Desert water collection and short-life plant micro-irrigation system based on humidity triggering

Abstract

The invention relates to the technical field of desert irrigation, in particular to a desert water collection and short-life plant micro-irrigation system based on humidity triggering, comprising: an intelligent water collection module for collecting desert atmosphere dispersed water vapor and sporadic rainfall; the humidity sensing module is used for monitoring soil humidity and atmospheric humidity of a root zone of a short-life plant in real time, setting a species exclusive humidity threshold value, and automatically generating an irrigation instruction and a moisture space distribution strategy when the humidity is lower than the threshold value; the micro-irrigation module is used for receiving an irrigation instruction and a water space distribution strategy, accurately conveying water to a root system area of a short-life plant through a porous capillary network, and controlling irrigation flow and duration; the ecological adaptation module is used for autonomously optimizing irrigation parameters and a moisture space distribution strategy according to desert surface microtopography and plant community distribution characteristics; the water storage regulation and control module is used for storing the collected water source, monitoring the water level in real time and dynamically adjusting the water supply amount according to irrigation requirements. Therefore, the problems of low water collection efficiency, poor irrigation accuracy and the like in the prior art are solved.

Description

Technical Field

[0001] This invention relates to the field of desert irrigation technology, specifically to a humidity-triggered desert water collection and micro-irrigation system for short-lived plants. Background Technology

[0002] Humidity-triggered desert water collection and micro-irrigation systems for short-lived plants are core technologies supporting desert ecological restoration and the conservation of short-lived plants. They directly determine the germination rate, biomass accumulation, and the effectiveness of desert vegetation community construction for short-lived plants (such as *Gnaphalium affine* and *Capsella bursa-pastoris*). However, long-term application has been hampered by environmental factors such as annual precipitation below 250 mm with extremely uneven spatial and temporal distribution, evaporation as high as 2000-3000 mm, and frequent wind and sand activity (more than 60 days of wind and sand per year). These factors easily lead to problems such as low water collection efficiency (less than 10% with traditional techniques), mismatch between irrigation timing and the water requirements of short-lived plants, and insufficient water use efficiency. This results in a germination failure rate exceeding 70% for short-lived plants, a sharp drop in population density, and even temporary stagnation of ecological restoration projects. As desert ecological restoration advances towards precision and intelligence, higher demands are placed on micro-irrigation systems in terms of dynamic humidity response speed, accuracy of water requirement adaptation for short-lived plants, resistance to interference in extreme environments, and water use efficiency. There is an urgent need for a humidity-triggered micro-irrigation system that can sense the multi-dimensional status of desert air humidity, soil humidity, and leaf surface humidity in real time, dynamically match the water requirement characteristics of short-lived plants at different growth stages, and quickly respond to complex environmental changes such as the dissipation of morning fog and short-term showers, so as to ensure the efficient survival and reproduction of short-lived plants in the special habitat of the desert.

[0003] However, traditional desert water collection and irrigation technologies for short-lived plants have inherent flaws: the water collection mechanism is passive, relying heavily on fixed structures such as fish-scale pits and static biomimetic condensation plates, without adjusting the water collection posture (such as the angle of the condensation plate and the tension of the fog net) in accordance with real-time humidity changes. This makes it difficult to capture transient humidity resources such as morning fog and short-term light rain, resulting in a usable water loss rate exceeding 70%. Furthermore, the irrigation trigger logic is rigid, relying solely on a single soil moisture threshold (such as a fixed 12% vol) or timed control, failing to adapt to the phased water requirement patterns of short-lived plants: "water-sensitive during germination (requiring stable soil moisture of 12%-15%), concentrated water requirement during the growth period (daily water consumption 2-3 times that of germination), and reduced water requirement during the fruiting period (requiring a reduction to 8%-10% soil moisture)." This easily leads to under-irrigation during the germination period (…). The system suffers from several challenges: low root rot rate (<30%) or over-irrigation during the growing season (>25%), resulting in irrigation water utilization efficiency of less than 40%; poor hardware adaptability to harsh environments such as desert sandstorms (wind speeds often reach 15-25 m / s) and extreme temperature differences (-10℃~45℃); humidity sensor accuracy drift (error >5%RH); and water collection plate blockage rate exceeding 50%, leading to system adjustment response delays of over 10 minutes; furthermore, the lack of redundant protection design means that a single water collection module or irrigation pipe failure can cause irrigation interruption, and since short-lived plants have a growing season of only 60-90 days, a one-week interruption can cause the plants to wither, resulting in an average annual vegetation restoration failure rate of over 35% due to equipment failure. Overall, the technology faces multiple challenges, including low water collection efficiency, poor irrigation accuracy, weak environmental adaptability, and insufficient fault tolerance. Summary of the Invention

[0004] This application provides a humidity-triggered desert water collection and short-lived plant micro-irrigation system to solve the problems of low water collection efficiency and poor irrigation accuracy in the prior art.

[0005] The first aspect of this application provides a humidity-triggered desert water collection and micro-irrigation system for short-lived plants, comprising: an intelligent water collection module, a humidity sensing module, a micro-irrigation module, an ecological adaptation module, and a water storage and regulation module; wherein, the intelligent water collection module is used to collect atmospheric diffuse water vapor and sporadic precipitation in the desert; the humidity sensing module is used to monitor the soil humidity and atmospheric humidity in the root zone of short-lived plants in real time, set a species-specific humidity threshold, and automatically generate irrigation instructions and water spatial allocation strategies when the humidity is lower than the threshold; the micro-irrigation module is used to receive the irrigation instructions and the water spatial allocation strategy, and accurately deliver water to the root zone of short-lived plants through a porous capillary network, controlling the irrigation flow and duration; the ecological adaptation module is used to autonomously optimize irrigation parameters and the water spatial allocation strategy according to the desert surface micro-topography and plant community distribution characteristics; the water storage and regulation module is used to store the collected water source, monitor the water level in real time, and dynamically adjust the water supply according to irrigation needs.

[0006] Preferably, the intelligent water collection module includes an atmospheric water vapor collection unit and a scattered precipitation collection unit. The atmospheric water vapor collection unit uses an adsorption-type condensation plate, which uses the diurnal temperature difference to drive water vapor adsorption and condensation to collect diffused water vapor in the desert atmosphere. The scattered precipitation collection unit uses a folded rain collection bucket with a built-in filter to filter sand and dust and collect scattered precipitation in the desert.

[0007] Preferably, the humidity sensing module includes a root zone humidity sensor group, an atmospheric humidity monitoring unit, a threshold setting unit, and an instruction generation unit. The root zone humidity sensor group uses embedded probes to monitor the soil volumetric water content at different depths of the short-lived plant roots. The atmospheric humidity monitoring unit collects relative humidity and dew point temperature data in the desert environment in real time. The threshold setting unit presets species-specific humidity thresholds based on the water requirements of different short-lived plants. When the monitored humidity is below the set threshold, the instruction generation unit automatically generates an irrigation start instruction and generates a water spatial allocation strategy based on the plant community distribution density.

[0008] Preferably, the micro-irrigation module includes a capillary network unit, a flow control unit, and an irrigation execution unit. The porous capillary network unit is laid along the root distribution direction of short-lived plants to form a mesh-like water transport channel. The flow control unit adjusts the water flow rate through a solenoid valve to control the irrigation flow rate. The irrigation execution unit receives irrigation instructions and controls the irrigation duration of the corresponding area network according to the water spatial allocation strategy to accurately deliver water to the root area.

[0009] Preferably, the ecological adaptation module includes a micro-topography analysis unit, a community distribution identification unit, and a parameter optimization unit. The micro-topography analysis unit identifies the micro-topographic features of desert surface depressions and slopes through lidar scanning. The community distribution identification unit distinguishes between sparse and dense areas of short-lived plants through image recognition. The parameter optimization unit autonomously optimizes irrigation flow rate, duration, and water spatial allocation strategy based on micro-topographic features and community distribution density.

[0010] Preferably, the water storage and control module includes a water storage unit, a water level monitoring unit, and a water supply regulation unit. The water storage unit is an underground pressurized water tank that stores the collected water source. The water level monitoring unit is used to monitor the water level in the tank in real time and triggers a water replenishment warning when the water level is lower than 10% of the total volume. The water supply regulation unit is used to dynamically adjust the water supply and water pressure according to the optimized water spatial allocation strategy, irrigation demand, and water level data to ensure stable water supply in the pipeline network.

[0011] The second aspect of this application provides a humidity-triggered desert water collection and micro-irrigation method for short-lived plants, comprising: acquiring desert atmospheric diffuse water vapor data, sporadic precipitation data, and micro-topography and community distribution data of the short-lived plant growth area; based on the desert atmospheric diffuse water vapor data and sporadic precipitation data, monitoring the soil moisture and atmospheric humidity in the root zone of the short-lived plants in real time, and setting a dynamically adaptive species-specific humidity threshold in conjunction with a short-lived plant water stress prediction algorithm; automatically generating irrigation instructions and a water spatial allocation strategy when the humidity is lower than the threshold; based on the irrigation instructions and the water spatial allocation strategy, precisely delivering water to the root zone of the short-lived plants through a porous capillary network, controlling the irrigation flow rate and irrigation duration; optimizing irrigation parameters and the water spatial allocation strategy based on the micro-topography of the short-lived plant growth area and the community distribution data; storing the collected water source in an underground pressurized water tank and monitoring the water level in real time; and dynamically adjusting the water supply in conjunction with a predictive scheduling algorithm and the water level data and the water spatial allocation strategy.

[0012] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the program to implement a humidity-triggered desert water collection and micro-irrigation method for short-lived plants as described in the above embodiments.

[0013] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement a humidity-triggered desert water collection and short-lived plant micro-irrigation method as described in the above embodiments.

[0014] The fifth aspect of this application provides a computer program product, including a computer program or instructions, for implementing a humidity-triggered desert water collection and short-lived plant micro-irrigation method as described in the above embodiments.

[0015] Therefore, this application has the following beneficial effects:

[0016] This application embodiment utilizes an intelligent water collection module with an adsorption-type condensation plate and a built-in filter in a folded rainwater collection bucket to efficiently collect atmospheric diffuse water vapor and sporadic precipitation in desert areas, broadening the channels for obtaining scarce water resources and avoiding the passive inefficiency and insufficient water utilization of traditional water collection methods. The humidity sensing module, relying on multi-depth humidity monitoring in the root zone, atmospheric temperature and humidity data collection, and species-specific threshold configuration, automatically generates irrigation instructions and water spatial allocation strategies, accurately capturing the water demand timing of short-lived plants, breaking the limitations of relying on fixed humidity thresholds or experience-based judgments that lead to poor water demand adaptability. The micro-irrigation module, through targeted delivery via a porous capillary network, flow control by solenoid valves, and time-series irrigation regulation, delivers water to the root zone... Precision water supply ensures that water delivery always matches the root system's water requirements, solving the problems of high water loss and low accuracy in traditional irrigation. The ecological adaptation module combines lidar micro-topography analysis and image recognition of community distribution to autonomously optimize irrigation flow, duration, and water allocation strategies. It performs comprehensive ecological adaptation from micro-topography to community density response, improving the ecological fit of irrigation strategies. The water storage and regulation module utilizes underground pressurized water storage, water level warnings, and dynamic water supply regulation to ensure stable water supply through the pipeline network. It promptly triggers water replenishment warnings when water levels are insufficient, minimizing the risk of irrigation interruptions, improving the rationality of water storage and utilization, and increasing the germination and survival rates of short-lived plants. Thus, it solves the problems of low water collection efficiency and poor irrigation accuracy in existing technologies.

[0017] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0018] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0019] Figure 1 This is a schematic diagram of a humidity-triggered desert water collection and short-lived plant micro-irrigation system according to an embodiment of this application;

[0020] Figure 2 This is a schematic diagram of an intelligent water collection module according to an embodiment of this application;

[0021] Figure 3 This is a schematic diagram of a humidity sensing module according to an embodiment of this application;

[0022] Figure 4 This is a schematic diagram of a micro-irrigation module provided according to an embodiment of this application;

[0023] Figure 5 This is a schematic diagram of an ecosystem adaptation module provided according to an embodiment of this application;

[0024] Figure 6 This is a schematic diagram of a water storage control module according to an embodiment of this application;

[0025] Figure 7 This is a schematic diagram of a humidity-triggered desert water collection and short-lived plant micro-irrigation system according to an embodiment of this application;

[0026] Figure 8 A flowchart of a humidity-triggered desert water collection and short-lived plant micro-irrigation method according to an embodiment of this application;

[0027] Figure 9 This is a schematic diagram of a humidity-triggered desert water collection and short-lived plant micro-irrigation method according to an embodiment of this application;

[0028] Figure 10 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application. Detailed Implementation

[0029] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0030] The following description, with reference to the accompanying drawings, illustrates a humidity-triggered desert water harvesting and micro-irrigation system for short-lived plants, based on an embodiment of this application. Addressing the low response speed issue mentioned in the background section, this application provides a humidity-triggered desert water harvesting and micro-irrigation system for short-lived plants. In this system, an intelligent water harvesting module employs an adsorption-type condensation plate and a folded rainwater collection bucket with a built-in filter to efficiently collect atmospheric diffused water vapor and sporadic precipitation in the desert, broadening the channels for obtaining scarce water resources and avoiding the passive inefficiency and insufficient water utilization of traditional water harvesting methods. The humidity sensing module, relying on multi-depth humidity monitoring in the root zone, atmospheric temperature and humidity data collection, and species-specific threshold configuration, automatically generates irrigation instructions and water spatial allocation strategies, accurately capturing the water demand timing of short-lived plants, breaking the limitations of relying on fixed humidity thresholds or experience-based judgments leading to poor water demand adaptability. The micro-irrigation module utilizes porous capillaries... Targeted pipeline delivery, solenoid valve flow control, and time-series irrigation regulation ensure precise water supply to the root zone, guaranteeing that water delivery always matches the root system's water requirements. This solves the problems of high water loss and low accuracy in traditional irrigation. The ecological adaptation module combines lidar micro-topography analysis and image recognition of community distribution to autonomously optimize irrigation flow, duration, and water allocation strategies. It performs comprehensive ecological adaptation from micro-topography to community density response, improving the ecological fit of irrigation strategies. The water storage and regulation module utilizes underground pressurized water storage, water level warnings, and dynamic water supply regulation to ensure stable pipeline water supply. It promptly triggers water replenishment warnings when water levels are insufficient, minimizing the risk of irrigation interruptions, improving the rationality of water storage and utilization, and increasing the germination and survival rates of short-lived plants. This solves the problems of low water collection efficiency and poor irrigation accuracy in existing technologies.

[0031] Figure 1 This is a schematic diagram of a humidity-triggered desert water collection and short-lived plant micro-irrigation system provided in an embodiment of this application.

[0032] This application provides a humidity-triggered desert water collection and short-lived plant micro-irrigation system, the system 10 comprising:

[0033] Intelligent water collection module 100, humidity sensing module 200, micro-irrigation module 300, ecological adaptation module 400, and water storage and regulation module 500.

