Cotton water demand sensing and predicting system

Abstract

The invention relates to a cotton water demand sensing and predicting system, which relates to the technical field of agricultural water-saving irrigation and is characterized in that the influence of water and air environment quality on cotton water demand characteristics is studied by setting a project through numerical analysis, statistical analysis and other methods, so that the cotton water demand sensing and predicting system is established, and a favorable guarantee is provided for efficient utilization of water resources. The method comprises the following steps of: inputting main factors influencing the cotton water demand characteristics by analyzing the influence of a water-air environment on the cotton water demand characteristics and applying an artificial neural network prediction method in artificial intelligence to establish a cotton water demand perception and prediction system, thereby laying a foundation for further establishing a cotton irrigation decision system; the cotton water demand sensing and predicting system is utilized, an artificial intelligence method is adopted, and the application of the artificial intelligence technology in irrigation agriculture is better played. The research provides an important scientific basis for the influence of the moisture environment on the water demand characteristic of cotton, further promotes the development of intelligent agriculture, improves the utilization of water resources, and provides a theoretical basis for intelligent irrigation decision-making research.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural water-saving irrigation, and in particular to a cotton water demand sensing and prediction system. Background Art

[0002] In recent years, with the rapid development of the economy, people's demand for water has continued to increase, and water shortage has become a global problem. The efficient and rational use of water resources has become the focus of global attention. Among various water-using sectors, agriculture is the largest water-consuming industry in the world, and water use occupies a dominant position, especially irrigation water, which accounts for about 70% of global water use. In particular, the environmental changes in water quality and fine particulate matter pollution in air pollution, as well as the health issues of humans, animals and plants, are of great concern to everyone. Therefore, studying the effects of recycled water (recycled water) after sewage treatment and air quality index, PM2.5 and PM10 on cotton water demand in the northern slope economic zone of Tianshan Mountains in Xinjiang has important practical significance for saving water resources.

[0003] Cotton is not only the largest cash crop among my country's agricultural products, but also a special commodity that is related to the national economy and people's livelihood. Cotton has important uses in agriculture, medicine, national defense, textile industry, etc. Water is an indispensable condition for cotton growth. Xinjiang is an important cotton production base in my country. In the current context of increasingly scarce water resources, it is particularly important to realize intelligent precision irrigation of cotton and formulate a set of intelligent management decision-making systems suitable for efficient water use of cotton in my country.

[0004] In addition, to realize intelligent precision irrigation, modern artificial intelligence technology (AI) is needed. Since the beginning of the 21st century, the world has entered a new stage of information development marked by big data. The key in the big data value chain is to analyze the data and deeply mine the various information hidden in the data to assist in decision-making. The research focus of agricultural big data at home and abroad is mainly on agricultural big data collection and integration, storage management, monitoring, data mining, intelligent decision-making and platform construction. Among them, important methods of big data analysis include artificial intelligence technology, and important methods of artificial intelligence technology include machine learning. Deep learning (DL) in machine learning is one of the most popular methods today. Deep learning is developed from artificial neural network (ANN) and is a new field in machine learning. It is highly valued by people. At present, smart agriculture has also become the future development direction of agriculture. In the process of crop planting, people can understand field information in real time and make the best decision, which plays an important role in realizing intelligent decision-making and management of agriculture.

[0005] (1) Current status of research on the impact of water and air quality on crop water demand characteristics

[0006] Crop water requirement characteristics, namely actual crop water requirement (cropevaportranspiration, ET), is one of the key elements of the soil-plant-atmosphere continuum (SPAC), and is also an important basic data for irrigation project planning, design and management, and is the basis of most intelligent irrigation decision-making systems. The actual water requirement characteristics of the same crop planted in the same soil environment at the same growth stage are mainly affected by meteorological environmental factors, water quality environmental factors and other factors.

