CN115796366B - Forest pest prediction method and pest prediction forest map system - Google Patents
Forest pest prediction method and pest prediction forest map system Download PDFInfo
- Publication number
- CN115796366B CN115796366B CN202211537475.5A CN202211537475A CN115796366B CN 115796366 B CN115796366 B CN 115796366B CN 202211537475 A CN202211537475 A CN 202211537475A CN 115796366 B CN115796366 B CN 115796366B
- Authority
- CN
- China
- Prior art keywords
- pest
- data
- layer
- forest
- information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 241000607479 Yersinia pestis Species 0.000 title claims abstract description 289
- 238000000034 method Methods 0.000 title claims abstract description 59
- 238000012544 monitoring process Methods 0.000 claims abstract description 59
- 238000011161 development Methods 0.000 claims abstract description 19
- 230000007613 environmental effect Effects 0.000 claims abstract description 19
- 238000013500 data storage Methods 0.000 claims abstract description 15
- 241000238631 Hexapoda Species 0.000 claims description 62
- 238000012549 training Methods 0.000 claims description 39
- 230000006870 function Effects 0.000 claims description 36
- 230000018109 developmental process Effects 0.000 claims description 18
- 238000010586 diagram Methods 0.000 claims description 14
- 239000002689 soil Substances 0.000 claims description 12
- 230000008447 perception Effects 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 10
- 238000013528 artificial neural network Methods 0.000 claims description 9
- 230000005484 gravity Effects 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 8
- 230000000007 visual effect Effects 0.000 claims description 8
- 238000003916 acid precipitation Methods 0.000 claims description 4
- 238000004891 communication Methods 0.000 claims description 4
- 230000000694 effects Effects 0.000 claims description 4
- 239000011159 matrix material Substances 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 4
- 238000010792 warming Methods 0.000 claims description 4
- 241000894006 Bacteria Species 0.000 claims description 3
- 241000233866 Fungi Species 0.000 claims description 3
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 244000052769 pathogen Species 0.000 claims description 3
- 239000002023 wood Substances 0.000 claims description 3
- 230000001418 larval effect Effects 0.000 claims 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims 1
- 230000008859 change Effects 0.000 abstract description 9
- 238000003745 diagnosis Methods 0.000 abstract description 4
- 230000036541 health Effects 0.000 abstract description 4
- 238000012545 processing Methods 0.000 abstract description 4
- 238000005516 engineering process Methods 0.000 description 10
- 238000007726 management method Methods 0.000 description 6
- 230000015654 memory Effects 0.000 description 5
- 201000010099 disease Diseases 0.000 description 4
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 4
- 238000005286 illumination Methods 0.000 description 4
- 238000009792 diffusion process Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000009395 breeding Methods 0.000 description 1
- 230000001488 breeding effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 230000004083 survival effect Effects 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of image data processing, in particular to a forest pest prediction method and a pest prediction forest map system, wherein the method comprises the following steps: acquiring and displaying geographic information data based on a forest geographic information layer; acquiring and displaying environmental characteristic data based on the environmental characteristic information layer; acquiring and displaying pest characteristic data based on the pest characteristic information layer; based on the sensing node distribution map layer, initializing global monitoring and displaying are carried out on a forest pest monitoring area, key monitoring and displaying are carried out on a pest high-incidence area, and mobile information of mobile sensing nodes and acquired data are acquired and displayed; based on the pest trend prediction layer, based on the improved deep confidence model, predicting pest development trend and geographic position; based on the data storage module, geographic information data, environmental characteristic data, pest characteristic data are stored. By adopting the method, the forest health diagnosis and the monitoring and early warning of ecological diversity change can be realized.
Description
Technical Field
The invention relates to the technical field of image data processing, in particular to a forest pest prediction method and a pest prediction forest map system.
Background
Forest pest control work is highly valued by China all the time, and because forest fine management can not only greatly improve effective circulation of forest resources, but also effectively guarantee ecological safety. Therefore, the accurate prediction of forest insect pests according to the sensing results of climate change, forest growth speed, insect pest breeding process and the like is widely paid attention to. The method aims at accurately predicting the development trend of the insect pests by means of modern high technologies such as a geographic information system, a sensor sensing technology and the like, fully playing the efficiency of the method, improving the forest fine management and comprehensively solving a plurality of problems existing in the forest insect pest management at present. The requirements on the existing pest prediction technical means are higher and higher due to the change of natural climate, the diversity of forest species and the increase of pest and disease damage types, and the accuracy of the forest pest and disease damage prediction can be ensured only on the premise of comprehensively mastering the forest basic information, the climate change trend and the pest and disease damage evolution process.
The natural climate assessment, pest reproduction process, etc. involved in modern pest prediction systems are all dependent on sensor data collection. However, due to the diversity of forest resources and the rapid spread of plant diseases and insect pests, the traditional and tedious pest prediction means can not meet the requirements of forest precise control. Therefore, the fusion of the internet of things perception of the ecosystem with digital twinning technology is being widely studied.
