk prediction process in this embodiment includes three stages, i.e., data preparation, target detection, and risk prediction.
1. And (4) preparing data.
The data preparation stage is to prepare training data for training the detection model of the farm and other influence factor information, such as water network information, village information and the like.
Specifically, as shown in fig. 3, a 18-level 1:50000 satellite map of Guangdong province on common map software can be downloaded, wherein 32 of the maps serve as training data, 2 serve as verification data, and the others serve as test data; meanwhile, the longitude and latitude of villages, road networks (including expressways, national roads and the like) and water networks in Guangdong province on the map can be analyzed, and the longitude and latitude are used for assisting epidemic situation risk prediction after conversion; carrying out data annotation on 34 remote sensing satellite maps by utilizing ArcGIS geographic satellite annotation software to mark a specific breeding house, generating a geographic satellite picture with annotations as an experiment reference, and generating an annotation mask with a completely black background as experiment input; cutting out the labeled maps, wherein 32 training maps are cut into a plurality of small blocks with the size of 1024 x 1024 in an overlapping mode, and other maps are cut into a plurality of small blocks with the size of 8192 x 8192 in a non-overlapping mode; and selecting the small blocks with the farms and the marked masks of the small blocks as positive case training data, and selecting the small blocks without the farms as negative case training data.
And carrying out preliminary training by adopting positive training data, carrying out fine tuning by adopting negative training data, calculating each objective index, evaluating the model, and selecting the optimal model as a final farm detection model.
2. And detecting the target.
In the target detection stage, a farm detection model is mainly adopted to detect farms in the map of each area, and a breeding house mask map is obtained.
3. And (4) risk prediction.
For the farm to be predicted, a range around the farm to be predicted is determined, then a clustering algorithm is adopted to perform clustering calculation on the breeding information obtained in the target detection stage and other risk influence factor information, and various density values in the range are obtained, wherein the density values specifically comprise breeding density, high-speed density, village density, water network density and national road density. And inputting the density values into a risk prediction model, and finally outputting a risk prediction result of the farm to be predicted.
Further, based on the first and second embodiments, a third embodiment of the method for predicting risk of a farm according to the present invention is provided, in this embodiment, the other influence factor information includes water network information, and the step S30 includes:
step S301, determining a waterway network density characteristic value of the preset range according to the waterway network information and the preset range;
further, in this embodiment, the other influence factor information may be water network information. And predicting the risk condition of the farm to be predicted according to the breeding information and the waterway network information in the preset range.
Specifically, the characteristic value of the density of the waterway network in the preset range is determined according to the information of the waterway network and the preset range. The characteristic value of the density of the waterway network may be a result obtained by dividing the total length of the waterway network within the preset range by the total area of the preset range, or a result obtained by dividing the total area of the waterway network within the preset range by the total area of the preset range. The area and length calculation process may be as follows: the total area of the preset range can be determined according to the position information of the preset range, and if the total area of the preset range can be calculated according to the longitude and latitude interval of the preset range; determining the length of each water path network according to the position information of each water path network in the water path network information, for example, one river or one road can be represented by the longitude and latitude of two points, the length of the river or the road is calculated according to the longitude and latitude of the two points, and the lengths of all the rivers or the roads are added to obtain the length of a target water path network; the total area of the water network may also be determined based on the position information of each water network in the water network information, for example, one river or one road may be represented by latitude and longitude intervals, in which case the area of the river or the road is calculated based on the latitude and longitude intervals, and the total area of the water network is obtained by adding the areas of all the rivers or the roads.
Step S302, determining a farm density characteristic value in the preset range according to the breeding information and the preset range;
and determining the density characteristic value of the farm within the preset range according to the breeding information and the preset range. Specifically, the farm density characteristic value may be a result of dividing a total area of the farm within a preset range by the total area of the preset range. The total area calculation method of the farm in the preset range can be as follows: each farm can be represented by a longitude and latitude interval respectively, the area of each farm is obtained through calculation according to the longitude and latitude intervals, and then the areas are added to obtain the total area of the target farm.