[0034] Among them, the intelligent water collection module 100 is used to collect atmospheric diffuse water vapor and sporadic precipitation in the desert; the humidity sensing module 200 is used to monitor the soil humidity and atmospheric humidity in the root zone of short-lived plants in real time, set species-specific humidity thresholds, and automatically generate irrigation instructions and water spatial allocation strategies when the humidity is lower than the threshold; the micro-irrigation module 300 is used to receive irrigation instructions and water spatial allocation strategies, and accurately deliver water to the root zone of short-lived plants through a porous capillary network, controlling the irrigation flow and duration; the ecological adaptation module 400 is used to autonomously optimize irrigation parameters and water spatial allocation strategies according to the micro-topography and plant community distribution characteristics of the desert surface; and the water storage and regulation module 500 is used to store the collected water source, monitor the water level in real time, and dynamically adjust the water supply according to irrigation needs.

[0035] It is understood that in this embodiment, the intelligent water collection module employs an adsorption-type condensation plate and a folded rainwater collection bucket with a built-in filter to efficiently collect atmospheric diffuse water vapor and sporadic precipitation in the desert, broadening the channels for obtaining scarce water sources and avoiding the passive inefficiency and insufficient water utilization of traditional water collection methods. The humidity sensing module, relying on multi-depth humidity monitoring in the root zone, atmospheric temperature and humidity collection, and species-specific threshold configuration, automatically generates irrigation instructions and water spatial allocation strategies, accurately capturing the water demand timing of short-lived plants, breaking the limitations of poor water demand adaptability caused by relying on fixed humidity thresholds or experience judgment. The micro-irrigation module uses a porous capillary network for targeted delivery, electromagnetic valve flow control, and time-series irrigation regulation. Precise water supply to the root zone ensures that water delivery always matches the root system's water requirements, solving the problems of high water loss and low accuracy in traditional irrigation. The ecological adaptation module combines lidar micro-topography analysis and image recognition of community distribution to autonomously optimize irrigation flow, duration, and water allocation strategies. This comprehensive ecological adaptation, from micro-topography to community density response, enhances the ecological fit of irrigation strategies. The water storage and regulation module utilizes underground pressurized water storage, water level warnings, and dynamic water supply adjustment to ensure stable water supply through the pipeline network. It promptly triggers water replenishment warnings when water levels are insufficient, minimizing the risk of irrigation interruptions, improving the rationality of water storage and utilization, and increasing the germination and survival rates of short-lived plants. Thus, it solves the problems of low water collection efficiency and poor irrigation accuracy in existing technologies.

[0036] In this embodiment of the application, the intelligent water collection module 100 includes: Figure 2 As shown, there are atmospheric water vapor collection units and sporadic precipitation collection units.

[0037] Among them, the atmospheric water vapor collection unit adopts an adsorption condensation plate, which drives water vapor adsorption and condensation through the diurnal temperature difference to collect diffuse water vapor in the desert atmosphere; the sporadic precipitation collection unit adopts a folded rain collection bucket with built-in filter screen to filter sand and dust and collect sporadic precipitation in the desert.

[0038] It is understood that the atmospheric water vapor collection unit in this application adopts an adsorption-type condenser plate, utilizing the significant diurnal temperature variation characteristic of the desert environment to convert gaseous water, which is difficult to capture using traditional water collection methods, into usable liquid water. This breaks through the limitation of relying solely on precipitation for water collection and provides an additional stable water source. The sporadic precipitation collection unit uses a foldable rainwater collection bucket with a built-in filter. The foldable structure can flexibly adapt to the complex desert surface, and the built-in filter can accurately filter sand and dust impurities, preventing sand and dust from clogging subsequent pipes or polluting the water source, ensuring maximum collection and purification of short-term, small amounts of sporadic precipitation in the desert. Through targeted design adapted to the desert environment, water collection efficiency and equipment anti-interference capabilities are improved.

[0039] For example, in the conservation area for short-lived plants on the edge of the Taklamakan Desert in Northwest China, the daytime temperature in this area can reach over 40°C in summer, while at night, due to the low heat capacity of sand particles and strong radiative heat dissipation, the temperature can plummet to below 10°C, creating a significant diurnal temperature range. Furthermore, the average daily relative humidity is often as low as 20%, making it difficult to obtain effective water sources using traditional water collection methods. The atmospheric water vapor collection unit deployed here uses an adsorption-condensation plate array loaded with MOF material, leveraging the extreme diurnal temperature range of the desert to drive the adsorption-condensation cycle: when the temperature drops at night, the MOF material efficiently adsorbs diffuse water vapor in the air due to its well-developed porous structure, capturing 0.5-0.7 liters of water vapor per kilogram of material; the following day, heated by solar radiation, the MOF material desorbs and releases water vapor, which condenses into liquid water on the surface of the condensation plate coated with a radiative cooling coating, and is collected in a collection tank and channeled into the water storage system. Even in extremely dry weather with a relative humidity of only 18%, the unit can still produce an average of 0.12 liters of water per square meter per day, providing a stable water source for the germination period of short-lived plants such as sagebrush and desert bulrush, effectively solving the problem of water resource acquisition during non-rainfall periods in desert areas.

[0040] In this embodiment, the humidity sensing module 200 includes: Figure 3 As shown, the system includes a root zone humidity sensor group, an atmospheric humidity monitoring unit, a threshold setting unit, and an instruction generation unit.

[0041] The root zone humidity sensor group uses embedded probes to monitor the soil volumetric water content at different depths of the roots of short-lived plants; the atmospheric humidity monitoring unit is used to collect relative humidity and dew point temperature data in the desert environment in real time; the threshold setting unit presets species-specific humidity thresholds based on the water requirements of different short-lived plants; and the instruction generation unit automatically generates an irrigation start instruction when the monitored humidity is lower than the set threshold, and generates a water spatial allocation strategy based on the plant community distribution density.

[0042] Understandably, this application embodiment uses a root zone humidity sensor group to monitor the soil volumetric water content at different root depths using embedded probes, accurately capturing the actual water needs of short-lived plants and avoiding misjudgments caused by only monitoring surface soil moisture while the roots are dehydrated. The atmospheric humidity monitoring unit collects relative humidity and dew point temperature data in real time, predicting the trend of desert atmospheric water vapor changes in advance and providing additional environmental basis for selecting irrigation timing. The threshold setting unit presets exclusive thresholds according to the water requirements of different short-lived plants, breaking the limitation of traditional fixed thresholds for adapting to multiple species. The instruction generation unit automatically generates irrigation instructions when the humidity is below the threshold and generates a spatial allocation strategy based on the plant community distribution density, allowing for on-demand water supply and allocation without manual intervention. This avoids germination failure caused by under-irrigation and root rot caused by over-irrigation, improving the accuracy and automation level of irrigation decisions.

[0043] It should be noted that the threshold setting unit presets species-specific humidity thresholds based on the water requirements of different short-lived plants. For example, for shallow-rooted plants like *Gnaphalium affine* that are sensitive to water during germination, the preset soil moisture threshold for its root zone is 12%-15%, while for *Capsella bursa-pastoris*, which has deeper roots and is more drought-tolerant, it is preset at 10%-13%. Dynamic adjustments are made based on growth stages; for instance, the threshold for *Capsella bursa-pastoris* during its peak water requirement period is set at 9%-12%, while the upper limit is controlled at 13%-16% during the fruiting period when water requirements decrease. Additionally, habitat characteristics are considered; for species in gravelly areas with poor water retention, the threshold is increased by 2%-3% to compensate for water loss, while for species in sandy soils, the lower limit is emphasized to prevent root dehydration. This translates plant physiological characteristics and desert habitats into quantifiable irrigation standards, providing a precise benchmark for subsequent monitoring and instruction generation. This reduces the problems of under-irrigation and over-irrigation for some species caused by traditional fixed thresholds, ensuring that each short-lived plant completes its life cycle under suitable humidity.

[0044] When the monitored humidity is below a set threshold, the instruction generation unit automatically generates an irrigation start instruction and, in conjunction with the plant community distribution density, generates a water spatial allocation strategy. When the soil moisture at the key depth of the root zone is continuously below the species-specific threshold, and atmospheric humidity monitoring shows no possibility of natural humidification, an irrigation start instruction is automatically generated to ensure timely water replenishment before plants suffer from drought. At the same time, a spatial allocation strategy is formulated based on the community distribution density, increasing the irrigation flow or duration in dense areas, reducing the water volume and narrowing the scope in sparse areas, and closing the valve in blank areas. This not only solves the lag of traditional manual judgment but also avoids water waste or local water shortage caused by uneven plant distribution, significantly improving water use efficiency and ensuring the stable growth of short-lived plant communities.

[0045] For example, in the Gansu Hexi Corridor Desert Short-lived Plant Conservation Area, two typical short-lived plants, *Echinochloa crus-galli* and *Capsella bursa-pastoris*, are distributed in this area, which also has two differentiated habitats: gravel-covered areas and sandy soil areas. The threshold setting unit has formulated refined and specific thresholds for different species, habitats, and growth stages: For *Echinochloa crus-galli*, which has shallow roots (roots concentrated at 5-10cm) and is sensitive to water during germination, the soil moisture threshold in its root zone during germination is preset at 12%-15% to avoid seed germination failure due to insufficient moisture; after entering the growth period, as the water requirement for leaf growth increases, the threshold is lowered to 10%-12% to ensure that water supply matches growth needs; and during the fruiting period, to prevent root rot caused by excessive moisture, the upper limit of humidity is controlled at 15%. For *Capsella bursa-pastoris*, which has deeper roots (8-12cm) and stronger drought resistance, the threshold during germination is set at 10%-13%, lower than that of *Echinochloa crus-galli* to adapt to its drought resistance characteristics. Meanwhile, considering habitat differences, the threshold for *Gnaphalium affine* in gravel-covered areas is increased by 2% to 14%-17% due to poor soil water retention, compensating for rapid water loss; while for *Capsella bursa-pastoris* in sandy soil areas, the lower threshold is protected with a setting of 9% to prevent root dehydration caused by short-term drought. Customized threshold settings ensure that different short-lived plants complete their life cycle under suitable humidity conditions.

[0046] In this embodiment of the application, the micro-irrigation module 300 includes: Figure 4 As shown, the capillary network unit, flow control unit, and irrigation execution unit are shown.

[0047] Among them, the porous capillary network unit is used to lay along the distribution direction of the root system of short-lived plants to form a network of water transport channels; the flow control unit adjusts the water flow speed through the solenoid valve to control the irrigation flow rate; the irrigation execution unit is used to receive irrigation instructions, control the irrigation duration of the corresponding area network according to the water spatial distribution strategy, and accurately deliver water to the root area.

[0048] It is understood that in this embodiment, capillary network units are laid along the root distribution direction of short-lived plants to form a mesh-like water transport channel, which allows water to directly penetrate into the root absorption area, avoiding water loss to deeper soil or areas without plants in traditional flood irrigation; the flow control unit uses solenoid valves to precisely adjust the water flow rate, and can dynamically control the irrigation flow according to the community density and water shortage level in different areas to prevent excessive or insufficient water supply to individual plants; after receiving the instruction, the irrigation execution unit controls the irrigation duration of the corresponding area network according to the water spatial allocation strategy, and performs differentiated water supply in different areas to improve water use efficiency. Through the dual regulation of flow and duration, it avoids root rot caused by over-irrigation or plant wilting caused by under-irrigation, and improves germination rate and survival rate.

[0049] It should be noted that the irrigation execution unit is used to receive irrigation instructions, control the irrigation duration of the corresponding area's pipe network according to the water spatial allocation strategy, and accurately deliver water to the root zone. It receives standardized irrigation instructions from the instruction generation unit in real time through an industrial-grade IoT link. The instructions not only include the core instruction to start irrigation, but also the precise coordinates of the target irrigation area, the basic irrigation duration, and the plant community density level of the area, providing clear parameter basis for subsequent execution. Secondly, a deep analysis of water spatial allocation strategies is conducted, dynamically linking community density with irrigation duration. For densely populated areas (such as low-lying desert areas where competition for water among plants is intense), the basic irrigation duration is extended by 15%-20% (e.g., from 45 minutes to 54 minutes), while ensuring that all water outlets in the capillary network of that area are opened simultaneously to prevent localized water shortages due to close proximity of plants. For sparsely populated areas (such as desert slopes where plants are scattered), the basic irrigation duration is shortened to 60%-70% (e.g., from 30 minutes to 21 minutes), and zoned valves are used to control the irrigation network, opening only in areas with plant distribution to reduce water infiltration in areas without plants. For blank areas without plants, the electromagnetic valves of the corresponding irrigation network are directly closed. In addition, during the irrigation process, the system continuously receives feedback data from the root zone humidity sensor: if the soil moisture in a certain area has risen to the species-appropriate threshold (such as 12% for tooth grass) when it reaches 70% of the preset time, the irrigation in that area will be automatically stopped in advance to avoid over-irrigation and root rot; if the humidity in a certain area rises slowly due to poor soil water retention (such as sandy soil), the irrigation will be extended by 5-10 minutes within a safe range to ensure that the roots absorb water fully.

[0050] For example, in the Hexi Corridor desert short-lived plant conservation area, considering the different root characteristics of *Gnaphalium affine* (roots concentrated in the 5-10cm soil layer) and *Capsella bursa-pastoris* (roots reaching 8-12cm depth), a customized laying scheme of depth adaptation + mesh coverage was adopted for the capillary network unit. In the *Gnaphalium affine* growing area, polyethylene capillary networks with an inner diameter of 5mm and 0.5mm micropores were buried at a depth of 8cm, arranged in a zigzag pattern along the plant row spacing, with the water outlet spacing set at 15cm to ensure that there are 2-3 stable water outlets around the roots of each *Gnaphalium affine* plant. In the *Capsella bursa-pastoris* growing area, the burial depth of the network was adjusted to 10cm, and the water outlet spacing was reduced to 12cm to accommodate its denser fibrous root system. During irrigation, water slowly seeps out through capillaries at a rate of 0.15 L / h, directly penetrating to the root absorption core area, avoiding the loss of more than 60% of water to deep sandy soil or diffusion to unplanted areas as in traditional flood irrigation. After one growth cycle monitoring, this capillary network unit increased water use efficiency from 38% in traditional irrigation to 86%, maintained soil moisture in the root zone of *Cephalotaxus fortunei* at a suitable threshold of 12%-15%, increased germination rate from 29% to 82%, and reduced leaf wilting rate of *Capsella bursa-pastoris* from 22% to below 5% during the growing season.

[0051] In this embodiment, the ecosystem adaptation module 400 includes, for example: Figure 5 As shown, there are three units: micro-topography analysis unit, community distribution identification unit, and parameter optimization unit.

[0052] Among them, the micro-topography analysis unit identifies the micro-topographic features of desert surface depressions and slopes through lidar scanning; the community distribution identification unit distinguishes between sparse and dense areas of short-lived plants through image recognition; and the parameter optimization unit autonomously optimizes irrigation flow, duration, and water spatial allocation strategies based on micro-topographic features and community distribution density.