The enhanced twin map for forest pest prediction is a novel pest twin map with higher precision and richer data information compared with a common forest pest distribution system, and natural climate change, ecological diversity change, larva reproduction process, pest monitoring mode and the like can be presented in the digital twin system so as to meet the requirement of pest accurate prediction. To construct a complete pest prediction enhanced forest twin map, the reasonable design of a pest prediction model is a non-negligible problem.
However, most pest prediction data are isolated and not perfect at present, and the problems of redundancy, low precision, poor applicability and the like of the data structure exist, and even the information quantity of some pest perception data is small. These deficiencies severely restrict the improvement of pest prediction precision, and further form serious threat to the ecological environment, and influence the quality of forest resources.
Disclosure of Invention
The embodiment of the invention provides a forest pest prediction method and a pest prediction forest map system. The technical scheme is as follows:
in one aspect, a forest pest prediction method is provided, the method is realized based on a pest prediction forest map system, and the pest prediction forest map system comprises a forest geographic information layer, an environment characteristic information layer, a pest characteristic information layer, a perception node distribution diagram layer, a pest trend prediction layer and a data storage module;
the method comprises the following steps:
acquiring and displaying geographic information data based on the forest geographic information layer;
acquiring and displaying environmental characteristic data based on the environmental characteristic information layer, wherein the environmental characteristic data comprises natural climate information and soil information;
acquiring and displaying pest characteristic data based on the pest characteristic information layer, wherein the pest characteristic data comprises pest period distance and pest kind;
based on the sensing node distribution map layer, initializing global monitoring and displaying are carried out on a forest pest monitoring area, key monitoring and displaying are carried out on a pest high-incidence area, and mobile information of mobile sensing nodes and acquired data are acquired and displayed;
based on the pest trend prediction layer and an improved depth confidence model, predicting pest development trend and geographic position through geographic information data, environment characteristic data, pest characteristic data and data acquired by the mobile sensing nodes, and forming a visualized pest prediction map;
and storing geographic information data, environment characteristic data, insect pest characteristic data and data acquired by the mobile sensing node based on the data storage module.
Optionally, the natural climate information comprises the temperature, humidity, atmospheric pollution, acid rain, climate warming, illumination and rainwater related information of the environment;
the soil information includes desertification land data and wetland data.
Optionally, the pest period distance refers to the time interval between pest periods, i.e. the number of days from development of pest larvae to the later period;
the insect pest species include bacteria, fungi and bacteria-like pathogens which damage wood.
Optionally, the initializing the forest pest monitoring area for global monitoring and displaying includes:
and deploying mobile sensing nodes according to a space segmentation coverage algorithm to obtain pest distribution current information in a forest pest monitoring area.
Optionally, deploying mobile sensing nodes according to a space division coverage algorithm to obtain pest distribution current status information in a forest pest monitoring area, including:
s51, initializing the coverage radius, the communication radius and the insect pest monitoring area of the mobile sensing node;
s52, randomly disposing a plurality of mobile sensing nodes in the forest pest monitoring area;
s53, constructing a plurality of polyhedrons based on a space division algorithm, and calculating the center of gravity point (x i ,y i ,z i )；
S54, controlling the mobile sensing node to move to the gravity center point of each polyhedron;
s55, calculating coverage rate of the mobile sensing node, and executing a step S56 if the coverage of the whole forest pest monitoring area is preliminarily completed; otherwise, executing the step S53;
s56, outputting a result to finish pest initialization global monitoring.
Optionally, the monitoring and displaying the pest high-incidence area with emphasis includes:
simulating the superposition effect of a plurality of pest damage sources by using a Gaussian basis function based on a Gaussian estimation method, and obtaining a pest damage degree distribution map;
and determining a high-incidence area of the insect damage according to the insect damage degree distribution diagram, controlling the mobile sensing node to perform key coverage on the high-incidence area of the insect damage, and obtaining detailed data of the serious-incidence area of the insect damage.
Optionally, the gaussian estimation method simulates the superposition effect of a plurality of pest loss sources by using a gaussian basis function, and obtains a pest damage degree distribution map, including:
s71 using bounded and continuous density functionsEstimating the value of each point in the forest pest monitoring area, wherein +.>Beta represents a first weight coefficient as a bounded continuous basis function;
s72, adopting a group of three-dimensional Gaussian functions which are evenly distributed in a task area as basis functions, wherein each basis function is represented by the following formula (1):
wherein,,indicating the extent of influence of the pest source->Represents the lateral distance of the pest source,/-)>Represents the longitudinal distance of the pest source,/-)>Representing the vertical distance, x, of the pest source 0 The abscissa, y, representing the spatial center point of the basis function 0 Representing the ordinate, z, of the spatial center point of the basis function 0 Vertical coordinates of a spatial center point of the basis function are represented, x represents an abscissa of the pest source, y represents an ordinate of the pest source, and z represents a vertical coordinate of the pest source;
s73, obtaining an estimation result of a density function by using an adaptive parameter adjustment method by adopting the following formula (2):
wherein,,representing the result of the estimation of the density function, T representing the matrix transpose,/->Representing a second weight coefficient;
s74, if the estimated result of the density function of each point in the forest pest monitoring area is satisfied Determining the estimation result as a global optimal estimation;
and S75, determining a pest damage degree distribution map according to the global optimal estimation.
