Farmland liquid nitrogen freezing weeding control method, device, equipment and medium

Abstract

The invention relates to a farmland liquid nitrogen freezing weeding control method and device, equipment and a medium, and the method comprises the steps: calling a preset weed type-liquid nitrogen spraying parameter mapping table to query the weed type of weeds to be removed and reference liquid nitrogen spraying parameters corresponding to the growth stage; acquiring real-time temperature data and real-time humidity data in the farmland to be weeded, calculating and determining a spraying compensation factor of the weeds to be weeded, and respectively updating the first spraying pressure final value and the first spraying speed final value to determine a second spraying pressure final value and a second spraying speed final value; inputting the second spraying pressure final value and the second spraying speed final value into a liquid nitrogen spraying actuator, and driving the liquid nitrogen spraying actuator to perform liquid nitrogen freezing weeding on weeds in the farmland to be weeded; and circularly executing the steps until the weeds in the farmland to be weeded are completely removed. According to the application, liquid nitrogen spraying parameters can be dynamically optimized, incomplete weeding caused by indifference spraying is avoided, and it is ensured that various weeds can be effectively killed.

Description

Technical Field

[0001] This application relates to the field of agricultural weed control, and more particularly to a method for controlling weed control in farmland using liquid nitrogen freezing, a corresponding device, electronic equipment, and a computer-readable storage medium. Background Technology

[0002] Farmland weed infestations are one of the major biological disasters restricting high and stable agricultural yields in my country. Weeds compete with crops for resources such as water, fertilizer, and light, and serve as hosts for pests and diseases, posing a serious threat to crop yield and quality. For a long time, chemical weeding has dominated farmland weed control due to its high efficiency and convenience. However, the resulting problems, such as accelerated weed population succession, rapid development of herbicide resistance, herbicide residues in soil, and herbicide damage to crops, are becoming increasingly prominent, seriously hindering the sustainable development of agriculture.

[0003] Mechanical weeding, as a green alternative technology, uses a power mechanism to drive weeding wheels and shovels to effectively suppress weed growth through root cutting, burying, or pulling. This technology also loosens the topsoil and improves aeration, making it increasingly important in farmland ecological management. However, its practical application still faces challenges such as high seedling damage rates, limited operational efficiency, and cost control, requiring further optimization. In addition, thermophysical weeding technologies, represented by electric shock weeding and laser weeding, are gradually emerging. Their principle is to use high-energy current or laser beams to instantaneously act on weed tissues, destroying their cellular structure, thereby achieving non-chemical, non-contact control. These methods offer high precision, but are limited by low technological maturity, insufficient operational efficiency, high equipment costs, and poor compatibility with traditional production models, preventing large-scale application. Currently, the problems of severe weed infestation, difficulty in control, and high costs in the fields remain prominent.

[0004] In summary, given the increasingly prominent problems in existing chemical weeding technologies, such as accelerated weed population succession, rapid development of herbicide resistance, herbicide residues in soil, and crop damage, as well as the high precision of thermophysical weeding technology, which is limited by low technological maturity, insufficient operational efficiency, high equipment costs, and poor compatibility with traditional production models, the applicant has made corresponding explorations to address these issues. Summary of the Invention

[0005] The purpose of this application is to solve the above-mentioned problems by providing a method for controlling weeding in farmland by liquid nitrogen freezing, a corresponding device, electronic equipment and computer-readable storage medium.

[0006] To achieve the various objectives of this application, the following technical solution is adopted:

[0007] A method for controlling weeds in farmland by freezing with liquid nitrogen, proposed to meet one of the purposes of this application, includes:

[0008] The mobile platform is driven to acquire images of farmland containing weeds to be removed. Based on a farmland weed recognition model that has been trained to a convergent state, the target detection is performed on the images of the farmland to be removed to determine the three-dimensional spatial coordinates of the growth point, the type of weed, and the growth stage of the weeds to be removed.

[0009] Call the preset weed type-liquid nitrogen spraying parameter mapping table to query the weed type and the corresponding benchmark liquid nitrogen spraying parameters for the growth stage of the weed to be removed. The benchmark liquid nitrogen spraying parameters include the final value of the first spray pressure and the final value of the first spraying speed of the liquid nitrogen spraying actuator.

[0010] The real-time temperature and humidity data of the farmland to be weeded are obtained. The spraying compensation factor of the weeds to be removed is calculated based on the real-time temperature and humidity data. The final value of the first spray pressure and the final value of the first spraying speed are updated based on the spraying compensation factor to determine the final value of the second spray pressure and the final value of the second spraying speed.

[0011] The final value of the second spray pressure and the final value of the second spray speed are input to the liquid nitrogen spraying actuator to drive the liquid nitrogen spraying actuator to perform liquid nitrogen freezing weeding on the weeds in the field to be weeded;

[0012] Repeat the above steps until the weeds in the field to be weeded are removed, so as to complete the control of liquid nitrogen freezing weeding in the field.

[0013] Optionally, based on a farmland weed recognition model trained to convergence, the step of performing target detection on the image of the farmland to be weeded to determine the three-dimensional spatial coordinates of the growth point corresponding to the weed to be removed, the weed species, and the growth stage includes:

[0014] Receive real-time images of farmland to be weeded from the image acquisition system on the mobile platform;

[0015] The trained farmland weed recognition model is invoked, and the farmland weed recognition model is used to perform target detection on the farmland image to be weeded, and output the pixel coordinates of the growth point of the weed to be removed, the weed type and the growth stage of the weed in the farmland image to be weeded.

[0016] Based on the binocular vision stereo matching algorithm, the pixel coordinates of the growth point of the weed to be removed are combined with the camera calibration parameters, the real-time position data of the mobile platform in the farmland world coordinate system, and the real-time attitude data to determine the three-dimensional spatial coordinates of the growth point of the weed to be removed in the farmland world coordinate system.

[0017] Optionally, the step of calling a preset weed type-liquid nitrogen spraying parameter mapping table to query the weed type and corresponding baseline liquid nitrogen spraying parameters for the growth stage of the weeds to be removed includes:

[0018] Obtain the weed species and growth stage corresponding to the weeds to be removed, as well as a preset weed species-liquid nitrogen spraying parameter mapping table. The growth stage is determined based on the weed physiological model and the analysis of weed morphological characteristics, covering the key growth stages within the weed growth cycle. The weed species-liquid nitrogen spraying parameter mapping table is constructed through previous farmland test data and includes the benchmark liquid nitrogen spraying parameters corresponding to different weed species and different growth stages. The benchmark liquid nitrogen spraying parameters include at least the final value of the first spray pressure and the final value of the first spraying speed of the liquid nitrogen spraying actuator.

[0019] For broadleaf weeds with thick cell walls and strong cold resistance, a higher final value of the first spray pressure and a larger final value of the first spray speed are matched to the corresponding weed species and growth stage in the weed species-liquid nitrogen spraying parameter mapping table, so that liquid nitrogen can penetrate the weed leaves and reach the preset freezing depth.

[0020] For tender, narrow-leaved weeds, a lower final value for the first spray pressure and a smaller final value for the first spray speed are matched to the corresponding weed species and growth stage in the weed species-liquid nitrogen spraying parameter mapping table, so as to accurately freeze and kill the weeds and reduce the consumption of liquid nitrogen.

[0021] Optionally, the step of acquiring real-time temperature and humidity data in the farmland to be weeded, and calculating and determining the spraying compensation factor for the weeds to be removed based on the real-time temperature and humidity data, includes:

[0022] Acquire real-time temperature and humidity data of the farmland to be weeded, as well as reference spraying temperature and reference spraying humidity of the weeds to be removed;

[0023] Calculate and determine a first difference between the real-time temperature data and the reference spraying temperature, and calculate and determine a first product between the temperature adjustment coefficient and the first difference;

[0024] Calculate and determine a second difference between the real-time humidity data and the reference spray humidity, and calculate and determine a second product between the humidity adjustment coefficient and the second difference;

[0025] Calculate the first sum among the determined value, the first product, and the second product, and use the first sum as the spraying compensation factor for the weeds to be removed.

