CN117935081B - Cultivated land change monitoring method and system based on remote sensing satellite data - Google Patents
Cultivated land change monitoring method and system based on remote sensing satellite data Download PDFInfo
- Publication number
- CN117935081B CN117935081B CN202410322861.5A CN202410322861A CN117935081B CN 117935081 B CN117935081 B CN 117935081B CN 202410322861 A CN202410322861 A CN 202410322861A CN 117935081 B CN117935081 B CN 117935081B
- Authority
- CN
- China
- Prior art keywords
- soil
- data
- analysis
- information
- cultivated
- 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
- 238000000034 method Methods 0.000 title claims abstract description 93
- 230000008859 change Effects 0.000 title claims abstract description 57
- 238000012544 monitoring process Methods 0.000 title claims abstract description 40
- 239000002689 soil Substances 0.000 claims abstract description 322
- 238000004458 analytical method Methods 0.000 claims abstract description 155
- 238000007726 management method Methods 0.000 claims abstract description 110
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 89
- 238000003066 decision tree Methods 0.000 claims abstract description 31
- 238000012706 support-vector machine Methods 0.000 claims abstract description 24
- 238000007637 random forest analysis Methods 0.000 claims abstract description 23
- 238000000513 principal component analysis Methods 0.000 claims abstract description 22
- 238000003064 k means clustering Methods 0.000 claims abstract description 19
- 238000013528 artificial neural network Methods 0.000 claims abstract description 18
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 18
- 230000035558 fertility Effects 0.000 claims abstract description 18
- 238000010183 spectrum analysis Methods 0.000 claims abstract description 18
- 238000012937 correction Methods 0.000 claims abstract description 17
- 238000005516 engineering process Methods 0.000 claims abstract description 16
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 16
- 201000010099 disease Diseases 0.000 claims abstract description 15
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims abstract description 15
- 230000005855 radiation Effects 0.000 claims abstract description 15
- 230000012010 growth Effects 0.000 claims description 93
- 238000009826 distribution Methods 0.000 claims description 55
- 238000011156 evaluation Methods 0.000 claims description 53
- 230000036541 health Effects 0.000 claims description 50
- 238000001228 spectrum Methods 0.000 claims description 33
- 230000003595 spectral effect Effects 0.000 claims description 31
- 230000000694 effects Effects 0.000 claims description 26
- 238000002310 reflectometry Methods 0.000 claims description 25
- 238000004162 soil erosion Methods 0.000 claims description 24
- 230000002262 irrigation Effects 0.000 claims description 20
- 238000003973 irrigation Methods 0.000 claims description 20
- 241000607479 Yersinia pestis Species 0.000 claims description 18
- 238000010521 absorption reaction Methods 0.000 claims description 18
- 238000000605 extraction Methods 0.000 claims description 17
- 238000012549 training Methods 0.000 claims description 17
- 230000004720 fertilization Effects 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 16
- 238000005457 optimization Methods 0.000 claims description 15
- 230000015654 memory Effects 0.000 claims description 14
- 239000000126 substance Substances 0.000 claims description 13
- 239000002028 Biomass Substances 0.000 claims description 11
- 229930002875 chlorophyll Natural products 0.000 claims description 11
- 235000019804 chlorophyll Nutrition 0.000 claims description 11
- ATNHDLDRLWWWCB-AENOIHSZSA-M chlorophyll a Chemical compound C1([C@@H](C(=O)OC)C(=O)C2=C3C)=C2N2C3=CC(C(CC)=C3C)=[N+]4C3=CC3=C(C=C)C(C)=C5N3[Mg-2]42[N+]2=C1[C@@H](CCC(=O)OC\C=C(/C)CCC[C@H](C)CCC[C@H](C)CCCC(C)C)[C@H](C)C2=C5 ATNHDLDRLWWWCB-AENOIHSZSA-M 0.000 claims description 11
- 230000009467 reduction Effects 0.000 claims description 9
- 230000006872 improvement Effects 0.000 claims description 8
- 239000005416 organic matter Substances 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 claims description 8
- 238000001514 detection method Methods 0.000 claims description 7
- 238000007477 logistic regression Methods 0.000 claims description 6
- 238000011084 recovery Methods 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 230000003628 erosive effect Effects 0.000 claims description 5
- 238000012732 spatial analysis Methods 0.000 claims description 5
- 230000005540 biological transmission Effects 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000013210 evaluation model Methods 0.000 claims description 4
- 238000000556 factor analysis Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 4
- 239000011159 matrix material Substances 0.000 claims description 4
- 230000009897 systematic effect Effects 0.000 claims description 4
- 238000011217 control strategy Methods 0.000 claims description 3
- 238000013136 deep learning model Methods 0.000 claims description 3
- 230000009528 severe injury Effects 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 230000009525 mild injury Effects 0.000 claims description 2
- 230000006806 disease prevention Effects 0.000 claims 1
- 230000007787 long-term memory Effects 0.000 abstract description 4
- 230000006403 short-term memory Effects 0.000 abstract description 3
- 238000012271 agricultural production Methods 0.000 description 10
- 238000013468 resource allocation Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 3
- 238000007619 statistical method Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000010485 coping Effects 0.000 description 2
- 230000006378 damage Effects 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 230000001737 promoting effect Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000034373 developmental growth involved in morphogenesis Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000003337 fertilizer Substances 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000005802 health problem Effects 0.000 description 1
- 230000003862 health status Effects 0.000 description 1
- 238000010921 in-depth analysis Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Abstract
The invention relates to the technical field of agricultural monitoring, in particular to a farmland change monitoring method and system based on remote sensing satellite data, and the method comprises the following steps: based on remote sensing satellite data, a K-means clustering algorithm is adopted, and the processed graphic information comprising radiation correction and atmosphere correction is combined to divide land types into cultivated land and non-cultivated land, so that cultivated land area division information is generated. In the invention, a cultivated area is identified by combining a K-means clustering algorithm and a remote sensing technology, a support vector machine and spectrum analysis are used for evaluating soil fertility, pH value and organic matters, a long-term and short-term memory network and a normalized difference water index enhancement model are combined for monitoring soil humidity, early diseases are evaluated by a random forest algorithm, disaster influence is analyzed by a convolutional neural network and a decision tree algorithm, soil variability is revealed by a Kriging interpolation method and principal component analysis, and a multidimensional strategy is customized for cultivated area resource management by combining application of a multi-standard decision analysis and an artificial neural network.
Description
Technical Field
The invention relates to the technical field of agricultural monitoring, in particular to a farmland change monitoring method and system based on remote sensing satellite data.
Background
The technical field of agricultural monitoring is focused on utilizing remote sensing technology, a geographic information system, satellite image data and a data analysis method to monitor and evaluate the agricultural environment and the growth condition of crops in real time, and by analyzing batch data acquired from satellite remote sensing equipment, the agricultural monitoring technology can provide precious information about land coverage, soil humidity, crop health condition and changes thereof, so that decision makers, agricultural producers and environment managers are assisted in making decisions based on data, optimizing resource allocation, improving agricultural production efficiency and protecting the environment.
The method for monitoring the change of the cultivated land based on the remote sensing satellite data accurately monitors the change condition of the cultivated land by analyzing the remote sensing satellite data, and comprises the expansion or reduction of the cultivated land area, the change of the crop planting condition and the change of the using mode of the cultivated land. The method provides a high-efficiency and accurate monitoring tool for supporting the formulation of agricultural policies, the management of land resources and the sustainable development of agriculture, and improves the response speed and the processing capacity to the ecological change of agriculture through continuous monitoring of the change of cultivated land, thereby promoting the reasonable utilization of resources and the scientific management of agricultural production.
The traditional farmland change monitoring method is dependent on a single technology or algorithm in actual operation, lacks deep analysis and comprehensive utilization of data, limits the accuracy and coping capacity of monitoring, is difficult to process large-scale remote sensing data, causes insufficient comprehensive soil characteristic analysis, influences the accuracy of soil management, depends on visual evaluation or simple remote sensing data analysis, fails to realize early identification and fine management of diseases, and the defect of disaster influence evaluation limits the timeliness and effectiveness of disaster coping measures, causes poor resource allocation, prevents improvement of agricultural production efficiency and has insufficient environmental protection.
Disclosure of Invention
The invention aims to overcome the defects of the traditional technology and provides a farmland change monitoring method and system based on remote sensing satellite data.