[0053] Understandably, the micro-topography analysis unit in this application uses lidar scanning to accurately identify micro-topographic features such as low-lying areas and slopes, clarifying the differences in soil water retention capacity in different regions; the community distribution identification unit uses image recognition algorithms to distinguish between sparse and dense areas of short-lived plants, understanding the differences in actual total water demand in different regions; the parameter optimization unit is used to specifically optimize irrigation parameters and allocation strategies: for low-lying dense areas, the irrigation flow rate is appropriately reduced and the duration is shortened to avoid water accumulation and root rot; for sparse sloping areas, the irrigation flow rate is increased and the duration is extended to compensate for rapid water loss; for flat dense areas, the flow rate is kept stable and the uniformity of pipeline coverage is expanded, so that limited water resources are tilted towards areas with urgent water demand and high utilization efficiency, improving irrigation accuracy and water use efficiency, conforming to the characteristics of desert micro-ecology, and providing suitable growth conditions for short-lived plants with different micro-topography and distribution densities.

[0054] It should be noted that the community distribution identification unit distinguishes between sparse and dense areas of short-lived plants through image recognition. First, a 4K high-definition camera with an IP67 dustproof shell collects images matching the monitoring area using GPS positioning during the daily stable light period (9:00-10:00). The contrast between plants (gray-green) and sand (yellow-white) is enhanced through HSV spatial transformation to highlight plant pixels. Then, an improved U-Net model is used to segment plant pixels (distinguishing the independent outlines of overlapping plants), and interference such as small stones is removed through shape matching. Subsequently, the images are divided into 1m×1m grids, and the number of plants per grid is calculated based on the average pixel area of ​​a single plant. Finally, thresholds are set according to species characteristics (e.g., ≥8 plants / m² for *Gnaphalium affine* is a dense area, ≤3 plants / m² is a sparse area, and ≥6 plants / m² for *Capsella bursa-pastoris* is a dense area) to determine the level. The results, along with coordinates, are transmitted to the parameter optimization unit in real time, providing accurate spatial distribution data for differentiated irrigation.

[0055] The HSV space transformation formula is:

[0056]

[0057]

[0058]

[0059]

[0060]

[0061] in, To make the red channel of the image The normalized value; The original pixel values ​​of the red channel of the image; To make the image green channel The normalized value; The original pixel values ​​of the green channel of the image; To extract the blue channel of the image The normalized value; The original pixel values ​​of the blue channel of the image; The brightness of the HSV color space; In order to seek , , The maximum value of the three; The saturation level of the HSV space; In order to seek , , The minimum value of the three; The hue in the HSV color space; This is an intermediate variable in the following text; Angle coefficients calculated for hue; This is an intermediate variable in the following text; mod is the modulo operator.

[0062] Improved U-Net model algorithm formula:

[0063]

[0064]

[0065] in, Cross-entropy loss; The total number of pixel samples involved in the calculation; For the first The real label of each pixel; It is the natural logarithm function; To improve the U-Net model for the first The predicted probability that a pixel belongs to a plant; For the first One pixel indicates a non-plant label; For the model to predict the first The probability that a pixel does not belong to a plant; The Dice loss function; For real plant areas and model prediction of vegetation regions The number of pixels in the intersection; For smoothing terms; For real plant areas The total number of pixels; Predicting plant regions for the model The total number of pixels.

[0066] Hu's rectangular matching algorithm formula:

[0067]

[0068]

[0069]

[0070]

[0071]

[0072] in, The geometric moments are of order pq; Image width; Image height; The x-coordinate is the pixel coordinate. The vertical coordinate is the pixel coordinate. Let be the order of the moment; The x-coordinate is raised to the power of p. The y-axis is raised to the power of q. For pixels Grayscale value; These are the centroid coordinates of the image; It is a 10th order geometric moment; The geometric moments are of order 0 and 1. The geometric moment is of order 00; The central moments are of order pq; This represents the offset of the pixel's horizontal coordinate relative to the centroid. This represents the offset of the pixel's vertical coordinate relative to the centroid. The normalized central moments of order pq; The central moment is of order 00; For the order condition; This is the first component of the invariant moment; This is the second component of the invariant moment; This is the third component of the invariant moment; The normalized central moment is of order 20; The normalized central moments are of order 0 and 2. The normalized central moment is of order 11. The central moment is the 30th order normalized value; The central moment is the 12th order normalized moment; The normalized central moments are of order 21. It is the 0th and 3rd order normalized central moment.

[0073] In this embodiment of the application, the water storage and control module 500 includes, as follows: Figure 6 As shown, there are a water storage unit, a water level monitoring unit, and a water supply regulation unit.

[0074] The water storage unit uses an underground pressurized water tank to store the collected water source; the water level monitoring unit is used to monitor the water level in the tank in real time, and triggers a water replenishment warning when the water level is lower than 10% of the total volume; the water supply regulation unit is used to dynamically adjust the water supply and water pressure according to the optimized water spatial allocation strategy, irrigation demand and water level data to ensure the stability of the pipeline water supply.

[0075] It is understood that the water storage unit in this embodiment adopts an underground pressurized water tank, which avoids the thermal impact of extreme temperature differences on the water source and the erosion of the water tank by strong winds and sandstorms. It also provides basic pressure support for the pipeline water supply through the pressurized structure, avoiding the defects of above-ground water tanks that are easily damaged and have insufficient pressure. The water level monitoring unit tracks the water level in the tank in real time. When the water level is lower than 10% of the total volume, it triggers a water replenishment warning, reminding maintenance personnel to replenish the water source in advance to prevent irrigation interruption due to water depletion. This solves the problems of traditional water storage without warning and difficulty in emergency response to water shortages. The water supply regulation unit dynamically adjusts the water supply and water pressure according to the optimized water spatial distribution strategy, irrigation demand and water level data to ensure uniform water output and stable pressure in the capillary network. This avoids the excessive consumption of water source when the pressure drops suddenly, the water output is uneven or the water level is insufficient during irrigation, which is caused by traditional fixed water supply. This ensures that the irrigation process is uninterrupted and the water supply parameters are adapted to the needs, improving the reliability and stability of irrigation for short-lived desert plants.

[0076] It should be noted that the system monitors the water level in the tank in real time. When the water level is below 10% of the total volume, a water replenishment warning is triggered. The system uses a corrosion-resistant 316L stainless steel shell and an immersion-type static pressure level sensor with temperature compensation to collect data from the bottom, middle, and top of the tank. Data is sampled every 30 seconds and outliers are filtered by moving average before being transmitted wirelessly to the control center. The preset warning threshold of 10% below the total volume is set based on the difficulty of water replenishment in desert areas and emergency needs (e.g., a 5m³ tank with 0.5m³ emergency water volume can support one irrigation of a 500㎡ area, with 24 hours of water replenishment time reserved). When the water level is below this threshold for three consecutive sampling cycles (1.5 minutes), an warning is immediately triggered—both a water replenishment notification with location and water level information is sent to maintenance personnel, the intelligent water collection module is activated to strengthen water collection, and the water supply regulation unit is prompted to limit the flow to protect critical areas, reducing the problems of traditional water storage without warning and only discovering water shortages.

[0077] The water supply regulation unit is used to dynamically adjust the water supply volume and pressure according to the optimized water spatial allocation strategy, irrigation demand and water level data to ensure stable water supply in the pipeline network. First, it receives irrigation demand data (target flow, duration, priority for each area, updated every 10 seconds) and water level data (remaining water volume, rate of decline, whether an alert is issued) in real time. Then, it dynamically adjusts the water supply according to the optimized water spatial allocation strategy. When the water level is sufficient, it supplies water according to demand (water supply and pressure reach the upper limit in high-priority dense areas, and the lower limit in medium-priority sparse areas). When the water level is low, it prioritizes priority and limits the number of times (reduce the water volume of high-priority areas and suspend low-priority irrigation). When an alert is triggered, it only retains high-priority emergency water supply. Subsequently, it performs regulation by adjusting the speed of the variable frequency centrifugal pump and adjusting the opening of the zone electromagnetic regulating valve, combined with the pressure control of the pipeline network pressure sensor. At the same time, it receives feedback from the flow and root zone humidity sensors. When there is blockage, it briefly replenishes pressure and stops supply when the humidity reaches the standard.

[0078] This application proposes a humidity-triggered desert water harvesting and micro-irrigation system for short-lived plants. The system utilizes an intelligent water harvesting module with an adsorption-type condensation plate and a folded rainwater collection bucket with a built-in filter to efficiently collect atmospheric diffused water vapor and sporadic precipitation in the desert, broadening the channels for obtaining scarce water resources and avoiding the passive inefficiency and insufficient water utilization of traditional water harvesting methods. The humidity sensing module, based on multi-depth humidity monitoring in the root zone, atmospheric temperature and humidity data collection, and species-specific threshold configuration, automatically generates irrigation instructions and spatial water allocation strategies, accurately capturing the water demand timing of short-lived plants. This overcomes the limitations of relying on fixed humidity thresholds or experience-based judgments, which lead to poor water demand adaptability. The micro-irrigation module uses a porous capillary network for targeted delivery and a solenoid valve for flow control. Controlled and time-series irrigation regulation precisely supplies water to the root zone, ensuring that water delivery always matches the root system's water requirements, thus solving the problems of high water loss rate and low accuracy in traditional irrigation. The ecological adaptation module combines lidar micro-topography analysis and image recognition of community distribution to autonomously optimize irrigation flow, duration, and water allocation strategies. This comprehensive ecological adaptation, from micro-topography to community density response, enhances the ecological fit of irrigation strategies. The water storage and regulation module utilizes underground pressurized water storage, water level early warning, and dynamic water supply regulation to ensure stable water supply through the pipeline network. It promptly triggers water replenishment warnings when water levels are insufficient, minimizing the risk of irrigation interruptions, improving the rationality of water storage and utilization, and increasing the germination and survival rates of short-lived plants. Therefore, it solves the problems of low water collection efficiency and poor irrigation accuracy in existing technologies.

[0079] The following will illustrate a humidity-triggered micro-irrigation system for desert water collection and short-lived plants through a specific embodiment, such as... Figure 7 As shown, it includes:

[0080] Taking the western section of the Hexi Corridor in Gansu Province (39°20′-39°40′N, 94°10′-94°30′E) as the application scenario, this region has an average annual precipitation of only 40-60 mm and an average annual evaporation of 2800-3200 mm, making it a typical extremely arid desert environment. The region is dominated by two short-lived plants, *Eremopyrum orientale* and *Alyssum desertis*, which germinate and grow in spring (March-May), playing a crucial role in maintaining the fragile ecological balance of the region's desert. To address the core issues of water scarcity and low irrigation precision in the natural growth of these short-lived plants, a humidity-triggered desert water harvesting and short-lived plant micro-irrigation system is deployed.

[0081] Deployment and operation of intelligent water collection modules

[0082] The intelligent water collection module serves as the "water supply source" of the system. It is designed to address the extreme drought characteristics of desert areas (annual average precipitation of 40-60mm and evaporation of 2800-3200mm). Through the coordinated operation of atmospheric water vapor collection units and sporadic precipitation collection units, it maximizes the capture of scarce water resources in the desert. A single module can collect an average of 8-12L of water per day in spring (March-May, when the relative humidity at night can reach 60%-70%, which is the period with the highest humidity throughout the year). This is just enough to meet the basic irrigation needs of a 500㎡ short-lived plant growth area (mainly sagebrush and desert bulrush), thus making up for the lack of natural precipitation in the region from the source. The atmospheric water vapor collection unit uses a corrugated condenser plate made of 3003 aluminum alloy as its core component. This material not only has a thermal conductivity of up to 158W / (m・K), enabling it to quickly respond to the diurnal temperature difference in the desert (the surface temperature drops to 5-8℃ at night in spring and rises to 25-30℃ during the day), but also has excellent resistance to sand and dust erosion. Field tests showed that even after 6 months of continuous use in an environment with an average of 2-3 hours of sand and dust weather per day, there were no obvious signs of corrosion. The condenser plate is designed with a single piece size of 1.2m×0.8m and a thickness of 2mm. Compared with the traditional flat plate structure, the corrugated structure increases the water vapor contact area by 40%, significantly improving the condensation efficiency. The condenser plates are installed at an angle 1.5m above the ground using steel supports. The bottom of the supports is embedded in a 0.8m deep concrete foundation, which prevents long-term accumulation of sand and dust that could bury the supports and ensures installation stability. The 30° installation angle is precisely calculated to match the solar altitude angle in the Hexi Corridor during spring (approximately 45° at noon), effectively preventing the condenser plates from overheating due to direct sunlight at midday and affecting nighttime water vapor condensation. Each system is arranged in a staggered pattern with 6 condenser plates per 500㎡, ensuring that atmospheric water vapor within the coverage area can evenly contact the condensation surface without any blind spots. The unit's operation relies entirely on the diurnal temperature variation in the desert: from 8:00 PM to 6:00 AM the next day, as the surface temperature decreases, the temperature of the condenser plate also decreases. Water vapor dispersed in the atmosphere condenses into water droplets with a diameter of 1-2 mm on the plate surface. To accelerate the collection of water droplets, the surface of the condenser plate is sprayed with WD-40 special hydrophobic agent, which makes the water droplet contact angle reach 110°. The droplets can quickly roll down along the corrugated groove to the food-grade PE water collection tank at the bottom (cross-sectional dimensions 10cm×8cm, with a 0.5% slope at the bottom to ensure that the water flow is free of stagnation). The end of the water collection tank is connected to a Φ32mm PE water pipe with an inner wall smoothness of Ra0.8μm, which reduces water flow resistance and allows the condensate to be transported at a uniform speed to the underground water storage unit by gravity, with a water transport efficiency of over 95%.Meanwhile, the unit is equipped with a SEN0234 miniature wind speed sensor, which can monitor the wind speed at night in real time. When the wind speed exceeds 5m / s (at this speed, the air flow on the surface of the condenser plate will be accelerated, causing the plate temperature to rise by 5-8℃, making it difficult for water vapor to condense), the system will automatically control the condenser plate support to retract within 10 seconds, lowering the condenser plate to 0.5m above the ground surface—utilizing the microenvironment where the wind speed within 1m of the ground surface is reduced by 30%-40% to maintain condensation efficiency; when the wind speed drops below 3m / s, the support will slowly rise back to its original height to ensure a stable and uninterrupted nighttime condensation and water collection process. The sporadic rainfall collection unit is specifically designed for the occasional spring rainfall in the Hexi Corridor, characterized by single rainfall amounts of only 2-5mm, which is easily evaporated from the surface (evaporation rate up to 0.5mm / h). Its core component is a foldable rain collection hopper made of 0.5mm thick PVC-coated canvas with an IPX8 waterproof rating, ensuring no leakage even in heavy rain (daily rainfall exceeding 20mm). When folded, its volume is only 1 / 5 of its unfolded state, facilitating transport to desert sites for installation. Its lifespan is over 3 years, far exceeding the 1-year lifespan of ordinary canvas. When unfolded, the rain collection hopper is circular, with an effective collection area of ​​4㎡ (diameter 2.26m) and an edge height of 15cm, preventing water overflow during rainfall and avoiding water waste. The rain collection hopper is fixed by a retractable aluminum alloy bracket. When the bracket is retracted, the rain collection hopper can be completely folded and stored in a concrete storage trough 0.3m below the ground surface. The trough opening is covered with a sand cover plate flush with the ground surface, which can prevent wind and sand from burying the rain collection hopper and keep the ground surface flat, without affecting the original desert ecology. When the atmospheric humidity monitoring unit detects that "the relative humidity exceeds 80% for 1 hour and the difference between the dew point temperature and the ground surface temperature is less than 2℃" (under this condition, the probability of spring rainfall in the Hexi Corridor is more than 70%), the system will automatically control the bracket to extend within 30 seconds, raising the rain collection hopper to 0.8m above the ground surface and unfolding to form a stable rain collection state. To address the pipe blockage problem caused by frequent sand and dust (with a sand content of up to 0.5 g / L) accompanying rainfall in the Hexi Corridor, a 0.2 mm pore size stainless steel filter screen is installed 5 cm from the bottom inside the rainwater collection hopper. This filter screen can filter more than 98% of sand and dust particles. The filter screen is designed to be detachable for easy cleaning and replacement. Below the filter screen, an inclined water collection plate with a 5% slope is installed. After being filtered by the filter screen, rainwater flows quickly along the water collection plate into the Φ50 mm PE water collection pipe at the bottom. This water collection pipe and the water supply pipe of the atmospheric water vapor collection unit merge underground to form a Φ63 mm main water supply pipe, which is connected to the water storage unit to achieve unified storage of the two types of water sources.In addition, an HX711 weight sensor is installed at the bottom of the rain collection hopper. This sensor has an accuracy of ±1g and can monitor the weight of sand and dust accumulation on the filter surface in real time. When the accumulated weight exceeds 1kg (at which point the filter pore blockage rate reaches 30%, which will affect the rain collection efficiency), the system will immediately send a "filter cleaning reminder" to the maintenance personnel's mobile APP. At the same time, it will control the rain collection hopper to tilt 15° via the bracket, using gravity to shake off 30%-40% of the surface sand and dust, extending the service life of the filter. After a single rainfall event (the atmospheric humidity monitoring unit detects that the relative humidity has dropped below 40% and there has been no rainfall for 30 minutes), the system will control the rain collection hopper to fold and store it back in the underground tank to avoid prolonged exposure to strong ultraviolet rays (the ultraviolet intensity in the Hexi Corridor reaches UVB3.0-4.0 in spring), which can cause the canvas to age and crack.