[0026] Optionally, the step of updating the first final value of spray pressure and the first final value of spray speed according to the spray compensation factor to determine the second final value of spray pressure and the second final value of spray speed includes:

[0027] Obtain the spraying compensation factor of the weeds to be removed, the final value of the first spraying pressure of the weeds to be removed, and the final value of the first spraying speed;

[0028] The first product between the first final spray pressure and the spraying compensation factor is calculated to determine the second final spray pressure, and the second product between the first final spray speed and the spraying compensation factor is calculated to determine the second final spray speed.

[0029] Optionally, the step of inputting the second final spray pressure value and the second final spray speed value into the liquid nitrogen spraying actuator to drive the liquid nitrogen spraying actuator to perform liquid nitrogen freezing weeding on the weeds in the field to be weeded includes:

[0030] Receive the final value of the second spray pressure and the final value of the second spray speed corresponding to the weeds to be removed from the liquid nitrogen spraying control and decision-making system;

[0031] During the spraying process of the liquid nitrogen spraying actuator nozzle, based on the second final spray pressure value, the pressure of the liquid nitrogen delivery pipeline is adjusted by the pressure regulating valve of the liquid nitrogen storage and supply unit. The pressure sensor built into the pipeline collects the liquid nitrogen pressure in front of the nozzle in real time, compares the collected actual pressure with the second final spray pressure value, and uses a fuzzy PID control algorithm to generate a pressure adjustment signal to dynamically correct the pipeline pressure to the second final spray pressure value.

[0032] The actual spraying speed is collected in real time by a flow sensor. The actual spraying speed is compared with the final value of the second spraying speed. The duty cycle of the pulse width modulation signal is finely adjusted by fuzzy PID control to ensure that the actual spraying speed stably tracks the final value of the second spraying speed.

[0033] By combining the three-dimensional spatial coordinates of the growth point of the weed to be removed output by the weed identification and positioning system, the two-degree-of-freedom precision gimbal of the liquid nitrogen spraying actuator is driven to adjust the nozzle spray angle, so that the liquid nitrogen spray can be directionally covered to cover the growth point of the weed to be removed, thereby completing the timed, quantitative, and pressure-controlled liquid nitrogen freezing weeding operation on the weed to be removed.

[0034] Optionally, the farmland weed identification model includes the YOLO series detection models.

[0035] A liquid nitrogen freezing weed control device for farmland, provided for another purpose of this application, includes:

[0036] The weed identification module is configured to drive the mobile platform to acquire images of farmland containing weeds to be removed, and to perform target detection on the farmland images based on a farmland weed identification model that has been trained to convergence, so as to determine the three-dimensional spatial coordinates of the growth point, the type of weed, and the growth stage of the weeds to be removed.

[0037] The spraying parameter query module is set to call a preset weed type-liquid nitrogen spraying parameter mapping table to query the weed type and growth stage corresponding to the weed to be removed, wherein the benchmark liquid nitrogen spraying parameters include the final value of the first spray pressure and the final value of the first spraying speed of the liquid nitrogen spraying actuator.

[0038] The spraying parameter update module is configured to acquire real-time temperature data and real-time humidity data in the farmland to be weeded, calculate and determine the spraying compensation factor of the weeds to be removed based on the real-time temperature data and real-time humidity data, and update the first final value of spraying pressure and the first final value of spraying speed based on the spraying compensation factor to determine the second final value of spraying pressure and the second final value of spraying speed.

[0039] The liquid nitrogen weeding module is configured to input the final value of the second spray pressure and the final value of the second spray speed to the liquid nitrogen spraying actuator, and drive the liquid nitrogen spraying actuator to perform liquid nitrogen freezing weeding on the weeds in the farmland to be weeded;

[0040] The loop control module is set to repeatedly execute the above steps until the weeds in the farmland to be weeded are removed, so as to complete the liquid nitrogen freezing weeding control of the farmland.

[0041] An electronic device provided for another purpose of this application includes a central processing unit and a memory, the central processing unit being configured to invoke and run a computer program stored in the memory to perform the steps of the liquid nitrogen freezing weed control method for farmland described in this application.

[0042] A computer-readable storage medium is provided for another purpose of this application, which stores, in the form of computer-readable instructions, a computer program implemented according to the farmland liquid nitrogen freezing weed control method, which, when invoked by a computer, executes the steps included in the corresponding method.

[0043] Compared to existing technologies, this application addresses the increasingly prominent problems of accelerated weed population succession, rapid development of herbicide resistance, herbicide soil residues, and crop damage caused by chemical weeding. It also addresses the high precision of thermophysical weeding technology, which is limited by low technological maturity, insufficient operational efficiency, high equipment costs, and poor compatibility with traditional production models. This application offers the following beneficial effects, including but not limited to:

[0044] Firstly, in response to the problems of increased herbicide resistance in weeds, soil residues, and crop damage caused by chemical weeding, this application uses liquid nitrogen cryogenic freezing as its core principle, without the use of any chemical agents, thus preventing the evolution of herbicide resistance in weeds from the root and eliminating chemical pollution of soil and crops. Moreover, after liquid nitrogen vaporizes into nitrogen gas, there is no residue and no damage to the soil ecology, which fully meets the needs of sustainable agricultural development and solves the dual harm of chemical weeding to the environment and crops.

[0045] Secondly, the liquid nitrogen freezing weeding control method for farmland proposed in this application can significantly reduce the rate of seedling damage and improve the accuracy of operations: Compared with the problem that mechanical weeding and root cutting and burying can easily damage the crop roots, this application uses a farmland weed identification model to locate the three-dimensional coordinates of the weed growth point, and combines it with a liquid nitrogen spraying actuator for directional spraying, which only targets the weeds and avoids physical and low-temperature damage to the surrounding crops. It upgrades from indiscriminate operation to precise targeted killing, solving the pain point of high seedling damage rate of mechanical weeding.

[0046] Thirdly, in response to the problems of "high equipment cost, low operating efficiency, and difficulty in adapting to traditional farmland" in electric shock weeding or laser weeding, this application adopts mature industrial Dewar bottle liquid storage and solenoid valve nozzle components, which have lower equipment costs; and relies on a mobile platform for autonomous cyclic operation, adapting to the continuous operation needs of farmland of different sizes, solving the problem that thermophysical weeding is difficult to promote on a large scale.

[0047] Fourth, in response to the difficulty of weed control caused by the large variety of weeds and their different growth stages, this application uses a weed type-spraying parameter mapping table to match benchmark parameters for broadleaf weeds (high pressure and high speed to ensure penetration and freezing), narrowleaf weeds (low pressure and low speed for precise killing), and weeds at different growth stages. In addition, the liquid nitrogen spraying parameters are dynamically optimized by combining spraying compensation factors to avoid indiscriminate spraying that leads to incomplete weed control and to ensure that all types of weeds can be effectively killed.

[0048] Fifth, on the one hand, by optimizing parameters through compensation factors (such as increasing parameters to offset vaporization loss during high-temperature drying and reducing parameters to save liquid nitrogen during low-temperature and high-humidity conditions) and using fuzzy PID control of speed and pressure to achieve quantitative spraying, liquid nitrogen waste is reduced; on the other hand, the mobile platform's autonomous cyclic operation reduces manual intervention, lowers labor costs, and solves the problem of high input costs in traditional weeding.

[0049] Sixth, considering the scattered distribution of weeds in farmland and the need for batch processing, this application adopts a design that executes the process cyclically until the weeds are completely removed, avoiding omissions in a single operation; and in each cycle, images are re-acquired and weeds are identified. If the first spraying does not completely eliminate the weeds (such as insufficient parameter compensation), a second spraying can be triggered to ensure that the weeds in the entire farmland are completely removed, solving the problems of incomplete coverage and repeated operations required by traditional weeding methods, and improving operational efficiency and effectiveness. Attached Figure Description

[0050] 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:

[0051] Figure 1 This is an exemplary network architecture diagram used in the farmland liquid nitrogen freezing weeding control system in the embodiments of this application;

[0052] Figure 2 This is a flowchart illustrating the liquid nitrogen freezing weed control method for farmland in the embodiments of this application;

[0053] Figure 3 This is a schematic diagram of the liquid nitrogen freezing weeding control device for farmland in the embodiments of this application;

[0054] Figure 4 This is a schematic diagram of the structure of the computer device in the embodiments of this application. Detailed Implementation

[0055] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0056] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this application means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.