The characteristic value of the farm density may also be a result of dividing the total feeding amount of the farm in the preset range by the total area of the preset range, that is, the characteristic value of the farm density may be the feeding density. The total feeding amount of the target farm can be calculated in the following mode: and determining the area of each farm according to the position information of each farm in the breeding information, namely the total area calculation process of the farms. The correspondence between the area of the farm and the amount of feed can be set in advance, for example, 1.5 square meters is set to produce one pig. And determining the total feeding amount of each farm according to the area of each farm obtained by calculation and the corresponding relation. For example, if the area of a farm is 150 square meters, the total feeding amount of each farm is 100 pigs according to the above example; dividing the total feeding amount by the total area, and obtaining the result as a density characteristic value of the farm.
Step S303, inputting the characteristic value of the density of the waterway network and the characteristic value of the density of the farm into the risk prediction model to obtain a risk prediction result of the farm to be predicted.
And inputting the density characteristic value of the water way network and the density characteristic value of the farm into a risk prediction model to obtain a risk prediction result of the farm to be predicted. In this embodiment, the risk prediction result of the farm to be predicted is calculated by calculating the farm density characteristic value and the waterway network density characteristic value in the preset range around the farm to be predicted, inputting the farm density characteristic value and the waterway network density characteristic value into the risk prediction model, and calculating the risk prediction result of the farm to be predicted, so that the risk of the farm is predicted according to the density characteristics of the farm and the waterway network in the preset range, and the actual infectious characteristics of animal infectious diseases are combined, so that the predicted risk prediction result is more accurate.
Further, the risk prediction model includes preset weight values corresponding to the density characteristic value of the waterway network and the density characteristic value of the farm, and the step S303 includes:
step S3031, inputting the waterway network density characteristic value and the farm density characteristic value into the risk prediction model, and calling the risk prediction model to calculate a risk coefficient of the farm to be predicted based on the weight value, the waterway network density characteristic value and the farm density characteristic value;
further, the risk prediction model may be a linear model, and may set weights corresponding to the farm density characteristic values, the water network density characteristic values, and the road network density characteristic values, respectively.
After the water path network density characteristic value and the farm density characteristic value are obtained, the water path network density characteristic value and the farm density characteristic value are input into a risk prediction model, and a risk coefficient of the farm to be predicted is calculated by calling the risk prediction model based on the weight values, the water path network density characteristic value and the farm density characteristic value. Specifically, the density characteristic value of the farm may be multiplied by a corresponding weight, the density characteristic value of the water network may be multiplied by a corresponding weight, the multiplication results are added, and the added result is used as a risk coefficient, that is, the nature of the risk prediction model may be a linear model. It should be noted that each weight value in the risk prediction model may be set according to specific experience.
Step S3032, the risk coefficient is used as a risk prediction result of the farm to be predicted, or after a risk grade is determined according to the risk coefficient and a preset corresponding relation between the coefficient and the grade, the risk grade is used as the risk prediction result of the farm to be predicted.
After the risk coefficient is obtained, the risk coefficient can be directly used as a risk prediction result of the farm to be predicted. The corresponding relationship between the risk coefficients and the levels may also be preset, for example, the risk coefficients are sorted from low to high, the lowest risk coefficient 1/4 is divided into a first level, 1/4-1/2 is divided into a second level, 1/2-3/4 is divided into a third level, and the highest risk coefficient 1/4 is divided into a fourth level. It should be noted that the number of levels and the boundary of each level can be selected according to specific experience. And determining which risk grade the risk coefficient corresponds to according to the obtained risk coefficient and the corresponding relation between the coefficient and the grade, and taking the determined risk grade as a risk prediction result of the farm to be predicted. According to different actual requirements of users, risk prediction results can be different, and risk levels are possibly more intuitive for some users, so that the users can make a targeted prevention and control scheme.
In addition, an embodiment of the present invention further provides a farm risk prediction device, and referring to fig. 4, the farm risk prediction device includes:
the input module 10 is configured to input the map to be detected into a farm detection model to obtain farm information in the map to be detected;
the determining module 20 is configured to determine a farm to be predicted according to the farm information, determine breeding information in a preset range around the farm to be predicted, and acquire other influence factor information except the farm in the preset range;
and the prediction module 30 is configured to input the breeding information and the other influence factor information into a preset risk prediction model to obtain a risk prediction result of the farm to be predicted.