[0083] Configuration and threshold triggering mechanism of humidity sensing module

[0084] The humidity sensing module, as the core of the system's decision-making trigger, is responsible for real-time capture of key humidity parameters for the growth of short-lived plants, precise setting of species-specific thresholds, and automatic generation of irrigation instructions and water spatial allocation strategies. All its components have undergone extreme environment testing and can operate stably within a temperature range of -30℃ (extreme low temperatures in winter in the Hexi Corridor) to 70℃ (high surface temperatures in summer), perfectly adapting to the extreme temperature difference environment of the desert and ensuring the reliability of data acquisition and instruction generation. Among them, the root zone humidity sensor group uses the TDR-300 time-domain reflectometry sensor. This sensor calculates volumetric water content by measuring the soil dielectric constant. Compared with traditional capacitive sensors, its biggest advantage is that it is not affected by the salt content in desert sandy soil (the salt content of desert soil in the Hexi Corridor is about 0.3%-0.5%). The measurement range covers 0-100%, with an accuracy of ±1% and a resolution of 0.1%, which can accurately capture subtle changes in soil moisture in the root zone. To address the differences in root distribution between *Gnaphalium affine* (whose roots are concentrated in the 5-10cm soil layer and rely primarily on shallow roots for water absorption) and *Capsella bursa-pastoris* (whose roots extend to 8-12cm, exhibiting greater drought resistance and requiring monitoring of deeper soil moisture), a "layered burial" approach was adopted for the sensors. In the *Gnaphalium affine* growing area, one set of sensors was placed every 10 square meters, buried at three depths: 5cm (shallow root absorption zone), 8cm (middle main root zone), and 10cm (deep root reserve zone), with one probe at each depth, forming a three-dimensional root moisture monitoring chain of "shallow-middle-deep" to ensure comprehensive understanding of the moisture status of different soil layers. In the *Capsella bursa-pastoris* growing area, the sensor burial depth was adjusted to 8cm, 10cm, and 12cm, focusing on monitoring the moisture in the main root distribution area to avoid misjudgments of water demand due to improper monitoring depth. To prevent sandy soil particles (0.1-0.5mm in diameter) from clogging the sensor probe and affecting measurement accuracy, the probe is wrapped with a 0.1mm quartz sand filter sleeve. This filter sleeve not only blocks soil particles but also allows soil moisture to freely permeate to the probe surface. It is also resistant to soil corrosion and has a service life consistent with the sensor (approximately 5 years). The probe is connected to the CR1000X data acquisition unit via an RVVP2×0.5mm² shielded cable. This cable employs a double-layer shielding structure: an outer layer of galvanized steel wire braided mesh and an inner layer of aluminum foil, effectively resisting abrasion from wind and sand in desert areas. Despite the static interference generated by wiping and the electromagnetic interference from surrounding equipment, the data transmission error was still controlled within ±0.5% even under strong interference conditions. The sampling cycle of the data acquisition unit is set to 30 seconds / time, which can reflect humidity changes in real time while avoiding energy waste caused by oversampling. The acquisition unit uploads real-time humidity data to the system control center through the RA-02 LoRa wireless transmission module. This module has a transmission distance of up to 3km, which can achieve stable data transmission even in desert areas without wired network coverage. It also has low power consumption and can work continuously for 3 months on a single charge.To prevent the sensors from being buried by sandstorms (the average monthly sandstorm accumulation thickness in the Hexi Corridor reaches 2-3 cm in spring), each sensor is surrounded by a PE protective ring with a diameter of 30 cm and a height of 15 cm. The top of the protective ring is covered with a plastic mesh with a 5 mm aperture. The mesh aperture has been calculated to prevent sand and gravel larger than 5 mm from entering while not affecting the infiltration of rainwater and irrigation water, ensuring that the soil moisture around the sensor is consistent with the actual root zone. The atmospheric humidity monitoring unit uses an SHT35 digital temperature and humidity sensor. This sensor has a relative humidity measurement range of 0-100% with an accuracy of ±2%RH (20℃-60℃ range, which is the range for most of the desert in spring), and a temperature measurement range of -40℃-125℃. It can simultaneously collect relative humidity and dew point temperature data. The dew point temperature is calculated by the sensor's built-in high-precision algorithm with an accuracy of ±0.5℃, which can accurately predict the atmospheric humidification potential. The sensor is mounted on a 1.5m high meteorological observation bracket, with the base fixed to a 0.5m deep concrete foundation, ensuring stability even in gale-force winds (common in spring in the Hexi Corridor). The sensor housing features an IP67-rated dustproof design, with louvered ventilation openings on the sides. Dustproof mesh (0.05mm aperture) is affixed to the inside of the ventilation openings, ensuring airflow to the sensor probe for measurement accuracy while preventing direct impact and damage from sand and dust. The sensor collects data at a frequency of 1 minute and transmits it via a 485 bus to a nearby root zone humidity sensor data acquisition unit. The root zone humidity data is then synchronously uploaded to the control center, avoiding data delays caused by separate transmissions from multiple modules. Atmospheric humidity data is mainly used in two key scenarios: First, to help determine the trend of humidity changes in the root zone—when the humidity in the root zone is close to a set threshold (e.g., 12% during the germination period of *Gnaphalium affine*) and the dew point temperature is consistently more than 2°C lower than the surface temperature, it indicates that there is no possibility of natural humidification in the next 24 hours. The system will start irrigation preparation 10 minutes in advance, such as pre-starting the variable frequency centrifugal pump of the water supply regulation unit to ensure that the irrigation command can be executed immediately after it is issued; Second, to trigger the deployment of the rain collection bucket of the sporadic precipitation collection unit—when the relative humidity exceeds 80% for 1 hour and the difference between the dew point temperature and the surface temperature is less than 2°C, it is determined that "the probability of rainfall is high," and the rain collection bucket is automatically controlled to deploy to avoid missing the brief opportunity for rainfall.The threshold setting unit is based on the water requirements of the two short-lived plants and the water retention capacity of the sandy soil in the Hexi Corridor (the field water holding capacity of sandy soil is about 8%-12%, the water infiltration rate is fast, and the water retention time is short). The parameter configuration interface of the system control center presets the species-specific humidity threshold, and the threshold is finely set according to the dual dimensions of "growth stage + species" to ensure that it is fully adapted to the differences in water requirements at different stages: for *Gnaphalium affine*, during the germination period (mid to late March, when seeds need sufficient water for germination and the root system has not yet formed the ability to absorb water). The threshold for volumetric water content in the 5-8cm soil layer of the root zone is set at 12%-15%. Irrigation is triggered immediately when it falls below 12% to ensure seed germination rate. During the growing season (April, when leaves grow rapidly and water demand increases, but excessive moisture in sandy soil can easily lead to root rot due to oxygen deficiency), the threshold is lowered to 10%-12% to meet growth needs while preventing root rot. During the fruiting period (early May, when water supply needs to be reduced to promote seed maturation and prevent excessive plant growth), the threshold is further lowered to 8%-10% to ensure seed plumpness. For *Capsella bursa-pastoris*, because its drought tolerance is stronger than that of *Gnaphalium affine* (thicker leaf cuticle, 20% lower transpiration rate), the threshold is set at 10%-13% during the germination period (late March), triggering irrigation when it falls below 10%. The threshold is 8%-10% during the growing season (mid-to-late April) and 6%-8% during the fruiting period (mid-May), further reducing water demand. In addition, the threshold setting unit supports a "dynamic adjustment" function—maintenance personnel can view historical humidity data (humidity change curves for the past 7 and 30 days) and plant growth status in real time via a mobile app (by visually judging whether plants show signs of water shortage such as wilting and yellowing leaves through images uploaded by the community distribution identification unit). If wilting is found in plants at a certain growth stage (e.g., when the humidity of *Gnaphalium affine* reaches 10% but the leaves are curled), the threshold for that stage can be remotely increased by 1%-2% (e.g., from 10%-12% to 11%-13%). After adjustment, the system will immediately perform monitoring according to the new threshold, ensuring that the threshold always accurately matches the actual water requirement. The instruction generation unit, as the "decision core" of the humidity sensing module, realizes a fully automated process of "threshold determination - strategy generation - instruction issuance" through the embedded algorithm built into the control center, without manual intervention. Its judgment logic has been optimized through extensive field testing: when the data uploaded by the root zone humidity sensor group for three consecutive sampling cycles (a total of 1.5 minutes) are all below the threshold of the current growth stage of the corresponding species (such as less than 10% during the germination period of desert purslane), and the atmospheric humidity monitoring unit reports that "the dew point temperature is more than 2°C lower than the surface temperature" (ensuring that there is no possibility of natural humidification and avoiding accidental irrigation), the unit will automatically determine that "irrigation needs to be started". This design can effectively avoid accidental irrigation caused by instantaneous sensor errors (such as a sudden drop in humidity caused by sand and dust briefly covering the probe).After the determination, the unit immediately generates two types of key information: First, irrigation instructions, including the target irrigation area (determined by GPS positioning of the root zone humidity sensor, accurate to a 10m×10m partition, such as "the 10m×10m area covered by sensor 3 in the germination zone of *Gnaphalium affine*"), basic irrigation flow rate (referencing the species water requirement characteristics of the threshold setting unit, with a basic flow rate of 1.2L / h·m² during the germination period of *Gnaphalium affine* and 1.0L / h·m² during the germination period of *Capsella bursa-pastoris*, the flow rate setting is based on the average daily water requirement of a single plant), and basic irrigation duration (calculated according to the degree of water shortage: mild water shortage—humidity 1%-2% below the threshold, 30 minutes; moderate water shortage—humidity 2%-3% below the threshold, 45 minutes; severe water shortage—humidity more than 3% below the threshold, 60 minutes, to avoid insufficient irrigation in a short period of time); Second, a water spatial allocation strategy, which is generated in conjunction with the plant density data uploaded by the community distribution identification unit to ensure that water resources are allocated to areas with more urgent water needs. Area adjustment: For densely populated areas (≥8 plants / m² for *Gnaphalium affine*, ≥6 plants / m² for *Capsella bursa-pastoris*, where water competition among plants is intense), increase the irrigation flow rate by 20%-30% (e.g., from 1.2 L / h·m² to 1.4-1.5 L / h·m² in densely populated areas), and extend the irrigation duration by 15%-20% (e.g., from 30 minutes to 34.5-36 minutes) to ensure each plant receives sufficient water; for sparsely populated areas (*Gnaphalium affine*...)... For areas with ≤3 plants / m² and ≤2 plants / m² of *Capsella bursa-pastoris* (due to weak water competition and easy infiltration into sparsely populated areas), the irrigation flow rate is reduced to 60%-70% of the baseline flow rate (e.g., from 1.0 L / h·m² to 0.6-0.7 L / h·m² in sparsely populated areas), and the irrigation duration is shortened to 60%-70% of the baseline duration (e.g., from 30 minutes to 18-21 minutes), reducing water waste. For sparsely populated areas, "no irrigation required" is directly marked to avoid ineffective water supply. The generated irrigation instructions and water spatial allocation strategies are transmitted in real-time to the micro-irrigation module and water storage control module via the EC20 4G IoT module. This module supports full network connectivity, ensuring instruction delivery even in areas with weak signals at the edge of the desert via SMS backup, ensuring coordinated response from all modules and preventing execution delays.