[0057] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0058] Those skilled in the art will understand that although the various methods in this application are described based on the same concept and thus present commonality among them, they can be performed independently unless otherwise specified. Similarly, the various embodiments disclosed in this application are all based on the same inventive concept; therefore, concepts expressed in the same way, as well as concepts that are appropriately changed for convenience but are expressed differently, should be understood equivalently.

[0059] Unless otherwise expressly stated, the various embodiments disclosed in this application can be combined in a cross-cutting manner to flexibly construct new embodiments, as long as such combination does not depart from the inventive spirit of this application and can meet the needs of the prior art or solve a certain deficiency in the prior art. Those skilled in the art should be aware of such modifications.

[0060] Please see Figure 1 The liquid nitrogen freezing weed control method for farmland in this application can be based on, for example... Figure 1 The farmland liquid nitrogen freezing weeding control system shown is implemented. This system includes a mobile platform, an image acquisition system, a weed identification and positioning system, a liquid nitrogen storage and supply unit, a liquid nitrogen spraying control and decision-making system, a central control and communication module, and an energy and safety system, etc.

[0061] In some embodiments, the mobile platform 1 comprises a chassis, a power unit, and a navigation module. The mobile platform carries these components and moves along the rows of crops, accurately positioning the spray nozzles above the target weeds. The navigation module utilizes a combination of BeiDou navigation / RTK and visual guidance to achieve field positioning. During movement, the mobile platform works in conjunction with the identification and spraying systems: after the platform stops at a designated location, the control system triggers liquid nitrogen spraying; after spraying, it continues to the next target. The mobile platform is a tracked or large-diameter four-wheeled electric vehicle, providing excellent field mobility. It is equipped with a high-precision BeiDou navigation receiver and an inertial measurement unit (IMU), forming a positioning and navigation module capable of centimeter-level accuracy in autonomous path tracking and recording of work points.

[0062] In some embodiments, the image acquisition system 2 is installed on the mobile platform for real-time acquisition of images of farmland scenes. The image acquisition system 2 consists of two high-resolution global industrial color cameras, mounted at a certain angle on a shock-resistant gimbal at the front of the platform. It can perceive the rows of crops in front in three dimensions, with a field of view covering at least two rows of crops, which helps in the subsequent three-dimensional positioning of weeds.

[0063] In some embodiments, the weed identification and localization system includes an image processing algorithm and a classification model. The image acquisition system is responsible for acquiring real-time images of crops and weeds in the farmland and sending the image data to the next-level module. The identification and localization system analyzes and processes the images, identifies different weed species, and determines their growth point coordinates. The identification algorithm can employ deep learning or traditional machine vision methods to classify common field weeds such as grasses and broadleaf weeds. The identification results include weed category and spatial location, providing a basis for spraying decisions. The core of the weed identification and localization system 3 is an embedded industrial computer equipped with a GPU. Its built-in deep learning model is based on the YOLOv8 architecture and trained on a large dataset of labeled farmland weed images. It can identify and distinguish crops from at least ten common weeds (such as barnyard grass, crabgrass, amaranth, and goosegrass) in real time. The system uses a binocular vision stereo matching algorithm, combined with camera calibration parameters and platform pose, to convert the identified weed growth points from image pixel coordinates to three-dimensional coordinates (X, Y, Z) in the field world coordinate system.

[0064] In some embodiments, the liquid nitrogen storage and supply unit includes a liquid nitrogen tank, a pressure regulating valve, and insulated delivery pipelines for storing and supplying liquid nitrogen. The core of the liquid nitrogen storage and supply unit is a standard 60-liter industrial Dewar flask with a high-vacuum, multi-layered insulated structure that effectively reduces liquid nitrogen evaporation loss. The unit is equipped with a precision pressure regulating valve and a pressure sensor to maintain stable pressure in the delivery pipelines. All pipelines use vacuum-insulated flexible hoses to minimize liquid nitrogen vaporization during delivery.

[0065] In some embodiments, the liquid nitrogen precision spraying actuator 6 is mounted on the mobile platform and connected to the liquid nitrogen spraying control and decision system. It receives control commands and, according to the commands, performs directional, quantitative, and constant-pressure liquid nitrogen spraying on the identified weed growth points. For community weeds, continuous or zoned spraying can be selected to cover all targets; for single weeds, a point-to-point impact spraying method is used. The liquid nitrogen precision spraying actuator 6 includes a two-degree-of-freedom precision gimbal and a high-speed response solenoid valve (response time <10ms) and a fan-shaped atomizing nozzle mounted on it. The gimbal can adjust the spray angle according to the precise position of the weeds to ensure precise coverage of the growth points by the liquid nitrogen spray. The solenoid valve is controlled by a PWM signal from the liquid nitrogen spraying control and decision system 5 to achieve precise control of the nozzle opening duration and effective opening degree. The specific control flow is as follows: The system generates a control signal based on the optimized spraying parameters. The pressure regulating valve maintains the pipeline pressure based on a fuzzy PID algorithm. The solenoid valve is controlled by a PWM signal, where the duty cycle is mapped to the target opening degree according to the optimized spraying parameters, and the duration is set to t. Under the coordination of the central control and communication module 7, the actuator precisely initiates spraying when the mobile platform positions the nozzle above the weed growth point, ensuring that liquid nitrogen covers the target area.

[0066] In some embodiments, the central control and communication module coordinates data interaction and control between various functional units, including the image acquisition system, weed identification and positioning system, liquid nitrogen spraying control system, and mobile platform. The central control and communication module 7, acting as the system's overall coordination center, employs a high-performance industrial PLC or embedded industrial computer. It receives position and attitude information from the mobile platform via a CAN bus, receives image and weed positioning data via gigabit Ethernet, and interacts in real-time with the liquid nitrogen spraying control and decision-making system 5 using high-speed I / O ports. This module is responsible for the logical scheduling of the entire workflow, ensuring that the series of actions from image acquisition to liquid nitrogen spraying are executed precisely in the correct timing.

[0067] In some embodiments, the energy and safety system includes a power supply, a battery, and safety protection devices. The liquid nitrogen system must be equipped with insulation and pressure relief safety devices to prevent liquid nitrogen leakage or overpressure, ensuring operational safety. It also includes a power supply to power the control system, cameras, and valve control equipment. The energy and safety system 8 includes a 48V lithium battery pack and a DC-DC power conversion module to power all system equipment. For safety, the liquid nitrogen Dewar flask is equipped with dual redundant pressure sensors, temperature sensors, and a mechanical safety valve. The central control module monitors these sensor data in real time. Once an abnormal increase in pressure or suspected leakage (such as a sudden drop in temperature) is detected, the main supply valve will be immediately shut off, and then an audible and visual alarm will be triggered to ensure field operation safety.

[0068] In a further embodiment, the mobile platform is an autonomous navigation vehicle or a tractor-mounted platform, which integrates a positioning and navigation module for walking along a preset path and recording the work position.

[0069] In a further embodiment, the image acquisition system includes at least one high-resolution color camera or depth camera, facing the crop rows at a certain downward angle to ensure that the morphological characteristics of crops and weeds can be clearly captured.

[0070] In a further embodiment, the weed identification and positioning system has a built-in image recognition model based on deep learning. This model is trained to distinguish between crops and various common weeds and can output the pixel coordinates of the weeds in the image. Combined with the camera calibration parameters and the positioning information of the mobile platform, the three-dimensional coordinates of the weed growth point in the farmland world coordinate system are calculated.