Optionally, the predicting pest development trend and geographic location based on the improved deep confidence model through geographic information data, environmental characteristic data, pest characteristic data and data collected by the mobile sensing node, and forming a visualized pest prediction map includes:
forming the environmental characteristic data into an original data set;
according to the original data set, a forest environment information database which is regular with insect pest occurrence is established, and data in the forest environment information database are normalized to form a training data set and a testing data set;
constructing an initial improved deep confidence model, wherein the initial improved deep confidence model consists of three layers of unsupervised limited boltzmann machines and 1 back propagation neural network;
training the initial improved deep confidence model based on the training data set to obtain a trained improved deep confidence model;
inputting the test data set into the trained improved deep confidence model, and calculating the predicted result of the pest development trend;
and forming a visual pest prediction map according to the pest prediction result.
Optionally, training the initial improved deep confidence model based on the training data set to obtain a trained improved deep confidence model, including:
inputting the training data set into a first layer of unsupervised limited Boltzmann machine for unsupervised training, and adding insect pest priori knowledge P in the training process i Obtaining weight parameters by using the obtained weight parameters as the input of a next-layer unsupervised limited boltzmann machine, and the like, extracting pest depth characteristics layer by layer to obtain an initialized parameter model;
adding a label layer to the top layer through forward propagation, obtaining model parameters through unsupervised learning, and using reverse propagation to propagate errors from top to bottom to each layer of unsupervised limited boltzmann machine, and obtaining second pest characteristics from the training data set;
acquiring output characteristics corresponding to the training data set through a back propagation neural network arranged behind the third-layer unsupervised limited boltzmann machine;
according to the output characteristics corresponding to the training data set, adjusting parameters in the initial improved depth confidence model, and iterating;
and stopping iteration when the iteration times reach the preset times, and determining the trained improved depth confidence model.
In another aspect, a pest prediction forest map system is provided, the system including a forest geographic information layer, an environmental characteristic information layer, a pest characteristic information layer, a perception node distribution map layer, a pest trend prediction layer, and a data storage module; wherein:
the forest geographic information layer is used for acquiring and displaying geographic information data;
the environment characteristic information layer is used for acquiring and displaying environment characteristic data, wherein the environment characteristic data comprises natural climate information and soil information;
based on the pest characteristic information layer, the pest characteristic information layer is used for acquiring and displaying pest characteristic data, wherein the pest characteristic data comprises pest period distance and pest species;
the sensing node distribution map layer is used for carrying out initialized global monitoring and displaying on a forest pest monitoring area, carrying out key monitoring and displaying on a pest high-incidence area, and obtaining and displaying mobile information of mobile sensing nodes and acquired data;
the pest trend prediction layer is used for predicting pest development trend and geographic position based on the improved deep confidence model through geographic information data, environment characteristic data, pest characteristic data and data acquired by the mobile sensing nodes, and forming a visual pest prediction map;
the data storage module is used for storing geographic information data, environment characteristic data, insect pest characteristic data and data acquired by the mobile sensing node.
In another aspect, an electronic device is provided that includes a processor and a memory having at least one instruction stored therein that is loaded and executed by the processor to implement the forest pest prediction method described above.
In another aspect, a computer readable storage medium having stored therein at least one instruction that is loaded and executed by a processor to implement the forest pest prediction method described above is provided.
The technical scheme provided by the embodiment of the invention has the beneficial effects that at least:
according to the embodiment of the invention, the existing sensor sensing technology is used as a basis, and the insect pest prediction enhanced forest twin map is formed by combining the digital twin technology with the requirements of accurate forest insect pest prediction. The enhanced insect pest prediction twin map mainly comprises five layers of a forest geographic information layer, an environment characteristic information layer, an insect pest characteristic information layer, a perception node distribution diagram layer, an insect pest trend prediction layer and a data storage module. The enhanced insect pest prediction twin map provided by the invention has the advantages of high accuracy, full information quantity, intelligent management, real-time feedback and the like, and can realize monitoring and early warning of forest health diagnosis and ecological diversity change to a great extent, thereby reducing forest insect pest diffusion and improving forest ecological environment.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a forest pest prediction method provided by an embodiment of the invention;
FIG. 2 is a schematic illustration of a polyhedron provided in accordance with an embodiment of the present invention;
FIG. 3 is a schematic illustration of an initial improved deep confidence model provided by an embodiment of the present invention;
FIG. 4 is a block diagram of a pest-predicted forest map system provided by an embodiment of the present invention;
fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the technical problems, technical solutions and advantages to be solved more apparent, the following detailed description will be given with reference to the accompanying drawings and specific embodiments.
The embodiment of the invention provides a forest pest prediction method which can be realized by a pest prediction forest map system. As shown in the flow chart of the forest pest prediction method in fig. 1, the process flow of the method can comprise the following steps:
s1, acquiring and displaying geographic information data based on a forest geographic information layer.
In a possible implementation mode, the forest resource geographic position is acquired based on the existing geographic information system technology, and is organized into geographic information data and displayed. The obtaining and displaying of the geographic information data may be performed by a method commonly used in the prior art, which is not described in detail in the embodiments of the present invention.