Further, the input module 10 includes:
the system comprises a first input unit, a first storage unit and a second input unit, wherein the first input unit is used for inputting a map to be detected into the farm detection model to obtain the classification category of each pixel point in the map to be detected, and the classification category is used for indicating whether the corresponding pixel point belongs to the farm;
and the first determining unit is used for determining the farm information in the map to be detected according to the classification type.
Further, the determining unit includes:
the first determining subunit is used for determining target pixel points belonging to a farm in the map to be detected according to the classification category;
the clustering subunit is used for clustering the target pixel points by adopting a preset clustering algorithm to obtain each clustering group;
and the second determining subunit is used for determining the farm information in the map to be detected according to each cluster group.
Further, the plant risk prediction device further includes:
the system comprises a preliminary training module, a data processing module and a data processing module, wherein the preliminary training module is used for carrying out preliminary training on a model to be trained by adopting regular training data, and the regular training data comprises a plurality of pre-collected remote sensing satellite maps containing farms and farm marking data corresponding to the remote sensing satellite maps;
the adjusting module is used for adjusting the model to be trained after the initial training by adopting negative example training data or adopting the negative example training data and the positive example training data, wherein the negative example training data comprise a plurality of remote sensing satellite maps which are collected in advance and do not comprise a farm;
a defining module, configured to, when it is detected that the adjusted model to be trained meets a preset model condition, use the adjusted model to be trained as the farm detection model, and otherwise, execute the step again based on the adjusted model to be trained: and (5) performing primary training on the model to be trained by adopting the normal training data. .
Further, the preliminary training module includes:
the augmentation unit is used for carrying out data augmentation operation on each remote sensing satellite map in the formal training data to obtain an augmentation map, wherein the data augmentation operation at least comprises distortion operation, turning operation and noise adding operation;
and the preliminary training unit is used for carrying out preliminary training on the model to be trained by adopting the formal training data and the augmentation map.
Further, the other influence factor information includes water network information, and the prediction module 30 includes:
the second determining unit is used for determining the waterway net density characteristic value of the preset range according to the waterway net information and the preset range;
the second determining unit is used for determining a farm density characteristic value in the preset range according to the breeding information and the preset range;
and the second input unit is used for inputting the density characteristic value of the waterway network and the density characteristic value of the farm into the risk prediction model to obtain a risk prediction result of the farm to be predicted.
Further, the risk prediction model includes preset weight values corresponding to the density characteristic value of the waterway network and the density characteristic value of the farm, and the second input unit includes:
the input subunit is used for inputting the waterway network density characteristic value and the farm density characteristic value into the risk prediction model so as to call the risk prediction model to calculate a risk coefficient of the farm to be predicted based on the weight value, the waterway network density characteristic value and the farm density characteristic value;
and the third determining subunit is used for taking the risk coefficient as a risk prediction result of the farm to be predicted, or taking the risk grade as a risk prediction result of the farm to be predicted after determining the risk grade according to the risk coefficient and a preset corresponding relation between the coefficient and the grade.
The expansion content of the specific implementation of the plant risk prediction device of the present invention is basically the same as that of each embodiment of the plant risk prediction method, and is not described herein again.
Furthermore, an embodiment of the present invention further provides a computer-readable storage medium, in which a farm risk prediction program is stored, and when executed by a processor, the farm risk prediction program implements the steps of the farm risk prediction method as described below.
For the embodiments of the plant risk prediction device and the computer-readable storage medium of the present invention, reference may be made to the embodiments of the plant risk prediction method of the present invention, and details are not described herein.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal device (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present invention.
The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (10)
1. A risk prediction method for a farm, characterized by comprising the steps of:
inputting a map to be detected into a farm detection model to obtain farm information in the map to be detected;
determining a farm to be predicted according to the farm information, determining culture information in a preset range around the farm to be predicted, and acquiring other influence factor information except the farm in the preset range;
and inputting the breeding information and the other influence factor information into a preset risk prediction model to obtain a risk prediction result of the farm to be predicted.