[0071] In a further embodiment, the liquid nitrogen spraying control and decision-making system pre-stores a mapping table of weed species and liquid nitrogen spraying parameters. For broadleaf weeds with thick cell walls and strong cold resistance, the mapping table corresponds to a higher final spray pressure and a larger final spraying speed to ensure that liquid nitrogen can penetrate the leaves and reach sufficient freezing depth. For tender, narrow-leaf weeds, the mapping table corresponds to a lower final spray pressure and a lower final spraying speed to achieve precise targeting and save liquid nitrogen consumption. Furthermore, the liquid nitrogen spraying control and decision-making system can identify and adapt to different growth stages of weeds, achieving a further leap from "indiscriminate coverage" to "precise targeted eradication." Through multi-source data fusion and a built-in weed physiological model, the system divides the weed growth cycle into several key stages and formulates an optimal liquid nitrogen spraying strategy for each stage, aiming to maximize weed control efficiency and resource utilization efficiency.

[0072] In a further embodiment, the liquid nitrogen precision spraying actuator includes one or more independently controllable solenoid valve nozzles, and the liquid nitrogen spraying control and decision-making system controls the nozzle opening duration, opening degree, and liquid nitrogen pressure through the following specific processing steps:

[0073] Parameter mapping and target value setting: Based on weed species, growth stage, and environmental compensation results, the system queries the target spray pressure, target spray speed, and target on-time from the pre-stored decision model. The spray speed is positively correlated with the nozzle opening; a mapping relationship is established through calibration experiments.

[0074] Pressure regulation: The system controls the pressure in the liquid nitrogen supply pipeline via a pressure regulating valve. A pressure sensor monitors the liquid nitrogen pressure in front of the nozzle in real time and compares it with the target spray pressure. A fuzzy PID control algorithm is used to generate the pressure regulation signal.

[0075] Nozzle Control: The system controls the opening duration and degree of the solenoid valve using a pulse width modulation (PWM) signal. First, the target opening degree is calculated based on the target spraying speed and nozzle characteristic curve. Then, during spraying, the system generates a PWM signal with a duty cycle proportional to the target opening degree and a duration set to t. Simultaneously, the flow sensor provides real-time feedback on the actual spraying speed, and the system fine-tunes the PWM signal's duty cycle using fuzzy PID control to ensure the actual spraying speed tracks the target spraying speed.

[0076] In a further embodiment, the central control and communication module uses an industrial-grade microprocessor or programmable logic controller (PLC) as its core, integrating multiple communication interfaces (such as CAN, Ethernet, RS485, etc.). It is responsible for receiving real-time position and attitude data of the mobile platform from the positioning and navigation module, raw image data sent by the image acquisition system, and the three-dimensional coordinates of the weeds calculated by the weed identification and positioning system. Based on preset operational logic, this module sends control commands containing weed coordinates and spraying parameters to the liquid nitrogen spraying control and decision-making system, and synchronously coordinates the movement, start, and stop of the mobile platform to ensure precise data synchronization and action timing between the various systems.

[0077] In a further embodiment, the energy and safety system includes an on-board battery pack or generator as the main power source, providing stable power to the control system, sensors, and actuators via a voltage conversion module. The liquid nitrogen storage and supply unit uses a high-vacuum multi-layer insulated Dewar flask for insulation and integrates pressure sensors, temperature sensors, a safety relief valve, and a liquid nitrogen level monitoring device. When the system detects an abnormal increase in pipeline pressure or a risk of leakage, the control module will immediately cut off the liquid nitrogen supply and trigger an audible and visual alarm to ensure the safety and reliability of field operations.

[0078] In a further embodiment, the system also includes an environmental monitoring module for real-time monitoring of ambient temperature and humidity; the liquid nitrogen spraying control and decision-making system can compensate and optimize the liquid nitrogen spraying parameters according to the ambient temperature and humidity, for example, appropriately adjusting the pressure to increase the liquid nitrogen spraying volume in a high-temperature and dry environment to compensate for vaporization loss.

[0079] Based on the above exemplary scenario, please refer to Figure 2 In one embodiment of the liquid nitrogen freezing weed control method for farmland of this application, the method includes:

[0080] Step S10: Drive the mobile platform to acquire an image of farmland containing weeds to be removed. Based on the farmland weed recognition model that has been trained to a convergent state, perform target detection on the image of the farmland to be removed to determine the three-dimensional spatial coordinates of the growth point, the type of weed, and the growth stage of the weeds to be removed.

[0081] The liquid nitrogen freezing weed control system in the terminal device can drive a mobile platform to acquire images of farmland containing weeds to be removed. Based on a farmland weed recognition model that has been trained to a convergent state, the system performs target detection on the farmland images to determine the three-dimensional spatial coordinates of the growth point, the weed species, and the growth stage of the weeds to be removed. The farmland weed recognition model includes the YOLO series detection model.

[0082] In some embodiments, the step of performing target detection on the image of the farmland to be weeded based on a farmland weed recognition model trained to a convergent state, in order to determine the three-dimensional spatial coordinates of the growth point corresponding to the weed to be removed, the weed species, and the growth stage, includes:

[0083] Step S101: Receive real-time images of the farmland to be weeded from the image acquisition system on the mobile platform;

[0084] Step S102: Call the farmland weed recognition model that has been trained to convergence, and use the farmland weed recognition model to perform target detection on the farmland image to be weeded, and output the pixel coordinates of the growth point of the weed to be removed, the weed type and the growth stage of the weed in the farmland image to be weeded.

[0085] Step S103: Based on the binocular vision stereo matching algorithm, the pixel coordinates of the growth point of the weed to be removed are combined with the camera calibration parameters, the real-time position data of the mobile platform in the farmland world coordinate system, and the real-time attitude data to determine the three-dimensional spatial coordinates of the growth point of the weed to be removed in the farmland world coordinate system.

[0086] As can be seen from steps S101 to S103 above, receiving real-time images from the mobile platform image acquisition system can cover the entire farmland to be weeded by leveraging the platform's mobility. Furthermore, the image acquisition system (such as a high-resolution camera) can clearly capture the morphological characteristics of crops and weeds, providing high-quality raw data for subsequent identification and avoiding missed weed detections due to image blur or limited acquisition range, thus ensuring the integrity and accuracy of the basic identification data. Calling the converged farmland weed identification model to detect images serves two purposes: firstly, the converged model, trained on a large amount of data, can accurately distinguish between crops and various weeds, and the output weed species and growth stages provide a basis for matching liquid nitrogen spraying parameters, avoiding indiscriminate spraying; secondly, directly outputting the pixel coordinates of growth points lays the foundation for converting to three-dimensional spatial coordinates, reducing subsequent data processing steps and improving identification efficiency. By combining binocular vision stereo matching algorithm with multi-dimensional data to determine three-dimensional coordinates, and correcting image distortion through camera calibration parameters, the pixel coordinates are converted into absolute field coordinates using the real-time position and attitude data of the mobile platform in the farmland world coordinate system. This eliminates positioning errors caused by platform tilt and camera angle deviation, ensuring accurate positioning of weed growth points and providing precise spatial guidance for subsequent liquid nitrogen directional spraying, thus avoiding accidental spraying of crops or missed spraying of weeds.

[0087] Step S20: Call the preset weed type-liquid nitrogen spraying parameter mapping table to query the weed type and the corresponding benchmark liquid nitrogen spraying parameters for the growth stage of the weed to be removed. The benchmark liquid nitrogen spraying parameters include the final value of the first spray pressure and the final value of the first spraying speed of the liquid nitrogen spraying actuator.

[0088] The mobile platform is driven to acquire images of farmland containing weeds to be removed. Based on a farmland weed recognition model that has been trained to a convergent state, the target detection is performed on the images of the farmland to be removed to determine the three-dimensional spatial coordinates of the growth point, the type of weed, and the growth stage of the weeds to be removed.

