302 results about "Soil water" patented technology

The invention relates to the field of power system supply and demand interaction, and discloses an intelligent greening irrigation system and a water-saving control method, the system comprises a multi-source data acquisition module, a space-time tensor modeling module, a dynamic decision generation module, an edge calculation acceleration module, a distributed execution control module and a drip irrigation module; the method comprises the following steps: collecting multispectral, meteorological and soil moisture data of a plant canopy in real time to generate an original data set; a fusion data set is obtained through space-time alignment and normalization preprocessing; the water demand is predicted based on an LSTM-Transform model, and an irrigation safety interval is screened; the irrigation uniformity and water pressure distribution are simulated and evaluated through multi-physics field coupling; a double-delay DDPG algorithm is used for optimizing valve opening parameters; the model and the execution error are corrected through Bayesian modeling, and an NDVI feedback optimization strategy is combined. According to the invention, a double-delay depth deterministic strategy gradient algorithm is introduced for dynamic optimization of an irrigation strategy, so that the capability of quickly adapting to environmental changes is achieved.

The invention discloses a soil water content inversion method and system based on L-band dual-polarization data and a medium, and belongs to the field of remote sensing technology and environment monitoring. The system comprises an SAR data acquisition module, an optical data processing module, a vegetation scattering separation module, an inversion model construction module, a spatial distribution calculation module and a risk assessment module, wherein the SAR data acquisition module is used for regularly acquiring L-band dual-polarized SAR data of a target area; the optical data processing module assists in identifying and separating vegetation scattering components by calculating vegetation indexes; the vegetation scattering separation module is used for extracting pure soil back scattering signals; the inversion model building module is used for training and optimizing a soil water content inversion model based on the pure soil back scattering signals and field sampling data; the spatial distribution calculation module is used for calculating a spatial distribution diagram of the surface soil water content of the target area; and the risk assessment module is used for carrying out landslide risk grade division and early warning management.

The invention relates to a crop irrigation demand prediction method based on an intelligent integrated prefabricated pump station, and belongs to the technical field of artificial intelligence. The method comprises the following steps: collecting a crop image in a prefabricated pump station irrigation scene, obtaining auxiliary data, calculating a daily accumulated reference evapotranspiration value, and obtaining a time-synchronized multi-modal data set; labeling and dividing the multi-modal data set; constructing an irrigation demand prediction model, wherein the irrigation demand prediction model comprises a spectral reflectivity enhancement module, a double-flow separation convolution module, a cross gating fusion module, a region sensitive pyramid module and a coupling prediction head module; inputting data in the divided training set into the model, and training the model through weighted total loss to obtain a trained model; inputting a to-be-detected image into the trained model to obtain a soil water content and transpiration prediction result; and performing irrigation decision based on the soil water content and transpiration prediction result. The accuracy of irrigation prediction can be improved.

The invention discloses a soil water content inversion method based on improved combination roughness, and relates to the field of remote sensing, and the method comprises the steps: synchronously obtaining a Sentinel-1 radar image and a Sentinel-2 optical image, and extracting different polarization backscattering coefficients, local incident angles and normalized water body indexes through preprocessing; the method comprises the following steps: generating a bare soil simulation backscattering coefficient data set, removing vegetation scattering contribution by using a water cloud model, obtaining a real bare soil backscattering coefficient, constructing a training set and a verification set containing actually measured soil water content, constructing a lookup table based on double models, and calculating the water content of the bare soil by minimizing a root-mean-square error between simulation and the real backscattering coefficient. Global search is carried out in a preset parameter space to determine an optimal earth surface root mean square height and a correlation length, a novel polynomial combination roughness is constructed, a physical correlation between the roughness and a backscattering coefficient is established, a dual-polarization empirical equation set is constructed, and simultaneous solution is carried out after parameters are optimized according to a criterion; according to the method, a more reasonable inversion result of the soil water content in a large range can be obtained.