2. The method for predicting risk of a farm according to claim 1, wherein the step of inputting the map to be detected into a farm inspection model to obtain farm information in the map to be detected comprises:
inputting the map to be detected into the farm detection model to obtain the classification category of each pixel point in the map to be detected, wherein the classification category is used for indicating whether the corresponding pixel point belongs to the farm;
and determining the farm information in the map to be detected according to the classification type.
3. The method of predicting risk of a farm according to claim 2, wherein the step of determining farm information in the map to be tested according to the classification category comprises:
determining target pixel points belonging to a farm in the map to be detected according to the classification category;
clustering the target pixel points by adopting a preset clustering algorithm to obtain each cluster group;
and determining the farm information in the map to be detected according to each cluster group.
4. The method for predicting risk of a farm according to claim 1, wherein before the step of inputting the map to be detected into a farm inspection model to obtain farm information in the map to be detected, the method further comprises:
performing primary training on a model to be trained by adopting the formal training data, wherein the formal training data comprises a plurality of pre-collected remote sensing satellite maps containing farms and farm marking data corresponding to each remote sensing satellite map;
adjusting the model to be trained after the initial training by adopting negative example training data or adopting the negative example training data and the positive example training data, wherein the negative example training data comprise a plurality of remote sensing satellite maps which are collected in advance and do not comprise a farm;
when the adjusted model to be trained is detected to meet the preset model condition, taking the adjusted model to be trained as the farm detection model, otherwise, executing the steps again based on the adjusted model to be trained: and (5) performing primary training on the model to be trained by adopting the normal training data.
5. The farm risk prediction method of claim 4, wherein the step of initially training the model to be trained using normative training data comprises:
carrying out data augmentation operation on each remote sensing satellite map in the formal training data to obtain an augmented map, wherein the data augmentation operation at least comprises distortion operation, turning operation and noise adding operation;
and performing preliminary training on the model to be trained by adopting the sound training data and the augmentation map.
6. The method according to any one of claims 1 to 5, wherein the other influence factor information comprises water network information, and the step of inputting the breeding information and the other influence factor information into a preset risk prediction model to obtain a risk prediction result of the farm to be predicted comprises:
determining a waterway net density characteristic value of the preset range according to the waterway net information and the preset range;
determining a farm density characteristic value in the preset range according to the breeding information and the preset range;
and inputting the density characteristic value of the waterway network and the density characteristic value of the farm into the risk prediction model to obtain a risk prediction result of the farm to be predicted.
7. The method according to claim 6, wherein the risk prediction model includes preset weight values corresponding to the waterway network density characteristic value and the farm density characteristic value, respectively, and the step of inputting the waterway network density characteristic value and the farm density characteristic value into the risk prediction model to obtain the risk prediction result of the farm to be predicted includes:
inputting the waterway network density characteristic value and the farm density characteristic value into the risk prediction model to call the risk prediction model, and calculating to obtain a risk coefficient of the farm to be predicted based on the weight value, the waterway network density characteristic value and the farm density characteristic value;
and taking the risk coefficient as a risk prediction result of the farm to be predicted, or taking the risk grade as a risk prediction result of the farm to be predicted after determining the risk grade according to the risk coefficient and a preset corresponding relation between the coefficient and the grade.
8. A farm risk prediction device, characterized by comprising:
the input module is used for inputting the map to be detected into a farm detection model to obtain farm information in the map to be detected;
the determining module is used for determining a farm to be predicted according to the farm information, determining breeding information in a preset range around the farm to be predicted, and acquiring other influence factor information except the farm in the preset range;
and the prediction module is used for inputting the breeding information and the other influence factor information into a preset risk prediction model to obtain a risk prediction result of the farm to be predicted.