The aim of the invention is achieved by the following technical measures:
A farmland change monitoring method based on remote sensing satellite data comprises the following steps:
S1: based on remote sensing satellite data, adopting a K-means clustering algorithm, combining the processed graphic information including radiation correction and atmospheric correction, dividing land types into cultivated land and non-cultivated land, and generating cultivated land area division information;
S2: based on the ploughing area division information, a spectrum analysis method is adopted, spectrum information in a remote sensing image and a soil spectrum database are combined, characteristic values including reflectivity and absorption characteristics are extracted from the spectrum data, soil characteristic analysis including soil fertility, pH value and organic matters is carried out through a support vector machine algorithm, soil types are classified, and ploughing quality basic information is generated;
S3: based on the farmland coverage area division information, a long-period and short-period memory network is adopted, the change trend of soil humidity is analyzed through time sequence data, and drought and overwet areas are identified by combining a normalized difference moisture index enhancement model to generate farmland humidity distribution information;
s4: based on the farmland coverage area division information, adopting a random forest algorithm to analyze remote sensing data in a crop growth period, identifying various indexes of crop health conditions, including chlorophyll content and biomass, providing multidimensional assessment of crop growth states, including early disease detection, and generating a crop growth state assessment result;
s5: based on the cultivated area division information and the crop health monitoring information, analyzing disaster influence, including influence of the disaster on crop coverage and soil erosion conditions, by combining pre-disaster and post-disaster remote sensing image data through a convolutional neural network and a decision tree algorithm, and generating a disaster influence analysis result;
s6: estimating soil characteristics of an un-sampled point by adopting a kriging interpolation method based on the tilling quality basic information, carrying out spatial analysis on the physical and chemical characteristics of the soil by combining principal component analysis, identifying key factors affecting the quality of the soil, revealing the spatial distribution and variability of the soil characteristics, and generating a soil characteristic analysis result;
S7: based on the farmland humidity distribution information, the crop growth state evaluation result, the disaster influence analysis result and the soil characteristic analysis result, a plurality of agricultural management activities including irrigation plans, fertilization schemes and pest control measures are analyzed by adopting a multi-standard decision analysis and an artificial neural network, a farmland management strategy is identified through training data, and a management strategy conforming to the farmland condition is matched to generate a farmland resource management scheme.
As a further aspect of the present invention, the cultivated area division information includes cultivated area boundary coordinates, cultivated area total area and non-cultivated area characteristics, the cultivated area quality basic information includes soil type classification results, key fertility index values and soil improvement schemes, the cultivated area humidity distribution information includes drought risk areas, over-wet risk areas and normal moisture areas, the crop growth state evaluation results include healthy crop areas, affected crop areas and growth trend prediction information, the disaster effect analysis results include affected crop areas, soil erosion degrees and estimated recovery times, the soil characteristic analysis results include soil physical characteristic profiles, chemical characteristic profiles and soil improvement priority areas, and the cultivated area resource management schemes include irrigation optimization schemes, fertilization optimization schemes and pest control strategies.
As a further scheme of the invention, based on remote sensing satellite data, a K-means clustering algorithm is adopted, the processed graphic information comprising radiation correction and atmosphere correction is combined, the land types are divided into cultivated lands and non-cultivated lands, and the step of generating cultivated land area division information comprises the following steps:
S101: based on remote sensing satellite data, correcting the systematic deviation of the sensor by adopting an absolute radiometric calibration method, eliminating the influence of atmospheric scattering and absorption, including the scattering effect of fog and cloud by using an MODTRA model, and generating a corrected remote sensing image;
s102: based on the corrected remote sensing image, analyzing image data by adopting a K-means clustering algorithm, setting the clustering quantity, dividing land coverage types including water bodies, buildings and altitudes, and generating earth surface type identification information;
S103: based on the earth surface type identification information, a support vector machine algorithm is adopted to identify cultivated land and non-cultivated land areas, including forest, water area, grassland and cultivated land types, and the cultivated land area division information is generated by combining cadastral data and land use records, verifying the identification result.
As a further scheme of the invention, based on the cultivated area division information, a spectrum analysis method is adopted, spectrum information in a remote sensing image and a soil spectrum database are combined, characteristic values including reflectivity and absorption characteristics are extracted from the spectrum data, soil characteristic analysis including soil fertility, pH value and organic matters is performed through a support vector machine algorithm, the soil type is classified, and the step of generating cultivated quality basic information is specifically as follows:
S201: based on the ploughing area division information, extracting spectral characteristics and spectral information in a soil spectral database for comparison by adopting a spectral analysis method through remote sensing image information, wherein the spectral characteristics comprise reflectivity and absorption characteristics, and identifying spectral characteristic values of various soil types to generate soil spectral characteristic data;
S202: based on the soil spectrum characteristic data, a support vector machine algorithm is adopted, the extracted characteristic values are input into a support vector machine model, soil characteristic analysis is carried out, various soil characteristics including soil fertility, pH value and organic matters are identified, and a soil characteristic classification result is generated;
S203: based on the soil characteristic classification result, a classification decision process is adopted, soil types in the region are classified by analyzing soil characteristic data of the cultivated region, the soil cultivated land quality in the region is identified, and cultivated soil quality basic information is generated.
As a further scheme of the invention, based on the farmland coverage area division information, a long-period memory network is adopted, the change trend of soil humidity is analyzed through time sequence data, and drought and over-wet areas are identified by combining a normalized difference water index enhancement model, so that farmland humidity distribution information is generated specifically by the steps of:
S301: based on the farmland coverage area division information, a long-period memory network is adopted, and the time sequence remote sensing data of the corresponding farmland area are combined to simulate and analyze the change trend of soil humidity along with time so as to generate soil humidity change trend data;
S302: based on the soil humidity change trend data, applying a normalized difference moisture index enhancement model, analyzing the soil humidity at a plurality of time points by comparing the reflectivity differences of the differential wavelengths, identifying the moisture conditions of soil and vegetation, and generating soil humidity analysis data;
S303: based on the soil humidity analysis data, a threshold analysis method and a data classification technology are adopted, drought, normal and over-wet soil humidity conditions are distinguished by setting humidity thresholds, soil humidity states of a plurality of areas are identified by comparing soil humidity indexes with preset drought and over-wet thresholds, and cultivated land humidity distribution information is generated.
As a further scheme of the invention, based on the farmland coverage area division information, a random forest algorithm is adopted to analyze remote sensing data in a crop growth period, various indexes of crop health conditions are identified, including chlorophyll content and biomass, multidimensional assessment of crop growth states is provided, detection of early diseases is included, and the step of generating a crop growth state assessment result specifically comprises the following steps:
S401: based on the farmland coverage area division information, analyzing remote sensing data in a crop growth period by adopting a random forest algorithm, wherein the remote sensing data comprise reflectivity and spectrum indexes, and identifying key characteristics of crops, including vegetation indexes and moisture indexes, so as to generate crop growth characteristic data;
S402: based on the crop growth characteristic data, applying logistic regression analysis, analyzing the health state of crops according to crop growth key characteristics, including normal, mild damage and severe damage, evaluating the health state of crops, and generating crop health analysis information;
S403: based on the crop health analysis information, adopting decision tree analysis and principal component analysis, combining a plurality of health indexes, analyzing the crop growth state, obtaining crop health scores, identifying the overall growth condition of crops and the farmland problem area, including early diseases, and generating a crop growth state evaluation result.
As a further scheme of the invention, based on the cultivated area division information and the crop health monitoring information, the disaster influence is analyzed by combining the pre-disaster post-disaster remote sensing image data through a convolutional neural network and a decision tree algorithm, and the steps of generating disaster influence analysis results comprise:
s501: based on the cultivated area division information and the crop health monitoring information, performing image feature extraction by adopting a convolutional neural network, identifying the influence of disasters on the change of crop coverage, and generating disaster feature extraction data;
s502: based on the disaster feature extraction data, analyzing and classifying factors influencing crop coverage by adopting a random forest algorithm, analyzing the influence of various disaster features on the crop coverage, including vegetation index change and soil exposure, identifying key influencing factors, and generating a crop coverage influencing factor analysis result;
S503: based on the analysis result of the crop coverage factor influence factors, a decision tree algorithm is adopted, the influence of disasters on soil erosion is estimated by analyzing the relevance between crop coverage reduction and soil erosion conditions, and the analysis result of the disasters influence is generated, wherein the influence comprises erosion degree and distribution areas.
As a further scheme of the invention, based on the basic information of the tilling quality, the method adopts a Kriging interpolation method to estimate the soil characteristics of the non-sampling points, combines the principal component analysis to carry out space analysis on the physical and chemical characteristics of the soil, identifies key factors influencing the quality of the soil, reveals the space distribution and variability of the soil characteristics, and specifically comprises the following steps of:
S601: based on the tilling quality basic information, a Kriging interpolation method is adopted, soil characteristics of an un-sampled position, including pH value, organic matter content and soil density, are estimated according to the space distance and similarity through existing soil sample data, and soil characteristic interpolation data are generated;
S602: based on the soil characteristic interpolation data, adopting principal component analysis, performing feature selection and feature extraction on the physicochemical characteristics of the soil by calculating a covariance matrix of the soil characteristic data, and identifying key features by combining contribution rates of various soil features to generate a soil key characteristic data set;
S603: based on the soil key characteristic data set, spatial mapping and trend analysis are carried out by using a geographic information system technology, the spatial distribution mode and influence factors of soil quality are analyzed, key factors influencing the soil quality are identified, the spatial distribution and variability of soil characteristics are obtained, and a soil characteristic analysis result is generated.