The invention relates to the technical field of bamboo forest ecological environment monitoring, in particular to a bamboo forest ecological environment collaborative observation method based on multi-source Internet of Things sensing, and the method comprises the steps: synchronously collecting a multi-source collaborative observation parameter set of a bamboo forest slope through an Internet of Things sensor network, and then constructing a point-surface scale fusion bamboo forest moisture dynamic model; correcting evapotranspiration point cluster data by using a random forest gradient lifting algorithm to generate terrain weighted evapotranspiration, aggregating runoff data according to terrain weight to obtain a total amount of runoff sub-regions, then executing dynamic water balance constraint calculation, optimizing infiltration capacity in combination with a variation assimilation algorithm, redistributing infiltration capacity space proportion according to soil hydraulic conductivity when a residual error exceeds a threshold value, and calculating the total amount of the runoff sub-regions. Finally, according to the residual space distribution thermodynamic diagram, the evapotranspiration point cluster is migrated to the micro-topography representative area, and observation network adaptive optimization is achieved, the problems that in the prior art, the data scale is disjointed, the observation precision is insufficient, and the network is difficult to adapt to change are solved, and the bamboo forest ecological observation precision and reliability are improved.

The invention belongs to the technical field of irrigation management and control, and provides a multi-source sensing data-driven irrigation area water demand prediction and accurate water distribution method. The method comprises the steps of obtaining multi-source data of a target irrigation area when a preset irrigation condition is met; performing quantitative evaluation on the first soil water content data and the second soil water content data based on the meteorological data to determine respective confidence degrees, and giving dynamic weights based on the confidence degrees, the temporal and spatial variation characteristics and the actual condition of the irrigation area; performing fusion processing on the first soil water content data and the second soil water content data based on the dynamic weight to obtain third soil water content data; on the basis of the third soil water content data and the crop growth stage data of the target irrigation area, the water demand of the stage is predicted; and generating a control instruction of the intelligent valve of the water delivery canal system based on the water demand. The accuracy of water demand prediction can be remarkably improved, the problems of water resource waste and insufficient supply are reduced, and efficient utilization and scientific management of water resources in an irrigation area are achieved.

The invention discloses an arid region groundwater reserve estimation method and system based on multi-source remote sensing data, and relates to the technical field of groundwater monitoring, and the method comprises the steps: collecting land water reserve data and remote sensing image data of a target arid region, obtaining a land water reserve change sequence according to the land water reserve data, carrying out the region weighting of the target arid region, and obtaining a target arid region groundwater reserve change sequence; obtaining a land water reserve change sequence of the target arid region; based on the remote sensing image data of the target arid region, using a digital elevation model to obtain a surface water reserve change sequence of the target arid region, using a land surface process model to obtain a soil water reserve change sequence of the target arid region, and establishing an underground water reserve model of the target arid region; and performing error analysis on an output result of the groundwater reserve model by using a Monte Carlo simulation method. According to the method, through collaborative inversion of multi-source remote sensing data and integration of the groundwater reserve models, accurate quantification of spatial-temporal dynamic of the groundwater reserves in the arid region is realized.

The invention relates to the technical field of data processing, in particular to an agricultural water efficiency dynamic monitoring and optimizing system based on machine learning, which comprises a processor and a memory, the processor executes a computer program of the memory to realize the following steps: acquiring historical soil water content data and multi-dimensional historical environment data of any to-be-monitored position at each moment in a preset historical time period; according to the data change characteristics of the historical soil water content data and the multi-dimensional historical environment data, obtaining the correlation degree of each environment and the soil water content; when an LSTM prediction algorithm is used for prediction, a prediction time period is obtained, and the weight of each piece of historical soil water content data is obtained according to the correlation degree of each environment and the soil water content and the difference between each piece of multi-dimensional historical environment data in the prediction time period and the multi-dimensional environment data at the current moment; and the soil water content data at the next moment of the current moment is predicted, so that the accuracy of predicting the soil water content is improved.