[0007] As one of the main sources of irrigation water, the solute composition of natural surface water in the circulation process is affected by many factors such as basin geology, geography, climate, hydrology, and human activities. In recent decades, water pollution has become a hot issue in the field of environment, especially the health issues of the environment and human beings, animals and plants involved in water quality changes. ElBilali A. et al. used Boosting, random forest and other systems to predict and evaluate irrigation water quality indicators, and established a water quality prediction system suitable for similar areas; Lermontov A. et al. used artificial intelligence-based computing methods to create a new water quality indicator based on fuzzy logic, namely the fuzzy water quality index (FWQI), and evaluated the applicability of this method. Chen J. et al. pointed out that the water quality of the basin has a great impact on agricultural irrigation, and used machine learning methods to establish a water quality intelligent monitoring system based on the fusion random vector function chain network (RVFL) and group data processing (GMDH) system, which has good performance. Li Xueqing and others in China predicted water quality in a large spatial range. The prediction method used machine learning techniques such as support vector machines, decision trees, and artificial neural networks to establish a prediction system for regional water quality levels. Compared with existing water quality prediction methods, machine learning methods have fast calculation speed, do not need to extract statistical features of data, are simple to operate, and are easier to use in water environment management. Zang Chuankai and others studied the color changes of water bodies and the identification of suspected polluted water bodies through UAV hyperspectral remote sensing images. A water quality parameter inversion system was constructed based on measured data. The study shows that the inversion results of water body color parameters and water quality parameters with high spatiotemporal continuity obtained by UAV hyperspectral images provide reliable technical support for river and lake water environment surveys while saving costs. At present, research on irrigation water sources mainly focuses on water environment prediction, monitoring, simulation and evaluation. There is still a lack of research on the mechanism of the influence of irrigation water quality on crop water demand characteristics. Revealing the relationship between them is of great significance for improving irrigation water utilization efficiency and conducting intelligent regulation and management.

[0008] In addition to water environment quality, atmospheric environment quality will also have an impact on crops. Atmospheric pollution can affect the growth of humans, animals and plants. The US Environmental Protection Agency believes that the content of particulate pollutants in the air may become the most important air quality health indicator in the future. Therefore, the study of particulate matter pollution in air pollution has become a hot issue. Among them, artificial intelligence has been studied and applied in water quality monitoring and air quality monitoring in the agricultural field at home and abroad. Zhao C. et al. used a random forest system to estimate PM2.5 based on satellite, meteorological and terrain data, and considered the impact of meteorological disasters. The system performed well. Pak U. et al. proposed a CNN-LSTM system that combines the network CNN in deep learning with the long short-term memory network (LSTM) to predict the daily average concentration of PM2.5. The system not only considers the linear correlation between the target parameters and the observed parameters, but also considers the nonlinear correlation. The results show that the system has good prediction performance. Liu Yuanyuan constructed interactive spatiotemporal features between multiple sites, built a deep spatiotemporal system combining CNN and LSTM, found out the features that have a greater impact on the air quality index (AQI) and focused on them, and constructed an AQI prediction system that integrates CNN-LSTM and attention mechanism. The system has better performance in predicting the air quality index than the base system. Zhang Ailin et al. selected the two factors that have the greatest impact on air pollution for modeling, and used a multivariate fractional gray system to study the air quality in Zhengzhou, with good prediction results. Compared with the national atmospheric environmental capacity, Xinjiang's atmospheric self-purification capacity is relatively low, and this feature is particularly obvious during the high-pollution season. In addition, there is little research on the relationship and mechanism between air environmental quality and crop water demand. The application of these theoretical results to explore the relationship between them can not only manage the irrigation intelligent decision-making system, but also further save irrigation water and improve water use efficiency.

[0009] (2) Current status of research on crop water requirement prediction

[0010] Reasonable irrigation schedules and plans can be determined by predicting ET. ET can be measured directly, but these devices are expensive to install and maintain, and are complex to operate, so their use is limited. If ET is estimated by establishing a mathematical system, the input parameters of the system are not easy to obtain, and some systems also require a large amount of meteorological data. Therefore, we have to seek new ways.