As a further scheme of the invention, based on the farmland humidity distribution information, the crop growth state evaluation result, the disaster influence analysis result and the soil characteristic analysis result, a plurality of agricultural management activities including irrigation plans, fertilization schemes and pest control measures are analyzed by adopting a multi-standard decision analysis and an artificial neural network, and the farmland management strategies are identified through training data, and the management strategies conforming to the farmland conditions are matched, so that the farmland resource management scheme is generated specifically by the steps of:
S701: based on the farmland humidity distribution information, the crop growth state evaluation result, the disaster influence analysis result and the soil characteristic analysis result, adopting a hierarchical analysis method, and optimizing the accuracy and reliability of evaluation by constructing a hierarchical structure of an evaluation standard, performing pairwise comparison calculation weight and consistency test, and generating a farmland management scheme evaluation result;
s702: based on the evaluation result of the cultivated land management scheme, an artificial neural network is adopted, and a multilayer perceptron model is trained and optimized by inputting quantitative indexes of irrigation, fertilization and pest control schemes, so as to generate a cultivated land management strategy set;
s703: based on the tilling management strategy set, a decision support system is adopted to evaluate the tilling management strategy effect, a comprehensive evaluation model is applied, the tilling state and the management target are combined, and the management strategy is adjusted and matched through a weighted sum model and priority ranking, so that a tilling resource management scheme is generated.
The system comprises an image preprocessing module, a cultivated area dividing module, a soil characteristic and humidity analysis module, a crop growth state evaluation module, a disaster influence analysis module and a cultivated land resource management strategy module;
The image preprocessing module adopts an absolute radiation calibration method and a medium resolution atmospheric transmission model based on remote sensing satellite data, eliminates the atmospheric scattering and absorption influence of an image, optimizes the accuracy and reliability of the data and generates a corrected remote sensing image;
The cultivated land area dividing module is used for identifying land types, including cultivated lands and non-cultivated lands, based on the corrected remote sensing images, applying a K-means clustering algorithm, and identifying boundaries and areas of the cultivated lands through iterative optimization to generate cultivated land area dividing information;
The soil characteristic and humidity analysis module is used for identifying a soil characteristic value based on the cultivated area division information by adopting a spectrum analysis method, analyzing the change trend of soil humidity by utilizing a long-period memory network, and evaluating the physical and chemical properties and the moisture condition of soil by combining a normalized difference moisture index enhancement model to generate a soil characteristic and humidity analysis result;
The crop growth state evaluation module analyzes remote sensing data in a crop growth period, including reflectivity and spectrum index, based on the ploughing area division information by adopting a random forest algorithm, evaluates the health state of crops, including chlorophyll content and biomass, and generates a crop growth state evaluation result;
The disaster influence analysis module is used for analyzing the influence of disasters on crop coverage and soil erosion conditions by adopting a convolutional neural network and a decision tree algorithm based on the cultivated area division information and the crop growth state evaluation result to generate a disaster influence analysis result;
The cultivated land resource management strategy module adopts multi-standard decision analysis and an artificial neural network to analyze and optimize agricultural management activities, including irrigation, fertilization and pest control measures, based on soil characteristics and humidity analysis results, crop growth state assessment results and disaster influence analysis results, and generates a cultivated land resource management scheme through deep learning model training and optimization.
Compared with the prior art, the invention has the advantages and positive effects that:
According to the invention, the cultivated land and non-cultivated land areas are identified by combining a K-means clustering algorithm with a remote sensing data processing technology, accurate basic data is provided for subsequent analysis, a support vector machine algorithm and a spectrum analysis method are adopted, a soil characteristic analysis result is optimized, soil fertility, pH value and organic matter indexes are evaluated, a long-period memory network and a normalized difference moisture index enhancement model are utilized, monitoring capability of soil humidity change trend is improved, arid and over-wet areas are identified, multidimensional assessment of crop health conditions including early disease detection is enhanced by a random forest algorithm, scientificity of crop management is improved, a convolutional neural network and a decision tree algorithm are utilized, influences of disasters are effectively analyzed by combining pre-disaster and post-disaster remote sensing images, and a scientific basis is provided for disaster response. The distribution and variability of soil characteristics are analyzed by adopting a Kriging interpolation method and principal component analysis, accurate data support is provided for land resource management, and a multi-dimensional strategy and scheme are provided for farmland management by combining multi-standard decision analysis and an artificial neural network.
Drawings
FIG. 1 is a schematic of the workflow of the present invention.
Fig. 2 is an S1 refinement flowchart of the present invention.
Fig. 3 is a S2 refinement flowchart of the present invention.
Fig. 4 is a S3 refinement flowchart of the present invention.
Fig. 5 is a S4 refinement flowchart of the present invention.
Fig. 6 is a S5 refinement flowchart of the present invention.
Fig. 7 is a S6 refinement flowchart of the present invention.
Fig. 8 is a S7 refinement flowchart of the present invention.
Fig. 9 is a system flow diagram of the present invention.
Detailed Description
Examples: referring to fig. 1, the present invention provides a technical solution: a farmland change monitoring method based on remote sensing satellite data comprises the following steps:
S1: based on remote sensing satellite data, adopting a K-means clustering algorithm, combining the processed graphic information including radiation correction and atmospheric correction, dividing land types into cultivated land and non-cultivated land, and generating cultivated land area division information;
s2: based on the cultivated area division information, a spectrum analysis method is adopted, spectrum information in a remote sensing image and a soil spectrum database are combined, characteristic values including reflectivity and absorption characteristics are extracted from the spectrum data, soil characteristic analysis including soil fertility, pH value and organic matters is carried out through a support vector machine algorithm, the soil types are classified, and cultivated quality basic information is generated;
s3: based on the cultivated land coverage area division information, a long-period memory network is adopted, the change trend of soil humidity is analyzed through time sequence data, and drought and overwet areas are identified by combining a normalized difference moisture index enhancement model to generate cultivated land humidity distribution information;
S4: based on the farmland coverage area division information, adopting a random forest algorithm to analyze remote sensing data in a crop growth period, identifying various indexes of crop health conditions, including chlorophyll content and biomass, providing multidimensional assessment of crop growth states, including early disease detection, and generating a crop growth state assessment result;
s5: based on the cultivated area division information and the crop health monitoring information, analyzing disaster influence, including influence of the disaster on crop coverage and soil erosion conditions, by combining pre-disaster and post-disaster remote sensing image data through a convolutional neural network and a decision tree algorithm, and generating a disaster influence analysis result;
S6: based on the tilling quality basic information, estimating the soil characteristics of the non-sampling points by adopting a kriging interpolation method, carrying out spatial analysis on the physical and chemical characteristics of the soil by combining with principal component analysis, identifying key factors affecting the quality of the soil, revealing the spatial distribution and variability of the soil characteristics, and generating a soil characteristic analysis result;
S7: based on the farmland humidity distribution information, the crop growth state evaluation result, the disaster influence analysis result and the soil characteristic analysis result, a plurality of agricultural management activities including irrigation plans, fertilization schemes and pest control measures are analyzed by adopting a multi-standard decision analysis and an artificial neural network, a farmland management strategy is identified through training data, and a farmland resource management scheme is generated by matching with the management strategy conforming to the farmland condition.
The cultivated area dividing information comprises cultivated area boundary coordinates, cultivated area total area and non-cultivated area characteristics, the cultivated area quality basic information comprises a soil type classification result, a key fertility index value and a soil improvement scheme, the cultivated area humidity distribution information comprises a drought risk area, an over-humidity risk area and a normal moisture area, the crop growth state assessment result comprises a healthy crop area, crop affected area and growth trend prediction information, the disaster effect analysis result comprises a disaster affected crop area, a soil erosion degree and a predicted recovery time, the soil characteristic analysis result comprises a soil physical characteristic distribution map, a chemical characteristic distribution map and a soil improvement priority area, and the cultivated area resource management scheme comprises an irrigation optimization scheme, a fertilization optimization scheme and a pest control strategy.
In the S1 step, remote sensing satellite data are analyzed through a K-means clustering algorithm, radiation correction and atmosphere correction processing are executed, and accuracy of image information is ensured. And (3) iteratively updating the clustering center until a stable state is reached by calculating Euclidean distance between data points, dividing land types into two major categories of cultivated land and non-cultivated land by an algorithm, and classifying based on spectrum signature differences of image pixel points. Classification refers to changes in surface reflectivity, and also includes the effects of terrain and vegetation cover factors. The generated cultivated land area division information is expressed in the form of a vector graph or a pixel classification graph, and the boundaries of cultivated land and non-cultivated land are clearly marked. The method realizes rapid and automatic identification of the large-scale land coverage type, provides basic data for subsequent soil and crop health analysis, and optimizes resource allocation and management decisions.
In the step S2, based on the ploughing area division information, the spectrum information in the remote sensing image is compared with a soil spectrum database by adopting a spectrum analysis method, and the reflectivity and the absorption characteristic value of the soil are extracted. The support vector machine algorithm accurately classifies soil types according to the characteristic values, and can process high-dimensional data and solve the problem of nonlinearity. By constructing a prediction model of soil fertility, pH value and organic matter parameters, high-accuracy classification is performed. And generating basic information of the cultivated quality, providing detailed information about the soil type and the fertility status of the soil, and providing scientific basis for accurate agriculture and soil management.
In the S3 step, a long-term memory network is adopted to process time sequence data, and the dynamic characteristics of the change of soil humidity along with time are captured. And identifying the change trend of the soil humidity by analyzing the remote sensing data of the continuous time points. And by combining the normalized difference moisture index enhancement model, drought or over-wet areas can be effectively identified. And the farmland humidity distribution information is generated, the soil humidity condition of the differentiated area is intuitively reflected, and accurate data is provided for irrigation management and water resource allocation.