[0089] In some embodiments, the step of calling a preset weed species-liquid nitrogen spraying parameter mapping table to query the weed species to be removed and the corresponding baseline liquid nitrogen spraying parameters for their growth stage includes:

[0090] Step S201: Obtain the weed species and growth stage corresponding to the weeds to be removed, as well as a preset weed species-liquid nitrogen spraying parameter mapping table. The growth stage is determined based on the weed physiological model and the analysis of weed morphological characteristics, covering the key growth stages within the weed growth cycle. The weed species-liquid nitrogen spraying parameter mapping table is constructed through previous farmland test data and includes the benchmark liquid nitrogen spraying parameters corresponding to different weed species and different growth stages. The benchmark liquid nitrogen spraying parameters include at least the final value of the first spray pressure and the final value of the first spraying speed of the liquid nitrogen spraying actuator.

[0091] Step S202: For broadleaf weeds with thicker cell walls and stronger cold resistance, match the corresponding weed species and growth stage with a higher final value of the first spray pressure and a larger final value of the first spray speed in the weed species-liquid nitrogen spraying parameter mapping table, so that the liquid nitrogen can penetrate the weed leaves and reach the preset freezing depth.

[0092] Step S203: For tender and narrow-leaved weeds, match the corresponding weed species and growth stage with a lower final value of the first spray pressure and a smaller final value of the first spray speed in the weed species-liquid nitrogen spraying parameter mapping table, so as to accurately freeze and kill the weeds and reduce the amount of liquid nitrogen consumed.

[0093] As can be seen from steps S201 to S203 above, obtaining the weed species-liquid nitrogen spraying parameter mapping table serves two purposes. First, the growth stage is determined based on the weed physiological model and morphological characteristics, ensuring that the judgment of weed growth status conforms to actual physiological laws and providing an accurate basis for parameter matching. Second, the mapping table is constructed from previous field trial data, avoiding the disconnect between theoretical parameters and actual operations, and clearly includes the final values ​​of pressure and velocity, providing a standardized benchmark for subsequent differentiated spraying, ensuring the scientific validity and reliability of parameter queries, and avoiding blindly setting parameters. Higher liquid nitrogen spraying parameters are matched for broadleaf weeds. Considering their thick cell walls and strong cold resistance, increasing pressure and velocity ensures that liquid nitrogen penetrates the leaves and reaches sufficient freezing depth, avoiding incomplete weeding due to insufficient parameters, ensuring the killing effect on stubborn weeds, and reducing the cost and time of secondary weeding. Lower liquid nitrogen spraying parameters are matched for narrow-leaved weeds, taking into account their delicate and easily killed characteristics. This achieves precise freezing and killing while reducing liquid nitrogen consumption and lowering operating costs. It also avoids accidental freezing damage to surrounding crops caused by high-parameter spraying, balancing weeding efficiency with resource conservation and crop protection, which meets the needs of green agriculture and precision agriculture.

[0094] Step S30: Obtain real-time temperature data and real-time humidity data in the farmland to be weeded; calculate and determine the spraying compensation factor for the weeds to be removed based on the real-time temperature data and real-time humidity data; update the first final value of spraying pressure and the first final value of spraying speed based on the spraying compensation factor to determine the second final value of spraying pressure and the second final value of spraying speed.

[0095] The system calls a preset weed type-liquid nitrogen spraying parameter mapping table to query the weed type and growth stage corresponding to the benchmark liquid nitrogen spraying parameters. The benchmark liquid nitrogen spraying parameters include the first final value of the liquid nitrogen spraying actuator (benchmark final value of the spraying pressure) and the first final value of the spraying speed (benchmark final value of the spraying speed). Then, the system obtains the real-time temperature data and real-time humidity data of the farmland to be weeded. The system calculates and determines the spraying compensation factor for the weed to be removed based on the real-time temperature data and real-time humidity data. The system updates the first final value of the spraying pressure and the first final value of the spraying speed based on the spraying compensation factor to determine the second final value of the spraying pressure and the second final value of the spraying speed.

[0096] In some embodiments, the step of acquiring real-time temperature data and real-time humidity data in the farmland to be weeded, and calculating and determining the spraying compensation factor for the weeds to be removed based on the real-time temperature data and real-time humidity data, includes:

[0097] Step S301: Obtain real-time temperature and humidity data of the farmland to be weeded, as well as reference spraying temperature and reference spraying humidity of the weeds to be removed;

[0098] Step S302: Calculate and determine the first difference between the real-time temperature data and the reference spraying temperature, and calculate and determine the first product between the temperature adjustment coefficient and the first difference;

[0099] Step S303: Calculate and determine the second difference between the real-time humidity data and the reference spraying humidity, and calculate and determine the second product between the humidity adjustment coefficient and the second difference;

[0100] Step S304: Calculate the first sum between the determined value, the first product, and the second product, and use the first sum as the spraying compensation factor for the weeds to be removed.

[0101] Specifically, the calculation formula for the spraying compensation factor of the weeds to be removed is expressed as follows:

[0102] K = 1 + α × ((TT) ref )+β×(HH ref ),

[0103] Where K represents the spraying compensation factor for the weeds to be removed; α represents the temperature adjustment coefficient, used to quantify the effect of temperature deviation on liquid nitrogen vaporization; T represents the real-time temperature data of the farmland to be weeded; T ref The reference spraying temperature is indicated by β; β represents the humidity adjustment coefficient, used to quantify the effect of humidity deviation on liquid nitrogen vaporization; H represents the real-time humidity data of the farmland to be weeded; H ref This indicates the reference spraying humidity for removing weeds.

[0104] Acquiring real-time temperature and humidity data from the farmland to be weeded serves two purposes. First, real-time data reflects the dynamic field environment (e.g., high temperatures at midday, high humidity in the early morning), ensuring that compensation factors align with current operating conditions. Second, reference temperature and humidity provide a benchmark for calculations, avoiding compensation deviations due to a lack of standardized data. The combination of these two factors lays a data foundation for subsequent accurate calculations, ensuring the objectivity of the compensation logic. Calculating temperature-related differences and products quantifies the impact of temperature deviations on liquid nitrogen (e.g., high temperatures accelerate liquid nitrogen vaporization) using a temperature adjustment coefficient. This transforms abstract temperature differences into calculable compensation terms, ensuring that the influence of temperature on spraying parameters is accurately quantified, avoiding parameter inaccuracies caused by adjustments based solely on experience. Similarly, calculating humidity-related differences and products quantifies the impact of humidity deviations (e.g., low humidity exacerbates liquid nitrogen evaporation) using a humidity adjustment coefficient, supplementing the compensation basis for the humidity dimension. This achieves full coverage of both temperature and humidity factors, avoiding incomplete compensation due to considering only temperature while ignoring humidity, and improving the accuracy of compensation factors. The compensation factor is obtained by summing, which integrates the effects of temperature and humidity into a unified coefficient. It can then be directly used to correct spraying parameters (such as pressure and speed), simplifying the parameter adjustment process. Moreover, the compensation factor is based on 1 (1 when there is no deviation), which intuitively reflects the degree of influence of the environment on spraying, making it easier for the system to respond quickly to environmental changes and ensuring the stability of liquid nitrogen spraying effect.

[0105] In a specific embodiment, the liquid nitrogen spraying control and decision-making system 5 is communicatively connected to both the weed identification and positioning system and the liquid nitrogen storage and supply unit. It receives information on the type and location of weeds and generates control commands based on a pre-stored decision model. This decision model defines the optimal final values ​​for spray pressure and spray speed for different weed species. The liquid nitrogen spraying control and decision-making system 5 receives information on weed type, location, and growth stage from the weed identification and positioning system 3, as well as temperature and humidity data from the environmental monitoring module. The system internally stores a detailed "weed-liquid nitrogen spraying parameter mapping table" and a "weed physiological model." Different nozzle specifications can be selected based on different weed species, growth cycles, and spraying heights. For example, experiments have shown that spraying liquid nitrogen at a nozzle height of approximately 10 to 30 centimeters above the ground, at a high pressure (approximately 0.041 MPa), and a spraying duration of 100 ms can effectively suppress crabgrass. An environmental temperature and humidity compensation algorithm calculates the compensation factor K in real time and dynamically adjusts the baseline parameters. For example, under high temperature and dry conditions in summer (T=35℃, H=30%), assuming α=0.02 / ℃ and β=0.005 / %, then K=1+0.02×(35-20)+0.005×(50-30)=1.4, the system will increase the spray pressure and spraying speed by 40% to offset the loss of liquid nitrogen vaporization.