S2, acquiring and displaying the environmental characteristic data based on the environmental characteristic information layer.
The environment characteristic data comprises natural climate information and soil information.
The natural climate information is used for reflecting the influence of natural climate factors on pest occurrence in a forest system, and the natural phenomena such as temperature and humidity, atmospheric pollution, acid rain, climate warming, illumination, rainwater and the like of the environment have different influences on pest propagation, so the natural climate information can comprise the related information such as temperature, humidity, atmospheric pollution, acid rain, climate warming, illumination, rainwater and the like of the environment.
The soil information is used for analyzing the influence of different types of soil environments on the development process of the forest tree and the influence of insect pest survival, and can comprise desertification land data, wetland data and the like.
In a possible implementation manner, the acquiring and displaying of the environmental characteristic data may be performed by a method commonly used in the prior art, and the embodiments of the present invention are not described herein.
S3, based on the pest characteristic information layer, obtaining and displaying pest characteristic data, wherein the pest characteristic data comprises pest period distance and pest kind.
The pest period distance refers to the time interval between pest periods, namely the days from the development of pest larvae to the later period, and the threats of pests in different periods to the attack of trees are different.
Insect pest species include bacteria, fungi, and bacteria-like pathogens that damage wood.
In a possible implementation manner, the pest characteristic data may be obtained and displayed by a method commonly used in the prior art, and the embodiments of the present invention will not be described herein.
S4, based on the sensing node distribution map layer, initializing global monitoring and displaying are conducted on the forest pest monitoring area, key monitoring and displaying are conducted on the pest high-incidence area, and movement information of the mobile sensing nodes and collected data are obtained and displayed.
In a possible embodiment, S4 may comprise the following steps S41-S42:
s41, initializing a forest pest monitoring area for global monitoring and displaying, wherein the method comprises the following steps:
and deploying mobile sensing nodes according to a space segmentation coverage algorithm to obtain pest distribution current information in a forest pest monitoring area.
The method specifically comprises the following steps:
s411, initializing the coverage radius, the communication radius and the pest monitoring area size of the mobile sensing node.
In a possible implementation, the coverage radius R of the mobile awareness node is initialized a Radius of communication R c ＝2R a And pest monitoring area size a×b×c.
S412, randomly disposing a plurality of mobile sensing nodes in the forest pest monitoring area.
In a possible implementation manner, n mobile sensing nodes s= { S are randomly deployed in the pest monitoring area 1 ,…,S k ,…S n }。
S413, constructing a polyhedron based on a space division algorithm, and calculating the center of gravity point (x i ,y i ,z i )。
In one possible embodiment, a plurality of polyhedrons are constructed based on a spatial segmentation algorithm, as shown in FIG. 2 as an exemplary, block of { D } 1 ,D 2 ,D 3 ,D 4 ,D 5 ,D 6 A polyhedron which may be an octahedron, and then calculating the center of gravity point (x i ,y i ,z i ) Wherein D is 1 -D 6 The vertex of the polyhedron is represented, and Q represents the center of gravity of the polyhedron.
And S414, controlling the movement sensing node to move towards the center of gravity point of the polyhedron.
S415, calculating coverage rate of the mobile sensing node, and executing S416 if the coverage of the whole forest pest monitoring area is preliminarily completed. Otherwise, step S413 is performed.
It should be noted that, the method for calculating the coverage rate of the mobile sensing node may be a calculation method commonly used in the prior art, which is not described in detail in the embodiment of the present invention.
S416, outputting a result to complete pest initialization global monitoring.
In a possible implementation, the information monitored by each mobile sensing node is output.
S42, carrying out key monitoring and displaying on the high-incidence areas of the insect pests, wherein the method comprises the following steps:
s421, simulating the superposition effect of a plurality of pest damage sources by using a Gaussian basis function based on a Gaussian estimation method, and obtaining a pest damage degree distribution map.
In a possible embodiment, a bounded and continuous density function is usedEstimating the value of the center of gravity of each polyhedron of the pest monitoring area A×B×C, wherein ∈> For a bounded continuous basis function, T represents the matrix transpose and β is the first weight coefficient.
A set of three-dimensional gaussian functions evenly distributed over the task area is used as basis functions, where each basis function is expressed as:
in the formula (1)Indicating the extent of influence of the pest source (x) 0 ,y 0 ,z 0 ) Is the spatial center point coordinates of the basis function, < +.>Represents the lateral distance of the pest source,/-)>Represents the longitudinal distance of the pest source,/-)>Representing the vertical distance, x, of the pest source 0 The abscissa, y, representing the spatial center point of the basis function 0 Representing the ordinate, z, of the spatial center point of the basis function 0 The vertical coordinate of the spatial center point of the basis function is represented, x represents the horizontal coordinate of the pest source, y represents the vertical coordinate of the pest source, and z represents the vertical coordinate of the pest source.
Obtaining an estimated value of the density function by using an adaptive parameter adjustment method:wherein T represents the matrix transpose, ">Is a second weight coefficient;
in the pest monitoring area, if the density function estimation result satisfies at every point in the area And the estimation result is called as global optimal estimation, and the pest damage degree distribution map is determined according to the global optimal estimation.