In the S4 step, remote sensing data in the crop growth period is analyzed through a random forest algorithm, and various indexes of the crop health condition are identified. The random forest algorithm can process a batch of data sets and improve the prediction accuracy by constructing a plurality of decision trees and performing integrated learning, and performs multidimensional assessment of the crop growth state by combining chlorophyll content and biomass indexes of crops. And the crop growth state evaluation is generated, early diseases are detected, and scientific basis is provided for crop growth management.
In the S5 step, the disaster influence is analyzed based on the cultivated area division information and the crop health monitoring information by combining a convolutional neural network and a decision tree algorithm. By analyzing the remote sensing image data before and after the disaster, the disaster influence area and degree are effectively identified. The decision tree algorithm iteratively analyzes the specific influence of disasters on crop coverage and soil erosion conditions, generates a disaster influence analysis result, and provides key information for disaster evaluation and agricultural recovery.
In the step S6, the soil characteristics of the non-sampling points are estimated by adopting a Kriging interpolation method and principal component analysis, and space analysis is performed. The Kriging interpolation method is an efficient spatial interpolation method, can estimate the spatial distribution of soil characteristics, combines principal component analysis, identifies key factors influencing the soil quality, reveals the spatial distribution and variability of the soil characteristics, and provides scientific basis for soil improvement and sustainable management as a result of the generated soil characteristic analysis.
In the S7 step, agricultural management activities are analyzed through multi-standard decision analysis and an artificial neural network by integrating farmland humidity distribution information, crop growth state evaluation results, disaster influence analysis results and soil characteristic analysis results. The multi-standard decision analysis provides a method for evaluating and selecting the differential management scheme, and the artificial neural network identifies the cultivated land management strategy through training data and matches the management scheme suitable for the current cultivated land condition. And an cultivated land resource management scheme is generated, so that optimal decision support is provided for agricultural production, and reasonable utilization of resources and efficient management of the agricultural production are promoted.
Referring to fig. 2, based on remote sensing satellite data, the method adopts a K-means clustering algorithm, combines the processed graphic information including radiation correction and atmosphere correction, and divides land types into cultivated land and non-cultivated land, and the method specifically comprises the following steps of:
S101: based on remote sensing satellite data, correcting the systematic deviation of the sensor by adopting an absolute radiometric calibration method, eliminating the influence of atmospheric scattering and absorption, including the scattering effect of fog and cloud by using an MODTRA model, and generating a corrected remote sensing image;
s102: based on the corrected remote sensing image, analyzing image data by adopting a K-means clustering algorithm, setting the clustering quantity, dividing land coverage types including water bodies, buildings and altitudes, and generating earth surface type identification information;
s103: based on the earth surface type identification information, a support vector machine algorithm is adopted to identify cultivated land and non-cultivated land areas, including forest, water area, grassland and cultivated land types, and the cultivated land area division information is generated by combining cadastral data and land use records, verifying the identification result.
In the S101 substep, an absolute radiometric calibration method is adopted to correct the remote sensing satellite data, and the influence of the system deviation of the sensor and the atmospheric condition on the quality of the remote sensing image is eliminated. The absolute radiation calibration method utilizes a known standard radiation source to calibrate signals received by a satellite sensor, ensures that image data reflects real surface radiation intensity, and processes the image data to eliminate the influence of atmospheric scattering and absorption, including the scattering effect of fog and cloud by combining with an MODTRA model. The spectrum values of the pixels in the image are finely adjusted to reflect the actual situation of the earth surface, and the generated corrected remote sensing image is closer to the actual observation object in terms of color and brightness, so that accurate basic data is provided for subsequent land type analysis.
In the S102 substep, based on the corrected remote sensing image, the image data is analyzed and the land coverage type is divided by using a K-means clustering algorithm. The K-means clustering algorithm automatically groups image data according to similarity among pixel points by setting a preset clustering quantity, wherein each group represents a land coverage type including water, buildings and altitudes. The method comprises the steps of classifying pixels into the nearest clustering center by calculating the distance between each pixel and the clustering center, identifying and distinguishing various land coverage types according to key characteristics of the surface characteristics by iterative optimization process until the clustering result is stable, and accurately reflecting actual coverage conditions of the surface by generated surface type identification information to lay a foundation for identification of cultivated lands and non-cultivated lands.
In the step S103, the cultivated land and the non-cultivated land area are identified by using a support vector machine algorithm based on the surface type identification information. The support vector machine algorithm is a supervised learning method, samples are classified by constructing one or more hyperplanes, and the selection of the hyperplanes is optimized to maximize the interval between the samples, so that efficient and accurate classification is performed. In this step, the algorithm uses spectral information and texture features in the remote sensing image as inputs, and combines the cadastral data and the land usage record as training samples to train a classification model to distinguish between cultivated land and non-cultivated land. The high accuracy and the strong generalization capability of the support vector machine algorithm ensure the accuracy and the reliability of the identification result, and the generated cultivated area division information provides key basic data for agricultural monitoring and land resource management.
Referring to fig. 3, based on the partition information of the cultivated area, a spectrum analysis method is adopted, spectrum information in a remote sensing image and a soil spectrum database are combined, characteristic values including reflectivity and absorption characteristics are extracted from the spectrum data, soil characteristic analysis including soil fertility, pH value and organic matters is performed through a support vector machine algorithm, the soil types are classified, and the step of generating basic information of the cultivated quality is specifically as follows:
S201: based on the ploughing area division information, extracting spectral characteristics and spectral information in a soil spectral database for comparison by adopting a spectral analysis method through remote sensing image information, wherein the spectral characteristics comprise reflectivity and absorption characteristics, and identifying spectral characteristic values of various soil types to generate soil spectral characteristic data;
s202: based on the soil spectrum characteristic data, a support vector machine algorithm is adopted, the extracted characteristic values are input into a support vector machine model, soil characteristic analysis is carried out, various soil characteristics including soil fertility, pH value and organic matters are identified, and a soil characteristic classification result is generated;
S203: based on the soil characteristic classification result, a classification decision process is adopted, soil types in the region are classified by analyzing soil characteristic data of the cultivated region, the soil cultivated land quality in the region is identified, and cultivated soil quality basic information is generated.
In the S201 substep, spectral features are extracted from the remote sensing image by a spectral analysis method and are compared with information in a soil spectrum database, the quality of image data is ensured by preprocessing of the remote sensing image including denoising, radiometric calibration and atmospheric correction, and spectral features extraction is carried out on the preprocessed image by utilizing a spectral analysis technology, including quantification of reflectivity and absorption characteristics of each pixel point in the image. The soil types with the same spectral characteristics are identified through similarity analysis by matching the spectral characteristics with the spectral information pre-classified and marked in the soil spectral database. And identifying spectral characteristic values of various soil types through accurate spectral matching, and generating soil spectral characteristic data. The method comprises the steps of including basic spectrum information of soil, reflecting diversity of soil types and providing detailed basic information for subsequent soil characteristic analysis.
In the S202 substep, the spectral characteristic data of the soil is analyzed by adopting a support vector machine algorithm, and specific characteristics of the soil, including soil fertility, pH value and organic matter content, are identified. The support vector machine algorithm classifies the input characteristic values by constructing a model which maximizes classification boundary in the characteristic space, firstly, the algorithm trains a model which can distinguish differentiated soil characteristics according to the soil spectrum characteristic data, and inputs the characteristic values into the model for classification. The high efficiency and accuracy of the support vector machine algorithm plays a key role in this step, so that the algorithm can accurately divide soil into different categories, and the soil characteristics of each category are described in detail. The generated soil characteristic classification result provides basic characteristic information of soil and provides scientific basis for comprehensive evaluation and management of soil quality.
In the S203 substep, based on the soil characteristic classification result, the soil characteristics of the cultivated land area are comprehensively analyzed and classified by adopting a classification decision process. Soil characteristic data, including fertility, pH and organic matter content, are analyzed to determine soil classification by evaluating the overall quality of the soil in the cultivated area. And the classification decision process comprehensively evaluates and classifies the soil through weight distribution and evaluation of the characteristics of the differential soil, and identifies the land with the differential quality grade in the cultivated area. The space distribution condition of the cultivated land quality is disclosed, accurate soil quality information is provided for agricultural production, and the generated cultivated land quality basic information becomes a key basis for determining a cultivation mode and management measures, so that the method has a key effect on improving the agricultural production efficiency and promoting the reasonable utilization of land resources.
Referring to fig. 4, based on the area division information of the cultivated land, a long-period memory network is adopted, a change trend of soil humidity is analyzed through time sequence data, and drought and over-wet areas are identified by combining a normalized difference water index enhancement model, so that the method specifically comprises the following steps of:
S301: based on the farmland coverage area division information, a long-period memory network is adopted, and the time sequence remote sensing data of the corresponding farmland area are combined to simulate and analyze the change trend of soil humidity along with time so as to generate soil humidity change trend data;
S302: based on the soil humidity change trend data, applying a normalized difference moisture index enhancement model, analyzing the soil humidity at a plurality of time points by comparing the reflectivity differences of the differential wavelengths, identifying the moisture conditions of soil and vegetation, and generating soil humidity analysis data;
s303: based on soil humidity analysis data, a threshold analysis method and a data classification technology are adopted, drought, normal and over-wet soil humidity conditions are distinguished by setting humidity thresholds, soil humidity states of a plurality of areas are identified by comparing soil humidity indexes with preset drought and over-wet thresholds, and cultivated land humidity distribution information is generated.