[0106] In a further embodiment, the step of updating the first final value of spray pressure and the first final value of spray speed according to the spray compensation factor to determine the second final value of spray pressure and the second final value of spray speed includes:

[0107] Step S3001: Obtain the spraying compensation factor of the weeds to be removed, the final value of the first spraying pressure of the weeds to be removed, and the final value of the first spraying speed.

[0108] Step S3002: Calculate and determine the first product between the first final value of the spray pressure and the spraying compensation factor to determine the second final value of the spray pressure, and calculate and determine the second product between the first final value of the spraying speed and the spraying compensation factor to determine the second final value of the spraying speed.

[0109] As can be seen from steps S3001 to S3002 above, the compensation factor and the initial pressure and speed final values ​​are obtained. The compensation factor is then associated with the previously determined first spray pressure final value and first spray speed final value to form a connection logic between basic parameters and environmental compensation. This ensures that parameter updates do not deviate from the characteristics of the weeds themselves (such as the initial parameter differences between broadleaf and narrowleaf weeds), laying the foundation for precise adjustments. The second spray pressure final value is determined by calculating the first product between the first spray pressure final value and the spray compensation factor, and the second spray speed final value is determined by calculating the second product between the first spray speed final value and the spray compensation factor. The operation is simple and the logic is intuitive. It can quickly convert the influence of environmental temperature and humidity on liquid nitrogen spraying (such as vaporization loss caused by high temperature) into specific pressure and speed adjustment values. When the spray compensation factor > 1 (high temperature and dryness), the pressure and speed are increased to offset the loss; when the spray compensation factor < 1 (low temperature and high humidity), the parameters are reduced to save liquid nitrogen. This quantitative update method avoids the errors of empirical adjustments, ensuring that the second parameter is both adapted to the characteristics of weeds and fits the real-time environment, taking into account both weed control effect and resource efficiency, and ensuring that liquid nitrogen spraying is accurate and controllable under different environments.

[0110] Step S40: Input the final value of the second spray pressure and the final value of the second spray speed into the liquid nitrogen spraying actuator, and drive the liquid nitrogen spraying actuator to perform liquid nitrogen freezing weeding on the weeds in the farmland to be weeded;

[0111] The system acquires real-time temperature and humidity data of the farmland to be weeded, calculates and determines the spraying compensation factor for the weeds to be removed based on the real-time temperature and humidity data, updates the first final value of the spraying pressure and the first final value of the spraying speed based on the spraying compensation factor, and after determining the second final value of the spraying pressure and the second final value of the spraying speed, the second final value of the spraying pressure and the second final value of the spraying speed are input to the liquid nitrogen spraying actuator to drive the liquid nitrogen spraying actuator to perform liquid nitrogen freezing weeding on the weeds in the farmland to be weeded.

[0112] In some embodiments, the step of inputting the second final spray pressure value and the second final spray speed value to the liquid nitrogen spraying actuator, and driving the liquid nitrogen spraying actuator to perform liquid nitrogen cryogenic weeding on the weeds in the field to be weeded, includes:

[0113] Step S401: Receive the final value of the second spray pressure and the final value of the second spray speed corresponding to the weeds to be removed from the liquid nitrogen spraying control and decision-making system.

[0114] Step S402: During the spraying process of the nozzle of the liquid nitrogen spraying actuator, based on the final value of the second spray pressure, the pressure of the liquid nitrogen delivery pipeline is adjusted by the pressure regulating valve of the liquid nitrogen storage and supply unit. The pressure sensor built into the pipeline is used to collect the liquid nitrogen pressure in front of the nozzle in real time. The collected actual pressure is compared with the final value of the second spray pressure. A pressure adjustment signal is generated by using a fuzzy PID control algorithm to dynamically correct the pipeline pressure to the final value of the second spray pressure.

[0115] Step S403: Use a flow sensor to collect the actual spraying speed in real time, compare the actual spraying speed with the final value of the second spraying speed, and fine-tune the duty cycle of the pulse width modulation signal through fuzzy PID control to ensure that the actual spraying speed stably tracks the final value of the second spraying speed.

[0116] Step S404: Combining the three-dimensional spatial coordinates of the growth point of the weed to be removed output by the weed identification and positioning system, drive the two-degree-of-freedom precision gimbal of the liquid nitrogen spraying actuator to adjust the nozzle spray angle so that the liquid nitrogen spray can be directionally covered to cover the growth point of the weed to be removed, so as to complete the timed, quantitative, and constant-pressure liquid nitrogen freezing weeding operation on the weed to be removed.

[0117] As shown in steps S401 to S404, by dynamically adjusting the pressure using fuzzy PID control and combining it with real-time feedback from the pressure sensor, pipeline pressure fluctuations (such as pressure drops caused by liquid nitrogen vaporization) can be quickly corrected. This ensures that the actual pressure stably matches the second final value, preventing incomplete weeding due to insufficient pressure or waste of liquid nitrogen due to excessive pressure, thus achieving precise constant-pressure control. By finely adjusting the duty cycle of the pulse width modulation signal (PWM) using fuzzy PID control and combining it with real-time monitoring by the flow sensor, spraying speed deviations (such as speed changes caused by nozzle wear) can be dynamically corrected, ensuring that the actual speed stably tracks the second final value. This precisely controls the amount of liquid nitrogen sprayed, achieving the "quantitative" target while balancing weeding effectiveness and resource conservation. By adjusting the gimbal angle using three-dimensional coordinates, liquid nitrogen can be directed to cover weed growth points, avoiding accidental spraying of crops or missed weeds, thus achieving directional operation. Simultaneously, by integrating constant-pressure, quantitative, and directional control, timed (matching spraying duration) weeding can be completed, ensuring precise and efficient operation, meeting the requirements of green agriculture for low damage and high precision.

[0118] Step S50: Repeat the above steps until the weeds in the farmland to be weeded are removed, so as to complete the control of liquid nitrogen freezing weeding in the farmland.

[0119] The above steps S10 to S40 are repeated until the weeds in the farmland to be weeded are completely removed, thus completing the liquid nitrogen freezing weeding control of the farmland. Specifically, the weeds in the farmland are not distributed in single points, scattered or clustered, but rather concentrated in a certain area. A single execution of the image acquisition, recognition, parameter calculation, and spraying process cannot cover the entire farmland. Therefore, it is necessary to cycle the mobile platform in units of "movement-operation": after the mobile platform travels a certain distance along the preset path, it repeatedly executes "acquiring the current area image → identifying weeds → calculating parameters → targeted spraying" until all areas of the farmland to be weeded have been traversed, avoiding omissions due to the scattered distribution of weeds and ensuring weeding coverage. The termination criterion is not simply "the mobile platform has traversed the path", but rather dynamically determined based on the weed recognition results: after each cycle of spraying, the image acquisition system will re-acquire the image of the area, and the weed recognition model will detect whether there are still effective weeds (such as weed targets with confidence ≥ preset threshold). If no effective weeds are detected in a certain area for 2 to 3 consecutive times, the weeding in that area is considered complete. When the mobile platform has traversed all the preset paths and all areas meet the condition of "no effective weeds", the weeding in the farmland to be weeded is considered complete, thus avoiding incomplete operation due to "the path has been completed but some weeds have not been removed".

[0120] On the one hand, the weeding effect is ensured through cyclic verification. If the weeds are not completely eradicated after the first spraying in a certain area (such as insufficient parameter compensation under low temperature conditions), the model will identify and trigger a second spraying in subsequent cycles until the weeds are cleared. On the other hand, the energy and safety system will continuously monitor the liquid nitrogen storage and equipment status during the cycle. If problems such as insufficient liquid nitrogen or abnormal pressure occur during the cycle, the system will pause the cycle and sound an alarm. It will restart after the fault is cleared, taking into account both the continuity of operation and safety, and ultimately achieving "full-area, thorough and safe" liquid nitrogen freezing weeding control in farmland.