The invention relates to a power transmission line operation state intelligent identification and analysis method, and belongs to the technical field of power transmission line state analysis. The method comprises the following steps: acquiring soil water content and foundation settlement data in real time, recognizing tower node displacement accumulation and a stress transfer path in combination with environmental factors such as temperature, humidity and wind speed, accurately judging a structural deformation abnormal region, and generating a stress concentration and transfer prediction chart through multi-source data fusion and historical data contrastive analysis. According to the method, the coordinates of potential safety hazard parts are dynamically calibrated, infinitesimal displacement monitoring errors are eliminated, finally, a tower overall structure safety evaluation report and maintenance management measures are output, through full-link data analysis and dynamic calibration, the monitoring precision and early warning capacity of tower stability under geological condition changes are remarkably improved, and safe operation of power transmission facilities is guaranteed.

The invention discloses a microbial fertilizer with growth-promoting and yield-increasing effects and a preparation method thereof. The microbial fertilizer is prepared by mixing a hydrogel fertilizer and a microbial agent, the hydrogel fertilizer is formed by taking chitosan as a base material and carrying out cross-linking reaction with 2, 2-dimethyl-5-hexenoic acid, N-(N-butoxymethyl) acrylamide and urea, and the microbial agent is prepared by loading bacteria N9 separated from farmland soil on biochar. The microbial fertilizer has good water retention performance and nutrient slow release effect, the contained microorganisms can effectively promote growth, tests show that the soil water retention rate can be remarkably increased, growth of pakchoi and other crops is promoted, the biological safety is high, and the microbial fertilizer is suitable for achieving crop growth promotion and yield increase in agricultural production.

The invention discloses a star soil water-ice in-situ quantitative analysis system and method, and belongs to the technical field of planetary science and space exploration, the star soil water-ice in-situ quantitative analysis system comprises a detection platform, and a star soil collection unit and a star soil water vapor heating extraction unit are arranged outside a cabin of the detection platform; a first detection unit, a second detection unit, a control unit and a gas distribution unit are arranged in a cabin of the detection platform; the gas distribution unit is used for distributing the water vapor extracted by the star soil water vapor heating and extracting unit to the first detection unit and the second detection unit for analyzing the content and isotope composition of the water vapor of the star soil; the first detection unit and the second detection unit adopt different detection analysis technologies; and the control unit is used for controlling the detection analysis process of the analysis system. According to the system, an in-situ sampling analysis mode of'quantitative sampling of water-containing star soil, low-loss sample transfer, efficient heating extraction and high-precision instrument analysis' is adopted, so that accurate quantitative analysis of star soil water ice is realized.

The invention belongs to the technical field of equipment planning, and discloses an automatic planning method for intelligent agricultural Internet of Things equipment, which comprises the following steps: dividing a target region into grid units, collecting soil moisture parameters and nutrient parameters of each grid, and establishing a regional digital twinborn model; obtaining a target crop of each grid, obtaining growth information from the crop knowledge base, and executing time sequence simulation to generate a water and fertilizer demand table of each grid; executing a water-fertilizer balance strategy based on the water-fertilizer demand table, and generating a water-fertilizer supply plan of each grid; analyzing spatial distribution characteristics of supply plans, and clustering adjacent grids with similar plans into irrigation management partitions; and generating a pipe network topological structure and an equipment planning scheme based on the partition layout. According to the method, a traditional uniformly distributed area is scientifically divided into management subareas with consistent internal requirements, so that unified management of the same subarea is realized, and the problem of supply and demand mismatching caused by neglect of spatial heterogeneity in a traditional method is solved.

The invention relates to the technical field of underground water state evaluation, and discloses an underground water level dynamic evaluation method based on multi-source data mapping and fluid model correction. The invention aims to solve the technical problems of insufficient underground water level monitoring precision and lack of space details caused by low spatial resolution of existing satellite gravity detection data and difficulty in accurately separating a deep underground water change signal from a land total water reserve signal. According to the method, through space-time reference unification of multi-source heterogeneous data, high-resolution surface texture features are utilized to guide a low-resolution gravitational field to carry out space downscaling reconstruction; surface water body and soil water reserve signals are removed layer by layer in combination with spectral characteristics and a thermal inertia principle; and introducing measured data to construct a spatial error correction curved surface of specific yield parameters, and identifying the underground water funnel by using an image connected domain analysis technology.