[0011] When we focus on the acquisition of ET values, that is, the system output values, the above problems can be overcome by using a black box empirical system. The black box performs a mathematical mapping between historical inputs and outputs without studying the physical information of the system. Among them, ANN has been widely used in fields such as hydrology and water resources engineering, including the estimation of ET. In 2013, Mohammad-Bagher Aghajanloo et al. used genetic algorithms to optimize the parameters of neural networks. Based on the meteorological data obtained at the Tabriz synoptic station from 2000 to 2005, multilayer perceptrons (MLP), radial basis functions (RBF) and Jordan–Elman networks were used, and they were optimized with genetic algorithms to predict potato ET. The results showed that the ANN method can better predict potato ET. In 2018, Dahua Tang et al. used genetic algorithms to optimize the ANN (three-layer back propagation neural network) system in Shouyang, Shanxi to predict the ET of corn under partial mulch coverage. The input factors selected meteorological data and crop data, and the crop data included leaf area index (LAI) and plant height (hc). Meteorological data include temperature, relative humidity, solar radiation, and wind speed. The results show that the GANN system can be used as a recommendation system for the ET system. The prediction system with only meteorological data as input is acceptable, but the performance is worse than the prediction system with meteorological data and crop data as input. In 2019, Nazanin Abrishami, Ali Reza Sepaskhah et al. used ANN to estimate the actual daily evapotranspiration of wheat and corn. The ANN system used a multi-layer feedforward neural network with one or two hidden layers, a sigmoid transfer function (logsigmoid and tansigmoid) in the middle layer, a linear transfer function in the output layer, and the trainBR and trainLM learning algorithms. Finally, the optimal system input and multi-layer feedforward network system for estimating the actual evapotranspiration of wheat and corn were obtained. The results show that artificial neural network is an effective method for estimating the actual evapotranspiration of wheat and corn. Studying the crop water demand prediction system and accurately predicting the crop water demand is the basis and prerequisite for better guiding irrigation decisions and saving irrigation water.

[0012] (3) Research status of irrigation decision-making system

[0013] Decision Support System (DSS) was proposed by Keen et al. in the 1970s. It is a comprehensive subject involving artificial intelligence, communication technology and other directions.

[0014] Foreign researchers have conducted relevant research on DSS systems or irrigation decision indicators, such as the DSSAT system (decision support system for agrotechnology transfer) that has been used in countries around the world for more than ten years; Pulido-Calvo et al. established a multivariate linear system in southern Spain to predict the water demand of corn in the region, and showed good results in real-time irrigation guidance. Dureun et al. predicted soil moisture changes through an artificial neural network system to make irrigation decisions. The irrigation decision system effectively improved crop water use efficiency and reduced system energy consumption. Cvejic et al. also pointed out that in Slovenia, farmers have changed from irrigation plans based on experience and assumptions to irrigation plans based on real-time soil-water monitoring, and conducted irrigation research based on irrigation decision support systems. In addition, one of the hot topics in the research of decision-making systems at home and abroad is to make decisions based on environmental factors that affect crop ET. In terms of soil moisture parameter indicators, the technology currently used for soil moisture monitoring in a large area abroad is mainly based on generalized electromagnetic methods, such as time domain reflectometry (TDR), frequency domain reflectometry (FDR), near infrared method and large-scale remote sensing technology. In terms of crop physiological signal indicators, foreign scholars such as Damiano and Tanner proposed that fruit tree sap flow, stem variation and canopy temperature are representative in the precision irrigation decision-making of orchards. In terms of environmental factor indicators, representative calculation methods are Turc radiation method, Blaney-Criddle temperature method and Penman comprehensive method. The United States, Israel and other countries have integrated advanced computer control technology, neural networks, and expert system control to study various forms of irrigation decisions based on transpiration, crop water demand systems, etc., taking into account irrigation technology and water demand of irrigation objects, taking into account factors such as temperature, humidity, wind speed and plant water demand laws, etc., so as to achieve automatic irrigation on demand and in quantity based on rainfall and plant water balance analysis, thus forming advanced precision irrigation decision-making technology. The latest ET system developed by Hunter Industries in the United States effectively prevents water waste caused by subjective factors in human management by tracking the on-site climate and automatically calculating the scientific irrigation volume every day, and scientifically achieves reasonable irrigation. Recently, Wong W.Y. et al. discussed the interaction between water, soil and air pollutants on mangrove ecosystems and the technical research progress of applying artificial intelligence in decision support systems.