[0121] As can be seen from the above embodiments, compared with the prior art, this application addresses the increasingly prominent problems of accelerated weed population succession, rapid development of herbicide resistance, herbicide soil residues and crop damage caused by chemical weeding, as well as the high precision of thermophysical weeding technology. However, it is limited by low technological maturity, insufficient operational efficiency, high equipment costs, and poor compatibility with traditional production models. This application includes, but is not limited to, the following beneficial effects:

[0122] Firstly, in response to the problems of increased herbicide resistance in weeds, soil residues, and crop damage caused by chemical weeding, this application uses liquid nitrogen cryogenic freezing as its core principle, without the use of any chemical agents, thus preventing the evolution of herbicide resistance in weeds from the root and eliminating chemical pollution of soil and crops. Moreover, after liquid nitrogen vaporizes into nitrogen gas, there is no residue and no damage to the soil ecology, which fully meets the needs of sustainable agricultural development and solves the dual harm of chemical weeding to the environment and crops.

[0123] Secondly, the liquid nitrogen freezing weeding control method for farmland proposed in this application can significantly reduce the rate of seedling damage and improve the accuracy of operations: Compared with the problem that mechanical weeding and root cutting and burying can easily damage the crop roots, this application uses a farmland weed identification model to locate the three-dimensional coordinates of the weed growth point, and combines it with a liquid nitrogen spraying actuator for directional spraying, which only targets the weeds and avoids physical and low-temperature damage to the surrounding crops. It upgrades from indiscriminate operation to precise targeted killing, solving the pain point of high seedling damage rate of mechanical weeding.

[0124] Thirdly, in response to the problems of "high equipment cost, low operating efficiency, and difficulty in adapting to traditional farmland" in electric shock weeding or laser weeding, this application adopts mature industrial Dewar bottle liquid storage and solenoid valve nozzle components, which have lower equipment costs; and relies on a mobile platform for autonomous cyclic operation, adapting to the continuous operation needs of farmland of different sizes, solving the problem that thermophysical weeding is difficult to promote on a large scale.

[0125] Fourth, in response to the difficulty of weed control caused by the large variety of weeds and their different growth stages, this application uses a weed type-spraying parameter mapping table to match benchmark parameters for broadleaf weeds (high pressure and high speed to ensure penetration and freezing), narrowleaf weeds (low pressure and low speed for precise killing), and weeds at different growth stages. In addition, the liquid nitrogen spraying parameters are dynamically optimized by combining spraying compensation factors to avoid indiscriminate spraying that leads to incomplete weed control and to ensure that all types of weeds can be effectively killed.

[0126] Fifth, on the one hand, by optimizing parameters through compensation factors (such as increasing parameters to offset vaporization loss during high-temperature drying and reducing parameters to save liquid nitrogen during low-temperature and high-humidity conditions) and using fuzzy PID control of speed and pressure to achieve quantitative spraying, liquid nitrogen waste is reduced; on the other hand, the mobile platform's autonomous cyclic operation reduces manual intervention, lowers labor costs, and solves the problem of high input costs in traditional weeding.

[0127] Sixth, considering the scattered distribution of weeds in farmland and the need for batch processing, this application adopts a design that executes the process cyclically until the weeds are completely removed, avoiding omissions in a single operation; and in each cycle, images are re-acquired and weeds are identified. If the first spraying does not completely eliminate the weeds (such as insufficient parameter compensation), a second spraying can be triggered to ensure that the weeds in the entire farmland are completely removed, solving the problems of incomplete coverage and repeated operations required by traditional weeding methods, and improving operational efficiency and effectiveness.

[0128] Please see Figure 3This application provides a liquid nitrogen freezing weed control device for farmland, comprising a weed identification module 1100, a spraying parameter query module 1200, a spraying parameter update module 1300, and a liquid nitrogen weeding module 1400. The weed identification module 1100 is configured to drive a mobile platform to acquire an image of farmland containing weeds to be removed, and to perform target detection on the image of the farmland containing weeds to be removed based on a farmland weed identification model trained to convergence, to determine the three-dimensional spatial coordinates of the growth point corresponding to the weeds to be removed, the weed species, and the growth stage. The spraying parameter query module 1200 is configured to call a preset weed species-liquid nitrogen spraying parameter mapping table to query the reference liquid nitrogen spraying parameters corresponding to the weed species and growth stage of the weeds to be removed, wherein the reference liquid nitrogen spraying parameters include the final value of the first spray pressure and the final value of the first spraying speed of the liquid nitrogen spraying actuator. The spraying parameter update module 1300 is configured to acquire the weed species of the farmland containing weeds to be removed. The system uses real-time temperature and humidity data from the field. Based on this data, a spraying compensation factor for the weeds to be removed is calculated. The final values ​​of the first spray pressure and the first spraying speed are updated based on this compensation factor to determine the final values ​​of the second spray pressure and the second spraying speed. A liquid nitrogen weeding module 1400 is configured to input the final values ​​of the second spray pressure and the second spraying speed to the liquid nitrogen spraying actuator, driving the actuator to perform liquid nitrogen freezing weeding on the weeds in the field. A cycle control module 1500 is configured to repeatedly execute the above steps until the weeds in the field are completely removed, thus completing the liquid nitrogen freezing weeding control of the farmland.

[0129] Based on any embodiment of this application, please refer to Figure 4 Another embodiment of this application also provides an electronic device, which can be implemented by a computer device, such as... Figure 4 The diagram shows the internal structure of a computer device. The computer device includes a processor, a computer-readable storage medium, a memory, and a network interface connected via a system bus. The computer-readable storage medium stores an operating system, a database, and computer-readable instructions. The database may store control information sequences. When the computer-readable instructions are executed by the processor, they enable the processor to implement a liquid nitrogen freezing weed control method for farmland. The processor of the computer device provides computing and control capabilities, supporting the operation of the entire computer device. The memory of the computer device may store computer-readable instructions. When these computer-readable instructions are executed by the processor, they enable the processor to execute the liquid nitrogen freezing weed control method for farmland of this application. The network interface of the computer device is used for communication with a terminal. Those skilled in the art will understand that… Figure 4The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0130] In this embodiment, the processor is used to execute... Figure 3 The specific functions of each module are defined within the device, and the memory stores the program code and various data required to execute these modules. A network interface is used for data transmission between the user terminal and the server. In this embodiment, the memory stores the program code and data required to execute all modules in the farmland liquid nitrogen freezing weeding control device of this application, and the server can call the server's program code and data to execute the functions of all modules.

[0131] This application also provides a storage medium storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the steps of the liquid nitrogen freezing weed control method for farmland described in any embodiment of this application.

[0132] This application also provides a computer program product, including a computer program / instructions that, when executed by one or more processors, implement the steps of the farmland liquid nitrogen freezing weed control method described in any embodiment of this application.

[0133] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. This computer program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The aforementioned storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0134] The above description is only a partial embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for controlling weeds in an agricultural field by liquid nitrogen freeze, characterized by, The method comprises the steps of: acquiring an image of a farmland to be weeded containing weeds to be removed, performing target detection on the image of the farmland to be weeded based on a farmland weed identification model trained to a convergent state, to determine the three-dimensional spatial coordinates of a growth point corresponding to the weeds to be removed, the weed species, and the growth stage of the weeds to be removed; calling a preset weed species-liquid nitrogen spraying parameter mapping table to query the reference liquid nitrogen spraying parameters corresponding to the weed species and the growth stage of the weeds to be removed, wherein the reference liquid nitrogen spraying parameters include a first spraying pressure final value and a first spraying speed final value of a liquid nitrogen spraying executor; acquiring real-time temperature data and real-time humidity data in the farmland to be weeded, calculating a spraying compensation factor of the weeds to be removed according to the real-time temperature data and the real-time humidity data, and updating the first spraying pressure final value and the first spraying speed final value respectively according to the spraying compensation factor to determine a second spraying pressure final value and a second spraying speed final value; inputting the second spraying pressure final value and the second spraying speed final value into the liquid nitrogen spraying executor, and driving the liquid nitrogen spraying executor to perform liquid nitrogen freezing weeding on the weeds in the farmland to be weeded; repeating the above steps until the weeds in the farmland to be weeded are removed, to complete the farmland liquid nitrogen freezing weeding control.

