Plant recognition method and device, electronic equipment and storage medium

Abstract

The application discloses a kind of plant identification method, device, electronic equipment and storage medium, belong to the technical field of monitoring and identification of coastal wetland invasive plants. Including: obtaining the detection data of the target area for the plant to be detected;Detection data includes the radial size data of underground rhizome, lateral deformation displacement data, morphological image data and rhizosphere soil physicochemical data;According to the detection data, the rhizome size morphological characteristics, rhizome growth rate characteristics and rhizosphere nutrient physicochemical characteristics of the underground rhizome of the plant to be detected are determined;Obtain the characteristic reference of target plant;According to the rhizome size morphological characteristics, rhizome growth rate characteristics, rhizosphere nutrient physicochemical characteristics of the plant to be detected and the characteristic reference of target plant, determine whether the plant to be detected in target area is target plant. Through the comprehensive judgment of underground rhizome multidimensional characteristics, the underground latent plant and ungerminated hidden patch of Spartina alterniflora can be accurately identified, and it is suitable for recurrence early warning and accurate prevention and control after Spartina alterniflora is removed.

Description

Technical Field

[0001] This invention belongs to the field of monitoring and identification technology for invasive plants in coastal wetlands, specifically relating to a plant identification method, a plant identification device, electronic equipment, and a computer-readable storage medium. Background Technology

[0002] Spartina alterniflora, a highly invasive coastal plant, spreads rapidly horizontally via its robust, creeping rhizomes. After surface clearing, the remaining rhizomes easily form dormant patches, which is the core reason for its recurring outbreaks and high relapse rate. Current monitoring technologies for Spartina alterniflora rely heavily on surface identification methods such as drone remote sensing and manual surveys. These methods can only identify established surface plants and cannot detect dormant rhizomes and undeveloped patches, resulting in monitoring delays and high false negative rates.

[0003] Existing underground monitoring technologies have significant shortcomings: single rhizome deformation sensors can only monitor rhizome growth rate and cannot distinguish plant species; simple root canal imaging identification is easily obscured by tidal flat silt and mixed vegetation interference; single rhizosphere physicochemical sensors lack support from plant characteristics, resulting in a high false positive rate. Current identification schemes cannot effectively distinguish the underground rhizomes of Spartina alterniflora from native tidal flat plants such as reeds, Suaeda salsa, and Rhizoma Cynanchi, making it difficult to accurately identify latent plants. Summary of the Invention

[0004] The purpose of this invention is to provide a plant identification method, a plant identification device, an electronic device, and a corresponding computer-readable storage medium, which can solve the problems of existing monitoring technologies for latent plants in Spartina alterniflora having a single monitoring dimension, low accuracy in classifying species, and inability to identify hidden patches.

[0005] To solve the above-mentioned technical problems, the present invention is implemented as follows: In a first aspect, embodiments of the present invention provide a method for identifying plants, the method comprising: Acquire detection data of the plants to be tested in the target area; the detection data includes radial dimension data, lateral deformation and displacement data, morphological image data, and rhizosphere soil physicochemical data of the underground rhizomes; Based on the detection data, the root size and morphological characteristics, root growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plants to be tested in the target area are determined. Obtain the characteristic benchmarks of the target plant; Based on the root and stem size and morphological characteristics, root and stem growth rate characteristics, rhizosphere nutrient physicochemical characteristics, and characteristic criteria of the target plant, it is determined whether the plant to be tested in the target area is the target plant.

[0006] Optionally, the characteristic criteria of the target plant include the root and stem size and morphological characteristics, root and stem growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the target plant. The step of determining whether a plant in the target area is a target plant based on the rootstock size and morphological characteristics, rootstock growth rate characteristics, rhizosphere nutrient physicochemical characteristics, and characteristic criteria of the target plant includes: The root and stem size and morphological characteristics of the plant to be tested are matched with the root and stem size and morphological characteristics of the target plant to determine the root and stem morphological matching score of the plant to be tested. The root and stem growth rate characteristics of the plant to be tested are matched with the root and stem growth rate characteristics benchmark of the target plant to determine the root and stem dynamic growth rate matching score of the plant to be tested. The rhizosphere nutrient physicochemical characteristics are matched with the rhizosphere nutrient physicochemical characteristics benchmark of the target plant to determine the rhizosphere physicochemical fingerprint matching score of the plant to be tested. The comprehensive matching score of the plant under test is determined based on the matching score of the root and stem morphology, the matching score of the root and stem dynamic growth rate, and the matching score of the rhizosphere physicochemical fingerprint. If the overall matching score of the plant to be tested is greater than or equal to a preset threshold, then the plant to be tested is determined to be the target plant.

[0007] Optionally, the radial dimension data and the lateral deformation displacement data are obtained by using a layered annular radial deformation sensor installed in the soil of the target area; The morphological image data was acquired by a miniature root canal imaging unit installed in the soil of the target area. The rhizosphere soil physicochemical data were collected by a gridded rhizosphere physicochemical sensor array set in the soil of the target area.

[0008] Optionally, determining the rootstock size and morphological characteristics, rootstock growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plants to be tested in the target area based on the detection data includes: Based on the morphological image data, the apparent diameter, cross-sectional structure, internode distribution morphology, and number of latent buds of the underground rhizome of the plant to be tested are determined. Based on the radial dimension data, the apparent diameter, and the cross-sectional structure, the thickness grade of the underground rhizome to be tested is determined; The cross-sectional structure, the internode distribution pattern, the number of latent buds, and the thickness grade are determined as the root and stem size and morphological characteristics of the underground rhizome of the plant to be tested.

[0009] Optionally, determining the rootstock size and morphological characteristics, rootstock growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plants to be tested in the target area based on the detection data further includes: Based on the radial dimension data of the underground rhizomes within a continuous time series, the average daily radial growth of the underground rhizomes of the plant under test is determined. Based on the lateral deformation displacement data within a continuous time series, the average daily lateral extension of the underground rhizomes of the plant under test is determined. The radial daily average growth and the lateral daily average extension are determined as the root growth rate characteristics of the underground rhizomes of the plant under test.

[0010] Optionally, the rhizosphere soil physicochemical data include at least one of the following: salt content, organic matter content, nitrogen and phosphorus content, electrical conductivity, and soil respiration rate. The step of determining the rootstock size and morphological characteristics, rootstock growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plants to be tested in the target area based on the detection data also includes: Based on at least one of the following in the rhizosphere soil physicochemical data: salt content, organic matter content, nitrogen and phosphorus content, electrical conductivity, and soil respiration rate, determine the nutrient enrichment coefficient and rhizosphere environmental fingerprint of the underground rhizome rhizosphere region to be tested; The nutrient enrichment coefficient and the rhizosphere environmental fingerprint are used to determine the rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plant under test.

[0011] Optionally, before determining the rootstock size and morphological characteristics, rootstock growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plants to be tested in the target area based on the detection data, the method further includes: Abnormal data caused by tidal and silt disturbances are removed from the detection data.

[0012] Optionally, after determining whether the plant to be detected in the target area is the target plant, the method further includes: Determine the location coordinates of the target plant.

[0013] Secondly, embodiments of the present invention provide a plant identification device, the device comprising: The detection data acquisition module is used to acquire detection data of the plants to be tested in the target area; the detection data includes radial dimension data of underground rhizomes, lateral deformation and displacement data, morphological image data, and rhizosphere soil physicochemical data; The feature determination module is used to determine the root size and morphological characteristics, root growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plant to be detected in the target area based on the detection data. The feature benchmark acquisition module is used to acquire the feature benchmarks of the target plant. The target plant determination module is used to determine whether the plant to be detected in the target area is the target plant based on the root and stem size and morphological characteristics, root and stem growth rate characteristics, rhizosphere nutrient physicochemical characteristics, and the characteristic criteria of the target plant.

[0014] Optionally, the characteristic criteria of the target plant include the root and stem size and morphological characteristics, root and stem growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the target plant. The target plant identification module includes: The root and stem morphology matching score determination submodule is used to match the root and stem size and morphology features of the plant to be tested with the root and stem size and morphology feature benchmark of the target plant to determine the root and stem morphology matching score of the plant to be tested. The root and stem dynamic growth rate matching score determination submodule is used to match the root and stem growth rate characteristics of the plant to be tested with the root and stem growth rate characteristic benchmark of the target plant to determine the root and stem dynamic growth rate matching score of the plant to be tested. The rhizosphere physicochemical fingerprint matching score determination submodule is used to match the rhizosphere nutrient physicochemical characteristics with the rhizosphere nutrient physicochemical characteristic benchmark of the target plant to determine the rhizosphere physicochemical fingerprint matching score of the plant to be tested. The comprehensive matching score determination submodule is used to determine the comprehensive matching score of the plant under test based on the matching score of the root and stem morphology, the matching score of the root and stem dynamic growth rate, and the matching score of the rhizosphere physicochemical fingerprint. The target plant determination submodule is used to determine the plant to be detected as the target plant if the comprehensive matching score of the plant to be detected is greater than or equal to a preset threshold.

[0015] Optionally, the radial dimension data and the lateral deformation displacement data are obtained by using a layered annular radial deformation sensor installed in the soil of the target area; The morphological image data was acquired by a miniature root canal imaging unit installed in the soil of the target area. The rhizosphere soil physicochemical data were collected by a gridded rhizosphere physicochemical sensor array set in the soil of the target area.

[0016] Optionally, the feature determination module includes: The first determining submodule is used to determine the apparent diameter, cross-sectional structure, internode distribution morphology, and number of latent buds of the underground rhizome of the plant to be detected based on the morphological image data. The second determining submodule is used to determine the thickness grade of the underground rhizome to be tested based on the radial dimension data, the apparent diameter, and the cross-sectional structure. The rootstock size and morphology characteristic determination submodule is used to determine the cross-sectional structure, the internode distribution pattern, the number of latent buds, and the thickness grade as the rootstock size and morphology characteristics of the underground rootstock of the plant to be tested.

[0017] Optionally, it also includes: The third determining submodule is used to determine the average daily radial growth of the underground rhizome of the plant to be tested based on the radial dimension data of the underground rhizome in a continuous time series. The fourth determination submodule is used to determine the average daily lateral extension of the underground rhizomes of the plant under test based on the lateral deformation displacement data in a continuous time series. The root and stem growth rate characteristic determination submodule is used to determine the radial average daily growth and the lateral average daily extension as the root and stem growth rate characteristics of the underground root and stem of the plant to be tested.

[0018] Optionally, the rhizosphere soil physicochemical data include at least one of the following: salt content, organic matter content, nitrogen and phosphorus content, electrical conductivity, and soil respiration rate. The feature determination module further includes: The fifth determination submodule is used to determine the nutrient enrichment coefficient and rhizosphere environmental fingerprint of the underground rhizome rhizosphere region to be tested based on at least one of the salt content, organic matter content, nitrogen and phosphorus content, electrical conductivity and soil respiration rate in the rhizosphere soil physicochemical data. The rhizosphere nutrient physicochemical characteristics determination submodule is used to determine the nutrient enrichment coefficient and the rhizosphere environmental fingerprint as the rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plant to be tested.

[0019] Optionally, it also includes: The anomaly removal module is used to remove abnormal data caused by tidal and silt disturbances from the detection data.

[0020] Optionally, after determining whether the plant to be detected in the target area is the target plant, the method further includes: The coordinate determination module is used to determine the position coordinates of the target plant.

[0021] Thirdly, embodiments of the present invention provide an electronic device including a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the method described in the first aspect.

[0022] Fourthly, embodiments of the present invention provide a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0023] The embodiments of the present invention have the following advantages: In this embodiment of the invention, a plant identification method is introduced, comprising: acquiring detection data of the plant to be detected in a target area; the detection data includes radial dimension data, lateral deformation and displacement data, morphological image data, and rhizosphere soil physicochemical data of underground rhizomes; collecting four types of heterogeneous data in parallel, simultaneously covering three independent dimensions of root static structure, dynamic activity, and microenvironmental influence, breaking through the limitation of surface monitoring that only sees the above-ground parts and not the underground parts, and bringing underground latent rhizomes into the monitoring range. Based on the detection data, determining the root size and morphological characteristics, root growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plant to be detected; enabling subsequent identification to be based on more stable feature vectors, with the three types of features characterizing the root state from three independent perspectives of structure, activity, and environment, forming multiple verifications. Obtaining the feature benchmark of the target plant; by pre-establishing a three-dimensional feature benchmark system for the underground rhizomes of the target plant, the qualitative description is transformed into a programmable quantitative standard, giving the in-situ monitoring data a directly comparable scale, realizing the transformation from manual experience judgment to standardized machine comparison. Based on the root and stem size and morphological characteristics, root and stem growth rate characteristics, rhizosphere nutrient physicochemical characteristics, and characteristic benchmarks of the target plant, it is determined whether the plant in the target area is the target plant. Through a three-level progressive judgment, the false judgment rate is reduced compared to single-dimensional judgment, and the judgment does not completely rely on surface features. Even if the target plant is in a dormant state, it can still be accurately identified as long as the underground root and stem are active and match the environmental fingerprint. It can accurately identify dormant plants and ungerminated hidden patches in Spartina alterniflora, and is suitable for early warning and precise control of recurrence after Spartina alterniflora control. Attached Figure Description

[0024] Figure 1 This is a flowchart of the steps of a plant identification method provided in an embodiment of the present invention; Figure 2 This is a flowchart of another plant identification method provided in an embodiment of the present invention; Figure 3 This is a structural block diagram of a plant identification device provided in an embodiment of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below 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.

[0026] The terms "first," "second," etc., used in the specification and claims of this invention are used to distinguish similar objects and are not used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention can be implemented in orders other than those illustrated or described herein. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0027] The following description, in conjunction with the accompanying drawings, details a plant identification method, a plant identification device, an electronic device, and a computer-readable storage medium provided by the embodiments of the present invention through specific examples and application scenarios.

[0028] Reference Figure 1 The diagram illustrates a flowchart of a plant identification method provided by an embodiment of the present invention. The method may specifically include the following steps: Step 101: Obtain detection data for the plants to be tested in the target area; the detection data includes radial dimension data, lateral deformation displacement data, morphological image data, and rhizosphere soil physicochemical data of the underground rhizomes; Radial dimension data refers to the variation in thickness of the underground rhizome in the radial direction, used to characterize the diameter of the rhizome and its change over time. Lateral deformation and displacement data refers to the lateral positional movement of the underground rhizome in the horizontal plane due to growth and spread, used to characterize the lateral expansion dynamics of the rhizome. Radial dimension data and lateral deformation and displacement data together constitute the rhizome growth dynamic monitoring data, which can reflect the activity state and growth rate of the rhizome.

[0029] Morphological image data refers to the physical visual image data of underground rhizomes, which is used to characterize the appearance and morphological features of rhizomes, including but not limited to visual information such as the diameter of the rhizome, cross-sectional structure, internode distribution, node bud protrusion, fleshy texture and sprouting status.

[0030] Rhizosphere soil physicochemical data refers to the soil physicochemical parameters in the rhizosphere microenvironment of the plant to be tested, which are used to characterize the impact of root and stem activity on the surrounding soil environment, including but not limited to soil salinity, electrical conductivity, organic matter content, available nitrogen and phosphorus nutrient content, and soil respiration rate.

[0031] The above four types of detection data constitute the original data foundation for subsequent multi-dimensional coupling identification from three dimensions: static morphological characteristics of rhizomes, dynamic growth characteristics, and rhizosphere environment characteristics.

[0032] In one embodiment, the radial dimension data and the lateral deformation displacement data are acquired by a layered annular radial deformation sensor installed in the soil of the target area; Specifically, in the intertidal mudflat soil of the target area, layered annular diameter variation sensing units were deployed at three key soil depths of 10cm, 30cm, and 50cm below the surface. Each layered annular diameter variation sensing unit consists of a high-sensitivity annular micro-displacement sensor. The sensors are horizontally distributed around the circumference of the rootstock to be monitored, with the sensor probes in contact with the soil medium at the location of the rootstock. The sensors employ resistance strain gauge or capacitive micro-displacement sensing principles, with a measurement range of 0–5cm and a resolution of no less than 0.01mm.

[0033] During data collection, each sensor continuously records the radial thickness variation of the rootstock (i.e., radial dimension data) and the lateral deformation data of the rootstock in the horizontal plane (i.e., lateral deformation displacement data) at a sampling frequency of once per hour. The collection period is no less than 15 days to obtain continuous temporal dynamic data on rootstock growth. The morphological image data was acquired by a miniature root canal imaging unit installed in the soil of the target area. Specifically, micro-root canal imaging units are deployed at three soil depths, the same as those of the aforementioned diameter variation sensing unit. Each micro-root canal imaging unit comprises a transparent root canal and an embedded micro-multispectral root canal camera. The transparent root canal is made of high-transmittance quartz glass or polycarbonate, with a diameter of 2-3 cm, and is vertically embedded in the soil, with the outer surface of the canal wall in close contact with the surrounding soil.

[0034] The embedded miniature multispectral root canal camera is installed inside a transparent root canal. It can be driven by a stepper motor to move up and down along the root canal axis and can achieve 360° circumferential rotation for shooting via a rotating mechanism. The camera integrates a visible light imaging module and a near-infrared multispectral imaging module, with an image resolution of no less than 1920×1080 pixels.

[0035] During the data collection, the camera was positioned at three different soil depths to capture in situ images of the rhizome on the outer side of the root canal wall, obtaining visual morphological image data such as the diameter of the rhizome, internode structure, fleshy texture, and sprouting characteristics. The rhizosphere soil physicochemical data were collected by a gridded rhizosphere physicochemical sensor array set in the soil of the target area.

[0036] Specifically, a rhizosphere physicochemical sensor array is deployed in a grid pattern in the target area. The gridded rhizosphere physicochemical sensor array consists of multiple sensing nodes, each arranged in a 2m × 2m grid on a horizontal plane. Each node synchronously integrates the following sensors: a salinity sensor (range 0–100 ppt, accuracy ±1% FS), a conductivity sensor (range 0–200 mS / cm, accuracy ±1% FS), a soil organic matter sensor (using near-infrared reflectance spectroscopy, detection range 0–50%), and a nitrogen and phosphorus nutrient sensor (using ion-selective electrode method to detect NH4). + -N, NO3 - -N and PO4³ - -P content) and soil respiration sensor (using infrared CO2 detection method, range 0-5000ppm).

[0037] Each sensor probe is buried in the rhizosphere soil at a depth of 10-15 cm below the surface, with a distance of no more than 5 cm from the rootstock to be monitored, in order to accurately reflect the impact of rootstock activity on the rhizosphere microenvironment. All sensors synchronously collect data on rhizosphere soil salinity, electrical conductivity, organic matter content, available nitrogen and phosphorus content, and soil respiration rate every 6 hours.

[0038] Furthermore, all exposed components of the aforementioned layered annular diameter variable sensing unit, miniature root canal imaging unit, and gridded rhizosphere physicochemical sensing array are designed with moisture-proof and corrosion-resistant sealing: the outer shell is made of 316L stainless steel or engineering plastic, with a protection rating of IP68; electrical connections are sealed with epoxy resin potting; and sealing caps are added to both ends of the transparent root canal. These designs ensure that each device can operate stably for extended periods in the tidal flat environment characterized by repeated tidal submersion and high salt corrosion, supporting 24 / 7 unattended data acquisition.

[0039] Step 102: Based on the detection data, determine the root size and morphological characteristics, root growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plants to be tested in the target area. The morphological characteristics of the rhizomes refer to physical parameters that characterize the static geometric properties of underground rhizomes, including the radial thickness (i.e., diameter), the type of internal structure (solid or hollow cross-section), the length of internodes and the state of bud protrusion, the continuous distribution pattern of rhizomes on the horizontal plane, and the branching density of rhizomes. These morphological parameters can be extracted by performing image recognition processing on the morphological image data and verifying it in conjunction with the radial dimension data.

[0040] The root and stem growth rate characteristics refer to the time-series variation parameters characterizing the dynamic growth activity of underground roots and stems, including the change in radial dimension of the root and stem per unit time (radial daily growth) and the amount of horizontal extension displacement of the root and stem per unit time (lateral daily extension). By performing continuous time-series analysis on the radial dimension data and lateral deformation displacement data, the change rate within each time window can be calculated to obtain the above-mentioned growth rate parameters.

[0041] The rhizosphere nutrient physicochemical characteristics refer to multi-dimensional indicators characterizing the chemical and physical properties of the soil microenvironment surrounding the rhizomes, including soil salinity, electrical conductivity, organic matter content, available nitrogen and phosphorus nutrient concentrations, and soil respiration rate. By integrating multiple parameters and extracting features from the rhizosphere soil physicochemical data, a physicochemical fingerprint vector reflecting the specific rhizosphere environment of Spartina alterniflora can be constructed.

[0042] In one embodiment, step 102 may include the following sub-steps: Sub-step S11: Based on the morphological image data, determine the apparent diameter, cross-sectional structure, internode distribution morphology, and number of latent buds of the underground rhizome of the plant to be tested. The micro-root canal imaging unit acquires in-situ images of underground rhizomes at preset time intervals. The images are preprocessed, including image enhancement and noise filtering, and then the contour of the rhizome region is extracted using an image segmentation algorithm. Based on the contour, multiple measurement sections are selected along the rhizome extension direction, the diameter at each section is calculated, and the average value is taken as the apparent diameter. Texture analysis is performed on the rhizome cross-sectional images to distinguish between solid and hollow structures, obtaining the cross-sectional structural information. Internode boundaries are identified by detecting node elevation features on the rhizome surface, and the internode length distribution is statistically analyzed to obtain the internode distribution morphology. The number of latent buds is obtained by identifying bud elevations or germination traces in the node region and counting the number of buds per unit length of rhizome.

[0043] Sub-step S12: Determine the thickness grade of the underground rhizome to be tested based on the radial dimension data, the apparent diameter, and the cross-sectional structure. The layered annular diameter variation sensing unit continuously monitors the radial micro-displacement changes of the underground rhizomes in a time-series manner, outputting radial dimension data. This radial dimension data is then fused with the apparent diameter and cross-sectional structure data to determine the thickness grade of the underground rhizomes to be detected, according to a preset thickness grading rule. The thickness grades include at least three levels: coarse rhizome, medium rhizome, and fine rhizome. Specifically: the coarse rhizome grade corresponds to rhizomes with a large apparent diameter and a solid cross-section; the medium rhizome grade corresponds to rhizomes with a medium apparent diameter or a hollow cross-section; and the fine rhizome grade corresponds to rhizomes with a small apparent diameter or a fibrous root structure.

[0044] Preferably, the threshold for determining the coarseness grade can be adaptively adjusted according to the local vegetation distribution in the area to be monitored, but the adjustment range should not exceed 20% above or below the aforementioned threshold.

[0045] Sub-step S13: The cross-sectional structure, the internode distribution pattern, the number of latent buds, and the thickness grade are determined as the root and stem size and morphological characteristics of the underground rhizome of the plant to be tested.

[0046] The cross-sectional structure, internode distribution morphology, and number of latent buds are correlated and combined with the thickness grade to form the rootstock size and morphology characteristics of the underground rhizome of the plant under test. The rootstock size and morphology characteristics are represented in the form of feature vectors, which include at least four dimensions: cross-sectional structure identifier, internode length statistics, number of buds per unit length, and thickness grade identifier.

[0047] Experiments have verified that the rhizome size and morphological characteristics extracted using the above method can effectively distinguish Spartina alterniflora from native tidal flat plants: the cross-sectional structure of the underground creeping rhizomes of Spartina alterniflora is solid, with evenly distributed internodes and an average internode length usually between 1.5-3.0 cm, and the number of latent buds per unit length is usually 3-8 / 10 cm, with a thickness grade of coarse rhizomes; the cross-sectional structure of the rhizomes of Phragmites australis is hollow, with relatively large internode lengths but a significant hollow structure; the rhizomes of Curcuma longa are of the fine rhizome grade, with a sparse number of latent buds (usually ≤1 / 10 cm); Suaeda salsa and Scutellaria barbata do not have coarse creeping rhizomes, are of the fine rhizome grade, and have no obvious internode structure.

[0048] In one embodiment, step 102 may further include the following sub-steps: Sub-step S21: Determine the average daily radial growth of the underground rhizomes of the plant to be tested based on the radial dimension data of the underground rhizomes within the continuous time series. Specifically, the layered annular diameter variation sensing unit continuously collects radial dimension data of underground rhizomes at each key soil layer according to a preset sampling frequency (e.g., once every 1 to 6 hours). The radial dimension data records the continuous change of rhizome diameter over time, and this change reflects the radial growth or contraction of the rhizome. Using days as the statistical unit, the daily average value of the rhizome radial dimension is calculated within each statistical period, and the difference between the daily average values ​​of adjacent statistical periods is taken as the daily radial growth of that period; the arithmetic mean of the daily radial growth of multiple consecutive statistical periods (e.g., 7 to 15 days) is taken to obtain the daily radial growth.

[0049] Sub-step S22: Based on the lateral deformation displacement data within the continuous time series, determine the average daily lateral extension of the underground rhizomes of the plant to be tested. Specifically, the layered annular diameter variation sensing unit synchronously collects lateral deformation displacement data of the rootstock by monitoring the relative displacement of each monitoring point on the rootstock surface in the horizontal direction. This data reflects the change in the horizontal extension length of the rootstock. Using days as the statistical unit, the cumulative change in the lateral extension length of the rootstock within each statistical period is calculated as the daily lateral extension amount for that period; the arithmetic mean of the daily lateral extension amounts for multiple consecutive statistical periods (e.g., 7-15 days) is taken to obtain the average daily lateral extension amount.

[0050] Sub-step S23: The radial daily average growth and the lateral daily average extension are determined as the root growth rate characteristics of the underground rhizomes of the plant to be tested.

[0051] Specifically, the daily radial growth and daily lateral extension are correlated and combined to form a root and stem growth rate feature vector.

[0052] Experiments have verified that the rhizome growth rate characteristics extracted using the above method can effectively distinguish between Spartina alterniflora and native tidal flat plants: Spartina alterniflora has highly active fleshy rhizomes that grow continuously throughout the year, with a daily radial growth of 0.2-0.5 mm and a daily lateral extension of 0.8-2 cm in spring and summer, and still shows slight growth during dormancy; Reed rhizomes have a high degree of lignification and extremely low growth rate, with a daily radial growth of <0.03 mm and weak lateral spreading ability; Native fibrous rooted plants such as Curcuma longa have daily radial and lateral growth rates below the measurement noise threshold of the sensing unit (0.01 mm / day and 0.01 cm / day, respectively) at the monitoring points, and therefore cannot detect effective growth rate characteristics.

[0053] The rhizome growth rate characteristics extracted using the above method can effectively distinguish Spartina alterniflora from native tidal flat plants. Experiments show that, within the same monitoring period, the average daily radial growth and average daily lateral extension at Spartina alterniflora monitoring sites are significantly higher than those at monitoring sites for Phragmites australis and native fibrous root plants, demonstrating good species differentiation capabilities.

[0054] In one embodiment, the rhizosphere soil physicochemical data includes at least one of the following: salt content, organic matter content, nitrogen and phosphorus content, electrical conductivity, and soil respiration rate; step 102 may further include the following sub-steps: Sub-step S31: Based on at least one of the salt content, organic matter content, nitrogen and phosphorus content, electrical conductivity and soil respiration rate in the rhizosphere soil physicochemical data, determine the nutrient enrichment coefficient and rhizosphere environmental fingerprint of the underground rhizome rhizosphere region to be tested. Specifically, the gridded rhizosphere physicochemical sensor array is deployed at monitoring points in each key soil layer to continuously collect rhizosphere soil physicochemical data at a preset sampling frequency (e.g., once per hour or set according to the tidal cycle). The rhizosphere soil physicochemical data includes at least one of the following: salt content (which can be characterized by soil leachate conductivity), organic matter content, available nitrogen content, available phosphorus content, in-situ soil conductivity, and soil respiration rate (e.g., characterized by CO2 release rate).

[0055] The nutrient enrichment coefficient is used to quantify the degree of soil nutrient enrichment in the rhizosphere region of the tested plant. The enrichment coefficient for each nutrient index is calculated by comparing the content differences of corresponding nutrient indicators in the rhizosphere soil region (i.e., the soil immediately adjacent to the rhizome) and the non-rhizosphere background soil region (i.e., the soil in the same layer far from the roots and unaffected by the roots). Specifically, the enrichment coefficient is calculated by comparing the content value of each nutrient index in the rhizosphere region with the background value of the corresponding index in the non-rhizosphere region. A coefficient greater than 1 indicates an enrichment effect, while a coefficient less than 1 indicates a deficiency effect.

[0056] Preferably, a comprehensive value (e.g., arithmetic mean or weighted average) of the enrichment coefficients of multiple nutrient indicators can be calculated as a comprehensive nutrient enrichment coefficient to comprehensively evaluate the overall enrichment level of rhizosphere nutrients.

[0057] The rhizosphere environmental fingerprint is a joint feature vector composed of multiple rhizosphere soil physicochemical parameters, used to characterize the specific combination of physicochemical features of the rhizosphere microenvironment of a particular plant species. Multiple or all of the parameters, including salt content, organic matter content, available nitrogen content, available phosphorus content, electrical conductivity, and soil respiration rate, are combined in a preset order to form a feature vector, which serves as the rhizosphere environmental fingerprint for that monitoring point.

[0058] Preferably, to further enhance the species differentiation ability of the rhizosphere environmental fingerprint, the difference or rate of change of each physicochemical parameter between the rhizosphere soil region and the non-rhizosphere background soil region can be calculated, and the difference vector or rate of change vector can be used as a supplementary feature of the rhizosphere environmental fingerprint to characterize the degree of modification of the rhizosphere microenvironment by the plant.

[0059] Sub-step S32: The nutrient enrichment coefficient and the rhizosphere environmental fingerprint are determined as the rhizosphere nutrient physicochemical characteristics of the underground rhizosphere of the plant to be tested.

[0060] Specifically, the nutrient enrichment coefficients (including individual enrichment coefficients and / or comprehensive enrichment coefficients) and the rhizosphere environmental fingerprint vector are correlated and combined to form a rhizosphere nutrient physicochemical feature vector for subsequent multi-dimensional coupled identification model rhizosphere physicochemical coupling verification. This feature vector includes the measured values, enrichment coefficients, and derived indices of each of the aforementioned individual physicochemical parameters.

[0061] The aforementioned rhizosphere nutrient physicochemical characteristics can reflect the differentiated ability of different plant species to modify the rhizosphere microenvironment. For example, the rhizosphere of Spartina alterniflora typically exhibits a combination of high salt, high organic matter, high nitrogen and phosphorus accumulation, and high soil respiration rate, which is significantly different from the rhizosphere physicochemical characteristics of native plants such as Phragmites australis, Scutellaria barbata, and Suaeda salsa. Therefore, it can serve as an effective basis for distinguishing Spartina alterniflora from native plants.

[0062] Preferably, the nutrient enrichment coefficient can be calculated as follows: using the moving average value within a preset time window (e.g., 24 hours) as the stable representative value of each physicochemical parameter, the representative values ​​of the rhizosphere and non-rhizosphere regions are calculated separately, and then their ratio is obtained. The rhizosphere region is defined as the soil within a certain radial distance (e.g., 0–2 cm) from the central axis of the rootstock, and the non-rhizosphere region is defined as the same layer of soil at a greater distance (e.g., 10–15 cm) from the central axis of the rootstock and without root distribution. Preferably, the rhizosphere environmental fingerprint can include six core physicochemical parameters (salt content, organic matter, available nitrogen, available phosphorus, electrical conductivity, and soil respiration rate), and parameters can be added or removed adaptively according to the local vegetation type of the monitoring area.