[0015] The research on irrigation decision support system started relatively late in my country and is in the initial stage of research and development and trial. After the 1990s, my country introduced a large number of intelligent irrigation equipment from abroad. With the development of science and technology and the diversification of monitoring methods, domestic scholars have gradually optimized the decision-making system. In recent years, in terms of decision-making system: Yu Guoxiong et al. studied the information acquisition and intelligent irrigation decision-making expert system of litchi orchard based on the Internet of Things; Ji Ronghua et al. studied the soil moisture prediction based on multi-valued neuron complex neural network, and the prediction results can be used as a theoretical basis for the formulation of water-saving irrigation strategies. In order to improve the irrigation water utilization efficiency of solar greenhouses and give full play to the guiding role of existing irrigation decision-making theories, Gu Zhe et al. constructed a real-time and precise irrigation decision and control system based on ET and water balance methods, realizing the combination of control and decision-making, and providing a successful example for practical application; Li et al. took alfalfa as the research object in Zhuozhou City, Hebei Province, and developed a real-time irrigation decision support system (IDSS) based on fuzzy reasoning and related software; Yu Helong et al. proposed an intelligent and precise irrigation decision system for ginseng planting under the Internet of Things, which collects environmental information such as soil moisture content, air temperature and humidity in real time, transmits information to the information processing module using the Zigbee network, and establishes a DSS. The results show that the designed system can achieve effective decision-making. At present, domestic decisions are mainly made based on crop evaporation and transpiration. Domestic precision irrigation decision-making indicators have also achieved certain results in terms of soil moisture parameters. The standing wave ratio (SWR) soil moisture sensor developed by China Agricultural University has been widely used. In terms of crop physiological signal indicators, domestic researchers have conducted relevant research on canopy temperature indicators. For example, Cai Jiabing and others designed a real-time irrigation decision-making system that can continuously monitor field crop canopy temperature, environmental information and soil moisture online, and installed it in farmland for actual operation and observation. The system determines the degree of drought and provides the required irrigation amount based on fuzzy logic algorithms. In terms of environmental factor indicators, domestic decisions are currently mainly made based on crop evaporation and transpiration.

[0016] Therefore, developing smart agriculture, predicting crop water requirements, and realizing intelligent irrigation are particularly important for saving water and solving the problem of water shortage. At the same time, the study applied the air quality index (AQI, PM2.5) and water quality indicators of reclaimed water irrigation (PH value, total phosphorus, total nitrogen, endocrine disruptors, antibiotics, resistance genes, etc.) to the irrigation intelligent decision-making management system, which is of great significance for the research on intelligent irrigation decision-making.

[0017] In summary, there is an urgent need for a cotton water demand sensing and prediction system to solve the above problems. Summary of the invention

[0018] The technical problem to be solved by the present invention is to provide a scientifically optimized cotton water demand perception prediction system, which can apply the air quality index and water quality indicators of recycled water irrigation to cotton intelligent irrigation, and provide a theoretical basis for intelligent irrigation decision-making research.

[0019] To achieve the above object, the present invention provides a cotton water demand perception prediction system, comprising the following steps:

[0020] Step S1, analyzing the impact of water vapor environment quality on crop water demand characteristics and setting up multiple experimental areas for verification experiments;

[0021] Step S2: Collect and measure agricultural big data (such as weather, irrigation water quality, soil, irrigation system, etc.), measure crop water requirements, and measure crop growth indicators;

[0022] Step S3, conducting economic benefit analysis on each experimental area;

[0023] Step S4: Establish a cotton water demand sensing and prediction system.

[0024] As a preferred embodiment, the specific implementation method of step S1 adopted in the present invention is: using meteorological stations, remote sensing photos or manual detection to obtain water quality and meteorological data, cooperating with field experiments, conducting data analysis and processing, parameter inversion, obtaining multispectral data maps, and analyzing the impact of water quality and meteorological data on crop water demand characteristics; setting up multiple verification experiments, in which cotton is planted using drip irrigation, and setting up two experimental area treatments: T1 uses reclaimed water irrigation throughout the growth period, T2 uses reclaimed water irrigation during the critical period of crop water demand, and well water irrigation during the non-critical period of water demand; setting the control CK as well water irrigation during the entire growth period of cotton.

[0025] As a preferred implementation, the crop growth indicators in step S2 include plant height, yarn diameter, crop dry weight, cotton yield, and quality.

[0026] As a preferred implementation, the crop water requirement in step S2 is measured by using a soil drying method combined with TDR.

[0027] As a preferred implementation, the specific implementation of step S3 adopted by the present invention is: using formula (1) to calculate WUE, using formula (2) to calculate IWUE, and using formula (3) to calculate PUE.