2. The method for controlling weeds by liquid nitrogen freezing in an agricultural field according to claim 1, characterized by, The step of performing target detection on the image of the farmland to be weeded based on a farmland weed identification model trained to a convergent state to determine the three-dimensional spatial coordinates of a growth point corresponding to the weeds to be removed, the weed species, and the growth stage of the weeds to be removed comprises the steps of: receiving an image of the farmland to be weeded collected by an image collection system on a mobile platform in real time; calling a farmland weed identification model trained to a convergent state, performing target detection on the image of the farmland to be weeded by the farmland weed identification model, and outputting the pixel coordinates of a growth point of the weeds to be removed, the weed species, and the growth stage of the weeds to be removed in the image of the farmland to be weeded; based on a binocular vision stereo matching algorithm, combining the pixel coordinates of the growth point of the weeds to be removed with camera calibration parameters, real-time position data, and real-time attitude data of the mobile platform in a farmland world coordinate system to determine the three-dimensional spatial coordinates of the growth point of the weeds to be removed in the farmland world coordinate system.

3. The method for controlling weeds in farmland using liquid nitrogen freezing according to claim 1, characterized in that, The step of calling a preset weed species-liquid nitrogen spraying parameter mapping table to query the reference liquid nitrogen spraying parameters corresponding to the weed species and the growth stage of the weeds to be removed comprises the steps of: acquiring the weed species and the growth stage corresponding to the weeds to be removed, and a preset weed species-liquid nitrogen spraying parameter mapping table, wherein the growth stage is determined based on a weed physiological model combined with weed morphological feature analysis, and covers key growth stages in the growth cycle of weeds, the weed species-liquid nitrogen spraying parameter mapping table is constructed through previous farmland test data, and contains reference liquid nitrogen spraying parameters corresponding to different weed species and different growth stages, and the reference liquid nitrogen spraying parameters at least include a first spraying pressure final value and a first spraying speed final value of a liquid nitrogen spraying executor. For the broad-leaved weeds with thick cell walls and strong cold resistance, a higher final value of the first spraying pressure and a larger final value of the first spraying speed corresponding to the weed species and growth stage are matched in the weed species-liquid nitrogen spraying parameter mapping table, so that the liquid nitrogen can penetrate the weed leaves and reach the preset freezing depth; For the tender and narrow-leaved weeds, a lower final value of the first spraying pressure and a smaller final value of the first spraying speed corresponding to the weed species and growth stage are matched in the weed species-liquid nitrogen spraying parameter mapping table, so as to accurately freeze and kill the weeds and reduce the consumption of liquid nitrogen.

4. The method for controlling weeds by liquid nitrogen freezing in an agricultural field according to claim 1, characterized by, The step of acquiring real-time temperature data and real-time humidity data in the farmland to be weeded and calculating a spraying compensation factor of the weeds to be removed based on the real-time temperature data and the real-time humidity data comprises: acquiring real-time temperature data and real-time humidity data in the farmland to be weeded, and a reference spraying temperature and a reference spraying humidity of the weeds to be removed; calculating a first difference between the real-time temperature data and the reference spraying temperature, and a first product between a temperature adjustment coefficient and the first difference; calculating a second difference between the real-time humidity data and the reference spraying humidity, and a second product between a humidity adjustment coefficient and the second difference; calculating a first sum value between the first product, the second product and a numerical value one, and taking the first sum value as the spraying compensation factor of the weeds to be removed.

5. The method of claim 4, wherein the liquid nitrogen is applied to the crop field at a rate of from about 0.1 gallons per acre to about 10 gallons per acre. The step of updating the first spraying pressure final value and the first spraying speed final value respectively according to the spraying compensation factor to determine a second spraying pressure final value and a second spraying speed final value comprises: acquiring the spraying compensation factor of the weeds to be removed, the first spraying pressure final value and the first spraying speed final value of the weeds to be removed; calculating a first product between the first spraying pressure final value and the spraying compensation factor to determine the second spraying pressure final value, and calculating a second product between the first spraying speed final value and the spraying compensation factor to determine the second spraying speed final value.

6. The method of claim 1, wherein the liquid nitrogen freezing weed control method is performed in an agricultural field. The step of inputting the second spraying pressure final value and the second spraying speed final value into the liquid nitrogen spraying executor and driving the liquid nitrogen spraying executor to perform liquid nitrogen freezing weeding on the weeds in the farmland to be weeded comprises: receiving the second spraying pressure final value and the second spraying speed final value corresponding to the weeds to be removed output by the liquid nitrogen spraying control and decision system; during the spraying process of the nozzle of the liquid nitrogen spraying executor, adjusting the pipeline pressure of the liquid nitrogen storage and supply unit through the pressure regulating valve based on the second spraying pressure final value, collecting the liquid nitrogen pressure in front of the nozzle in real time through the pipeline built-in pressure sensor, comparing the collected actual pressure with the second spraying pressure final value, generating a pressure adjusting signal through the fuzzy PID control algorithm, and dynamically correcting the pipeline pressure to the second spraying pressure final value; collecting the actual spraying speed in real time through the flow sensor, comparing the actual spraying speed with the second spraying speed final value, adjusting the duty cycle of the pulse width modulation signal through the fuzzy PID control, and ensuring that the actual spraying speed stably tracks the second spraying speed final value; The two-degree-of-freedom precision gimbal of the liquid nitrogen spraying executor is driven to adjust the spray angle of the nozzle, so that the liquid nitrogen spray covers the growth point of the weed to be removed, thereby completing the timed, quantitative, and constant-pressure liquid nitrogen freezing weeding operation on the weed to be removed.

7. The method according to any one of claims 1 to 6, wherein the method is a method for controlling weeds by liquid nitrogen freezing in an agricultural field. The farmland weed identification model comprises a YOLO series detection model.

8. An agricultural field liquid nitrogen freeze weed control apparatus, characterized by, The method comprises the following steps: The weed identification module is configured to drive the mobile platform to obtain a farmland image containing weeds to be removed, and perform target detection on the farmland image based on a farmland weed identification model trained to a convergent state, to determine the three-dimensional spatial coordinates of the growth point of the weeds to be removed, the weed species, and the growth stage. The spraying parameter query module is configured to query the reference liquid nitrogen spraying parameters corresponding to the weed species and the growth stage of the weeds to be removed by calling a preset weed species-liquid nitrogen spraying parameter mapping table, wherein the reference liquid nitrogen spraying parameters comprise a first spray pressure final value and a first spray speed final value of the liquid nitrogen spraying executor. The spraying parameter update module is configured to obtain real-time temperature data and real-time humidity data in the farmland to be weeded, calculate a spraying compensation factor of the weeds to be removed according to the real-time temperature data and the real-time humidity data, and update the first spray pressure final value and the first spray speed final value according to the spraying compensation factor to determine a second spray pressure final value and a second spray speed final value. The liquid nitrogen weeding module is configured to input the second spray pressure final value and the second spray speed final value into the liquid nitrogen spraying executor, and drive the liquid nitrogen spraying executor to perform liquid nitrogen freezing weeding on the weeds in the farmland to be weeded. The central processing unit is configured to call and run a computer program stored in the memory to perform the steps of the method according to any one of claims 1 to 7.

9. An electronic device comprising a central processing unit and a memory, characterized in that The computer program is stored in the form of computer readable instructions and is implemented according to the method of any one of claims 1 to 7. When the computer program is called and run by a computer, the steps included in the corresponding method are performed.

10. A computer-readable storage medium, characterized in that, ​

Images