[0085] Laying out and precisely controlling micro-irrigation modules

[0086] As the water delivery terminal of the system, the micro-irrigation module is a key link in accurately delivering water to the root zone of short-lived plants. Its core objective is to ensure that each plant receives adequate water supply while minimizing water loss. A single module can cover a 500㎡ growing area. Field tests have shown that its water use efficiency is over 85%, far exceeding the 30%-40% efficiency of traditional flood irrigation, significantly improving the utilization value of desert water resources. The module includes a capillary network unit, a flow control unit, and an irrigation execution unit, which work together to achieve an irrigation process of "precise water delivery - on-demand quantity control - timely start and stop". As the "last mile" of water delivery, the capillary network unit uses a porous capillary network made of food-grade polyethylene (PE), which has excellent aging resistance and a service life of more than 5 years under strong ultraviolet radiation in the desert. It also has outstanding low-temperature anti-freezing and crack resistance, and will not break even at extreme low temperatures of -20℃, making it perfectly suited to the complex desert environment. The network specification is Φ5mm×0.8mm (inner diameter 5mm, wall thickness 0.8mm), which ensures a certain water delivery capacity while controlling the water flow rate through the small pipe diameter, avoiding uneven water output due to excessive pressure. Micropores with a diameter of 0.5 mm are evenly spaced at 15 cm intervals on the surface of the pipe network. This pore size has been optimized through fluid dynamics simulation and field testing: if the pore size is too small (<0.4 mm), it is easily blocked by sand and dust, and if the pore size is too large (>0.6 mm), it will lead to an excessively fast water flow rate, which will easily impact the soil and expose the roots. The water outlet method of the 0.5 mm micropores is "slow seepage", with a stable water flow rate of 0.15 L / h・m. The water can be evenly penetrated into the soil to form a moist zone with a radius of 10-15 cm centered on the pipe network, which just covers the root distribution range of short-lived plants. Based on the difference in root depth between *Erigeron cantorii* and *Capsella burlata*, a "depth-adaptive" burial scheme was adopted for the pipeline network: In the *Erigeron cantorii* growing area, the pipeline network was buried at a depth of 8cm, corresponding to the middle of the 5-10cm soil layer where its roots are concentrated, ensuring that water is directly transported to the main root absorption area; considering the small spacing between *Erigeron cantorii* plants (row spacing 20cm), the pipeline network was laid in a zigzag pattern along the row spacing, with one pipeline network laid on each side of each row of plants, so that each *Erigeron cantorii* plant has 2-3 stable micropore water outlets around its roots, avoiding blind spots in water coverage; in the *Capsella burlata* growing area, because its roots are deeper (8-12cm), the pipeline network burial depth was adjusted to 10cm, and the row spacing was 30cm (plant spacing is larger), so one pipeline network laid on one side of each row of plants is sufficient to meet the water demand, while the micropore spacing was reduced to 12cm to adapt to its denser fibrous root system, ensuring that the fibrous roots can fully absorb water.Before laying the pipeline, meticulous site pretreatment is required: First, excavate trenches on the ground with a depth of "burial depth + 5cm" according to the pipeline route (e.g., in the area of ​​*Gnaphalium affine*, the trench depth is 8cm + 5cm = 13cm). The trench width should be controlled at 10cm to avoid excessive backfilling. A 5cm thick quartz sand cushion layer (0.5-1mm particle size) should be laid at the bottom of the trench. This cushion layer has good filtration and buffering effects, filtering out fine sand and dust in the soil to prevent them from migrating with water and clogging the micropores of the pipeline, and buffering the direct contact between the pipeline and the soil to reduce soil pressure damage to the pipeline. The pipeline should be laid straight to avoid bending that could obstruct water flow. After laying, backfill with the original sandy soil generated during trench excavation in layers, each layer 5cm thick. Light compaction should be done manually during backfilling (85% compaction per layer) to prevent voids from forming in the trench, which could cause pipeline displacement and affect the uniformity of water output. To further prevent blockage of the micropores in the pipe network, a 2W-160-10 type drain valve is installed at the end of the pipe network (the farthest point of each zone). Maintenance personnel must open the drain valve once a month for 5-10 minutes each time to flush out any remaining traces of sand and dust in the pipe network using water pressure, ensuring unobstructed micropores. Testing has shown that regular flushing can keep the pipe network blockage rate below 5%. The flow control unit, acting as a "precise regulator" of irrigation flow, uses a 2W-160-15 type electromagnetic regulating valve as its core control component. This valve has a 15mm diameter, a working pressure range of 0.02-1.0MPa, fully covering the pressure requirements of the desert micro-irrigation system, and a response time of ≤1 second, enabling rapid response to flow adjustment commands. Each 10m×10m irrigation zone is equipped with one electromagnetic regulating valve, installed at the inlet end of the capillary network, and connected to the water supply pipeline of the water storage and control module via a Φ32mm PE main pipe, achieving independent flow control for each zone. The opening degree (0%-100%) of the electromagnetic regulating valve has a strict linear relationship with the irrigation flow rate. After calibration, a 50% opening corresponds to a flow rate of 0.6 L / h·m², and a 100% opening corresponds to a flow rate of 1.2 L / h·m². The system can precisely control the opening degree by sending a PWM (Pulse Width Modulation) signal to the valve, achieving stepless flow regulation. To ensure flow control accuracy, an FS300A miniature turbine flow sensor is installed at the outlet of each electromagnetic regulating valve. This sensor has a measurement range of 0.1-10 L / min and an accuracy of ±2%, and can collect actual outflow flow data in real time, feeding the data back to the control center via a 485 bus. When the deviation between the actual flow rate and the commanded flow rate exceeds 5% (e.g., the commanded flow rate is 1.2 L / h·m², but the actual flow rate is only 1.1 L / h·m²), the control center will automatically adjust the opening degree of the electromagnetic regulating valve (e.g., from 100% to 105%) until the flow deviation is controlled within ±3%, ensuring that the actual water supply to each zone perfectly matches the demand.In addition, the flow control unit also has an "anti-clogging protection" function—when the flow sensor detects a sudden drop in flow of more than 30% within 10 seconds (determined as a partial blockage in the pipeline), the control center will immediately control the electromagnetic regulating valve of that zone to quickly switch to the "fully open" state (for 10 seconds), using high-pressure water flow to flush out the blockage; if the flow does not recover after flushing, the system will send a "pipeline blockage warning" to the maintenance personnel, along with the GPS coordinates of the blocked zone, to facilitate accurate location and manual cleaning by the maintenance personnel, preventing the blockage from expanding and affecting the irrigation of the entire zone. The irrigation execution unit, as the "command execution terminal" of the micro-irrigation module, is responsible for translating irrigation commands into specific irrigation actions. Through the zone control logic of the control center, it achieves precise control of the irrigation duration for different areas, while also having real-time feedback and adjustment capabilities to ensure that the irrigation process is "neither too much nor too little irrigation". After receiving the instruction and generating the "target area - basic duration - density adjustment coefficient" data issued by the unit, the unit will automatically calculate the actual irrigation duration for each zone. The calculation formula is "actual duration = basic duration × density adjustment coefficient". For example, if the basic duration for a dense area of ​​*Gnaphalium affine* is 30 minutes and the density adjustment coefficient is 1.2 (the duration needs to be extended in dense areas), then the actual irrigation duration is 30 minutes × 1.2 = 36 minutes. After the calculation is completed, the unit sends an "open" instruction to the electromagnetic regulating valve of the corresponding zone, and irrigation officially starts. During irrigation, the unit continuously receives real-time feedback data from the root zone humidity sensor, forming a dynamic adjustment mechanism: If the root zone humidity has risen to the species-appropriate threshold (e.g., 70% of 36 minutes is 25.2 minutes) when a certain zone reaches 70% of the actual duration (e.g., 70% of 36 minutes is 25.2 minutes), the unit will immediately send a "close" command to the electromagnetic regulating valve of that zone to stop irrigation in advance and avoid over-irrigation leading to root rot. Tests have shown that this mechanism can reduce water waste by 15%-20%. If the humidity in a certain zone has not reached the threshold by the end of the actual duration due to poor soil water retention (e.g., sandy slopes with fast water infiltration), the unit will automatically request "extended irrigation" from the control center. The request includes current humidity data, soil type (provided by the micro-topography analysis unit), and other information. After the control center approves the request, it allows an extension of 5-10 minutes (the extension time shall not exceed 30% of the base duration to avoid excessive water consumption) to ensure that the roots absorb enough water to meet their growth needs.In addition, the irrigation execution unit also supports the "priority scheduling" function. When the water storage and control module reports "low water level (remaining volume 10%-30%)", the unit will automatically retrieve the irrigation priority of each zone (priority is determined by the species' growth stage, such as germination period > growth period > fruiting period, and *Gnaphalium affine* > *Capsella bursa-pastoris*). Priority will be given to irrigating high-priority areas (such as the germination area of ​​*Gnaphalium affine*) and suspending irrigation of low-priority areas (such as the fruiting area of ​​*Capsella bursa-pastoris*). After the water level in the water storage unit recovers to more than 30%, the low-priority areas will be irrigated according to the principle of "urgent first, then less urgent", ensuring that limited water resources are tilted towards the areas with the most urgent water needs and maximizing the value of water resource utilization.

[0087] Environmental perception and parameter optimization of the ecological adaptation module

[0088] The ecological adaptation module, serving as the dynamic optimization hub of the system, addresses water misallocation issues caused by neglecting micro-topographical differences (such as low-lying areas and slopes) and uneven plant community distribution in deserts (e.g., waterlogging and root rot in low-lying areas, and water shortage and wilting in slopes). Through autonomous optimization of irrigation parameters and spatial water allocation strategies, it further improves water use efficiency by 10%-15%. The module includes a micro-topographical analysis unit, a community distribution identification unit, and a parameter optimization unit, which together form a complete link of "environmental perception - data processing - strategy optimization," ensuring that the irrigation scheme is highly adapted to the actual desert environment. The micro-topography analysis unit, acting as a "surface environment detector," utilizes the RIEGLVZ-4000 3D lidar as its core sensing device. This radar boasts exceptionally high measurement accuracy and environmental adaptability: its measurement range reaches 4000m, fully covering the 500㎡ control range of a single system; its horizontal angular resolution is 0.004°, and its vertical angular resolution is 0.002°, enabling it to accurately capture centimeter-level elevation changes on the desert surface, clearly identifying even tiny depressions with a diameter of 1m and a height of 0.1m. The radar also possesses strong resistance to sandstorm interference; even in sandstorm weather with visibility of 500m, the measurement error remains within ±2cm, ensuring data reliability. The radar is mounted atop a 20m-high observation tower, which employs a steel truss structure and is fixed to a 1.5m-deep concrete foundation, with a wind resistance rating of up to level 12, capable of withstanding the 8-10 level gales commonly seen in the Hexi Corridor during spring. The tower surface is coated with anti-corrosion paint to prevent corrosion caused by prolonged exposure to sandstorms. The radar performs a three-dimensional scan of the system's coverage area once a week, with the scan time chosen to be between 9:00 and 10:00 AM every Wednesday (when the lighting is stable, surface shadows cause less interference, and data repeatability is good). Each scan lasts about 30 minutes, and the scan data is transmitted in real time to the terrain analysis server in the control center via gigabit Ethernet to avoid data storage delays.The server uses CloudCompare (a professional point cloud processing software) to standardize the scanned data: The first step is noise removal, using a statistical filtering algorithm to remove noise points (approximately 5%-8% of the total point cloud) caused by sand, birds, and dead wood fragments, retaining only the effective point cloud that reflects the surface elevation; the second step is point cloud interpolation, using the Kriging interpolation algorithm to generate a digital elevation model (DEM) with a resolution of 0.1m × 0.1m from the discrete point cloud data. This resolution accurately reflects the micro-topographic features of the desert surface, providing data for subsequent classification. The third step is micro-topographic classification, which calculates the slope and relative elevation of each grid based on DEM data: areas with a slope >5% are classified as "sloping land"—in the sandy soil of the Hexi Corridor, the water infiltration rate is more than 30% faster than in flat areas when the slope is >5%, making water loss more likely; areas with elevations more than 0.3m lower than the surrounding average elevation are classified as "low-lying areas"—the soil in these areas has strong water retention, and water tends to accumulate after irrigation, easily leading to root rot; the remaining areas are classified as "flat areas"—the soil has moderate water retention and water infiltration rate, and conventional irrigation parameters can be used. The micro-topographic classification results are stored in a structured data format of "topographic type-coordinate range (accurate to 0.1m)" and synchronized in real time to the parameter optimization unit, providing a surface environment basis for irrigation parameter optimization. The community distribution identification unit, acting as a "plant distribution mapper," is responsible for accurately identifying sparse and dense areas of short-lived plants, providing a basis for plant distribution in water spatial allocation. The unit uses a Hikvision DS-2CD8A26FWD-E 4K industrial camera as its image acquisition device. This camera has a resolution of 3840×2160, clearly capturing details of short-lived plants with a diameter of only 3-5cm; a frame rate of 25fps ensures no image blur; the lens focal length is 8mm, and the horizontal field of view is approximately 60°. A single camera at a height of 10m can cover an area of ​​20m×20m, with uniform imaging and no edge distortion. Each system is configured with one camera per 100㎡ (a total of 5 cameras). The cameras are mounted on a 10m high steel bracket, with the base of the bracket fixed to a concrete foundation. The camera angle is tilted downwards by 15° to ensure complete coverage of the designated area with no blind spots. The camera housing features an IP67-rated dustproof design, and a replaceable dustproof lens is installed in front of the lens. The lens surface is coated with an anti-fog coating to prevent fogging caused by temperature differences between day and night, which could affect image quality. The camera has a built-in automatic dust removal function, which automatically activates a brush to clean the lens every 24 hours to ensure image clarity. The camera collects images daily between 9:00 and 10:00 AM (when the sun's altitude angle is approximately 45°, the ground lighting is uniform, and the contrast between vegetation and sand is at its highest, resulting in an accuracy rate of over 95%). Each time, it collects 10 images from different angles (rotating once at 36° intervals). PoE (Power over Ethernet) technology integrates image data transmission and device power supply, simplifying wiring.After the image data is transmitted to the image processing server, plant distribution identification is completed through a five-step method: The first step is image preprocessing. First, the color image is converted to a grayscale image to reduce color interference. Then, a 5×5 kernel Gaussian filter algorithm is used to smooth image noise (mainly high-frequency noise caused by sand particles). Finally, a histogram equalization algorithm is used to enhance the grayscale contrast between plants (gray-green) and sand (yellow-white), making the plant outlines clearer—the contrast of the processed image can be improved by more than 40%. The second step is color space conversion. The preprocessed grayscale image is converted back to an RGB image, and then converted to the HSV (Hue-Saturation-Lightness) color space. The characteristics of plants and sand in the HSV space are used for preliminary screening: the saturation (S channel) of plants is usually >30%, and the hue (H channel) is concentrated in 120°-180° (green range), while the saturation of sand is <20%, and the hue is concentrated in 40°-60° (yellow range). By setting a threshold, more than 90% of the plant pixel areas can be preliminarily screened out, and most of the sand background can be excluded. The third step is semantic segmentation. An improved version of the U-Net model is used to perform fine segmentation on the selected images. This model adds an attention mechanism module on the basis of the traditional U-Net, which can more accurately capture the subtle edge features of short-lived plants (such as leaf serrations and stem details), solving the problem of low segmentation accuracy of traditional models for small target plants. The model is trained using local short-lived plant samples from the Hexi Corridor (a total of 10,000 labeled images were collected, covering different growth stages and different lighting conditions). After training, the segmentation accuracy of *Gnaphalium affine* and *Capsella bursa-pastoris* reached 96% and 94%, respectively. Finally, a binary segmentation map of "plant-background" is output (white represents plant pixels, and black represents sandy background). The fourth step is density calculation. The binarized segmented image is divided into several statistical units using a 1m×1m grid (consistent with the irrigation zone scale). An 8-neighborhood connected region analysis algorithm is used to identify independent plant pixel clusters within each grid (each cluster represents a single plant). Combined with the "average pixel area per plant" (approximately 180 pixels / plant for *Gnaphalium affine* and approximately 150 pixels / plant for *Capsella bursa-pastoris*) calibrated through previous field sampling, the actual number of plants in each grid, i.e., the plant density (plants / m²), is calculated. The fifth step is density level determination. Density thresholds are set based on the growth characteristics of the two plants: areas with ≥8 plants / m² for *Gnaphalium affine* are determined as "dense areas," and areas with ≤3 plants / m² are determined as "sparse areas"; areas with ≥6 plants / m² for *Capsella bursa-pastoris* are determined as "dense areas," and areas with ≤2 plants / m² are determined as "sparse areas." The determination results are synchronized to the parameter optimization unit in the data format of "density level-coordinate range-species type."The parameter optimization unit, acting as the "irrigation strategy optimizer," is the core of the ecological adaptation module. It is responsible for dynamically optimizing the initial irrigation parameters (flow rate, duration) and water spatial allocation strategies issued by the instruction generation unit, based on the "terrain type data" from the micro-terrain analysis unit and the "density level data" from the community distribution identification unit. The optimization logic is designed according to "terrain-density" combination scenarios to ensure that the irrigation scheme for each scenario can adapt to the environment and plant needs. The first scenario is "low-lying area + dense area" (such as the low-lying dense germination area of ​​*Gnaphalium affine*). In this scenario, the soil in the low-lying area has strong water retention, and water easily accumulates after irrigation, and the dense... The plant roots in the low-lying area are intertwined, and excessive water can easily lead to root rot due to lack of oxygen. Therefore, the optimization strategy is "reduce flow rate and shorten duration": reduce the initial basic flow rate by 20% (e.g., from 1.2 L / h·m² to 0.96 L / h·m²) to reduce the water supply per unit time; shorten the irrigation duration by 15% (e.g., from 30 minutes to 25.5 minutes) to avoid prolonged water accumulation; and reduce the pipeline pressure to 0.12 MPa (lower than 0.15 MPa in flat areas) to slow down the water infiltration rate and ensure that water is evenly distributed in the root zone. Tests have shown that this optimization can reduce the root rot rate in low-lying areas from 15% before optimization to below 4%. The second scenario is "slope area + sparse area" (such as the sparse growth period of *Capsella bursa-pastoris* on a slope). In this scenario, the soil in the downhill area has poor water retention, and the water infiltration and loss are rapid. In addition, the spacing between plants in the sparse area is large, and water easily diffuses to areas without plants. Therefore, the optimization strategy is "increase pressure - extend duration": increase the pipeline pressure by 15% (e.g., from 0.15MPa to 0.17MPa) to increase the water penetration depth and ensure that water can reach the deep root system; extend the irrigation duration by 20% (e.g., from 40 minutes to 48 minutes) to make up for the gap caused by rapid water loss; maintain the initial base value of flow rate (0.8L / h・㎡) to avoid excessive flow rate leading to increased water loss. After optimization, the soil wetting depth in the slope area can be increased from 8cm before optimization to 12cm, completely covering the root zone of *Capsella bursa-pastoris*. The third scenario is "flat area + dense area" (such as the flat and dense growth period of *Gnaphalium affine*). In this scenario, the soil conditions are moderate, but the plants in the dense area compete fiercely for water. Therefore, the optimization strategy is "stable flow rate - adjusted pressure": the flow rate is maintained at the initial base value (1.0 L / h·m²) to ensure sufficient total water supply; the pipeline pressure is adjusted to 0.15 MPa, at which pressure the water penetration range is uniform and can cover all the roots of the dense plants; the duration is calculated normally according to the density coefficient (such as a base of 30 minutes × 1.2 coefficient = 36 minutes) to ensure that each plant can obtain enough water. After optimization, the uniformity of plant growth in the flat and dense area is improved by 20%.The fourth scenario is a "flat area + sparse area" (such as the flat and sparse fruiting stage of *Capsella bursa-pastoris*). In this scenario, the plants require less water, and the sparse distribution easily leads to water waste. Therefore, the optimization strategy is "reduce flow rate - stabilize duration": the flow rate is reduced to 70% of the initial baseline value (e.g., from 0.6 L / h·m² to 0.42 L / h·m²) to reduce ineffective water supply; the duration is maintained at the initial baseline value (e.g., 30 minutes) to ensure that the roots can fully absorb water, while preventing water from spreading to the blank areas. After optimization, the water waste rate in the flat and sparse area is reduced by 30%. The optimized irrigation parameters are sent in real time from the control center to the micro-irrigation module (adjusting the opening of the electromagnetic regulating valve and the irrigation duration) and the water storage control module (adjusting the water supply and water pressure), directly replacing the original baseline parameters to ensure that the optimization strategy is executed immediately. Meanwhile, the parameter optimization unit generates an "optimization effect report" every week. The report includes: water use efficiency of each scenario before and after optimization (calculated as "irrigation water volume / root zone humidity compliant area"), plant growth status (indicators such as plant height, number of leaves, and germination rate through image analysis by the community distribution identification unit), and soil moisture distribution uniformity (calculated using root zone humidity sensor data as the coefficient of variation). If the optimization effect of a certain scenario is not good (such as the root rot rate still exceeding 5% after optimization in low-lying dense areas), the unit will automatically adjust the optimization coefficient of that scenario (such as further reducing the flow rate by 5%, from 0.96 L / h·m² to 0.91 L / h·m²), and implement a new optimization strategy the following week. Through the cycle of "optimization-evaluation-adjustment", the optimization strategy is ensured to continuously adapt to the dynamic changes in the desert environment and plant growth.