S422, determining a high-incidence area of the insect damage according to the insect damage degree distribution diagram, controlling the mobile sensing node to perform key coverage on the high-incidence area of the insect damage, and obtaining detailed data of the serious-incidence area of the insect damage.
In one possible embodiment, the area where the pest damage level is greater than the preset threshold is determined as the pest high incidence area according to the pest damage level distribution map.
It should be noted that, the method for controlling the mobile sensing node to perform key coverage on the pest high-emission area and the method for obtaining detailed data of the pest serious area according to the mobile sensing node may be a method commonly used in the prior art, which is not described in detail in the embodiment of the present invention.
S5, based on the pest trend prediction layer and based on the improved deep confidence model, predicting the pest development trend and the geographic position through geographic information data, environment characteristic data, pest characteristic data and data acquired by the mobile sensing nodes, and forming a visual pest prediction map.
In a possible embodiment, the step S5 may specifically include the following steps S51 to S55:
s51, forming the environment characteristic data into an original data set.
In a possible embodiment, the environmental characteristic data (soil temperature, humidity, soil moisture, rain season, air humidity, illumination intensity, etc.) are formed into a raw data set E o 。
S52, establishing a forest environment information database according to the original data set and the occurrence rule of insect pests, and carrying out normalization processing on data in the forest environment information database to form a training data set and a testing data set.
S53, constructing an initial improved deep confidence model, wherein the initial improved deep confidence model consists of three layers of unsupervised limited Boltzmann machines and 1 back propagation neural network.
In a possible embodiment, as shown in fig. 3, first, the raw data D is preprocessed o Inputting into a first layer of unsupervised limited Boltzmann machine for unsupervised training, and adding insect pest priori knowledge P in the training process i Obtaining weight parameters by = { i=1, 2,3}, taking the output as input of a next-layer unsupervised limited boltzmann machine, repeatedly training in this way, extracting pest depth characteristics layer by layer, and obtaining an initialized parameter model;
then, adding a label layer on the top layer through forward propagation, obtaining model parameters through unsupervised learning, and using reverse propagation to propagate errors from top to bottom to each layer of unsupervised limited boltzmann machine, so as to obtain more abstract pest characteristics from the original data;
finally, setting a back propagation neural network behind the third-layer unsupervised limited boltzmann machine to obtain a final output characteristic; the top layer counter propagation neural network comprises an input layer, an hidden layer and an output layer and is used for predicting the pest development trend.
And S54, training the initial improved deep confidence model based on the training data set to obtain the trained improved deep confidence model.
S55, inputting the test data set into the trained improved deep confidence model, and calculating a predicted result of the pest development trend.
S56, forming a visual pest prediction map according to the pest prediction result.
According to the embodiment of the invention, the existing sensor sensing technology is used as a basis, and the insect pest prediction enhanced forest twin map is formed by combining the digital twin technology with the requirements of accurate forest insect pest prediction. The enhanced insect pest prediction twin map mainly comprises five layers of a forest geographic information layer, an environment characteristic information layer, an insect pest characteristic information layer, a perception node distribution diagram layer, an insect pest trend prediction layer and a data storage module. The enhanced insect pest prediction twin map provided by the invention has the advantages of high accuracy, full information quantity, intelligent management, real-time feedback and the like, and can realize monitoring and early warning of forest health diagnosis and ecological diversity change to a great extent, thereby reducing forest insect pest diffusion and improving forest ecological environment.
Fig. 4 is a diagram illustrating a pest prediction forest map system for implementing a forest pest prediction method according to an exemplary embodiment. As shown in fig. 4, the system comprises a forest geographic information layer, an environmental characteristic information layer, a pest characteristic information layer, a perception node distribution diagram layer, a pest trend prediction layer and a data storage module; wherein:
the forest geographic information layer is used for acquiring and displaying geographic information data;
the environment characteristic information layer is used for acquiring and displaying environment characteristic data, wherein the environment characteristic data comprises natural climate information and soil information;
the pest characteristic information layer is used for acquiring and displaying pest characteristic data, wherein the pest characteristic data comprises pest period distance and pest species;
the sensing node distribution map layer is used for carrying out initialization global monitoring and displaying on the forest pest monitoring area, carrying out key monitoring and displaying on the pest high-incidence area, and obtaining and displaying the mobile information of the mobile sensing nodes and the acquired data;
the pest trend prediction layer is used for predicting pest development trend and geographic position through geographic information data, environment characteristic data, pest characteristic data and data acquired by the mobile sensing nodes based on the improved deep confidence model, and forming a visual pest prediction map;
the data storage module is used for storing geographic information data, environment characteristic data, insect pest characteristic data and data acquired by the mobile sensing node.