In the S301 substep, the time series remote sensing data based on the farmland coverage area division information is processed and analyzed through a long-period memory network model, and the change trend of the soil humidity along with time is simulated and analyzed. The remote sensing data of the corresponding cultivated land area are arranged according to time sequence by capturing long-term dependency relations in the time sequence data, a time sequence data set is constructed, the data are learned and analyzed by using a long-term and short-term memory network, the network adjusts internal parameters through iterative training, and the prediction capability of the model on the soil humidity change trend is optimized. The regular variation of the soil humidity along with the change of seasons and climate conditions is identified and simulated, the generated soil humidity variation trend data accurately reflects the history and current state of the soil humidity of the cultivated land, the future variation trend is predicted, and the method has key reference value for decision makers to make irrigation plans and evaluate drought and waterlogging risks.
In the step S302, the soil humidity change trend data is analyzed through a normalized differential water index enhancement model. The normalized differential moisture index enhancement model effectively evaluates the moisture status of soil and vegetation by analyzing the reflectivity differences of the soil and vegetation, including the reflectivity changes at the target wavelengths. And calculating the difference of the reflectivity of soil and vegetation at the target wavelength by selecting proper wavelength data, and constructing a normalized difference moisture index. The change condition of the soil humidity at the differential time point is reflected, and the moisture conditions of the soil and vegetation are identified through the analysis of the sequence index. The generated soil humidity analysis data provides scientific basis for subsequent farmland management, including judgment of irrigation requirements and reasonable distribution of water resources, and the target data has significant significance for preventing crop damage and optimizing water resource use in seasons with drought or rainfall abnormality.
In the sub-step S303, soil moisture analysis data is classified by a threshold analysis method and a data classification technique to distinguish cultivated lands differentiating soil moisture conditions. Soil moisture is classified into three levels of drought, normal and over-wet by setting a threshold value for soil moisture. By comparing the soil humidity index of each area with preset drought and over-humidity thresholds, the soil humidity states of multiple areas are accurately identified. By adopting a threshold analysis method and a data classification technology, the soil humidity data in batches can be effectively processed, and the rapid and accurate soil humidity state classification can be performed. The generated farmland humidity distribution information is displayed in a graph or table form, the soil humidity condition of the differential area is intuitively reflected, accurate guidance is provided for irrigation management of the farmland and crop planting selection, agricultural producers can take corresponding measures to cope with the differential soil humidity condition, and healthy growth and yield stability of crops are ensured.
Referring to fig. 5, based on the area division information of the cultivated land, a random forest algorithm is adopted to analyze the remote sensing data in the growth period of the crops, identify various indexes of the health condition of the crops, including chlorophyll content and biomass, provide multidimensional assessment of the growth state of the crops, including detection of early diseases, and the steps of generating the assessment result of the growth state of the crops are as follows:
S401: based on the cultivated land coverage area division information, analyzing remote sensing data in a crop growth period by adopting a random forest algorithm, wherein the remote sensing data comprise reflectivity and spectrum indexes, and identifying key characteristics of crops, including vegetation indexes and moisture indexes, so as to generate crop growth characteristic data;
S402: based on the crop growth characteristic data, applying logistic regression analysis, analyzing the health state of crops according to the crop growth key characteristics, including normal, slight damage and severe damage, evaluating the health state of the crops, and generating crop health analysis information;
S403: based on crop health analysis information, adopting decision tree analysis and principal component analysis, combining a plurality of health indexes, analyzing the growth state of crops, obtaining crop health scores, identifying the whole growth condition of crops and the farmland problem area including early diseases, and generating a crop growth state evaluation result.
In the S401 substep, analyzing the cultivated land coverage area division information and remote sensing data in the crop growth period through a random forest algorithm, wherein reflectivity and spectral indexes in the key attention data are used for identifying key growth characteristics of crops, including vegetation indexes and moisture indexes. The random forest algorithm is an integrated learning method, and is used for training a data set by constructing a plurality of decision trees and summarizing the results to improve the prediction accuracy. In this step, the algorithm first segments the remote sensing dataset into a plurality of sub-sample sets, and constructs a decision tree for each sub-sample set, each tree referencing a randomly selected portion of the features when split. The risk of overfitting is reduced, and the generalization capability of the model is improved. By integrating the prediction results of each decision tree, a comprehensive data set reflecting the key characteristics of crop growth is generated, and the data accurately reveal the key growth parameters of crops in the differential growth stage, so that a key basis is provided for subsequent health state analysis.
In the sub-step S402, the crop growth characteristic data is subjected to in-depth analysis by applying logistic regression analysis, and the health status of the crop is estimated by the key characteristics. Logistic regression is a statistical method widely used for classifying tasks, the probability of occurrence of an event is predicted by fitting a logistic function, key growth characteristics of crops, including vegetation indexes and moisture indexes, are used as independent variable input models, the models output probability distribution of health states of the crops, and the training process of the logistic regression model comprises selecting proper parameters so that the fitting degree of the models to training data is highest, the current health state of the crops is estimated, and health problems of the crops are predicted. The generated crop health analysis information provides real-time and accurate health monitoring results for agricultural production, and is beneficial to timely taking corresponding agricultural management measures.
In the step S403, the decision tree analysis and the principal component analysis are combined, and the multi-dimensional assessment of the growth state of the crop is performed by comprehensively referring to a plurality of health indexes in the crop health analysis information, including chlorophyll content and biomass. The decision tree analysis classifies or predicts data through simple rules, the principal component analysis is a statistical method for reducing data dimension, the data set is simplified through extracting principal components in the data, meanwhile, the information of the original data is reserved as much as possible, the health state of crops is primarily classified through the decision tree analysis, the weight distribution and optimization are carried out on multiple indexes in classification results through the principal component analysis, and the overall growth condition of the crops is comprehensively evaluated. The advantages and problem areas of crop growth, including the possibility of early disease occurrence, are accurately identified, and the generated crop growth state evaluation result provides scientific basis for agricultural production management and decision making, thereby being beneficial to improving crop yield and quality.
Referring to fig. 6, based on the cultivated area dividing information and the crop health monitoring information, the disaster influence is analyzed by combining the pre-disaster post-disaster remote sensing image data through a convolutional neural network and a decision tree algorithm, including the influence of the disaster on the crop coverage and the soil erosion condition, and the steps of generating a disaster influence analysis result are specifically as follows:
S501: based on the cultivated area division information and the crop health monitoring information, performing image feature extraction by adopting a convolutional neural network, identifying the influence of disasters on the change of crop coverage, and generating disaster feature extraction data;
S502: based on disaster feature extraction data, analyzing and classifying factors influencing crop coverage by adopting a random forest algorithm, analyzing the influence of various disaster features on the crop coverage, including vegetation index change and soil exposure, identifying key influencing factors, and generating a crop coverage influencing factor analysis result;
S503: based on the analysis result of crop coverage influence factors, a decision tree algorithm is adopted, the influence of disasters on soil erosion is evaluated by analyzing the relevance between crop coverage reduction and soil erosion conditions, and the analysis result of the disasters is generated, wherein the influence comprises erosion degree and distribution areas.
In the step S501, image feature extraction is performed on the cultivated area division information and the crop health monitoring information through the convolutional neural network, and the image data is processed through a multi-layer structure of the convolutional neural network, including a convolutional layer, a pooling layer and a full connection layer, so as to extract advanced features associated with the change of crop coverage. In the convolution layer, the model automatically identifies local features in the image through a learning filter, the pooling layer reduces the space dimension of the features, reduces the calculated amount, improves the robustness of the features, and finally integrates the features through the full-connection layer to output high-level analysis on the influence of crop coverage. The generated disaster characteristic extraction data comprise quantitative information about crop coverage change, including the reduction ratio of the coverage before and after disaster, and provide a basis for subsequent analysis.
In the step S502, the disaster feature extraction data is analyzed by adopting a random forest algorithm, factors influencing the crop coverage are identified and classified, the accuracy and the stability of prediction are improved by constructing a plurality of decision trees and summarizing the prediction results, and the algorithm analyzes various disaster features including vegetation index change and soil exposure and specific influences on the crop coverage. Each decision tree is trained based on a random subset of disaster feature extraction data, enabling the impact of disaster features on crop coverage to be assessed from a differential perspective. By the integrated learning method, the model identifies the most obvious factors affecting the crop coverage, and evaluates the criticality of multiple factors. The generated crop coverage influence factor analysis result details the specific contribution of multiple influence factors to the crop coverage, and provides scientific basis for formulating effective post-disaster recovery strategy and disaster prevention and reduction measures.