[0028] WUE=Y / ET(1)

[0029] IWUE=Y / I(2)

[0030] PUE=Y / P(3)

[0031] Where:

[0032] WUE is water use efficiency (based on economic yield Y);

[0033] Y(kghm -2 ) is economic output;

[0034] ET (mm) is the total field water consumption during the whole growth period of cotton;

[0035] IWUE is irrigation water use efficiency;

[0036] I (mm) is the irrigation amount during the whole growth period of cotton;

[0037] PUE is the precipitation production efficiency;

[0038] P (mm) is the rainfall in the experimental field during the entire growth period of cotton.

[0039] As a preferred embodiment, the specific implementation method of step S4 adopted in the present invention is: through the above-mentioned measurement results, the collected data is analyzed using data analysis software such as Excel, Origin or Spass, and the relationship between various decision-making control indicators selected under cotton drip irrigation and cotton water demand is established; thereby establishing a cotton water demand perception prediction system under different decision-making indicators; performing data normalization processing on the selected irrigation decision indicators, using intelligent algorithms (deep learning artificial neural networks, optimized deep learning artificial neural networks), establishing a cotton intelligent perception water demand prediction system, and using the root mean square error (Root Mean Square Error, RMSE), mean absolute error (mean absolute error, MAE) and coefficient of determination (coefficient of determination, R2) to analyze the prediction results of each system, and select a high-precision cotton water demand prediction system.

[0040] As a preferred implementation, based on the most accurate cotton water demand intelligent perception prediction system, a deep learning-based irrigation decision system is established in combination with soil moisture content and meteorological data to predict the irrigation date and amount. Based on soil sensor monitoring data and rainfall data, the actual irrigation system is obtained using the farmland water balance equation. Finally, the results of irrigation using the decision system are evaluated with the results of irrigation using the actual calculated irrigation system, and the evaluation indicators are selected as irrigation water utilization efficiency, cotton yield, quality, and water saving rate.

[0041] As a preferred implementation, based on the established crop water demand perception and prediction system, combined with soil sensor monitoring data and meteorological forecast data, a multi-objective optimization irrigation decision-making intelligent algorithm is established. Based on the principle of water balance, the trial algorithm or deep learning neural network is used to optimize the irrigation decision algorithm in real time, thereby establishing a real-time optimized irrigation decision system.

[0042] Compared with the prior art, the technical solution of this application has the following beneficial effects:

[0043] 1. The cotton water demand perception prediction system demonstrated by the present invention studies the influence of water vapor environment quality on cotton water demand characteristics by setting up projects through numerical analysis, statistical analysis and other methods, thereby establishing a cotton demand perception prediction system to provide favorable guarantees for the efficient use of water resources; through the analysis of the influence of water vapor environment on cotton water demand characteristics, the artificial neural network prediction method in artificial intelligence is used to input the main factors affecting cotton water demand characteristics, and a cotton water demand perception prediction system is established to lay the foundation for further establishing a cotton irrigation decision-making system; the cotton water demand perception prediction system is used to adopt artificial intelligence methods to better play the application of artificial intelligence technology in irrigation agriculture. This study provides an important scientific basis for the influence of water vapor environment on cotton water demand characteristics, further promotes the development of smart agriculture, improves the utilization of water resources, and provides a theoretical basis for the study of intelligent irrigation decision-making.

[0044] 2. The cotton water demand perception prediction system demonstrated by the present invention reveals the influence of recycled water irrigation and air quality on cotton water demand through the collection of agricultural big data, and selects appropriate irrigation indicators for the cotton irrigation decision-making system; and selects economically reasonable irrigation methods through the analysis of the impact on cotton growth and quality, and improves economic benefits; by selecting different irrigation decision-making indicators and artificial intelligence algorithms, a high-precision cotton water demand perception prediction system is established, and the system is evaluated to provide a reference for intelligent water-saving irrigation. At the same time, the relationship and mechanism between air environment quality and crop water demand are studied, and the relationship between them is explored. The application of these theoretical results can not only manage the irrigation intelligent system, but also further save irrigation water and improve water use efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 The present invention is a schematic diagram of the process of a cotton water demand sensing and prediction system. DETAILED DESCRIPTION

[0046] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0047] Embodiment 1:

[0048] A cotton water demand perception prediction system. In order to achieve efficient and economical irrigation water conservation, this project attempts to explore the intelligent water-saving irrigation intelligent control and management system suitable for my country's arid areas. The continuously increasing population size in the northern foothills of Tianshan Mountain in Xinjiang has caused tremendous pressure on the ecological environment, including water resource tension, water resource and atmospheric environment pollution. Therefore, the research area of ​​this project is planned to be selected in the northern foothills of Tianshan Mountain. The cotton in the economic belt of the northern slope of Tianshan Mountain in Xinjiang was selected as the research object. Through the collection of agricultural big data, the influence of medium-water irrigation and air quality on cotton water demand was revealed, and suitable irrigation indicators were selected for the cotton water demand intelligent perception prediction system. And through the analysis of the impact on cotton growth and quality, an economically reasonable irrigation method was selected to improve economic benefits. By selecting different irrigation decision-making indicators and artificial intelligence algorithms, a high-precision cotton water demand perception prediction system was established, and the system was evaluated to provide a reference for Xinjiang intelligent water-saving irrigation.

[0049] Step S1, analyzing the impact of water vapor environment quality on crop water demand characteristics and setting up multiple experimental areas for verification experiments;

[0050] The specific implementation method is: use meteorological stations, remote sensing photos or manual detection to obtain water quality and meteorological data, cooperate with field experiments, conduct data analysis and processing, parameter inversion, obtain multispectral data maps, and analyze the impact of water quality and meteorological data on crop water demand characteristics; set up multiple verification experiments, in which cotton is planted using drip irrigation, and set up two experimental area treatments: T1 uses reclaimed water irrigation throughout the growth period, T2 uses reclaimed water irrigation during the critical period of crop water demand, and well water irrigation during the non-critical period of water demand; set the control CK as well water irrigation for the entire growth period of cotton.

[0051] Step S2, collect and measure agricultural big data (meteorological, irrigation water quality, soil, irrigation system and other information), measure crop water requirement (measured by soil drying method combined with TDR), and measure crop growth indicators (plant height, warp diameter, crop dry weight, cotton yield, quality, etc.); through field experiments, collect water quality, soil, crop, and meteorological data of reclaimed water, analyze the effects of reclaimed water irrigation, well water irrigation, and reclaimed water irrigation during the critical period of crop water requirement on cotton: and the impact of air quality on cotton water requirement.

[0052] Step S3, conducting economic benefit analysis on each experimental area;

[0053] Use formula (1) to calculate WUE, formula (2) to calculate IWUE, and formula (3) to calculate PUE.

[0054] WUE=Y / ET(1)

[0055] IWUE=Y / I(2)

[0056] PUE=Y / P(3)

[0057] Where:

[0058] WUE is water use efficiency (based on economic yield Y);

[0059] Y(kghm -2 ) is economic output;

[0060] ET (mm) is the total field water consumption during the whole growth period of cotton;

[0061] IWUE is irrigation water use efficiency;

[0062] I (mm) is the irrigation amount during the whole growth period of cotton;

[0063] PUE is the precipitation production efficiency;

[0064] P (mm) is the rainfall in the experimental field during the whole growth period of cotton;

[0065] Through the analysis of cotton yield, quality, and economic benefits, the cotton drip irrigation mode with the highest economic benefits is proposed for subsequent research. The three experimental treatments of well water irrigation, recycled water irrigation, and recycled water irrigation during the critical period of crop water demand are discussed, as well as the impact of air quality on cotton water demand, growth, yield, and quality. The economic benefits of the three experimental treatments are analyzed.

[0066] Step S4, establishing a cotton water demand perception prediction system;

[0067] The specific implementation method is: through the above-mentioned measurement results, use data analysis software such as Excel, Origin or Spass to analyze the collected data, and establish the relationship between various decision-making control indicators selected under cotton drip irrigation and cotton water demand; thereby establishing a cotton water demand perception prediction system under different decision-making indicators; normalize the data of the selected irrigation decision indicators, and use intelligent algorithms (deep learning artificial neural network, optimized deep learning artificial neural network) to establish a cotton intelligent perception water demand prediction system, and use the root mean square error (Root Mean Square Error, RMSE), mean absolute error (mean absolute error, MAE) and coefficient of determination (coefficient of determination, R2) to analyze the prediction results of each system, and select a high-precision cotton water demand prediction system.