According to the embodiment of the invention, the existing sensor sensing technology is used as a basis, and the insect pest prediction enhanced forest twin map is formed by combining the digital twin technology with the requirements of accurate forest insect pest prediction. The enhanced insect pest prediction twin map mainly comprises five layers of a forest geographic information layer, an environment characteristic information layer, an insect pest characteristic information layer, a perception node distribution diagram layer, an insect pest trend prediction layer and a data storage module. The enhanced insect pest prediction twin map provided by the invention has the advantages of high accuracy, full information quantity, intelligent management, real-time feedback and the like, and can realize monitoring and early warning of forest health diagnosis and ecological diversity change to a great extent, thereby reducing forest insect pest diffusion and improving forest ecological environment.
Fig. 5 is a schematic structural diagram of an electronic device 500 according to an embodiment of the present invention, where the electronic device 500 may have a relatively large difference due to different configurations or performances, and may include one or more processors (central processing units, CPU) 501 and one or more memories 502, where at least one instruction is stored in the memories 502, and the at least one instruction is loaded and executed by the processors 501 to implement the steps of the forest pest prediction method described above.
In an exemplary embodiment, a computer readable storage medium, such as a memory comprising instructions executable by a processor in a terminal to perform the above forest pest prediction method is also provided. For example, the computer readable storage medium may be ROM, random Access Memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program for instructing relevant hardware, where the program may be stored in a computer readable storage medium, and the storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
The foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and scope of the invention are intended to be included within the scope of the invention.
Claims (8)
1. The forest pest prediction method is characterized by being realized based on a pest prediction forest map system, wherein the pest prediction forest map system comprises a forest geographic information layer, an environment characteristic information layer, a pest characteristic information layer, a perception node distribution diagram layer, a pest trend prediction layer and a data storage module;
the method comprises the following steps:
acquiring and displaying geographic information data based on the forest geographic information layer;
acquiring and displaying environmental characteristic data based on the environmental characteristic information layer, wherein the environmental characteristic data comprises natural climate information and soil information;
acquiring and displaying pest characteristic data based on the pest characteristic information layer, wherein the pest characteristic data comprises pest period distance and pest kind;
based on the sensing node distribution map layer, initializing global monitoring and displaying are carried out on a forest pest monitoring area, key monitoring and displaying are carried out on a pest high-incidence area, and mobile information of mobile sensing nodes and acquired data are acquired and displayed;
based on the pest trend prediction layer and an improved depth confidence model, predicting pest development trend and geographic position through geographic information data, environment characteristic data, pest characteristic data and data acquired by the mobile sensing nodes, and forming a visualized pest prediction map;
based on the data storage module, geographic information data, environment characteristic data, insect pest characteristic data and data acquired by the mobile sensing node are stored;
the improved deep confidence model predicts the pest development trend and the geographic position through geographic information data, environment characteristic data, pest characteristic data and data acquired by the mobile sensing node, and forms a visualized pest prediction map, and the method comprises the following steps:
forming the environmental characteristic data into an original data set;
according to the original data set, a forest environment information database which is regular with insect pest occurrence is established, and data in the forest environment information database are normalized to form a training data set and a testing data set;
constructing an initial improved deep confidence model, wherein the initial improved deep confidence model consists of three layers of unsupervised limited boltzmann machines and 1 back propagation neural network;
training the initial improved deep confidence model based on the training data set to obtain a trained improved deep confidence model;
inputting the test data set into the trained improved deep confidence model, and calculating the predicted result of the pest development trend;
forming a visual insect pest prediction map according to the insect pest prediction result;
the training of the initial improved deep confidence model based on the training data set to obtain a trained improved deep confidence model comprises the following steps:
inputting the training data set into a first layer of unsupervised limited Boltzmann machine for unsupervised training, and adding insect pest priori knowledge P in the training process i Obtaining weight parameters by using the obtained weight parameters as the input of a next-layer unsupervised limited boltzmann machine, and the like, extracting pest depth characteristics layer by layer to obtain an initialized parameter model;
adding a label layer to the top layer through forward propagation, obtaining model parameters through unsupervised learning, and using reverse propagation to propagate errors from top to bottom to each layer of unsupervised limited boltzmann machine, and obtaining second pest characteristics from the training data set;
acquiring output characteristics corresponding to the training data set through a back propagation neural network arranged behind the third-layer unsupervised limited boltzmann machine;
according to the output characteristics corresponding to the training data set, adjusting parameters in the initial improved depth confidence model, and iterating;
and stopping iteration when the iteration times reach the preset times, and determining the trained improved depth confidence model.
2. The method of claim 1, wherein the natural climate information comprises information related to temperature, humidity, atmospheric pollution, acid rain, climate warming, light, rain water of the environment;
the soil information includes desertification land data and wetland data.
3. The method of claim 1, wherein the pest period distance refers to a time interval between pest individual pest periods, i.e., a number of days the pest larval period develops to a later stage;
the insect pest species include bacteria, fungi and bacteria-like pathogens which damage wood.
4. The method of claim 1, wherein the initializing the forest pest monitoring area for global monitoring and display comprises:
and deploying mobile sensing nodes according to a space segmentation coverage algorithm to obtain pest distribution current information in a forest pest monitoring area.