In the sub-step S503, the association of crop coverage reduction with soil erosion conditions is analyzed by a decision tree algorithm, thereby evaluating the impact of disasters on soil erosion. The decision tree algorithm simulates a decision process by constructing a simple decision rule, combines analysis results based on crop coverage influence factors, classifies data according to the degree of crop coverage reduction and soil erosion conditions, and selects the feature of uncertainty reduction in each classification. By progressive subdivision, the algorithm can reveal a direct link between crop coverage variation and soil erosion, including the type, extent, and affected geographical area of erosion. The generated disaster influence analysis result shows the concrete influence of the disaster on soil erosion in detail, and the concrete influence comprises erosion degree and a map of a distribution area, so that accurate guiding information is provided for post-disaster soil protection and restoration work, and resource allocation is optimized and ecological restoration process is accelerated.
Referring to fig. 7, based on the basic information of the cultivated quality, estimating the soil characteristics of the non-sampling points by using a kriging interpolation method, performing spatial analysis on the physical and chemical characteristics of the soil in combination with principal component analysis, identifying key factors affecting the quality of the soil, revealing the spatial distribution and variability of the soil characteristics, and generating a soil characteristic analysis result specifically comprises the following steps:
S601: based on the tilling quality basic information, a Kriging interpolation method is adopted, soil characteristics of an un-sampled position are estimated according to the space distance and the similarity through existing soil sample data, and the soil characteristics interpolation data are generated, wherein the soil characteristics comprise pH value, organic matter content and soil density;
S602: based on the soil characteristic interpolation data, adopting principal component analysis, performing feature selection and feature extraction on the physicochemical characteristics of the soil by calculating a covariance matrix of the soil characteristic data, and identifying key features by combining the contribution rates of various soil features to generate a soil key characteristic data set;
S603: based on the soil key characteristic data set, spatial mapping and trend analysis are carried out by using a geographic information system technology, the spatial distribution mode and influencing factors of the soil quality are analyzed, the key factors influencing the soil quality are identified, the spatial distribution and variability of the soil characteristics are obtained, and a soil characteristic analysis result is generated.
In the step S601, soil characteristics of the non-sampled position are estimated based on the existing soil quality basis information by the kriging interpolation method. The kriging interpolation is an optimal spatial interpolation method based on a statistical principle, and estimates the value of an unknown point by calculating spatial autocorrelation among a plurality of positions with reference to the spatial relationship between sample data points. According to the existing soil sample data including pH value, organic matter content, soil density and the like, a space autocorrelation model is constructed, and soil characteristics of an un-sampled position are predicted by using the model, so that soil characteristic interpolation data with space continuity is generated. Fills the information loss of the sampling blank area, provides a detailed and continuous soil characteristic space distribution map, and provides reliable basic information for deeply analyzing the space variability of the soil quality and carrying out accurate land management.
In the sub-step S602, the soil characteristic interpolation data is processed using principal component analysis in order to simplify the complexity of the dataset and extract key features. The principal component analysis is a statistical method, the principal component can analyze the variance of most of the original data set by converting the original variable into a group of linear independent variables, the characteristic value and the characteristic vector are extracted by calculating the covariance matrix of the soil characteristic data, the principal component is selected according to the contribution rate, the dimensionality of the data is reduced, meanwhile, key information is reserved, the generated soil key characteristic data set highlights key factors influencing the soil quality, the complexity of subsequent analysis work is simplified, and convenience is provided for identifying and analyzing the key factors influencing the soil quality.
In the sub-step S603, spatial mapping and trend analysis of soil characteristics are performed using geographic information system technology in combination with the soil key characteristic dataset. The key characteristic data of the soil obtained by the principal component analysis are mapped in space, the spatial distribution pattern of the soil quality is analyzed, the regions with high and low variation of the soil quality are identified, the influence of environment and management factors on the soil characteristics is estimated, the generated soil characteristic analysis result shows the spatial distribution condition of the soil quality, the variability of the soil characteristics is revealed, scientific basis is provided for agricultural production and land resource management, and reasonable utilization and sustainable development of resources are promoted.
Referring to fig. 8, based on the farmland humidity distribution information, the crop growth state evaluation result, the disaster impact analysis result, and the soil characteristic analysis result, multiple agricultural management activities including irrigation plans, fertilization schemes, and pest control measures are analyzed by adopting a multi-standard decision analysis and an artificial neural network, the farmland management strategies are identified through training data, and the management strategies conforming to the farmland conditions are matched, so that the farmland resource management scheme is generated specifically by the steps of:
S701: based on the farmland humidity distribution information, the crop growth state evaluation result, the disaster influence analysis result and the soil characteristic analysis result, adopting a hierarchical analysis method, and optimizing the evaluation accuracy and reliability by constructing a hierarchical structure of an evaluation standard, performing pairwise comparison calculation weight and consistency test, and generating a farmland management scheme evaluation result;
S702: based on the evaluation result of the cultivated land management scheme, an artificial neural network is adopted, and a multilayer perceptron model is trained and optimized by inputting quantitative indexes of irrigation, fertilization and pest control schemes, so as to generate a cultivated land management strategy set;
S703: based on the cultivated land management strategy set, a decision support system is adopted to evaluate the effect of the cultivated land management strategy, a comprehensive evaluation model is applied, the cultivated land state and the management target are combined, and the cultivated land resource management scheme is generated by adjusting and matching the management strategy through a weighted sum model and priority ordering.
In the step S701, the farmland management scheme is evaluated by means of a hierarchical analysis method with reference to the farmland humidity distribution information, the crop growth state evaluation result, the disaster impact analysis result, and the soil characteristic analysis result. By establishing a hierarchical structure of the evaluation standard, the complex decision problem is decomposed into a target layer, a criterion layer and a scheme layer, so that analysis and treatment of the problem are more systematic and organized, the relative key weights of multiple criteria and schemes are calculated by combining pair comparison, consistency check is performed by using consistency indexes and consistency ratios, consistency and reliability of evaluation are ensured, the generated evaluation result of the cultivated land management scheme lists the scores and comprehensive weights of multiple management schemes under the multiple criteria in detail, scientific basis and decision support are provided for the subsequent selection of the optimal management scheme, the decision process is more objective and accurate, and the scientificity and practicability of the selection of the cultivated land resource management scheme are improved.
In the step S702, based on the evaluation result of the cultivated land management scheme, an artificial neural network technology is adopted, and a multi-layer perceptron model is combined to analyze the agricultural management activity, including taking quantitative indexes of irrigation, fertilization and pest control schemes as input, optimizing the neural network model through training data, learning complex relations and modes among the input data, and identifying strategies for optimizing the cultivated land management association. The training process comprises the steps of forward propagation, loss function calculation, reverse propagation and weight updating, and the accuracy of identifying the cultivated land management strategy by the model is improved through iterative optimization of model parameters. The generated tilling management policy set includes a series of management policies that match the current tilling conditions, providing a decision maker with a plurality of management options.
In S703, based on the tilling management policy set, a policy effect is evaluated using a decision support system. And (3) applying a comprehensive evaluation model, combining the actual state of the cultivated land with the management target, refining and adjusting the management strategy through a weighted sum model and priority sequencing, and integrating various data sources and evaluation tools by the decision support system to provide multidimensional evaluation on the results of the management strategies. Through analysis and simulation of the system, the comprehensive effect generated by interaction and prediction among strategies is studied in detail, and a resource management scheme suitable for the current cultivated land condition is determined. The generated cultivated land resource management scheme refers to various factors such as water resource utilization, fertilizer use efficiency and pest control effect, and provides scientific decision basis for sustainable utilization of cultivated land resources and maximization of crop yield.
Referring to fig. 9, a cultivated land change monitoring system based on remote sensing satellite data is used for executing the cultivated land change monitoring method based on remote sensing satellite data, and the system comprises an image preprocessing module, a cultivated area dividing module, a soil characteristic and humidity analysis module, a crop growth state evaluation module, a disaster influence analysis module and a cultivated land resource management strategy module;
The image preprocessing module adopts an absolute radiation calibration method and a medium resolution atmospheric transmission model based on remote sensing satellite data, eliminates the atmospheric scattering and absorption influence of the image, optimizes the accuracy and reliability of the data and generates a corrected remote sensing image;
The cultivated area dividing module is used for identifying the types of the land, including cultivated land and non-cultivated land, based on the corrected remote sensing image, applying a K-means clustering algorithm, and identifying the boundary and the area of the cultivated land through iterative optimization to generate cultivated area dividing information;
The soil characteristic and humidity analysis module is used for identifying characteristic values of the soil characteristic by adopting a spectrum analysis method based on the ploughing area division information, analyzing the change trend of the soil humidity by utilizing a long-period memory network, and evaluating physical and chemical properties and the moisture condition of the soil by combining a normalized difference moisture index enhancement model to generate a soil characteristic and humidity analysis result;
The crop growth state evaluation module analyzes remote sensing data in a crop growth period, including reflectivity and spectral index, based on the ploughing area division information by adopting a random forest algorithm, evaluates the health state of crops, including chlorophyll content and biomass, and generates a crop growth state evaluation result;
The disaster influence analysis module is used for analyzing the influence of disasters on crop coverage and soil erosion conditions by adopting a convolutional neural network and a decision tree algorithm based on the cultivated area division information and the crop growth state evaluation result to generate a disaster influence analysis result;
The cultivated land resource management strategy module adopts multi-standard decision analysis and an artificial neural network to analyze and optimize agricultural management activities, including irrigation, fertilization and pest control measures, based on soil characteristics and humidity analysis results, crop growth state assessment results and disaster influence analysis results, and generates a cultivated land resource management scheme through deep learning model training and optimization.