[0068] Different decision-making indicators were selected, and artificial intelligence algorithms (deep learning artificial neural networks and optimized deep learning artificial neural networks) were used to establish an intelligent cotton water demand prediction system based on different decision-making indicators to predict cotton water demand. Through field experiments, the cotton water demand perception prediction system was evaluated and analyzed based on the root mean square error, mean absolute error, and Nash coefficient, and the cotton water demand perception prediction system with the highest prediction accuracy within the research scope was sought.

[0069] The project uses numerical analysis, statistical analysis and other methods to study the impact of water and air environment quality on cotton water demand characteristics, thereby establishing a cotton demand perception prediction system to provide favorable guarantees for the efficient use of water resources; through the analysis of the impact of water and air environment on cotton water demand characteristics, the artificial neural network prediction method in artificial intelligence is used to input the main factors affecting cotton water demand characteristics, and a cotton water demand perception prediction system is established to lay the foundation for further establishing a cotton irrigation decision-making system; using the cotton water demand perception prediction system and artificial intelligence methods, a cotton irrigation decision-making system is established to better play the application of artificial intelligence technology in irrigation agriculture. This study provides an important scientific basis for the impact of water and air environment on cotton water demand characteristics in Xinjiang, further promotes the development of smart agriculture, improves the utilization of water resources in Xinjiang, and provides a theoretical basis for the study of intelligent irrigation decision-making.

[0070] Furthermore, based on the most accurate cotton water demand intelligent perception prediction system, combined with soil moisture content and meteorological data, an irrigation decision system based on deep learning was established to predict the irrigation date and amount. Based on soil sensor monitoring data and rainfall data, the actual irrigation system was obtained using the farmland water balance equation. Finally, the results of irrigation using the decision system were evaluated with the results of irrigation using the actual calculated irrigation system, and the evaluation indicators were selected as irrigation water utilization efficiency, cotton yield, quality, and water saving rate.

[0071] Based on the established crop water demand perception and prediction system, combined with soil sensor monitoring data and meteorological forecast data, a multi-objective optimization irrigation decision-making intelligent algorithm is established. Based on the water balance principle, the trial algorithm or deep learning neural network is used to optimize the irrigation decision algorithm in real time, thereby establishing a real-time optimized irrigation decision-making system.

[0072] This system is of great significance for water-saving irrigation of farmland in arid areas:

[0073] (1) Promote ecological civilization construction and sustainable social development and improve economic efficiency

[0074] Changes in natural ecology play a decisive role in the rise and fall of human civilization. Water is the basis of survival and the foundation of ecological civilization construction, and the construction of water ecological civilization is also its core content. Saving and utilizing water resources is one of the important components of water ecological civilization construction. Implementing agricultural water-saving irrigation and intelligent intelligent regulation and management in Xinjiang cotton fields in my country's arid areas can improve irrigation water utilization efficiency, reduce irrigation costs, improve economic benefits, and achieve efficient agricultural water use, which will help promote my country's ecological civilization construction and promote sustainable economic and social development, and is of great significance to ensuring our ecological security.

[0075] (2) Promote the development of water-saving agriculture, improve crop quality, save irrigation water resources, and enrich water-saving irrigation technology

[0076] At present, there have been a lot of studies on water quality environment and air environment, but there are very few studies on the impact of water and air environment on crop water demand characteristics. Therefore, the results of this study provide a certain theoretical basis for further improving crop water demand characteristics, which is more conducive to promoting the development of water-saving agriculture in my country, and has important practical significance for improving crop quality, improving irrigation water utilization efficiency, saving irrigation water resources, and enriching water-saving irrigation technology.

[0077] (3) Promote the development of smart agriculture and provide a theoretical basis for research on smart irrigation decision-making

[0078] From the research background, it can be seen that promoting water-saving irrigation technology, implementing planned water use, conducting actual crop water demand research and precision intelligent water-saving irrigation, and accelerating the pace of agricultural modernization are urgent needs in my country. This study uses machine learning (ML) to predict cotton water demand and carry out intelligent precision irrigation. The application research of agricultural information technology, monitoring of air quality index and water quality indicators, provides a theoretical basis for subsequent intelligent precision irrigation decision-making and intelligent intelligent regulation and management system research. It plays an extremely important role and significance in promoting agricultural modernization, the development of smart agriculture, and improving the level of agricultural automation and work efficiency.