The embodiment of the invention provides a hydrological disaster early warning method and system based on big data, and relates to the technical field of hydrological disaster early warning, and the method comprises the steps: firstly obtaining a multi-source hydrological monitoring data set which comprises monitoring data including water level, flow, rainfall, soil water content and the like and geographic position coordinate data of each monitoring station; secondly, constructing a hydrological element space-time correlation evolution trajectory, recording dynamic changes of correlation relationships of various data among different sites, then tracking hydrological element space-time anomaly propagation paths, and defining a correlation relationship transmission process exceeding a conventional range; and according to the abnormal propagation path, generating a hydrological disaster dynamic early warning trigger condition including an associated change logic and a space transmission rule, and finally, based on the trigger condition, generating a hydrological disaster early warning instruction including an early warning level identifier and an affected station range, and sending the hydrological disaster early warning instruction to a target terminal, thereby comprehensively and accurately mastering hydrological changes. The accuracy and timeliness of early warning are improved, and disaster loss is effectively reduced.

The present invention discloses a method for agricultural-photovoltaic complementary rainwater harvesting and automatic irrigation on a tableland and relates to the technical field of photovoltaic power generation. The method for agricultural-photovoltaic complementary rainwater harvesting and automatic irrigation on a tableland of the present invention comprises the following steps: S1. Monitoring the soil water content of the tableland and creating a model for the daily water demand Q1 of the tableland crop; S2. Setting a minimum alarm value P1 and a maximum alarm value P2 of the soil water content of the tableland. When the soil water content of the tableland is lower than the minimum alarm value P1, the spray pipe starts spraying, and when the soil water content of the tableland reaches the maximum alarm value P2, the spray pipe stops spraying; S3.Monitoring the water level in the water collection tank and setting a first alarm level H1 and a second alarm level H2. When the water level is lower than H1, the water body is supplied to the water collection tank. When the water level is at H2, the water supply assembly stops supplying the water body to the water collection tank. The method for agricultural-photovoltaic complementary rainwater harvesting and automatic irrigation on a tableland of the present invention can use rainwater to automatically irrigate the tableland, ensure normal crop yield, and achieve win-win complementarity between agriculture and the photovoltaic industry.

The invention relates to the technical field of soil engineering model tests, and discloses a test system and method for simulating water and mud bursting of a subsea tunnel under the action of vibration, and the system comprises a multi-degree-of-freedom servo feedback vibration module which is used for providing a multidirectional vibration loading platform and feeding back a motor current signal in real time; the visual rock-soil water inrush simulation and multi-field loading module is mounted above the vibration table and is used for accommodating a sample and applying confining pressure and water pressure load; the hierarchical multi-source data acquisition module is communicated with the discharge port and the observation window, and is used for acquiring the outburst mixture and monitoring physical field data; and the intelligent control and risk assessment analysis module is connected with each module to regulate and control loading parameters and invert a catastrophe state based on feedback data. According to the method, a self-adaptive optimization loading mechanism based on the improved particle swarm optimization algorithm is constructed, so that collaborative progressive regulation and control of pore water pressure and vibration frequency parameters are realized, and a critical threshold point of water inrush and mud inrush catastrophe can be accurately captured.

The invention relates to the technical field of garden irrigation, in particular to a landscaping irrigation intelligent management system based on the Internet of Things, and the system comprises an environment monitoring module which is used for obtaining the real-time environment monitoring data, the real-time irrigation amount and the historical irrigation amount of a landscaping area; the data storage module is used for storing information of different vegetation units in the landscaping area and plant identity information corresponding to the vegetation units, the vegetation unit information comprises water and soil loss amount and soil water storage change amount, and the plant identity information comprises plant species and water amount required by plants; the irrigation prediction module is used for combining the real-time environment monitoring data, the real-time irrigation amount, the historical irrigation amount and the plant identity information through a prediction model to generate a predicted irrigation amount, adjusting the real-time irrigation amount based on the predicted irrigation amount, and generating an irrigation control instruction; the irrigation control module is used for issuing the irrigation control instruction to the corresponding irrigation equipment; and the self-maintenance diagnosis module is used for identifying the execution condition of the irrigation equipment according to the control instruction.