The image preprocessing module adopts an absolute radiation calibration method and a medium resolution atmospheric transmission model, so that the accuracy and reliability of remote sensing image data are remarkably improved, and high-quality basic data are provided for subsequent farmland monitoring and analysis. The cultivated land area dividing module accurately identifies cultivated lands and non-cultivated lands by using a K-means clustering algorithm, and provides scientific basis for cultivated land protection and reasonable utilization. The soil characteristics and humidity analysis module is combined with the spectrum analysis method and the long-term and short-term memory network to evaluate the physical and chemical properties and the moisture condition of the soil in a multidimensional manner, so that reliable data is provided for accurate agriculture. The crop growth state evaluation module analyzes remote sensing data in a crop growth period through a random forest algorithm, and accuracy and timeliness of crop disease early warning are improved. The disaster influence analysis module is combined with the convolutional neural network and the decision tree algorithm to effectively analyze the influence of disasters on crop coverage and soil erosion conditions and provide scientific basis for post-disaster recovery. The cultivated land resource management strategy module comprehensively analyzes soil characteristics, crop growth states and disaster influences, optimizes agricultural management activities by adopting multi-standard decision analysis and an artificial neural network, and improves the scientificity and effectiveness of agricultural resource management.
Claims (5)
1. A farmland change monitoring method based on remote sensing satellite data is characterized in that: the method comprises the following steps:
based on remote sensing satellite data, adopting a K-means clustering algorithm, combining the processed graphic information including radiation correction and atmospheric correction, dividing land types into cultivated land and non-cultivated land, and generating cultivated land area division information;
Based on the ploughing area division information, a spectrum analysis method is adopted, spectrum information in a remote sensing image and a soil spectrum database are combined, characteristic values including reflectivity and absorption characteristics are extracted from the spectrum data, soil characteristic analysis including soil fertility, pH value and organic matters is carried out through a support vector machine algorithm, soil types are classified, and ploughing quality basic information is generated;
Based on the farmland coverage area division information, a long-period and short-period memory network is adopted, the change trend of soil humidity is analyzed through time sequence data, and drought and overwet areas are identified by combining a normalized difference moisture index enhancement model to generate farmland humidity distribution information;
Based on the farmland coverage area division information, adopting a random forest algorithm to analyze remote sensing data in a crop growth period, identifying various indexes of crop health conditions, including chlorophyll content and biomass, providing multidimensional assessment of crop growth states, including early disease detection, and generating a crop growth state assessment result;
based on the cultivated area division information and the crop health monitoring information, analyzing disaster influence, including influence of the disaster on crop coverage and soil erosion conditions, by combining pre-disaster and post-disaster remote sensing image data through a convolutional neural network and a decision tree algorithm, and generating a disaster influence analysis result;
Estimating soil characteristics of an un-sampled point by adopting a kriging interpolation method based on the tilling quality basic information, carrying out spatial analysis on the physical and chemical characteristics of the soil by combining principal component analysis, identifying key factors affecting the quality of the soil, revealing the spatial distribution and variability of the soil characteristics, and generating a soil characteristic analysis result;
Based on the farmland humidity distribution information, the crop growth state evaluation result, the disaster influence analysis result and the soil characteristic analysis result, adopting multi-standard decision analysis and an artificial neural network to analyze various agricultural management activities including irrigation plans, fertilization schemes and pest control measures, identifying farmland management strategies through training data, and matching management strategies conforming to farmland conditions to generate a farmland resource management scheme;
Based on the cultivated area division information, a spectrum analysis method is adopted, spectrum information in a remote sensing image and a soil spectrum database are combined, characteristic values including reflectivity and absorption characteristics are extracted from the spectrum data, soil characteristic analysis including soil fertility, pH value and organic matters is carried out through a support vector machine algorithm, the soil types are classified, and the step of generating cultivated quality basic information specifically comprises the following steps:
Based on the ploughing area division information, extracting spectral characteristics and spectral information in a soil spectral database for comparison by adopting a spectral analysis method through remote sensing image information, wherein the spectral characteristics comprise reflectivity and absorption characteristics, and identifying spectral characteristic values of various soil types to generate soil spectral characteristic data;
Based on the soil spectrum characteristic data, a support vector machine algorithm is adopted, the extracted characteristic values are input into a support vector machine model, soil characteristic analysis is carried out, various soil characteristics including soil fertility, pH value and organic matters are identified, and a soil characteristic classification result is generated;
based on the soil characteristic classification result, classifying the soil types in the region by analyzing the soil characteristic data of the cultivated region by adopting a classification decision process, and identifying the soil cultivated land quality in the region to generate cultivated land quality basic information;
Based on the farmland coverage area division information, a long-period memory network is adopted, the change trend of soil humidity is analyzed through time sequence data, and drought and over-wet areas are identified by combining a normalized difference water index enhancement model, so that farmland humidity distribution information is generated specifically by the steps of:
Based on the farmland coverage area division information, a long-period memory network is adopted, and the time sequence remote sensing data of the corresponding farmland area are combined to simulate and analyze the change trend of soil humidity along with time so as to generate soil humidity change trend data;
Based on the soil humidity change trend data, applying a normalized difference moisture index enhancement model, analyzing the soil humidity at a plurality of time points by comparing the reflectivity differences of the differential wavelengths, identifying the moisture conditions of soil and vegetation, and generating soil humidity analysis data;
Based on the soil humidity analysis data, a threshold analysis method and a data classification technology are adopted, drought, normal and over-wet soil humidity conditions are distinguished through setting humidity thresholds, soil humidity states of a plurality of areas are identified through comparing soil humidity indexes with preset drought and over-wet thresholds, and cultivated land humidity distribution information is generated;
Based on the farmland coverage area division information, a random forest algorithm is adopted to analyze remote sensing data in a crop growth period, various indexes of crop health conditions are identified, including chlorophyll content and biomass, multidimensional assessment of crop growth states is provided, detection of early diseases is included, and the step of generating a crop growth state assessment result specifically comprises the following steps:
Based on the farmland coverage area division information, analyzing remote sensing data in a crop growth period by adopting a random forest algorithm, wherein the remote sensing data comprise reflectivity and spectrum indexes, and identifying key characteristics of crops, including vegetation indexes and moisture indexes, so as to generate crop growth characteristic data;
Based on the crop growth characteristic data, applying logistic regression analysis, analyzing the health state of crops according to crop growth key characteristics, including normal, mild damage and severe damage, evaluating the health state of crops, and generating crop health analysis information;
Based on the crop health analysis information, adopting decision tree analysis and principal component analysis, combining a plurality of health indexes, analyzing the crop growth state, obtaining crop health scores, identifying the overall growth condition of crops and the farmland problem area, including early diseases, and generating a crop growth state evaluation result;
Based on the cultivated area division information and the crop health monitoring information, the disaster influence is analyzed by combining the pre-disaster post-disaster remote sensing image data through a convolutional neural network and a decision tree algorithm, the disaster influence comprises the influence of the disaster on the crop coverage and the soil erosion condition, and the step of generating a disaster influence analysis result specifically comprises the following steps:
Based on the cultivated area division information and the crop health monitoring information, performing image feature extraction by adopting a convolutional neural network, identifying the influence of disasters on the change of crop coverage, and generating disaster feature extraction data;
Based on the disaster feature extraction data, analyzing and classifying factors influencing crop coverage by adopting a random forest algorithm, analyzing the influence of various disaster features on the crop coverage, including vegetation index change and soil exposure, identifying key influencing factors, and generating a crop coverage influencing factor analysis result;
Based on the analysis result of the crop coverage factor influence factor, adopting a decision tree algorithm, and evaluating the influence of disasters on soil erosion by analyzing the relevance between crop coverage reduction and soil erosion conditions, wherein the influence comprises erosion degree and distribution area, so as to generate a disaster influence analysis result;
Based on the tilling quality basic information, estimating the soil characteristics of the non-sampling points by adopting a kriging interpolation method, carrying out spatial analysis on the physical and chemical characteristics of the soil by combining principal component analysis, identifying key factors affecting the quality of the soil, revealing the spatial distribution and variability of the soil characteristics, and generating a soil characteristic analysis result specifically comprises the following steps:
Based on the tilling quality basic information, a Kriging interpolation method is adopted, soil characteristics of an un-sampled position, including pH value, organic matter content and soil density, are estimated according to the space distance and similarity through existing soil sample data, and soil characteristic interpolation data are generated;
Based on the soil characteristic interpolation data, adopting principal component analysis, performing feature selection and feature extraction on the physicochemical characteristics of the soil by calculating a covariance matrix of the soil characteristic data, and identifying key features by combining contribution rates of various soil features to generate a soil key characteristic data set;
Based on the soil key characteristic data set, spatial mapping and trend analysis are carried out by using a geographic information system technology, the spatial distribution mode and influence factors of soil quality are analyzed, key factors influencing the soil quality are identified, the spatial distribution and variability of soil characteristics are obtained, and a soil characteristic analysis result is generated.