5. The method of claim 4, wherein deploying mobile awareness nodes according to a spatial division coverage algorithm to obtain pest distribution presence information in a forest pest monitoring area comprises:
s51, initializing the coverage radius, the communication radius and the insect pest monitoring area of the mobile sensing node;
s52, randomly disposing a plurality of mobile sensing nodes in the forest pest monitoring area;
s53, constructing a plurality of polyhedrons based on a space division algorithm, and calculating the center of gravity point (x i ,y i ,z i )；
S54, controlling the mobile sensing node to move to the gravity center point of each polyhedron;
s55, calculating coverage rate of the mobile sensing node, and executing a step S56 if the coverage of the whole forest pest monitoring area is preliminarily completed; otherwise, executing the step S53;
s56, outputting a result to finish pest initialization global monitoring.
6. The method of claim 1, wherein the step of highlighting and displaying the high-incidence area of the pest comprises:
simulating the superposition effect of a plurality of pest damage sources by using a Gaussian basis function based on a Gaussian estimation method, and obtaining a pest damage degree distribution map;
and determining a high-incidence area of the insect damage according to the insect damage degree distribution diagram, controlling the mobile sensing node to perform key coverage on the high-incidence area of the insect damage, and obtaining detailed data of the serious-incidence area of the insect damage.
7. The method of claim 6, wherein simulating a superposition of multiple pest loss sources from a gaussian basis function based on a gaussian estimation method to obtain a pest damage level distribution map comprises:
s71 using bounded and continuous density functionsEstimating the value of each point in the forest pest monitoring area, wherein +.>Beta represents a first weight coefficient as a bounded continuous basis function;
s72, adopting a group of three-dimensional Gaussian functions which are evenly distributed in a task area as basis functions, wherein each basis function is represented by the following formula (1):
wherein,,indicating the extent of influence of the pest source->Represents the lateral distance of the pest source,/-)>Represents the longitudinal distance of the pest source,/-)>Representing the vertical distance, x, of the pest source 0 The abscissa, y, representing the spatial center point of the basis function 0 Representing the ordinate, z, of the spatial center point of the basis function 0 Vertical coordinates of spatial center points representing basis functions, x represents horizontal coordinates of pest sources, yRepresenting the ordinate of the pest source, z representing the ordinate of the pest source;
s73, obtaining an estimation result of a density function by using an adaptive parameter adjustment method by adopting the following formula (2):
wherein,,representing the result of the estimation of the density function, T representing the matrix transpose,/->Representing a second weight coefficient;
s74, if the estimated result of the density function of each point in the forest pest monitoring area is satisfied Determining the estimation result as a global optimal estimation;
and S75, determining a pest damage degree distribution map according to the global optimal estimation.
8. The system is characterized by comprising a forest geographic information layer, an environment characteristic information layer, a pest characteristic information layer, a perception node distribution diagram layer, a pest trend prediction layer and a data storage module; wherein:
the forest geographic information layer is used for acquiring and displaying geographic information data;
the environment characteristic information layer is used for acquiring and displaying environment characteristic data, wherein the environment characteristic data comprises natural climate information and soil information;
based on the pest characteristic information layer, the pest characteristic information layer is used for acquiring and displaying pest characteristic data, wherein the pest characteristic data comprises pest period distance and pest species;
the sensing node distribution map layer is used for carrying out initialized global monitoring and displaying on a forest pest monitoring area, carrying out key monitoring and displaying on a pest high-incidence area, and obtaining and displaying mobile information of mobile sensing nodes and acquired data;
the pest trend prediction layer is used for predicting pest development trend and geographic position based on the improved deep confidence model through geographic information data, environment characteristic data, pest characteristic data and data acquired by the mobile sensing nodes, and forming a visual pest prediction map;
the data storage module is used for storing geographic information data, environment characteristic data, insect pest characteristic data and data acquired by the mobile sensing node;
the improved deep confidence model predicts the pest development trend and the geographic position through geographic information data, environment characteristic data, pest characteristic data and data acquired by the mobile sensing node, and forms a visualized pest prediction map, and the method comprises the following steps:
forming the environmental characteristic data into an original data set;
according to the original data set, a forest environment information database which is regular with insect pest occurrence is established, and data in the forest environment information database are normalized to form a training data set and a testing data set;
constructing an initial improved deep confidence model, wherein the initial improved deep confidence model consists of three layers of unsupervised limited boltzmann machines and 1 back propagation neural network;
training the initial improved deep confidence model based on the training data set to obtain a trained improved deep confidence model;
inputting the test data set into the trained improved deep confidence model, and calculating the predicted result of the pest development trend;
forming a visual insect pest prediction map according to the insect pest prediction result;
the training of the initial improved deep confidence model based on the training data set to obtain a trained improved deep confidence model comprises the following steps:
inputting the training data set into a first layer of unsupervised limited Boltzmann machine for unsupervised training, and adding insect pest priori knowledge P in the training process i Obtaining weight parameters by using the obtained weight parameters as the input of a next-layer unsupervised limited boltzmann machine, and the like, extracting pest depth characteristics layer by layer to obtain an initialized parameter model;
adding a label layer to the top layer through forward propagation, obtaining model parameters through unsupervised learning, and using reverse propagation to propagate errors from top to bottom to each layer of unsupervised limited boltzmann machine, and obtaining second pest characteristics from the training data set;
acquiring output characteristics corresponding to the training data set through a back propagation neural network arranged behind the third-layer unsupervised limited boltzmann machine;
according to the output characteristics corresponding to the training data set, adjusting parameters in the initial improved depth confidence model, and iterating;
and stopping iteration when the iteration times reach the preset times, and determining the trained improved depth confidence model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211537475.5A CN115796366B (en) | 2022-12-02 | 2022-12-02 | Forest pest prediction method and pest prediction forest map system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211537475.5A CN115796366B (en) | 2022-12-02 | 2022-12-02 | Forest pest prediction method and pest prediction forest map system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115796366A CN115796366A (en) | 2023-03-14 |
| CN115796366B true CN115796366B (en) | 2023-08-01 |
Family
ID=85444922
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211537475.5A Active CN115796366B (en) | 2022-12-02 | 2022-12-02 | Forest pest prediction method and pest prediction forest map system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115796366B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117390110B (en) * | 2023-12-13 | 2024-02-06 | 四川省林业科学研究院 | Method and system for generating electronic map of plant diseases and insect pests |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020016676A1 (en) * | 2000-07-26 | 2002-02-07 | Christopher Sann | Climatic conditions based pest control management |
| CN103604880B (en) * | 2013-11-07 | 2015-01-14 | 同济大学 | Method for detecting triazophos pesticide in environmental samples of different types |
| CN105631526A (en) * | 2014-10-31 | 2016-06-01 | 航天恒星科技有限公司 | Forest pest disease outbreak risk prediction method and device |
| CN114067311B (en) * | 2021-08-11 | 2024-08-02 | 中国农业科学院茶叶研究所 | Tea plant disease and pest prediction method based on intelligent identification of crowd source data |
| CN114170513B (en) * | 2021-12-08 | 2024-06-14 | 广东省农业科学院植物保护研究所 | Insect condition monitoring method, system and storage medium for spodoptera frugiperda |
| CN113963262B (en) * | 2021-12-20 | 2022-08-23 | 中国地质大学（武汉） | Mining area land coverage classification method, equipment, device and storage medium |
| CN115017704B (en) * | 2022-06-07 | 2024-09-06 | 同济大学 | Dynamic-static node coupling pest key area sensing coverage method and system |
-
2022
- 2022-12-02 CN CN202211537475.5A patent/CN115796366B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115796366A (en) | 2023-03-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kamdem et al. | Anthropogenic habitat disturbance and ecological divergence between incipient species of the malaria mosquito Anopheles gambiae | |
| Hopper et al. | Mate finding, dispersal, number released, and the success of biological control introductions | |
| Simons et al. | Model sensitivity and robustness in the estimation of larval transport: a study of particle tracking parameters | |
| Nanninga et al. | Not finding Nemo: limited reef-scale retention in a coral reef fish | |
| WO2018099220A1 (en) | Method and device for predicting plant diseases and pests in planter | |
| Bone et al. | A fuzzy-constrained cellular automata model of forest insect infestations | |
| CN115169750B (en) | Lepidoptera pest control method, system and medium based on Internet of things | |
| Pérez et al. | ForestSimMPB: A swarming intelligence and agent-based modeling approach for mountain pine beetle outbreaks | |
| CN115796366B (en) | Forest pest prediction method and pest prediction forest map system | |
| Mi et al. | Combining occurrence and abundance distribution models for the conservation of the Great Bustard | |
| Nnko et al. | Potential impacts of climate change on geographical distribution of three primary vectors of African Trypanosomiasis in Tanzania’s Maasai Steppe: G. m. morsitans, G. pallidipes and G. swynnertoni | |
| Tonini et al. | Simulating the spread of an invasive termite in an urban environment using a stochastic individual-based model | |
| Vogt et al. | Do canopy disturbances drive forest plantations into more natural conditions?—A case study from Can Gio Biosphere Reserve, Viet Nam | |
| Lee et al. | Using crowdsourced images to study selected cultural ecosystem services and their relationships with species richness and carbon sequestration | |
| JP2000089664A (en) | Land use planning method | |
| Pages et al. | Modelling root system growth and architecture | |
| Redhead et al. | E-Planner: A web-based tool for planning environmental enhancement on British agricultural land | |
| Gauffre et al. | Agricultural intensification alters marbled newt genetic diversity and gene flow through density and dispersal reduction | |
| Benestan et al. | Contrasting influence of seascape, space and marine reserves on genomic variation in multiple species | |
| Harris et al. | Functional connectivity and matrix quality: network analysis for a critically endangered New Zealand lizard | |
| Chamberlain et al. | ImageCLEFcoral task: coral reef image annotation and localisation | |
| CN108244003A (en) | A kind of aquatic products plant disease epidemic trend prediction and methods of exhibiting and system | |
| CN116307879A (en) | Efficient cultivation method, system and medium for penaeus monodon larvae | |
| CN115759346A (en) | AO algorithm based carbon emission prediction method, device and equipment | |
| Valdés et al. | Microclimate influences plant reproductive performance via an antagonistic interaction |
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 |