The invention relates to the technical field of digital rural agricultural treatment, in particular to a digital rural agricultural treatment method and system based on artificial intelligence. Obtaining a water demand characteristic parameter of the type of a crop currently planted in the Badzhou field from a crop database; superposing the spatial distribution data of the soil moisture content and the spatial distribution map layer to generate a Badzhou field moisture state map layer; the accurate water supplementing amount of the Badzhou field is obtained through water demand analysis of the Badzhou field with the water stress; an irrigation path is planned according to the position proximity relation and gradient information of all the Buddha fields in the Buddha field moisture state graph layer, and differential irrigation schemes are generated by combining the precise water supplementing amount corresponding to each Buddha field; sending the differential irrigation scheme to irrigation execution equipment, and supplementing water to each palm field; the problem that precise irrigation is difficult due to scattered distribution and large environmental difference of the palm field in the mountainous area is effectively solved.

The invention provides a rock-soil moisture content and permeability testing system and method based on electromagnetic waves and infrared rays, and particularly relates to the technical field of geotechnical engineering testing and multi-source signal detection. According to the test system, construction of a rock-soil permeability experiment is realized through a rock-soil sample permeability test platform, the electromagnetic wave detection module realizes perception of a water-containing state in rock-soil, and the infrared thermal imaging module realizes perception of a wet boundary and seepage expansion characteristics of a rock-soil surface. The data acquisition and control module realizes integral control of a penetration experiment and acquisition of dielectric characteristics and infrared images, the signal fusion and inversion module realizes fusion of the dielectric characteristics and the infrared images and joint inversion of dynamic moisture content and penetration performance, and the result processing and visualization module realizes visual display of an inversion result; the method has the remarkable advantages of non-destructiveness, real-time performance, high spatial resolution, high adaptability, intelligent processing capacity and the like, and comprehensive characterization of the permeability behavior and the moisture content evolution process of the rock and soil sample is achieved.

The invention relates to the technical field of smart agriculture, in particular to an unmanned aerial vehicle multispectral soil water content inversion method fused with a deep learning algorithm. The soil water content inversion method specifically comprises the following steps: acquiring a multispectral image of a target area; estimating the vegetation coverage of the target area by using the multispectral image; based on the vegetation coverage, separating a soil reflectivity spectrum from the total surface reflectivity spectrum; based on the attenuation relation between the soil reflectivity spectrum and the soil water content, a first inversion sub-model is constructed, the input of the first inversion sub-model is the soil reflectivity spectrum, and the output of the first inversion sub-model is the first soil water content inversed based on the attenuation relation; based on the convolutional neural network architecture, constructing a second inversion sub-model, the output of the second inversion sub-model being a multispectral image, and the output being a second soil water content predicted based on the trained and verified convolutional network architecture; and fusing the output of the first inversion sub-model and the output of the second inversion sub-model to generate a third soil water content of the target area.

The invention discloses a multi-source soil water data fusion method based on a Bayesian neural network. Obtaining original data, and cleaning the original data to obtain a cleaned data set; on the basis of an inverse distance weighted interpolation method, calculating the value of data in the cleaning data set at the actual measurement site, and performing error correction on the multi-source soil water data by taking the soil water data of the actual measurement site as a reference to obtain a site data set; segmenting data points in the site data set to construct a training data set and a test data set; time characteristic data and geographical spatio-temporal data in the cleaned data set are used as input, a prior linear layer is embedded to generate a weight coefficient of multi-source soil aquatic product data, and a normal distribution prediction model is established for a weight generation process, network parameters and a prediction standard deviation based on a multilayer Bayesian method; and training parameters of the normal distribution prediction model, and judging the prediction precision of the normal distribution prediction model. According to the invention, reliability is ensured in a complex environment.

The invention provides a roof landform greening structure and a greening construction method thereof, and relates to the technical field of roof landform greening, the roof landform greening structure comprises extruded sheet layers which are arranged in a step shape with a high middle part and a low periphery, and the basic height of rugged landform greening is synergistically shaped through a formula of overlapping the extruded sheet layers with light nutrient soil for water retention, adsorption and ash reduction. According to the greening molding structure and the light nutrient soil formula used in cooperation with green land planting, the excellent water retention, fertilizer retention, dust falling and runoff pollutant adsorption effects are achieved under the combined action, and the total load of the structure can be greatly reduced in cooperation with a roof terrain greening molding structure.

The invention relates to the technical field of frozen soil hydrothermal coupling, in particular to a frozen soil hydrothermal coupling model based on a physical information neural network, which comprises the following steps of: 1, deducing a control equation, including deducing an energy conservation equation and deducing a mass conservation equation; 2, an auxiliary equation is established; 3, initial conditions describe physical field distribution in the whole domain at the beginning of calculation, and in practical application, the initial conditions are generally determined based on field observation data or theoretical analysis; 4, describing physical field constraints on the boundary of the computational domain according to boundary conditions, wherein the boundary conditions generally comprise a Dirichlet boundary condition, a Neumann boundary condition and a mixed boundary condition; and step 5, a physical information neural network method. According to the method, efficient meshless calculation is achieved, powerful support is provided for model parameter inversion and cold region engineering real-time prediction, and the practical value and application potential of the method are remarkably improved.

The invention discloses a nested drainage basin flood simulation method considering bidirectional coupling, and relates to the technical field of hydrological and hydrodynamic model coupling and nested drainage basin flood simulation, and the method comprises the steps: coupling hydrological and hydrodynamic modules to construct a small-scale flood model based on spatial structure characteristics and a water volume secondary infiltration participation soil water mechanism of a spill flooding area; based on the space, the time and the coupling boundary, nesting the small-scale flood model and a preset large-scale flood model to obtain a nested drainage basin flood simulation model; and utilizing the nested basin flood simulation model to simulate hydrology and hydrodynamic force of the basin to be detected to obtain a flood submerging area evaluation result. According to the constructed small-scale flood model, a plurality of hydrological and hydrodynamic modules are integrated, water overflowing to a flood area is supported to infiltrate again and participate in the soil water process, hydrological-hydrodynamic bidirectional coupling between a river channel and the flood area is achieved, and mechanism support is provided for improving local simulation precision in nested simulation.

The utility model relates to the technical field of soil monitoring, and provides a soil water content monitoring and processing device which comprises a bottom plate and two circular rings, the top of the bottom plate is fixedly connected with a supporting rod, and the two circular rings are fixedly connected to the outer surface of the supporting rod. According to the device, the protective net can be installed by rotating the screw rod with hands and limiting the arc-shaped block, the protective net has certain elasticity, when the wild boars and other animals run around in crop land, the wild boars and other animals are blocked and buffered through the protective net, when the wild boars and other animals touch the protective net, the first pull rope and the first bell shake, and therefore the wild boars and other animals can be prevented from falling off. The bell rings to drive animals such as wild boars, so that the situation that the animals such as the wild boars directly touch the monitoring and processing device, damage to the monitoring and processing device and bite damage of a part of exposed wires can be avoided, and the situation that the animals such as the wild boars are injured due to collision can also be avoided.

The invention provides a soil mixing and sowing method of typical herbaceous plants and a method for forming a biological ridge of a newly-built terrace edge ridge, which comprises the following steps of: 1) selecting qualified trifolium repens, ryegrass, alfalfa and vicia villosa seeds according to process requirements; 2) seed pretreatment: soaking trifolium repens and alfalfa seeds with warm water at 30-40 DEG C for 2-4 hours; the method comprises the following steps: soaking ryegrass and vetch seeds in cold water for 1-2 hours; (3) respectively and uniformly mixing the four seeds with soil according to the proportion required by the process; (4) uniformly sowing the seed-soil mixture on a target land parcel; and 5) management after sowing. According to the herbaceous vegetation soil mixing and sowing method, high germination and survival rate of the herbaceous vegetation can be guaranteed, the plant height, the above-ground biomass and the vegetation coverage of all the herbaceous vegetation are continuously increased in the growing season, and the root system biomass is in the trend of firstly decreasing and then increasing; and all herbaceous vegetation improves the water retention performance of the soil.

The invention discloses a water-absorbing material for soil water retention and fertilizer self-supply as well as a preparation method and application thereof, and relates to the technical field of soil water retention functional materials and preparation thereof. The water absorbing material is particulate matters with a graded micro-nano structure prepared by reacting natural agricultural and forestry wastes, animal wastes, acrylic acid or derivatives thereof, an initiator, a cross-linking agent and inorganic minerals; the mass ratio of the natural agricultural and forestry waste to the animal waste to the acrylic acid or the derivative thereof to the initiator to the cross-linking agent to the inorganic mineral is 100: (10-300): (500-6000): (5-100): (2-50): (10-200). The water absorbing material not only can be degraded under ultraviolet light, but also has high water absorption rate, can maintain water absorption performance after circulating freezing and thawing treatment, and spontaneously provides nutrients for soil.

The invention provides an intelligent drip irrigation and fertilization device for a garden green belt. The intelligent drip irrigation and fertilization device mainly comprises a supply mechanism, a conveying mechanism and a plurality of drip irrigation mechanisms. The supply mechanism is used for providing purified water and fertilizer liquid; the conveying mechanism is connected with the supply mechanism and is arranged in the green belt; the plurality of drip irrigation mechanisms are detachably arranged on the conveying mechanism, and the conveying mechanism is used for transferring the purified water and the fertilizer liquid provided by the supply mechanism to the plurality of drip irrigation mechanisms; the drip irrigation mechanism is used for mixing the purified water and the fertilizer liquid into irrigation liquid and adjusting the preset mixing proportion of the purified water and the fertilizer liquid and the preset distribution proportion of the different stratums according to the soil water potential of the different stratums in the green belt so as to achieve layered intelligent water supply. The drip irrigation mechanism serves as a modularized unit, and water and fertilizer can be pertinently supplemented to the upper soil and the lower soil in a local area according to the water potential and requirements of the upper soil and the lower soil. The problem of contradiction between salt accumulation on the upper layer of the garden green belt and long-term recessive water and fertilizer shortage of the lower layer is solved.

The utility model discloses a citrus orchard soil water retention device which comprises a plurality of water guide pipes, a communicating pipe is connected between every two adjacent water guide pipes, a supporting and fixing assembly is arranged at the joint of each water guide pipe and the corresponding communicating pipe, a water outlet pipe is arranged on the side wall of the middle of each communicating pipe, and a water outlet pipe is arranged on the side wall of the middle of each communicating pipe. A water outlet head is arranged at the end of the water outlet pipe, an electromagnetic valve is arranged on the water outlet pipe, an inserting rod is arranged on one side of the bottom of the electromagnetic valve, and a humidity sensor is arranged at the bottom of the inserting rod. The device can be freely assembled according to the distribution condition of citrus in an orchard, each citrus plant is provided with one water outlet pipe, water resources are greatly saved and fully utilized, it is guaranteed that water flow of each water outlet pipe flows into corresponding citrus soil, and under the cooperation of the humidity sensor and the electromagnetic valve, the humidity of the citrus soil can be accurately controlled. The purpose of water retention of citrus soil is achieved, the phenomenon of poor growth caused by water shortage of citrus is avoided, and therefore healthy growth of citrus is guaranteed.

The invention discloses a drought propagation time prediction method based on machine learning. The method comprises the following steps: collecting and preprocessing hydro meteorological data used for drought index calculation; calculating accumulated rainfall, potential evapotranspiration and soil water content under a given target time scale by adopting a day-by-day sliding method; fitting a probability distribution function of the cumulative data, and performing normal distribution inverse transformation on the cumulative probability to obtain meteorological drought and agricultural drought index sequences; identifying meteorological drought events and agricultural drought events, and counting soil water content states and meteorological drought event characteristics when the meteorological drought events occur as potential factors influencing drought propagation time; matching the identified meteorological drought event and agricultural drought event, and calculating the propagation time of each matched drought event pair; training a drought propagation time prediction model based on an extreme gradient lifting algorithm to realize prediction of drought propagation time; and a technical support is provided for agricultural drought prediction and drought-resistant water source optimization configuration.