[0089] Water storage and dynamic water supply in the water storage and regulation module

[0090] As the system's water source guarantee, the water storage and regulation module undertakes three key functions: water storage, water level monitoring, and dynamic water supply. Through the whole-process management of "safe storage - risk warning - on-demand water supply", it provides continuous and stable water source support for the entire micro-irrigation system, avoiding irrigation interruptions due to insufficient water source or unstable water supply. The module includes a water storage unit, a water level monitoring unit, and a water supply regulation unit. The three work together to ensure the reliability and stability of irrigation for short-lived desert plants. The water storage unit, acting as a "water reservoir," utilizes an underground pressurized water tank made of 304 stainless steel. 304 stainless steel boasts excellent corrosion resistance, resisting the erosion of salt and microorganisms in desert soil, and has a service life exceeding 10 years. The tank's volume is designed to be 5m³, calculated based on water balance: a single system collects an average of 8-12L of water per day, and the 5m³ tank can store approximately 400-600 days' worth of collected water (considering the high demand for irrigation in spring, it can actually meet the irrigation needs of one growth cycle (3 months),) avoiding frequent water replenishment. The tank's specifications are a diameter of 2m and a height of 1.6m. Its cylindrical structure ensures uniform stress distribution and strong pressure resistance, capable of withstanding soil pressure (approximately 15kPa) at a depth of 1.5m underground, preventing tank deformation. The water tank is buried underground at a depth of 1.5m (the top of the tank is 0.1m below the surface). This design has three major advantages: First, it avoids the impact of extreme temperature differences (-10℃ to 45℃) on the water source in the desert—the temperature at 1.5m underground remains stable at 10℃-15℃ year-round, preventing the tank from cracking due to freezing and expansion in winter, and also avoiding water evaporation caused by high temperatures in summer (the evaporation rate in the underground environment is less than 1% of that at the surface). Second, it prevents burial by wind and sand—only a 0.1m inspection opening is exposed on the surface, effectively preventing wind and sand accumulation from burying the tank in spring. Third, it reduces human-caused damage; the underground installation method provides strong concealment and reduces the risk of equipment theft. The inner wall of the water tank is coated with a food-grade epoxy resin coating (0.2mm thick), ensuring that the stored water meets the standards for plant irrigation (pH 6.5-7.5, total dissolved solids <1000mg / L); the smooth coating surface prevents the adhesion and growth of microorganisms in the water, reducing the frequency of tank cleaning. The top of the water tank has two functional interfaces: one is the water inlet interface (Φ50mm), which connects to the main water supply pipe of the intelligent water collection module through a Φ63mm PE pipe. The interface has a built-in H14W-16P type check valve, which has good sealing performance and can effectively prevent water in the water tank from flowing back into the water collection module, thus avoiding contamination of the water collection equipment; the other is the maintenance interface (Φ300mm), which has a sealing cap (rubber sealing ring). Maintenance personnel can enter the water tank through this interface (once per quarter) to clean and inspect it, ensuring that there is no sand or dust accumulation or rust inside the water tank.A 50mm outlet port is located at the bottom of the water tank, which connects to the inlet pipe of the water supply regulating unit via a 50mm stainless steel pipe. A Y-type filter (0.5mm pore size) is installed at the port to filter out any trace amounts of sand or dust that may remain in the water tank (such as fine particles that were not completely filtered during water collection), protecting downstream precision equipment such as the variable frequency centrifugal pump and electromagnetic regulating valve from clogging. Furthermore, the outside of the water tank is wrapped with a 5cm thick polyurethane insulation layer (thermal conductivity ≤0.024W / (m・K)). This insulation layer has excellent thermal insulation performance, maintaining the water temperature inside the tank above 5℃ even in winter environments of -10℃, completely preventing freezing. The outside of the insulation layer is wrapped with a waterproof membrane to prevent groundwater seepage that could cause the insulation layer to fail. As a "sentinel of water source safety," the water level monitoring unit is responsible for real-time monitoring of the water tank level and timely warning of water shortage and overfill risks. The unit adopts the E+HFMP51 submersible hydrostatic level sensor. The sensor's measurement principle is based on liquid hydrostatic pressure (the liquid level is proportional to the pressure), with a measurement range of 0-2m (fully covering the water tank height of 1.6m) and an accuracy of ±0.5%FS (full-scale error of only ±1cm). It can accurately reflect changes in the water level inside the tank. The sensor probe is encased in a corrosion-resistant 316L stainless steel shell, which is resistant to soil and water corrosion and has a service life consistent with that of the water tank. To achieve full water level monitoring, three sensors were installed, corresponding to three key nodes: low water level, medium water level, and high water level. The low water level sensor was installed 5cm from the bottom of the tank, corresponding to 10% of the total tank volume (5m³ × 10% = 0.5m³). This water level serves as a water replenishment warning line; if the water level falls below this level, water replenishment measures must be taken immediately. The medium water level sensor was installed 80cm from the bottom of the tank, corresponding to 50% of the total volume (2.5m³). This is the normal water level line, indicating sufficient water supply. The high water level sensor was installed 150cm from the bottom of the tank, corresponding to 95% of the total volume (4.75m³). This is the full water warning line; if the water level exceeds this level, water intake must be stopped to prevent the tank from overflowing. The sensor extends into the water tank from the top through a dedicated sealed interface, with the probe completely submerged in water to ensure measurement accuracy. The sensor signal cable (4-20mA analog signal) is protected by a galvanized steel pipe to prevent it from being corroded by soil or chewed by rodents. The cable is connected to the ADAM-4017 water level data acquisition unit. After the acquisition unit converts the analog signal into a digital signal, it is uploaded to the control center via Ethernet. The sensor sampling period is set to 30 seconds / time to ensure real-time monitoring of water level dynamics.The water level monitoring unit's early warning mechanism is divided into two categories: one is a water shortage warning. When the low water level sensor detects that the water level is below 0.5 m³ (10% of the total volume), the unit immediately triggers a water replenishment warning and simultaneously performs three actions: First, it sends a "water tank level warning" notification to the maintenance personnel's mobile APP. The notification includes the water tank location (GPS coordinates), the current water level (e.g., 0.45 m³), ​​the remaining water volume that can sustain irrigation for (calculated based on current irrigation needs, e.g., 24 hours), and a suggested water replenishment method ("Prioritize activating the atmospheric water vapor collection unit to extend condensation time, and assist with manual water replenishment"), ensuring a rapid response from maintenance personnel. Second, the system is linked to the intelligent water collection module, extending the nighttime working time of the condenser plate of the atmospheric water vapor collection unit from 10 hours (20:00-6:00) to 12 hours (19:00-7:00), while increasing the cleaning frequency of the hydrophobic coating of the condenser plate from once a month to once a week to maximize condensation and water collection efficiency. The system also keeps the rain collection hopper of the scattered precipitation collection unit in an extended state, preventing rainwater from being collected even if the relative humidity does not reach the predicted rainfall threshold, thus avoiding missing any potential precipitation. Third, a "flow restriction prompt" is sent to the water supply regulation unit to trigger an emergency water supply strategy, reducing the water supply to non-critical areas and extending the water source usage time. Additionally, a full water warning system is implemented. When the high water level sensor detects a water level higher than 4.75 m³ (95% of total volume), the unit triggers a full water warning, immediately closing the solenoid valve on the intelligent water collection module's inlet pipe to stop water intake. Once the water level drops to 4.5 m³ (90% of total volume), the valve automatically reopens to resume water intake, preventing water tank overflow and water waste. The water supply regulation unit, acting as the "water supply dispatching center," is responsible for dynamically adjusting the water supply volume and pressure based on water level and optimized irrigation needs. This ensures that the micro-irrigation module receives a stable and suitable water source. The unit uses a Grundfos CR3-5 variable frequency centrifugal pump as its core water supply equipment. This pump has a rated flow rate of 2 m³ / h and a rated head of 15 m, fully meeting the maximum water supply demand of a single system (approximately 1.5 m³ / h). The speed can be steplessly adjusted within the range of 960-1450 r / min, and the speed and flow rate have a strict linear relationship (1450 r / min corresponds to a flow rate of 2 m³ / h, and 960 r / min corresponds to a flow rate of 1.2 m³ / h), facilitating precise control of the water supply. The pump body is made of stainless steel, which is resistant to water corrosion and has low noise (operating noise <60 dB), making it suitable for installation in desert environments. The pump body is installed in an underground equipment well next to the water tank. The equipment well is 1.2m deep and 1m in diameter. The well wall is made of brick and the inner wall is plastered for waterproofing. The well cover is a cast iron well cover with ventilation holes (rainproof and dustproof). A temperature and humidity sensor is installed in the equipment well. When the temperature in the well exceeds 40℃, a small exhaust fan will be automatically started to cool down the pump body and ensure that the pump body operates at a suitable temperature (the efficiency is highest below 30℃).The pump's inlet is connected to the outlet at the bottom of the water tank via a Φ50mm stainless steel pipe. A gate valve is installed on the pipe (to facilitate shutting off the water supply during maintenance). The outlet is connected to a Φ50mm PE main pipe, which is then connected to the electromagnetic regulating valves of each irrigation zone via Φ32mm PE branch pipes, forming a water supply network of "main pipe-branch pipe-zone valve". The water supply regulation logic of the water supply regulation unit is based on dual data of "water level status - optimized irrigation demand" and is divided into three scenarios: The first scenario is when the water level is sufficient (remaining volume ≥30%, i.e. ≥1.5m³). At this time, the water source is sufficient and can fully meet the irrigation demand. The unit adjusts the speed of the variable frequency centrifugal pump according to the optimized irrigation flow demand. By receiving the total flow of each zone issued by the parameter optimization unit (e.g., 1.2m³ / h required by 5 zones), the pump speed is controlled to 960r / min to ensure that the total water supply matches the demand. At the same time, the water supply pressure is monitored in real time by the BOSCHBMP280 pressure sensor (measurement range 0-1MPa, accuracy ±1%) installed on the main pipeline. The pressure is stabilized at the optimized target value (e.g., 0.12MPa in low-lying areas, 0.17MPa in sloping areas, and 0.15MPa in flat areas). The pressure fluctuation is controlled within ±0.01MPa to ensure uniform water output from the capillary network and avoid uneven irrigation caused by pressure fluctuations. The second scenario involves low water levels (remaining volume 10%-30%, i.e., 0.5-1.5 m³). In this case, water resources are scarce, requiring water conservation while ensuring critical irrigation. The unit implements a "priority-based, limited-use" water supply strategy: First, reduce the speed of the variable frequency centrifugal pump, decreasing the total water supply to 70% of normal demand (e.g., from 1.5 m³ / h to 1.05 m³ / h), reducing water consumption per unit time. Second, based on irrigation priority (germination stage > growth stage > fruiting stage, *Gnaphalium affine* > *Capsella bursa-pastoris*), prioritize water supply to high-priority irrigation zones (e.g., *Gnaphalium affine* germination zone), suspending irrigation of low-priority zones (e.g., *Capsella bursa-pastoris* fruiting zone) until water levels recover. Third, appropriately reduce the pipeline pressure to 0.12 MPa (uniform pressure, simplified control), reducing water infiltration rate and improving water use efficiency. In this scenario, water usage time can be extended by more than 50%.The third scenario is a water level-triggered water replenishment warning (≤0.5m³). In this case, water resources are extremely scarce, and emergency water supply needs to be activated to avoid irrigation interruption. The unit implements three emergency measures: First, the speed of the variable frequency centrifugal pump is reduced to 720r / min, and the total water supply is reduced to 50% of the normal demand (e.g., 0.75m³ / h). Second, water is supplied only to the highest priority zone (e.g., the dense germination area of ​​*Gnaphalium affine*, which is related to the continuation of the plant population), and an "intermittent water supply" mode is adopted (water supply for 10 minutes, stop for 5 minutes). This intermittent water supply reduces water consumption and ensures that the roots do not wilt due to prolonged water shortage. Third, an "emergency water replenishment request" is sent to the control center. If the water level does not recover to above 0.5m³ within 24 hours, irrigation in non-critical areas is automatically stopped, and only emergency water supply is maintained in the 100㎡ core area (e.g., the *Gnaphalium affine* seed bank area) to ensure that the water source is not exhausted and to buy time for manual water replenishment. In addition, the water supply regulation unit also features a "flow feedback correction" function. An OMROND6F-PH turbine flow sensor is installed on the main pipeline. This sensor has a measurement range of 0.1-10 L / min and an accuracy of ±2%. It can collect total water supply data in real time and compare it with the sum of the demands of each zone. If the deviation between the actual total water supply and the sum of the demands exceeds 5% (e.g., demand 1.2 m³ / h, actual 1.14 m³ / h), the unit will automatically adjust the speed of the variable frequency centrifugal pump (e.g., fine-tuning from 960 r / min to 990 r / min) until the deviation is controlled within ±3%, ensuring accurate matching between water supply and demand and avoiding deviations caused by pump wear or changes in pipeline resistance. Simultaneously, the unit also has a self-diagnostic function. When faults such as pump overload, pipeline blockage, or abnormal pressure are detected, the pump will immediately stop operation, send a fault warning to maintenance personnel, and switch to a standby water supply mode (e.g., activating a small emergency water pump) to ensure uninterrupted irrigation in critical areas, further improving the reliability of the system's water supply.