2. The method for monitoring the change of cultivated land based on remote sensing satellite data according to claim 1, wherein the method comprises the following steps: the cultivated area dividing information comprises cultivated area boundary coordinates, cultivated area total area and non-cultivated area characteristics, the cultivated area quality basic information comprises a soil type classification result, a key fertility index value and a soil improvement scheme, the cultivated area humidity distribution information comprises a drought risk area, an over-wet risk area and a normal moisture area, the crop growth state assessment result comprises a healthy crop area, a pest affected area and growth trend prediction information, the disaster impact analysis result comprises a disaster affected crop area, a soil erosion degree and an estimated recovery time, the soil characteristic analysis result comprises a soil physical characteristic distribution map, a chemical characteristic distribution map and a soil improvement priority area, and the cultivated area resource management scheme comprises an irrigation optimization scheme, a fertilization optimization scheme and a pest and disease prevention and control strategy.
3. The method for monitoring the change of cultivated land based on remote sensing satellite data according to claim 1, wherein the method comprises the following steps: based on remote sensing satellite data, adopting a K-means clustering algorithm, combining the processed graphic information including radiation correction and atmosphere correction, dividing land types into cultivated land and non-cultivated land, and generating cultivated land area division information specifically comprises the following steps:
Based on remote sensing satellite data, correcting the systematic deviation of the sensor by adopting an absolute radiometric calibration method, eliminating the influence of atmospheric scattering and absorption, including the scattering effect of fog and cloud by using an MODTRA model, and generating a corrected remote sensing image;
Based on the corrected remote sensing image, analyzing image data by adopting a K-means clustering algorithm, setting the clustering quantity, dividing land coverage types including water bodies, buildings and altitudes, and generating earth surface type identification information;
Based on the earth surface type identification information, a support vector machine algorithm is adopted to identify cultivated land and non-cultivated land areas, including forest, water area, grassland and cultivated land types, and the cultivated land area division information is generated by combining cadastral data and land use records, verifying the identification result.
4. The method for monitoring the change of cultivated land based on remote sensing satellite data according to claim 1, wherein the method comprises the following steps: based on the farmland humidity distribution information, the crop growth state evaluation result, the disaster influence analysis result and the soil characteristic analysis result, a plurality of agricultural management activities including irrigation plans, fertilization schemes and pest control measures are analyzed by adopting a multi-standard decision analysis and an artificial neural network, a farmland management strategy is identified through training data, and a management strategy conforming to the farmland condition is matched, wherein the steps of generating a farmland resource management scheme specifically comprise:
based on the farmland humidity distribution information, the crop growth state evaluation result, the disaster influence analysis result and the soil characteristic analysis result, adopting a hierarchical analysis method, and optimizing the accuracy and reliability of evaluation by constructing a hierarchical structure of an evaluation standard, performing pairwise comparison calculation weight and consistency test, and generating a farmland management scheme evaluation result;
Based on the evaluation result of the cultivated land management scheme, an artificial neural network is adopted, and a multilayer perceptron model is trained and optimized by inputting quantitative indexes of irrigation, fertilization and pest control schemes, so as to generate a cultivated land management strategy set;
based on the tilling management strategy set, a decision support system is adopted to evaluate the tilling management strategy effect, a comprehensive evaluation model is applied, the tilling state and the management target are combined, and the management strategy is adjusted and matched through a weighted sum model and priority ranking, so that a tilling resource management scheme is generated.
5. The utility model provides a farmland change monitoring system based on remote sensing satellite data which characterized in that: the method for monitoring the change of cultivated land based on remote sensing satellite data according to any one of claims 1 to 4, wherein the system comprises an image preprocessing module, a cultivated area dividing module, a soil characteristic and humidity analysis module, a crop growth state evaluation module, a disaster influence analysis module and a cultivated land resource management strategy module;
The image preprocessing module adopts an absolute radiation calibration method and a medium resolution atmospheric transmission model based on remote sensing satellite data, eliminates the atmospheric scattering and absorption influence of an image, optimizes the accuracy and reliability of the data and generates a corrected remote sensing image;
The cultivated land area dividing module is used for identifying land types, including cultivated lands and non-cultivated lands, based on the corrected remote sensing images, applying a K-means clustering algorithm, and identifying boundaries and areas of the cultivated lands through iterative optimization to generate cultivated land area dividing information;
The soil characteristic and humidity analysis module is used for identifying a soil characteristic value based on the cultivated area division information by adopting a spectrum analysis method, analyzing the change trend of soil humidity by utilizing a long-period memory network, and evaluating the physical and chemical properties and the moisture condition of soil by combining a normalized difference moisture index enhancement model to generate a soil characteristic and humidity analysis result;
The crop growth state evaluation module analyzes remote sensing data in a crop growth period, including reflectivity and spectrum index, based on the ploughing area division information by adopting a random forest algorithm, evaluates the health state of crops, including chlorophyll content and biomass, and generates a crop growth state evaluation result;
The disaster influence analysis module is used for analyzing the influence of disasters on crop coverage and soil erosion conditions by adopting a convolutional neural network and a decision tree algorithm based on the cultivated area division information and the crop growth state evaluation result to generate a disaster influence analysis result;
The cultivated land resource management strategy module adopts multi-standard decision analysis and an artificial neural network to analyze and optimize agricultural management activities, including irrigation, fertilization and pest control measures, based on soil characteristics and humidity analysis results, crop growth state assessment results and disaster influence analysis results, and generates a cultivated land resource management scheme through deep learning model training and optimization.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410322861.5A CN117935081B (en) | 2024-03-21 | Cultivated land change monitoring method and system based on remote sensing satellite data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410322861.5A CN117935081B (en) | 2024-03-21 | Cultivated land change monitoring method and system based on remote sensing satellite data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117935081A CN117935081A (en) | 2024-04-26 |
| CN117935081B true CN117935081B (en) | 2024-07-05 |
Family
ID=
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117037076A (en) * | 2023-10-09 | 2023-11-10 | 中国农业科学院农田灌溉研究所 | Intelligent soil moisture content monitoring method based on remote sensing technology |
| CN117197662A (en) * | 2023-08-28 | 2023-12-08 | 贵州师范学院 | Rural land remote sensing monitoring method and system based on Internet of things |
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117197662A (en) * | 2023-08-28 | 2023-12-08 | 贵州师范学院 | Rural land remote sensing monitoring method and system based on Internet of things |
| CN117037076A (en) * | 2023-10-09 | 2023-11-10 | 中国农业科学院农田灌溉研究所 | Intelligent soil moisture content monitoring method based on remote sensing technology |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20180330247A1 (en) | Computer-implemented methods, computer readable medium and systems for a precision agriculture platform that integrates a satellite date model and an orchard data model | |
| Hu et al. | Integrating coarse-resolution images and agricultural statistics to generate sub-pixel crop type maps and reconciled area estimates | |
| Dehkordi et al. | Yield gap analysis using remote sensing and modelling approaches: Wheat in the northwest of Iran | |
| US10943173B2 (en) | Computer-implemented methods, computer readable medium and systems for generating a satellite data model for a precision agriculture platform | |
| CN115205695B (en) | Method and system for determining planting strategy according to planting data | |
| Li et al. | A county-level soybean yield prediction framework coupled with XGBoost and multidimensional feature engineering | |
| Jiang et al. | Monitoring the coastal environment using remote sensing and GIS techniques | |
| Tang et al. | Estimating wheat grain yield by assimilating phenology and LAI with the WheatGrow model based on theoretical uncertainty of remotely sensed observation | |
| CN117077868A (en) | Model-based grain yield prediction system and method | |
| Zhang et al. | Artificial intelligence in soil management: The new frontier of smart agriculture | |
| Rouze et al. | Evaluating unoccupied aerial systems (UAS) imagery as an alternative tool towards cotton-based management zones | |
| CN117935081B (en) | Cultivated land change monitoring method and system based on remote sensing satellite data | |
| Ayub et al. | Wheat Crop Field and Yield Prediction using Remote Sensing and Machine Learning | |
| WO2023131949A1 (en) | A versatile crop yield estimator | |
| CN117935081A (en) | Cultivated land change monitoring method and system based on remote sensing satellite data | |
| Azzari et al. | Understanding the requirements for surveys to support satellite-based crop type mapping | |
| Fadl et al. | Assessing soil productivity potential in arid region using remote sensing vegetation indices | |
| Taylor et al. | Improving spatial predictions of Eucalypt plantation growth by combining interpretable machine learning with the 3-PG model | |
| Bhimanpallewar et al. | A Survey of Automated Advisory Systems in Agriculture Field | |
| Parmar et al. | Remote Sensing and GIS Based Crop Acreage Estimation of the Rabi Season Growing Crop of the Middle Gujarat (India) | |
| Chen et al. | Forest age estimation using UAV-LiDAR and Sentinel-2 data with machine learning algorithms-a case study of Masson pine (Pinus massoniana) | |
| Grubert | A synergistic approach to sugarcane yield forecasting using machine learning, remote sensing, and process-based modeling | |
| Habyarimana et al. | Yield Prediction in Sorghum (Sorghum bicolor (L.) Moench) and Cultivated Potato (Solanum tuberosum L.) | |
| Estacio | Simulating the future of the Ifugao rice terraces through observations of the past | |
| CHAKRAWARTI et al. | A Potential Role of Geoinformatics for Food Security and Agriculture: An Introduction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |