Remote automatic management system and method for lotus root field cultivation

Abstract

The invention discloses a remote automatic management system and method for lotus root field cultivation in the technical field of Internet of Things control. The system comprises a stage recognition module, a water level adjustment module, a water and fertilizer management module and a pest control module. Wherein the stage recognition module is used for acquiring lotus root field image information, recognizing floating leaves, standing leaves and termination leaves on the water surface through an image analysis technology, recognizing the area proportion and the number proportion of the leaves of the corresponding types in the lotus root field image, and judging the current growth stage of lotus roots according to a standard growth stage strategy; the water level adjusting module is used for calling the water level threshold parameter of the corresponding stage according to the current growth stage of the lotus root and adjusting the water level of the lotus root field to a target range in real time; real-time environment parameters are obtained, and the adjustment amplitude of the lotus root field water level is dynamically corrected. According to the method, the growth stage is judged through quantitative analysis of leaf characteristics, and the water level adaptation timeliness is improved by means of automatic water level regulation and control and a dynamic correction mechanism.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) control technology, specifically to a remote automated management system and method for lotus root field cultivation. Background Technology

[0002] Lotus root (Nelumbo nucifera Gaertn.) is an important aquatic economic crop in my country, possessing edible, medicinal, and ecological value. Its tubers are rich in starch, protein, and various trace elements. With the upgrading of consumption and the advancement of agricultural industrialization, lotus root cultivation is gradually transforming from scattered cultivation by individual farmers to intensive and large-scale bases, which places higher demands on the precision and efficiency of planting management.

[0003] The entire growth cycle of lotus root involves four stages: germination, mature plant, tuber formation, and dormancy. The water level requirements for each stage differ significantly: during germination, 5-10cm of shallow water is needed to raise the soil temperature and promote the germination of the terminal bud; during mature plant, 20-40cm of deep water is needed to support the growth of upright leaves and suppress weeds; during tuber formation, the water level should be gradually reduced to 5-15cm to force the underground stems to swell and accumulate starch; during dormancy, a water level of 5-10cm or moist soil should be maintained to protect the seed tubers for overwintering.

[0004] Current lotus root cultivation still relies mainly on traditional methods, where water level control depends on manual judgment of the lotus root's growth stage before remote operation. This means that growers conduct field inspections, observe visual characteristics such as leaf morphology and stem growth, and combine this with experience to divide the growth stage. They then manually open / close the inlet valve and drain pump to adjust the water level (although some planting bases have tried to introduce simple water level monitoring equipment, most of them are single threshold alarm modes, which can only remind manual intervention when the water level exceeds the preset range). This method has the advantages of being easy to operate and low in cost in small-scale planting. However, in large-scale, intensive cultivation scenarios, many unavoidable defects have been exposed. For example, the division of growth stages relies entirely on manual experience and lacks unified quantitative standards, which can easily lead to misjudgments (such as mistaking small upright leaves in the mature stage for floating leaves in the budding stage), directly affecting water level adaptability and resulting in a decrease in yield. In large-scale bases, inspecting leaves and manually adjusting water levels is time-consuming. Inspection and control of a single plot can take several hours, and simultaneous operation of the entire field often takes more than a day, making it impossible to respond to growth needs in a timely manner. At the same time, there are time intervals in manual inspections. In the event of extreme weather such as heavy rain or high temperatures, or sudden changes in growth stages, failure to adjust water levels in a timely manner can easily lead to problems such as plant lodging and lotus cracking, resulting in significant losses.

[0005] Therefore, this invention proposes a remote automated management system and method for lotus root field cultivation to solve the above problems. Summary of the Invention

[0006] To address the aforementioned issues, this invention provides a remote automated management system and method for lotus root field cultivation. It uses quantitative analysis of leaf characteristics to determine the growth stage and relies on automated water level control and dynamic correction mechanisms to improve the timeliness of water level adaptation.

[0007] To achieve the above objectives, the technical solution of the present invention is as follows: a remote automated management system for lotus root field cultivation, comprising:

[0008] The stage identification module is used to acquire lotus field image information, identify floating leaves, upright leaves and terminal leaves through image analysis technology, and identify the area and number of the corresponding types of leaves in the lotus field image. According to the standard growth stage strategy, it determines the current growth stage of the lotus root. The lotus root growth stages include budding stage, mature plant stage, tuber formation stage and dormancy stage.

[0009] The water level adjustment module is used to retrieve the water level threshold parameters corresponding to the current growth stage of the lotus root, and adjust the water level of the lotus field to the target range in real time; and to obtain real-time environmental parameters to dynamically correct the adjustment range of the water level in the lotus field.

[0010] The water and fertilizer management module is used to regulate the nutrient content of the soil and water in the lotus pond;

[0011] The pest and disease control module is used to prevent and control the damage caused by pests and diseases to lotus roots.

[0012] Furthermore, in the stage recognition module, the lotus field image is preprocessed, and then the outlines and number of floating leaves, upright leaves and terminal leaves in the image are extracted by a convolutional neural network. The total area of ​​floating leaves, upright leaves and terminal leaves is calculated by pixel area conversion. Based on the preset lotus field parameters, the actual area data of the lotus field is obtained, and then the proportion of the total area of ​​each type of leaf in the lotus field area is obtained. Comparative data is generated, and then the current growth stage of the lotus root is determined according to the standard growth stage strategy.

[0013] Furthermore, standard growth stage strategies include:

[0014] The germination period determination strategy is as follows: when the total area of ​​floating leaves accounts for ≥60% of the lotus field area, the number of floating leaves accounts for 90%-100% of the total number of leaves, and the proportion of terminal leaves is 0, it is determined to be the germination period.

[0015] Strategy for determining the mature stage: When the total area of ​​upright leaves accounts for ≥70% of the lotus field area and no terminal leaves are detected, it is determined to be in the mature stage.

[0016] Strategy for determining the tuber formation period: When the total area of ​​the terminal leaves accounts for 1%-5% of the area of ​​the lotus field or the number of terminal leaves accounts for 5%-15% of the total number of leaves, and the number of upright leaves does not increase for two consecutive days, it is determined to be the tuber formation period.

[0017] Dormancy period determination strategy: When the total area of ​​floating leaves, upright leaves, and terminal leaves accounts for ≤10% of the lotus field area for 3 consecutive days, it is determined to be the dormancy period.

[0018] Furthermore, the determination strategy for each growth stage serves as a prerequisite for the next growth stage; and the determination of each growth stage requires that the determination conditions be met within a continuously preset time period before the determination result of the lotus root growth stage is confirmed.

[0019] Furthermore, in the water level adjustment module, the water level adjustment strategy includes:

[0020] When the seedlings are in their budding stage, the water level should be adjusted to 5-10cm.

[0021] When the plants are in the mature stage, the water level should be adjusted to 20-40cm.

[0022] When the lotus root is forming, the water level should be adjusted to 5-15cm.

[0023] When the water level is in a dormant period, adjust it to 5-10cm.

[0024] Furthermore, the water level adjustment module is also used to obtain meteorological forecast data for a future preset time period. Combined with the water level threshold parameters of the current lotus root growth stage, when the meteorological forecast data meets the preset adjustment conditions, a water level pre-adjustment instruction is generated to adjust the lotus field water level to the lower limit range of the corresponding growth stage water level threshold before the occurrence of a rainfall meteorological event, reserving water level buffer space.

[0025] Furthermore, a blade overlap area correction mechanism is added to the stage recognition module. When the convolutional neural network recognizes that there is contour overlap between floating blades, upright blades, and terminal blades, the area is calculated and the image is segmented according to the priority order of "terminal blade > upright blade > floating blade". The contour boundary of the high-priority blade is used as the segmentation benchmark, covering the pixel data of the low-priority blade in the overlap area. Only the area of ​​the overlap area of ​​the high-priority blade is retained and included in its total area. The area of ​​the low-priority blade is converted by the remaining contour pixels after deducting the overlap area.

[0026] Furthermore, the water and fertilizer management module is used to monitor dissolved oxygen, pH value, soil EC value, and nitrogen, phosphorus, and potassium content in the lotus field water. Combined with the nutrient requirement model of the current growth stage, it calculates the amount and ratio of water and fertilizer supplementation, and performs targeted fertilization and water quality regulation to achieve adaptive nutrient compensation.

[0027] Furthermore, the pest and disease control module is used to acquire hyperspectral images of lotus leaves through hyperspectral imaging technology, extract the spectral reflectance characteristics of lesions, the spectral outline and distribution density of insects in the hyperspectral images, and perform comparative analysis with a preset pest and disease spectral feature library to identify the type and level of pests and diseases. Based on the identification results, the corresponding control plan is matched, and targeted spraying is carried out on the affected leaf areas. The control time, dosage and subsequent spectral image changes are recorded to evaluate the control effect.

[0028] Furthermore, a remote automated management method for lotus root field cultivation includes the following steps:

[0029] Step 1, Growth Stage Identification: The lotus field images are collected and preprocessed through the stage identification module. Based on image recognition technology, the outlines and quantities of floating leaves, upright leaves, and terminal leaves are identified. The area of ​​each type of leaf is converted after overlapping area correction. The proportion of each type of leaf is calculated and matched with the standard growth stage strategy. Combined with the stage pre-requisite rules and continuous achievement time, the current growth stage is confirmed.

[0030] Step 2, water level regulation: The water level adjustment module regulates the water level according to the current growth stage and adjusts and reserves a buffer in advance based on weather forecasts; then, the water and fertilizer management module monitors the water and soil nutrient indicators and supplements water and fertilizer in a targeted manner according to the growth stage demand model.

[0031] Step 3, Pest Control: The pest control module identifies the type and level of pests and diseases through hyperspectral imaging, then carries out targeted control, records and evaluates the effect, and simultaneously uploads all data to the remote center to form a traceability archive.

[0032] The above approach has the following beneficial effects:

[0033] 1. Compared with existing technologies, this solution constructs a standardized judgment system through a stage recognition module. Utilizing image preprocessing and convolutional neural network technology, it accurately extracts the morphological features of floating leaves, upright leaves, and terminal leaves. Combined with a leaf overlap area correction mechanism, it effectively avoids interference from leaf occlusion on area and quantity statistics, ensuring the accuracy of leaf proportion data. Simultaneously, based on the quantitative correspondence between leaf proportion and growth stage, and combined with the judgment rules of "stage pre-dependency" and "continuous condition compliance," it transforms fuzzy empirical judgments into quantifiable objective standards, reducing the problem of human error. Accurate identification of growth stages provides a reliable basis for subsequent management, enabling measures such as water level control and fertilizer replenishment to precisely match the needs of lotus roots at different growth stages, reducing growth problems caused by stage misjudgment, and providing core technical support for standardized management in large-scale cultivation.

[0034] 2. Addressing the shortcomings of traditional management methods, such as lagging water level control and the disconnect between water and fertilizer management and pest and disease control, this solution establishes a closed-loop, full-cycle management system through multi-module collaboration. The water level adjustment module not only implements baseline control based on growth stages but also incorporates weather forecasts for proactive adjustments, reserving buffer space to cope with extreme weather and transforming passive response into active defense. The water and fertilizer management module, based on real-time monitoring of soil and water indicators and growth stage demand models, achieves adaptive nutrient compensation, avoiding resource waste or nutrient imbalance caused by indiscriminate fertilization. The pest and disease control module utilizes hyperspectral imaging technology for early identification and intervention, coupled with targeted application and effect evaluation mechanisms, improving control accuracy and reducing pesticide overuse. Data from all modules is linked in real-time and traceable throughout the entire process, reducing the time cost of manual inspections and operations, effectively mitigating production risks from extreme weather and pests, improving the management efficiency and stability of large-scale cultivation, and contributing to high-quality and high-yield lotus root production.

[0035] Additional aspects and advantages of the invention 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 the invention. Attached Figure Description

[0036] Figure 1 A flowchart illustrating an embodiment of the remote automated management system for lotus root field cultivation according to the present invention;

[0037] Figure 2 This is a schematic diagram of an embodiment of the remote automated management method for lotus root field cultivation according to the present invention. Detailed Implementation

[0038] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0040] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0041] The following detailed description illustrates the specific implementation method:

[0042] Example 1:

[0043] As attached Figure 1 As shown, a remote automated management system for lotus root field cultivation includes a stage identification module, a water level adjustment module, a water and fertilizer management module, and a pest and disease control module.

[0044] In the stage recognition module, the original lotus field image is first acquired through image acquisition technology (such as drone photography or high-definition industrial cameras deployed on supports around the lotus field). The original lotus field image is preprocessed, that is, an adaptive median filtering algorithm is used to eliminate noise interference caused by water mist, bird shadows and water surface reflection. Perspective transformation is used to correct geometric distortion caused by camera tilt (correction accuracy reaches the pixel level). Then, color gamut standardization processing is performed—converting the RGB color space to the HSV space, fixing the threshold of hue (H) and saturation (S) channels, eliminating the influence of light intensity differences at different times on the color characteristics of leaves, and ensuring that the color performance of the same type of leaves is consistent in different images.

[0045] The preprocessed lotus field images are fed into a lightweight convolutional neural network (using an improved MobileNet architecture, containing 12 convolutional layers and 3 pooling layers) for feature extraction: shallow convolutional layers capture the edge gradient and contour curve of leaves (such as the circular edge of floating leaves, the heart-shaped contour of upright leaves, and the curled edge of terminal leaves) through 3×3 convolutional kernels; middle convolutional layers extract texture features (such as the reflective texture of the waxy layer of floating leaves and the distribution density of veins of upright leaves); deep fully connected layers fuse morphological parameters (petiole length, leaf aspect ratio, and angle between the leaf surface and the water surface) to construct a dedicated feature vector library for three types of leaves, realizing the classification of floating leaves (thin petioles, leaves close to the water, angle ≤10°), upright leaves (hard petioles, upright leaves, angle ≥70°), and terminal leaves (curled leaf tips, located at the end of the plant, yellow and smaller than the diameter of upright leaves). Non-maximum suppression algorithm is used to remove interfering targets such as weeds and dead leaves, and the contour masks and quantity statistics of the three types of leaves are output. Simultaneously, a blade overlap area correction mechanism is added during this process. The intersection operation is performed on the three types of blade contour masks output by the convolutional neural network to automatically detect areas with pixel overlap (overlapping pixels accounting for ≥5% of the pixels in a single blade contour are considered valid overlap). When overlapping contours are detected in floating leaves, upright leaves, and terminal leaves, area calculation and image segmentation are performed according to the priority order of "terminal leaf > upright leaf > floating leaf." Using the contour boundary of the high-priority leaf as the segmentation benchmark, a pixel ownership determination algorithm completely allocates all pixel data in the overlapping area to the high-priority leaf, covering the pixel records of the low-priority leaf in that area. Only the overlapping area of ​​the high-priority leaf is retained in the total blade contour area. For low-priority leaves, based on the corrected contour mask, the number of pixels in the overlapping area is deducted before counting the total number of pixels in the remaining valid contour, ensuring that the area calculations for the three types of leaves are without duplication or omission.

[0046] To calculate the total leaf area, a conversion model of "pixel count - actual area" is first established using pre-set calibration board data (a black and white grid board of known actual size, placed in a typical area of ​​the lotus field for photography). The model achieves an accuracy of 0.01 cm² for the actual area per pixel. Then, based on the corrected contour mask, the total pixel area of ​​the three types of leaves is statistically analyzed and substituted into the model to convert it into the actual area. Simultaneously, pre-stored lotus field parameters (vector data such as field ridge boundaries and ditch locations measured in the field) are used to automatically deduct the area of ​​non-planted areas, obtaining the effective cultivated area. Finally, the proportion of the actual total area of ​​floating leaves, upright leaves, and terminal leaves to the effective cultivated area is calculated, generating three-dimensional comparative data of "area proportion - quantity proportion - leaf morphological characteristics." Finally, the comparative data is matched with a pre-set standard growth stage strategy library.

[0047] The standard growth stage strategy library includes: if the floating leaf area accounts for ≥60% and the number accounts for 90%-100%, there are no terminal leaves, and this condition is met for 2-3 consecutive days, then it is judged to be in the budding stage.

[0048] If the area of ​​upright leaves accounts for ≥70% and there are no terminal leaves, and the condition remains stable for 2-3 consecutive days, then it is determined to be in the mature stage.

[0049] When the total area of ​​terminal leaves accounts for 1%-5% of the lotus field area or the number of terminal leaves accounts for 5%-15% of the total number of leaves, and the number of upright leaves does not increase for 2-3 consecutive days, it is determined to be the lotus formation period.

[0050] When the total area of ​​floating leaves, upright leaves, and terminal leaves accounts for ≤10% of the lotus field area for 3 consecutive days, it is determined to be the dormant period.

[0051] Meanwhile, the determination of the growth stage follows the rule that "the previous stage is a prerequisite for the next stage" (e.g., the mature stage must be based on the budding stage) to avoid misjudgment across stages and ultimately output a definite result of the current growth stage, providing a precise basis for subsequent water level control and water and fertilizer management.

[0052] In the water level adjustment module, the current growth stage is first determined (budding stage, mature stage, rhizome formation stage, or dormancy stage), and then the water level adjustment strategy is invoked (which stores the benchmark water level range corresponding to each stage): the budding stage is matched with a shallow water threshold of 5-10cm to meet the needs of soil temperature increase and apical bud germination; the mature stage corresponds to a deep water threshold of 20-40cm to provide support for upright leaf growth and inhibit weed growth; the rhizome formation stage is set with a gradually shallow water threshold of 5-15cm to guide underground stem swelling and starch accumulation; and the dormancy stage maintains a water level threshold of 5-10cm to ensure the moist environment required for the overwintering of rhizomes.

[0053] Meanwhile, the water level adjustment module connects to the meteorological data interface to obtain real-time meteorological forecast data (including rainfall, rainfall intensity, wind force, etc.) for the next 24-48 hours. When the meteorological forecast data meets the preset adjustment conditions (such as the expected rainfall ≥20mm within 24 hours and the rainfall intensity reaching moderate rain or above), the pre-adjustment logic is immediately triggered. Combining the water level threshold parameters of the current growth stage, the pre-adjustment command is automatically generated to lower the water level in advance to the lower limit of the corresponding stage threshold (such as from the baseline 20-40cm to 20cm in the mature stage and from 5-10cm to 5cm in the budding stage), reserving sufficient water level buffer space to avoid problems such as plant lodging and root hypoxia caused by the water level exceeding the upper limit of the threshold after rainfall.

[0054] The water and fertilizer management module is used to monitor dissolved oxygen, pH value, soil EC value, and nitrogen, phosphorus and potassium content in lotus pond water. Combined with the nutrient requirement model of the current growth stage, it calculates the amount and ratio of water and fertilizer supplementation, and carries out targeted fertilization and water quality regulation to achieve adaptive nutrient compensation.

[0055] The pest and disease control module is used to acquire hyperspectral images of lotus leaves using hyperspectral imaging technology, extract the spectral reflectance characteristics of lesions, the spectral outline and distribution density of insects from the hyperspectral images, and compare and analyze them with a pre-set pest and disease spectral feature library to identify the type and level of pests and diseases. Based on the identification results, the module matches the corresponding control plan, links the targeted application device (such as the See&Spray system) to carry out point spraying on the affected leaf areas, and records the control time, dosage and subsequent spectral image changes to evaluate the control effect.

[0056] Example 2:

[0057] like Figure 2 As shown, a remote automated management method for lotus root field cultivation includes the following steps:

[0058] Step 1, Growth Stage Identification: The stage identification module acquires real-time panoramic images of the lotus root field (e.g., drone footage). First, multi-dimensional preprocessing is performed on the images. Then, a convolutional neural network is used to extract the contour boundaries and quantity information of floating leaves, upright leaves, and terminal leaves. A leaf overlap correction mechanism is activated, segmenting overlapping pixels according to the priority of "terminal leaf → upright leaf → floating leaf" to avoid counting errors. Based on the pixel-to-actual area conversion model established by the calibration board, combined with pre-stored effective cultivation area data of the lotus root field, the area and quantity proportions of each type of leaf are calculated. The generated comparison data is matched with the standard growth stage strategy. Simultaneously, the rule of "the previous stage is a prerequisite for the next stage" is strictly followed, and the judgment conditions must be continuously met for a preset time before the current growth stage of the lotus root (budding stage, mature stage, tuber formation stage, or dormancy stage) is finally confirmed.

[0059] Step two, water level regulation: The water level adjustment module retrieves the corresponding benchmark water level threshold based on the confirmed growth stage and adjusts the water level in the lotus root field to the target range. Simultaneously, it obtains a refined forecast for the next 24-48 hours. If a rainfall warning signal is detected, the water level is lowered to the lower limit of the current stage threshold to reserve buffer space. In addition, the water and fertilizer management module collects real-time data on dissolved oxygen, pH value, soil EC value, and nitrogen, phosphorus, and potassium content through sensors. Combined with nutrient requirement models for each growth stage, it accurately calculates the amount and ratio of water and fertilizer supplementation. Precise fertilization and water quality regulation are implemented through a directional irrigation device, achieving dynamic nutrient adaptation and compensation.

[0060] Step 3, Pest Control: The pest control module captures leaf spectral images using hyperspectral imaging equipment, extracts the spectral reflectance characteristics of lesions, the spectral outline and distribution density of insects, and compares them with a pre-set pest spectral feature database to quickly identify pest types and damage levels. Based on the identification results, a corresponding control plan is matched, and a targeted application device is activated to spray the affected area, automatically recording the application time, pesticide type, and dosage. Leaf spectral images are re-measured 24 hours after application, and the control effect is evaluated through changes in lesion shrinkage rate and insect density. Simultaneously, growth stage determination data, water level control records, water and fertilizer supply parameters, and pest control information are uploaded to a remote monitoring center, generating a full-cycle traceability archive with timestamps and location information, providing support for planting optimization and quality traceability.

[0061] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A remote automated management system for lotus root field cultivation, characterized in that, include: The stage identification module is used to acquire lotus field image information, identify floating leaves, upright leaves and terminal leaves through image analysis technology, and identify the area and number of the corresponding types of leaves in the lotus field image. According to the standard growth stage strategy, it determines the current growth stage of the lotus root. The lotus root growth stages include budding stage, mature plant stage, tuber formation stage and dormancy stage. The water level adjustment module is used to retrieve the water level threshold parameters corresponding to the current growth stage of the lotus root, and adjust the water level of the lotus field to the target range in real time; and to obtain real-time environmental parameters to dynamically correct the adjustment range of the water level in the lotus field. The water and fertilizer management module is used to regulate the nutrient content of the soil and water in the lotus pond; The pest and disease control module is used to prevent and control the damage caused by pests and diseases to lotus roots.

2. The remote automated management system for lotus root field cultivation according to claim 1, characterized in that: In the stage recognition module, the lotus field image is preprocessed, and then the outlines and number of floating leaves, upright leaves and terminal leaves in the image are extracted by a convolutional neural network. The total area of ​​floating leaves, upright leaves and terminal leaves is calculated by pixel area conversion. Based on the preset lotus field parameters, the actual area data of the lotus field is obtained, and then the proportion of the total area of ​​each type of leaf in the lotus field area is obtained. Comparative data is generated, and the current growth stage of the lotus root is determined according to the standard growth stage strategy.

3. The remote automated management system for lotus root field cultivation according to claim 2, characterized in that: Standard growth stage strategies include: The germination period determination strategy is as follows: when the total area of ​​floating leaves accounts for ≥60% of the lotus field area, the number of floating leaves accounts for 90%-100% of the total number of leaves, and the proportion of terminal leaves is 0, it is determined to be the germination period. Strategy for determining the mature stage: When the total area of ​​upright leaves accounts for ≥70% of the lotus field area and no terminal leaves are detected, it is determined to be in the mature stage. Strategy for determining the tuber formation period: When the total area of ​​the terminal leaves accounts for 1%-5% of the area of ​​the lotus field or the number of terminal leaves accounts for 5%-15% of the total number of leaves, and the number of upright leaves does not increase for two consecutive days, it is determined to be the tuber formation period. Dormancy period determination strategy: When the total area of ​​floating leaves, upright leaves, and terminal leaves accounts for ≤10% of the lotus field area for 3 consecutive days, it is determined to be a dormant period.

4. The remote automated management system for lotus root field cultivation according to claim 3, characterized in that: The strategy for determining each growth stage serves as a prerequisite for the next growth stage; and the determination of each growth stage requires that the determination conditions be met within a continuously preset time period before the determination result of the lotus root growth stage is confirmed.

5. The remote automated management system for lotus root field cultivation according to claim 4, characterized in that: The water level adjustment strategy in the water level adjustment module includes: When the seedlings are in their budding stage, the water level should be adjusted to 5-10cm. When the plants are in the mature stage, the water level should be adjusted to 20-40cm. When the lotus root is forming, the water level should be adjusted to 5-15cm. When the water level is in a dormant period, adjust it to 5-10cm.

6. The remote automated management system for lotus root field cultivation according to claim 5, characterized in that: The water level adjustment module is also used to obtain meteorological forecast data for a future preset time period. Combined with the water level threshold parameters of the current lotus root growth stage, when the meteorological forecast data meets the preset adjustment conditions, a water level pre-adjustment instruction is generated to adjust the water level of the lotus field to the lower limit range of the corresponding growth stage water level threshold before the occurrence of a rainfall meteorological event, reserving water level buffer space.

7. The remote automated management system for lotus root field cultivation according to claim 6, characterized in that: A blade overlap area correction mechanism is added to the stage recognition module. When the convolutional neural network recognizes that there is contour overlap between floating blades, upright blades and terminal blades, the area is calculated and the image is segmented according to the priority order of "terminal blade > upright blade > floating blade". The contour boundary of the high-priority blade is used as the segmentation benchmark, covering the pixel data of the low-priority blade in the overlapping area. Only the overlapping area of ​​the high-priority blade is retained and included in its total area. The area of ​​the low-priority blade is converted by the remaining contour pixels after deducting the overlapping area.

8. The remote automated management system for lotus root field cultivation according to claim 7, characterized in that: The water and fertilizer management module is used to monitor dissolved oxygen, pH value, soil EC value, and nitrogen, phosphorus and potassium content in lotus pond water. Combined with the nutrient requirement model of the current growth stage, it calculates the amount and ratio of water and fertilizer supplementation, and carries out targeted fertilization and water quality regulation to achieve adaptive nutrient compensation.

9. The remote automated management system for lotus root field cultivation according to claim 8, characterized in that: The pest and disease control module is used to acquire hyperspectral images of lotus leaves using hyperspectral imaging technology, extract the spectral reflectance characteristics of lesions, the spectral outline and distribution density of insects from the hyperspectral images, and perform comparative analysis with a pre-set pest and disease spectral feature library to identify pest and disease types and damage levels. Based on the identification results, the module matches the corresponding control plan, performs targeted spraying on the affected leaf areas, and records the control time, dosage and subsequent spectral image changes to evaluate the control effect.

10. A remote automated management method for lotus root field cultivation, comprising a remote automated management system for lotus root field cultivation according to any one of claims 1-9, characterized in that, Includes the following steps: Step 1, Growth Stage Identification: The lotus field images are collected and preprocessed through the stage identification module. Based on image recognition technology, the outlines and quantities of floating leaves, upright leaves, and terminal leaves are identified. The area of ​​each type of leaf is converted after overlapping area correction. The proportion of each type of leaf is calculated and matched with the standard growth stage strategy. Combined with the stage pre-requisite rules and continuous achievement time, the current growth stage is confirmed. Step 2, water level regulation: The water level adjustment module regulates the water level according to the current growth stage and adjusts and reserves a buffer in advance based on weather forecasts; then, the water and fertilizer management module monitors the water and soil nutrient indicators and supplements water and fertilizer in a targeted manner according to the growth stage demand model. Step 3, Pest Control: The pest control module identifies the type and level of pests and diseases through hyperspectral imaging, then carries out targeted control, records and evaluates the effect, and simultaneously uploads all data to the remote center to form a traceability archive.

Images