9. A farm risk prediction apparatus, characterized by comprising: a memory, a processor, and a farm risk prediction program stored on the memory and executable on the processor, the farm risk prediction program when executed by the processor implementing the steps of the farm risk prediction method according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium has stored thereon a farm risk prediction program that, when executed by a processor, implements the steps of the farm risk prediction method according to any one of claims 1 to 7.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010130767.1A CN111339237B (en) | 2020-02-28 | 2020-02-28 | Farm risk prediction method, device, equipment and storage medium |
PCT/CN2020/134032 WO2021169464A1 (en) | 2020-02-28 | 2020-12-04 | Method, apparatus and device for farm risk prediction, and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010130767.1A CN111339237B (en) | 2020-02-28 | 2020-02-28 | Farm risk prediction method, device, equipment and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111339237A true CN111339237A (en) | 2020-06-26 |
CN111339237B CN111339237B (en) | 2024-07-19 |
Family
ID=71185750
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010130767.1A Active CN111339237B (en) | 2020-02-28 | 2020-02-28 | Farm risk prediction method, device, equipment and storage medium |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN111339237B (en) |
WO (1) | WO2021169464A1 (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021169464A1 (en) * | 2020-02-28 | 2021-09-02 | 深圳前海微众银行股份有限公司 | Method, apparatus and device for farm risk prediction, and storage medium |
CN116957342A (en) * | 2023-09-18 | 2023-10-27 | 深圳市博安智控科技有限公司 | Intelligent management method, system and medium for farm based on Internet of things |
CN117787736A (en) * | 2023-11-17 | 2024-03-29 | 新光瑞合(青岛)生物科技有限公司 | Construction method and system of healthy cultivation safety system |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114545765B (en) * | 2022-01-24 | 2024-01-12 | 山东峰泰电力设备有限公司 | Transformer parameter determining method and system based on PID control |
CN116484226B (en) * | 2023-05-04 | 2023-10-03 | 乳山市海洋经济发展中心 | Method and device for building fishery culture model through neural network |
CN116313145A (en) * | 2023-05-10 | 2023-06-23 | 江西省农业科学院畜牧兽医研究所 | Intelligent epidemiological investigation system for animals |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101763095A (en) * | 2009-12-09 | 2010-06-30 | 中国人民解放军防化指挥工程学院 | Forecasting-early warning and emergency control method for moving chemical risk source |
CN103679558A (en) * | 2013-12-20 | 2014-03-26 | 国家电网公司 | Electric automobile charging and replacing station fire risk data evaluation method |
CN105205587A (en) * | 2015-08-15 | 2015-12-30 | 东北师范大学 | Northern grassland fire hazard risk grade rapid assessment method |
CN107705047A (en) * | 2017-11-08 | 2018-02-16 | 泰康保险集团股份有限公司 | A kind of adjuster Risk Forecast Method and device |
CN110059758A (en) * | 2019-04-24 | 2019-07-26 | 海南长光卫星信息技术有限公司 | A kind of remote sensing image culture pond detection method based on semantic segmentation |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109993432A (en) * | 2019-03-29 | 2019-07-09 | 柳州市水产畜牧兽医局 | A kind of pig farm cultivation quality safety methods of risk assessment |
CN111339237B (en) * | 2020-02-28 | 2024-07-19 | 深圳前海微众银行股份有限公司 | Farm risk prediction method, device, equipment and storage medium |
-
2020
- 2020-02-28 CN CN202010130767.1A patent/CN111339237B/en active Active
- 2020-12-04 WO PCT/CN2020/134032 patent/WO2021169464A1/en active Application Filing
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101763095A (en) * | 2009-12-09 | 2010-06-30 | 中国人民解放军防化指挥工程学院 | Forecasting-early warning and emergency control method for moving chemical risk source |
CN103679558A (en) * | 2013-12-20 | 2014-03-26 | 国家电网公司 | Electric automobile charging and replacing station fire risk data evaluation method |
CN105205587A (en) * | 2015-08-15 | 2015-12-30 | 东北师范大学 | Northern grassland fire hazard risk grade rapid assessment method |
CN107705047A (en) * | 2017-11-08 | 2018-02-16 | 泰康保险集团股份有限公司 | A kind of adjuster Risk Forecast Method and device |