[0091] In summary, this application embodiment maximizes the capture of atmospheric water vapor and sporadic precipitation through an intelligent water collection module (each unit collects 8-12L of water per day in spring), combined with a water storage and regulation module (a 5m³ water tank ensures water supply throughout the growth cycle) to achieve stable water storage; relying on a humidity sensing module (precisely setting thresholds according to the growth stage of the species) and a micro-irrigation module (water use efficiency exceeding 85%), water is precisely delivered to the root zones of *Gnaphalium affine* and *Capsella bursa-pastoris*, and then the irrigation strategy is dynamically adjusted by an ecological adaptation module (combining micro-topography and community distribution optimization parameters, such as reducing root rot rate in low-lying areas to below 4%), ultimately efficiently solving the problem of water shortage for short-lived plants, significantly improving water resource utilization efficiency, effectively maintaining the fragile balance of the regional desert ecosystem, and ensuring the spring germination and growth and population continuation of the two dominant short-lived plants.

[0092] Next, referring to the accompanying drawings, a humidity-triggered desert water collection and short-lived plant micro-irrigation method based on an embodiment of this application is described.

[0093] like Figure 8 As shown, this humidity-triggered desert water collection and short-lived plant micro-irrigation method includes the following steps:

[0094] In step S101, data on atmospheric diffuse water vapor in the desert, sporadic precipitation data, and micro-topography and community distribution data of short-lived plant growth areas are acquired.

[0095] It is understood that the embodiments of this application obtain atmospheric diffuse water vapor data (such as nighttime humidity and temperature difference) and sporadic precipitation data (such as rainfall probability and single rainfall amount) to accurately determine the operating time (such as condenser plate start-up and shutdown, rainwater collection bucket deployment) and water collection potential of intelligent water collection, avoiding blind energy consumption for water collection. Obtaining micro-topographic data (such as low-lying areas and slope distribution) can clarify the soil water retention and loss characteristics of different areas. Community distribution data (such as dense / sparse areas of toothgrass / desert bulrush) can reflect the actual water demand differences, providing a basis for subsequent optimization of irrigation parameters, avoiding water mismatch, accurately setting species-specific humidity thresholds, planning the burial depth and layout of capillary networks, ensuring the germination and growth of short-lived plants, and adapting to the extreme desert environment.

[0096] In step S102, based on desert atmospheric diffuse water vapor data and sporadic precipitation data, the soil moisture and atmospheric humidity in the root zone of short-lived plants are monitored in real time. Combined with the short-lived plant water stress prediction algorithm, a species-specific humidity threshold is set for dynamic adaptive adjustment. When the humidity is lower than the threshold, irrigation instructions and water spatial allocation strategies are automatically generated.

[0097] Among them, the short-lived plant water stress prediction algorithm is designed for the characteristics of desert short-lived plants (such as toothgrass and desert capernaum) and extreme drought environment. It integrates multi-dimensional data such as root zone soil moisture, transpiration rate, atmospheric relative humidity, and soil water potential to calculate the water stress index (WSI) in a weighted manner to quantify the stress degree, thereby predicting stress risk and providing a basis for water replenishment decision-making for micro-irrigation systems.

[0098] It is understood that the embodiments of this application accurately quantify the degree of stress by integrating multi-dimensional data such as root zone soil moisture, transpiration rate, atmospheric relative humidity, and soil water potential, and weighted calculation of water stress index. This avoids misjudgment caused by relying solely on a single soil moisture index and improves the accuracy of stress prediction in desert environments. Based on real-time atmospheric diffuse water vapor, sporadic precipitation data, and plant stress trends, it can dynamically and adaptively adjust the specific humidity thresholds for different growth stages of species such as *Hygrophorus spp.* and *Capsella bursa-pastoris*, avoiding the disconnect between the thresholds and actual water requirements. This allows for early prediction of water stress risks. When the stress level is determined to reach the warning line by WSI (e.g., WSI≥0.3), it can generate irrigation instructions and water spatial allocation strategies that better meet actual needs, ensuring that short-lived plants receive water in time to avoid wilting and reducing ineffective irrigation waste.

[0099] Algorithm for predicting water stress in short-lived plants:

[0100]

[0101] in, The water stress index; This represents the soil moisture stress weighting coefficient. This represents the current soil volumetric moisture content. This refers to the volumetric water content corresponding to the soil's field water holding capacity. This refers to the temperature stress weighting coefficient; The current ambient temperature; The highest temperature threshold suitable for plant growth; This is the relative humidity stress weighting coefficient; The current ambient relative humidity; The optimal relative humidity for plant growth; This is the water potential stress weighting coefficient; This represents the current water potential of the plant tissue. Water potential is used to indicate plant wilting; To control the water level in the fields.

[0102] For example, in the conservation of *Gnaphalium affine* during its germination period in extremely arid regions, this algorithm uses sensors to collect real-time data on atmospheric diffuse water vapor humidity (average 65% at night), sporadic precipitation (single 3mm light rain), soil moisture in the 5-10cm root zone (initial 11%), and transpiration rate (1.2 mmol / (m²・s)). It then integrates soil water potential (-1.2 MPa) and atmospheric relative humidity (30% during the day) parameters, and calculates the results using *Gnaphalium affine*-specific weights (α=0.4, β=0.2, etc.). The water stress index (WSI=0.32) triggered a mild stress warning. Simultaneously, the humidity threshold was dynamically adjusted (from 12% to 11.5%) by combining micro-topography (slope) and community density (dense area) data. This generated an irrigation command of "pressure increase delay + flow increase delay", which drove the capillary network to deliver water precisely. Ultimately, the germination rate of *Paspalum notatum* in the area increased by 28%, the wilting rate was controlled below 3%, and the water use efficiency reached 87%. Compared with traditional fixed threshold irrigation, ineffective water consumption was reduced by 15%.

[0103] In step S103, based on irrigation instructions and water spatial allocation strategies, water is precisely delivered to the root zone of short-lived plants through a porous capillary network, and irrigation flow and duration are controlled. Based on the micro-topography and community distribution data of the short-lived plant growth area, irrigation parameters and water spatial allocation strategies are optimized. Water collected from underground pressurized water tanks is stored and the water level is monitored in real time. Combined with predictive scheduling algorithms, water level data and water spatial allocation strategies are used to dynamically adjust the water supply.

[0104] Predictive scheduling algorithms refer to scheduling decision-making techniques that dynamically optimize resource allocation, task timing, or execution plans by integrating historical operational data, real-time monitoring data, and future trend prediction data (such as resource supply and demand, environmental changes, and equipment status) to proactively avoid risks, improve operational efficiency, and ensure system stability.

[0105] It is understood that the embodiments of this application regulate water supply and water allocation by integrating historical irrigation water consumption data, real-time data and future trend data. This not only adjusts the water supply strategy in advance based on predicted water level changes to avoid irrigation interruptions caused by temporary water shortages, but also optimizes irrigation flow and pressure parameters by combining the long-term water consumption characteristics of micro-topography and community distribution to reduce water resource and pipeline energy waste. Furthermore, it can identify risks in advance, avoid irrigation interruption and over-irrigation problems, improve water supply stability, improve water use efficiency, and reduce the pressure of operation and maintenance emergencies, providing continuous and adaptive water support for the stable growth of short-lived plants in extreme drought environments.

[0106] It should be noted that the formula for the predictive scheduling algorithm is as follows:

[0107]

[0108] in, For predicting traffic flow; This is the first weighting coefficient; Current traffic; This is the current water level; The water level is full. This is the weighting coefficient for the second term; Nutrient-related flows; This represents the probability of rainfall. This is a correction item indicating no impact from rainfall. This is the third weighting coefficient; Historical correction factor; Based on the basic traffic.

[0109] According to the embodiments of this application, a humidity-triggered desert water harvesting and micro-irrigation method for short-lived plants is proposed. This method utilizes an intelligent water harvesting module with an adsorption-type condensation plate and a folded rainwater collection bucket with a built-in filter to efficiently collect atmospheric diffuse water vapor and sporadic precipitation in the desert, broadening the channels for obtaining scarce water resources and avoiding the passive inefficiency and insufficient water utilization of traditional water harvesting methods. The humidity sensing module, relying on multi-depth humidity monitoring in the root zone, atmospheric temperature and humidity data collection, and species-specific threshold configuration, automatically generates irrigation instructions and water spatial allocation strategies, accurately capturing the water demand timing of short-lived plants, breaking the limitations of relying on fixed humidity thresholds or experience-based judgments that lead to poor water demand adaptability. The micro-irrigation module uses a porous capillary network for targeted delivery and a solenoid valve flow... The system employs quantity control and time-series irrigation regulation to precisely supply water to the root zone, ensuring that water delivery always matches the root system's water demand distribution. This solves the problems of high water loss and low accuracy associated with traditional irrigation. The ecological adaptation module combines lidar micro-topography analysis and image recognition of community distribution to autonomously optimize irrigation flow, duration, and water allocation strategies. This comprehensive ecological adaptation, from micro-topography to community density response, enhances the ecological fit of irrigation strategies. The water storage and regulation module utilizes underground pressurized water storage, water level warnings, and dynamic water supply regulation to ensure stable water supply through the pipeline network. It promptly triggers water replenishment warnings when water levels are insufficient, minimizing the risk of irrigation interruptions, improving the rationality of water storage and utilization, and increasing the germination and survival rates of short-lived plants. Thus, it solves the problems of low water collection efficiency and poor irrigation accuracy in existing technologies.

[0110] The following specific embodiment will illustrate a humidity-triggered micro-irrigation method for desert water harvesting and short-lived plants. Figure 9 As shown, it includes:

[0111] The experimental area was selected in the southern desert transition zone of Alxa Left Banner, Inner Mongolia. This area receives an average annual precipitation of 80-150 mm, concentrated between May and July, mostly as scattered showers. The average annual relative humidity is 30%-45%, and the diurnal temperature range is 15-20℃. The main short-lived plants are *Artemisia desertorum*, *Agriophyllum squarrosum*, and *Bassia dasyphylla*, with a natural survival rate of only 35%-40% and vegetation cover of 5%-12%. The micro-topography is mainly characterized by gentle slopes (2°-5°) and shallow depressions, with sandy soil, field water holding capacity of 12%-15%, and wilting humidity of 6%-8%. The implementation area was set at 50 mu (approximately 3.33 hectares), divided into 3 independent irrigation units (16.7 mu per unit) to facilitate zonal management and parameter optimization.

[0112] Each irrigation unit is equipped with two Licor LI-6400XT portable water vapor analyzers, mounted on windproof supports 1.5m high at the unit's edge (to avoid sand interference). The equipment operates within a temperature range of -30℃ to 60℃ and has an IP65 wind and sand resistance rating. Collected parameters include atmospheric water vapor partial pressure (accuracy ±0.01kPa) and water vapor diffusion rate (sampling frequency 1 time / 10 minutes). Data is uploaded to the central controller in real time via a LoRa wireless module (transmission distance 3km). The analyzer positions are adjusted monthly (moved 50m diagonally along the unit) to ensure data coverage of different microenvironments within the unit. The equipment is calibrated weekly using a standard humidity generator, with errors controlled within ±2%. One Tianjin Meteorological Instrument Factory TBQ-3 tipping bucket rain gauge is installed at the center of each of the three units. The rain gauge inlet is equipped with a dust filter (0.5mm aperture) to prevent sand and dust from clogging the tipping bucket. The device has a resolution of 0.1mm, a sampling frequency of 1 time / 5 minutes, and shares a LoRa transmission link with the water vapor analyzers. A 1m x 1m rain gauge is placed below the rain gauge in a PE collection tray (15° tilt), with a PE pipe (50mm diameter) connected to the bottom of the tray to guide the collected precipitation into an underground water tank for resource utilization. After each rainfall event (≥0.5mm), the filter screen is manually cleaned to ensure the accuracy of the next collection. A DJI M300RTK drone equipped with a Zenmuse P1 full-frame camera was used to conduct aerial surveys of a 50-acre area in late April (before the germination of short-lived plants). The flight altitude was 100m, and the ground resolution was 2cm, acquiring DSM (Digital Surface Model) and DOM (Digital Orthophoto). Micro-topographic parameters, including slope (dividing into ≤3° flat areas and 3°-5° gentle slope areas) and elevation difference (recording the distribution and depth of shallow depressions), were extracted using ArcGIS software. Simultaneously, the quadrat method was used to investigate the community distribution: 20 1m×1m quadrats were set up in each unit (evenly distributed in flat areas, gentle slope areas, and shallow depressions), and the species, number of plants, average plant height, and root distribution depth of short-lived plants in the quadrats were recorded (measured by soil drilling; root depth of Artemisia argyi was 18-22cm, that of Artemisia argyi was 12-15cm, and that of Artemisia argyi was 10-13cm). The data were entered into the community distribution database to provide a basis for the subsequent layout of irrigation pipelines.