[0079] (4) Provide technical support for water resources management decision-making, promote the integration of information technology and traditional industries, and accelerate the pace of agricultural modernization

[0080] We must accelerate the pace of agricultural information technology and steadily advance our country's development towards smart agriculture. Smart agriculture is an important part and symbol of modern agriculture. At the same time, reducing irrigation costs and improving economic benefits is not only the general direction of agricultural development in our country, but also in the world.

[0081] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A cotton water demand sensing and prediction system, characterized in that: The following steps are involved: Step S1, analyzing the impact of water vapor environment quality on crop water demand characteristics and setting up multiple experimental areas for verification experiments; Step S2: Collect and measure agricultural big data (such as weather, irrigation water quality, soil, irrigation system, etc.), measure crop water requirements, and measure crop growth indicators; Step S3, conducting economic benefit analysis on each experimental area; Step S4: Establish a cotton water demand sensing and prediction system.

2. A cotton water demand sensing and prediction system according to claim 1, characterized in that: The specific implementation method of step S1 is: using meteorological stations, remote sensing photos or manual detection to obtain water quality and meteorological data, cooperating with field experiments, conducting data analysis and processing, parameter inversion, obtaining multispectral data maps, and analyzing the impact of water quality and meteorological data on crop water demand characteristics; setting up multiple verification experiments, in which cotton is planted using drip irrigation, and setting up two experimental area treatments: T1 uses reclaimed water irrigation during the entire growth period, T2 uses reclaimed water irrigation during the critical period of crop water demand, and well water irrigation during the non-critical period of water demand; setting the control CK to well water irrigation during the entire growth period of cotton.

3. A cotton water demand sensing and prediction system according to claim 2, characterized in that: The crop growth indicators in step S2 include plant height, diameter of warp, crop dry weight, cotton yield and quality.

4. A cotton water demand sensing and prediction system according to claim 3, characterized in that: The crop water requirement in step S2 is measured by using a soil drying method combined with TDR.

5. A cotton water demand sensing and prediction system according to claim 4, characterized in that: The specific implementation of step S3 is: Use formula (1) to calculate WUE, formula (2) to calculate IWUE, and formula (3) to calculate PUE. WUE=Y / ET(1) IWUE=Y / I(2) PUE=Y / P(3) Where: WUE is water use efficiency (based on economic yield Y); Y(kghm -2 ) is economic output; ET (mm) is the total field water consumption during the whole growth period of cotton; IWUE is irrigation water use efficiency; I (mm) is the irrigation amount during the whole growth period of cotton; PUE is the precipitation production efficiency; P (mm) is the rainfall in the experimental field during the entire growth period of cotton.

6. A cotton water demand sensing and prediction system according to claim 5, characterized in that: The specific implementation method of step S4 is: through the above-mentioned measurement results, the collected data is analyzed using data analysis software such as Excel, Origin or Spass, and the relationship between various decision-making control indicators selected under cotton drip irrigation and cotton water demand is established; thereby establishing a cotton water demand perception prediction system under different decision-making indicators; performing data normalization processing on the selected irrigation decision indicators, using intelligent algorithms (deep learning artificial neural networks, optimized deep learning artificial neural networks), establishing a cotton intelligent perception water demand prediction system, and using root mean square error (Root Mean Square Error, RMSE), mean absolute error (mean absolute error, MAE) and coefficient of determination (coefficient of determination, R2) to analyze the prediction results of each system, and select a high-precision cotton water demand prediction system.

7. A cotton water demand sensing and prediction system according to claim 6, characterized in that: Based on the most accurate intelligent perception prediction system for cotton water demand, an irrigation decision-making system based on deep learning is established in combination with data such as soil moisture content and meteorological data to predict the irrigation date and amount of irrigation. Based on soil sensor monitoring data and rainfall data, the farmland water balance equation is used to obtain the actual irrigation system. The results of irrigation using the decision-making system are evaluated with the results of irrigation using the actual calculated irrigation system, and the evaluation indicators include irrigation water utilization efficiency, cotton yield, quality and water saving rate.

Images