CN110059758A (en) * | 2019-04-24 | 2019-07-26 | 海南长光卫星信息技术有限公司 | A kind of remote sensing image culture pond detection method based on semantic segmentation |
Non-Patent Citations (3)
Title |
---|
江炜炜: "规模化猪场疾病防治存在问题及对策", 《 畜牧兽医科学(电子版)》 * |
邓振民等: "基于信息技术的家禽肿瘤性疾病预警研究", 《山东农业大学学报(自然科学版)》, vol. 46, no. 3, pages 1 - 6 * |
陈俊松等: "基于GF-2影像的平原河网区规模化生猪养殖场提取方法研究", 《生态与农村环境学报》, vol. 36, no. 11, pages 1 - 2 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021169464A1 (en) * | 2020-02-28 | 2021-09-02 | 深圳前海微众银行股份有限公司 | Method, apparatus and device for farm risk prediction, and storage medium |
CN116957342A (en) * | 2023-09-18 | 2023-10-27 | 深圳市博安智控科技有限公司 | Intelligent management method, system and medium for farm based on Internet of things |
CN116957342B (en) * | 2023-09-18 | 2023-12-08 | 深圳市博安智控科技有限公司 | Intelligent management method, system and medium for farm based on Internet of things |
CN117787736A (en) * | 2023-11-17 | 2024-03-29 | 新光瑞合(青岛)生物科技有限公司 | Construction method and system of healthy cultivation safety system |
Also Published As
Publication number | Publication date |
---|---|
WO2021169464A1 (en) | 2021-09-02 |
CN111339237B (en) | 2024-07-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111339237B (en) | Farm risk prediction method, device, equipment and storage medium | |
Pontius et al. | Recommendations for using the relative operating characteristic (ROC) | |
Kang et al. | Making the most of spatial information in health: a tutorial in Bayesian disease mapping for areal data | |
CN111339965A (en) | Farm site selection method, farm site selection device, farm site selection equipment and computer readable storage medium | |
CN116304600B (en) | Foreign invasive species early warning method and system based on big data analysis | |
CN111048214A (en) | Early warning method and device for spreading situation of foreign livestock and poultry epidemic diseases | |
CN111126720A (en) | Farm risk prediction method, device, equipment and storage medium | |
CN116362359A (en) | User satisfaction prediction method, device, equipment and medium based on AI big data | |
Hadorn et al. | Establishing a cost-effective national surveillance system for Bluetongue using scenario tree modelling | |
CN109121133B (en) | Location privacy protection method and device | |
Wardius et al. | Terrestrial laser scanning vs. manual methods for assessing complex forest stand structure: a comparative analysis on plenter forests | |
Chadœuf et al. | Modelling unbiased dispersal kernels over continuous space by accounting for spatial heterogeneity in marking and observation efforts | |
CN113408374A (en) | Yield estimation method, device and equipment based on artificial intelligence and storage medium | |
CN115063021B (en) | Method, system, equipment and medium for identifying influence of reservoir dam engineering on environment | |
Ye et al. | A wavelet‐based approach to evaluate the roles of structural and functional landscape heterogeneity in animal space use at multiple scales | |
CN110808947B (en) | Automatic vulnerability quantitative evaluation method and system | |
CN116384844A (en) | Decision method and device based on geographic information cloud platform | |
CN111404835A (en) | Flow control method, device, equipment and storage medium | |
CN114580979A (en) | High-temperature disaster prevention index detection method, device, equipment, storage medium and product | |
CN114648688A (en) | Method, system and equipment for evaluating landscape level along high-speed rail and readable storage medium | |
Miller et al. | Assessing Ecological Divergence and Speciation Scenarios of the Paragalago zanzibaricus Species Complex Through Climatic Niche Modeling | |
CN114493346A (en) | Rural industry centralized layout method, system, device and storage medium | |
CN113849773A (en) | Method and system for quantifying relation between aerosol optical thickness and landscape index | |
Sellman et al. | Realistic assumptions about spatial locations and clustering of premises matter for models of foot-and-mouth disease spread in the United States | |
CN110569475A (en) | Evaluation method, device, equipment and storage medium for netizen influence |
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 |