[0113] Each irrigation unit is equipped with 7 monitoring terminals, arranged in a layout of 3 points in flat areas, 2 points in gentle slope areas, and 2 points in shallow depressions. Each terminal includes one soil moisture sensor (Beijing Soil Instrument Factory TDR-300 model, measurement range 0-100%vol, accuracy ±1%vol) and one temperature and humidity sensor (Swiss Sensirion SHT31, humidity accuracy ±2%RH). The soil moisture sensor is buried at the root depth of the species: 18cm in Artemisia argyi area, 15cm in Amur oleifera area, and 12cm in Acer palmatum area. The area around the sensor is backfilled and compacted with fine sand to avoid air gaps affecting the readings. The temperature and humidity sensors are installed on a bracket 50cm above the ground and collect data synchronously with the soil sensors (once every 5 minutes). The data is transmitted to the central controller via a 4G module, forming a real-time humidity monitoring network. Based on historical data of the region from 2022 to 2023 (including humidity, precipitation, and plant growth status), a water stress prediction model is constructed using a random forest algorithm. The model's input parameters are "average soil moisture in the root zone over the past 72 hours, minimum atmospheric humidity over the past 24 hours, and probability of precipitation in the next 48 hours," and the output parameters are "plant water stress level (0 - no stress, 1 - mild stress, 2 - moderate stress, 3 - severe stress)." The model was trained using the Python sklearn library, with the training set comprising 70% and the test set 30%, achieving an accuracy of 89%. The model was embedded in a central controller to receive monitoring data in real time and output the stress level, providing a basis for threshold setting. Based on the model output and plant growth stages, basic thresholds were set as follows: for seedling stage (late April to early May), the soil moisture threshold for the root zone of *Artemisia annua* was 10% vol, and the atmospheric humidity threshold was 25% RH; for *Agropyron cristatum*, the soil moisture threshold was 9% vol, and the atmospheric humidity threshold was 23% RH; and for *Cephalotaxus fortunei*, the soil moisture threshold was 8% vol, and the atmospheric humidity threshold was 20% RH. During the growing season (mid-May to late June), due to increased transpiration, the threshold is increased by 15% (e.g., the soil moisture threshold for Artemisia annua rises to 11.5% vol); during the flowering period (early July to late July), it is increased by 20%, and during the fruiting period (early August to late August), it is decreased by 10%. Simultaneously, adjustments are made based on real-time stress levels: when the stress level reaches level 1, the threshold is temporarily decreased by 5% (triggering irrigation earlier); when it reaches level 3, the threshold is increased by 10% (extending irrigation duration), forming a dynamic adaptive threshold system.

[0114] Anti-aging PE capillary network (Hebei Weixing Pipe Industry, inner diameter 16mm, wall thickness 2mm, working pressure 0.1-0.3MPa) is used, and the network is laid out according to the micro-topography and community distribution: in flat areas, the network is laid along the direction of the vegetation rows at a spacing of 1.2m; in gently sloping areas, it is laid along the contour lines at a spacing of 1.5m (to avoid water accumulation); in shallow depressions, where water easily accumulates, the spacing is increased to 1.8m, and the burial depth of the network is increased by 2cm (20cm in Artemisia argyi areas and 17cm in Milo areas). 0.3mm diameter permeable holes are opened on the sidewalls of the network (10cm spacing, quincunx arrangement) to ensure uniform water output (variance coefficient ≤5%). Each irrigation unit is equipped with one solenoid valve (Shanghai Julang ZCS type, DN20), which controls the water supply to the network within the control unit. The solenoid valve is linked to the central controller and opens / closes upon receiving irrigation commands. Irrigation parameters are set based on a threshold trigger signal: when the monitored value is below the basic threshold, the solenoid valve opens, and the initial flow rate is set according to the species (0.4 L / h·m for Artemisia arenaria, 0.35 L / h·m for Artemisia arenaria, and 0.3 L / h·m for Artemisia argyi). The irrigation duration is calculated based on the humidity difference (e.g., in Artemisia arenaria, the duration is set to 2 hours if the soil humidity is below the threshold by 1% vol; and 3.5 hours if it is below 2% vol). Pulse width modulation (PWM) technology is used to control the flow rate, and the solenoid valve opening (0-100%) is adjusted by the central controller to achieve precise flow rate regulation. During irrigation, soil humidity data is collected in real time. When the humidity rises back to the threshold + 0.5% vol, the flow rate is halved; when it rises back to the threshold + 1% vol, the solenoid valve closes to avoid over-irrigation. Parameters are optimized monthly based on micro-topography and community growth data: In flat areas, if persistently low soil moisture (below the threshold of 2% vol) is detected, the spacing between pipe networks in that area is reduced to 1m, and the flow rate is increased by 10%; in gently sloping areas, if water infiltration is too rapid, a 0.2mm thick PE geomembrane (30cm wide) is laid under the pipe network to prolong the water retention time; in shallow depressions, if water accumulation occurs, the diameter of the permeable pores in the pipe network is increased to 0.4mm to accelerate drainage. Simultaneously, adjustments are made based on community density: in areas with ≥15 plants per plot, the flow rate is increased by 15%; in areas with ≤5 plants per plot, the flow rate is reduced by 20%, ensuring that water is concentrated in the dominant community.

[0115] Each irrigation unit is equipped with a 5m³ stainless steel pressurized water tank (Jiangsu Jinze stainless steel products, 3mm thick, working pressure 0.6MPa), buried at a higher elevation (1.2m above the unit's average elevation), utilizing gravity for gravity-fed water supply. The tank has a vent (with a dust cap) at the top and a drain valve at the bottom (for monthly drainage). Two level sensors (Shanghai submersible level gauges, 0-3m measurement range, ±1mm accuracy) are installed on the side walls to monitor the high water level (2.8m) and low water level (0.5m), respectively, with data transmitted in real-time to the central controller. The tank inlet is connected to a PE pipe in a collection tray and an atmospheric condensate collection device (using a PP material collection plate, 20㎡ / unit area; condensate is channeled into the tank through a guide channel at night), achieving multi-source water collection. Based on the predictive scheduling algorithm built into the central controller, water supply is scheduled using two types of data: first, meteorological data for the next 7 days (obtained through the China Meteorological Administration API, including precipitation probability, temperature, and wind speed); and second, real-time water tank level and irrigation demand. When the predicted precipitation probability for the next 48 hours is ≥60% (precipitation ≥3mm), the algorithm automatically reduces the next day's water supply by 30%. If the water tank level is ≥2.5m, water supply is suspended and the overflow valve is opened. When the predicted precipitation for the next 7 days is no, and the water tank level is ≥1.5m, water is supplied according to normal irrigation needs. If the water tank level is ≤0.8m, the algorithm prioritizes irrigation for the *Artemisia argyi* area (which has weak drought resistance), reducing the water supply for the *Artemisia argyi* and *Artemisia argyi* areas by 40%, while simultaneously activating emergency water sources (water trucks replenishing water, 10m³ per truck, replenishing to a level of 1.2m). During the scheduling process, a water supply report is generated every 2 hours, recording the water supply, water level changes, and humidity status of the irrigation units, facilitating traceability and optimization.

[0116] The central controller uses a Siemens S7-1200 PLC, integrating data acquisition, algorithm calculation, and command issuance functions. It is equipped with a 10-inch touchscreen (displaying real-time data and equipment status) and connects to a remote monitoring platform (built on Alibaba Cloud) via a 4G module. Administrators can view data and modify threshold parameters via a mobile app (with hierarchical access control). Routine maintenance includes: weekly cleaning of sensor dust (using a soft brush), checking the sealing of pipe network joints (repairing leaks with a PE hot melt machine); monthly calibration of moisture sensors and rain gauges (using standard solutions and measuring cups); and quarterly inspection of the water tank's anti-corrosion layer (reapplying epoxy resin paint if rust is found). After operation from April to August 2024, the survival rate of short-lived plants in the experimental area increased to 78%, soil moisture remained within the threshold range for 92% of the time, and irrigation water utilization reached 85%, verifying the feasibility and effectiveness of the method.

[0117] In summary, the embodiments of this application improve the survival rate of Artemisia arenaria, Artemisia argyi, and Phyllostachys edulis by accurately collecting atmospheric diffused water vapor and sporadic precipitation (with LoRa transmission and water collection basin resource utilization), aerial surveys and quadrat surveys of micro-topography-community data, combined with a random forest algorithm water stress prediction model and species-specific dynamic humidity thresholds (adaptively adjusted according to growth stage and stress level), and micro-topography-adaptive deployment of anti-aging PE capillary networks (differentiated spacing and burial depth in flat areas / gentle slope areas / shallow depressions) and precise PWM flow control. Meanwhile, relying on the multi-source (precipitation + atmospheric condensation) collection and predictive scheduling algorithm of the underground pressurized water tank, combined with real-time humidity monitoring and irrigation parameter optimization, the soil moisture compliance rate reaches 92% and the irrigation water utilization rate reaches 85%. This not only efficiently adapts to the environmental characteristics of sandy soil with low field water holding capacity and large diurnal temperature difference, but also reduces manual maintenance costs through Siemens PLC central controller and Alibaba Cloud remote monitoring platform. It also takes into account the water requirements of different micro-topography and community density, and improves the vegetation restoration capacity and water resource utilization efficiency in desert areas.

[0118] Figure 10 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include:

[0119] The memory 1001, the processor 1002, and the computer program stored on the memory 1001 and capable of running on the processor 1002.

[0120] When the processor 1002 executes the program, it implements a humidity-triggered desert water collection and micro-irrigation method for short-lived plants provided in the above embodiments.

[0121] Furthermore, electronic devices also include:

[0122] Communication interface 1003 is used for communication between memory 1001 and processor 1002.

[0123] The memory 1001 is used to store computer programs that can run on the processor 1002.

[0124] The memory 1001 may include high-speed RAM (Random Access Memory) memory, and may also include non-volatile memory, such as at least one disk storage.

[0125] If the memory 1001, processor 1002, and communication interface 1003 are implemented independently, then the communication interface 1003, memory 1001, and processor 1002 can be interconnected via a bus to complete communication between them. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 10 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0126] Optionally, in a specific implementation, if the memory 1001, processor 1002, and communication interface 1003 are integrated on a single chip, then the memory 1001, processor 1002, and communication interface 1003 can communicate with each other through an internal interface.

[0127] The processor 1002 may be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of this application.

[0128] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described humidity-triggered desert water collection and micro-irrigation method for short-lived plants.

[0129] In addition, this application also provides a computer program product, including a computer program or instructions, which, when executed, implement the above-described humidity-triggered desert water collection and short-lived plant micro-irrigation method.

[0130] In the description of this specification, the references to "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0131] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0132] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0133] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any of the following techniques known in the art, or a combination thereof: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0134] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0135] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A humidity-triggered desert catchment and ephemeral micro-irrigation system based on, characterized by, The application relates to a desert short-lived plant intelligent irrigation system. The intelligent water collection module is used for collecting atmospheric water vapor and sporadic precipitation. The humidity sensing module is used for monitoring the soil humidity of the root zone of the short-lived plants and the atmospheric humidity in real time, setting the humidity threshold value of the species, and automatically generating irrigation instructions and water space distribution strategies when the humidity is lower than the threshold value. The micro-irrigation module is used for receiving the irrigation instructions and the water space distribution strategies, accurately delivering water to the root zone of the short-lived plants through the porous capillary tube network, and controlling the irrigation flow and time. The ecological adaptation module is used for automatically optimizing the irrigation parameters and the water space distribution strategies according to the micro-terrain and the plant community distribution characteristics of the desert surface. The water storage and regulation module is used for storing the collected water source, monitoring the water level in real time, and dynamically adjusting the water supply according to the irrigation demand. The intelligent water collection module comprises an atmospheric water vapor collection unit and a sporadic precipitation collection unit.

2. A humidity-triggered micro-irrigation system for desert catchment and ephemeral plants according to claim 1, characterized in that, The humidity sensing module comprises a root zone humidity sensor group, an atmospheric humidity monitoring unit, a threshold setting unit and an instruction generating unit.

3. A humidity-triggered micro-irrigation system for desert catchment and ephemeral plants according to claim 1, characterized in that, The micro-irrigation module comprises a capillary tube network unit, a flow control unit and an irrigation execution unit.

4. A humidity-triggered micro-irrigation system for desert catchment and ephemeral plants according to claim 1, characterized in that, The ecological adaptation module comprises a micro-terrain analysis unit, a community distribution identification unit and a parameter optimization unit.

5. A humidity-triggered micro-irrigation system for desert catchment and ephemeral plants according to claim 1, characterized in that, The micro-terrain analysis unit identifies the low-lying and slope micro-terrain characteristics of the desert surface through laser radar scanning. The community distribution identification unit distinguishes the sparse area and the dense area of the short-lived plants through image recognition. The parameter optimization unit automatically optimizes the irrigation flow, time and water space distribution strategy according to the micro-terrain characteristics and the community distribution density.

6. A humidity-triggered micro-irrigation system for desert catchment and ephemeral plants according to claim 1, characterized in that, The water storage regulation module comprises a water storage unit, a water level monitoring unit and a water supply amount adjusting unit, wherein the water storage unit adopts an underground pressure-bearing water tank to store collected water sources; the water level monitoring unit is used for monitoring the water level in the tank in real time, and triggering a water replenishment warning when the water level is lower than 10% of the total volume; and the water supply amount adjusting unit is used for dynamically adjusting the water supply amount and water supply pressure according to the optimized water space distribution strategy, irrigation demand and water level data, so as to ensure the stability of pipe network water supply.

7. A method applied to a humidity-triggered desert catchment and ephemeral plant micro-irrigation system according to any one of claims 1-6, characterized in that, The method comprises: obtaining desert atmospheric dispersed water vapor data, sporadic precipitation data and microtopography and community distribution data of a short-lived plant growth area; According to the desert atmospheric dispersed water vapor data and sporadic precipitation data, the root zone soil moisture and atmospheric humidity of the short-lived plant are monitored in real time, and a dynamic self-adaptive adjustment of the species-specific humidity threshold value is set by combining a short-lived plant water stress prediction algorithm, when the humidity is lower than the threshold value, an automatic irrigation instruction and water space distribution strategy are generated; Based on the irrigation instruction and the water space distribution strategy, water is accurately transported to the root zone area of the short-lived plant through a porous capillary tube network, the irrigation flow and irrigation time are controlled, the irrigation parameters and the water space distribution strategy are optimized according to the microtopography of the short-lived plant growth area and the community distribution data, the collected water sources are stored by the underground pressure-bearing water tank and the water storage level is monitored in real time, and the water supply amount is dynamically adjusted by combining the predictive scheduling algorithm, the water level data and the water space distribution strategy.

8. An electronic device, comprising: The computer program or instructions are executed to realize the method.

9. A computer readable storage medium having stored thereon a computer program or instructions, characterized in that, The computer program or instructions are executed to realize the method.

10. A computer program product comprising computer programs or instructions, characterized in that, The computer program or instructions are executed to realize the method.

Images