[0063] The above preferred methods are merely examples. Those skilled in the art can make reasonable adjustments to the sampling area division, time window length, and parameter combination according to actual monitoring conditions and local plant species, all of which fall within the protection scope of this invention.

[0064] In one embodiment, the following steps may be included before step 102: Abnormal data caused by tidal and silt disturbances are removed from the detection data.

[0065] The abnormal data generated by the tidal disturbance includes: data loss due to the periodic flooding of the equipment by the tide or invalid data generated due to the acquisition environment exceeding the normal working conditions of the sensor, as well as measurement noise data generated by the high-frequency vibration of the tidal surge coupled to the sensing unit.

[0066] For invalid data caused by tidal flooding, such abnormal data is identified and removed according to the preset tidal flooding judgment rules. For measurement noise introduced by high-frequency vibration of tidal surges, the time-series data of the diameter variation sensing unit is filtered and denoised to eliminate the interference of high-frequency vibration noise on the root deformation trend signal. The filtered data eliminates false deformation signals caused by vibration and retains low-frequency trend information reflecting the true growth dynamics of the root.

[0067] The abnormal data generated by the silt disturbance include: quality defects such as partial occlusion, blurring, or severe lack of contrast in the acquired image data caused by silt particles adhering to the lens surface of the micro-root canal imaging unit; and systematic deviations in radial dimension measurement data caused by changes in the bonding medium between the annular diameter variation sensing unit and the rootstock due to silt deposition.

[0068] For image data, each frame undergoes a quality assessment, calculating a quality evaluation index that reflects image sharpness and effective information content. When an image's quality evaluation index falls below a preset quality threshold, the image is deemed to be affected by silt obstruction and is removed as abnormal data, not used as a basis for subsequent root and stem morphology feature extraction. When multiple consecutive frames are determined to be abnormal, a device cleaning prompt can be issued, notifying maintenance personnel to clean the lens, or the device's self-cleaning mechanism can be automatically triggered to perform a cleaning action. For diameter variation sensor data, trend change analysis is performed on time-series data to identify slow-drifting abnormal signals caused by silt deposition.

[0069] After the above-mentioned abnormal data removal process, the effective data of various detection data are comprehensively evaluated. When the amount of effective data meets the minimum data requirements for subsequent feature extraction, the preprocessed multi-source data is time-aligned to form an effective dataset with a consistent time benchmark, which is then input into the subsequent feature extraction module. When the amount of effective data is insufficient, a signal indicating insufficient data quality is generated to indicate that the data acquisition cycle needs to be extended or the equipment needs to be maintained.

[0070] Step 103: Obtain the characteristic benchmarks of the target plant; In this invention, the feature benchmark refers to a pre-established and stored reference dataset used for comparison with the measured features of the plant to be tested. The target plant is Spartina alterniflora, and the feature benchmark includes at least the standard rhizome size and morphology benchmark of Spartina alterniflora, the standard growth rate benchmark of Spartina alterniflora, and the standard rhizosphere physicochemical fingerprint benchmark of Spartina alterniflora.

[0071] Specifically, the characteristic benchmark is obtained in the following manner: In a target area known to be a pure Spartina alterniflora community, the layered annular diameter variation sensing unit, micro-root canal imaging unit, and gridded rhizosphere physicochemical sensing array are installed in the key soil layer according to the same layout as the aforementioned steps. Multi-source monitoring data covering at least one complete growth cycle are continuously collected, and the complete growth cycle includes at least the spring germination period, the summer vigorous growth period, and the autumn dormancy preparation period. Statistical analysis is performed on the collected multi-source monitoring data to extract typical characteristic parameters and their fluctuation ranges of the underground rhizomes of Spartina alterniflora in three dimensions: size and morphology, growth rate, and rhizosphere physicochemical environment. These typical characteristic parameters and their fluctuation ranges are stored as the characteristic benchmark.

[0072] As an alternative to the above implementation method, the feature benchmark can also be established by integrating and extracting the rhizomes and rhizosphere feature parameters of Spartina alterniflora from publicly available literature data. As a further improvement to the above implementation method, when the target monitoring area is a newly developed area, existing feature benchmarks of other similar tidal flat areas stored in the benchmark database are used as initial benchmarks; as monitoring data of the target monitoring area continues to accumulate, the initial benchmark is adaptively corrected according to the difference between the measured data and the initial benchmark to generate a unique feature benchmark for the area.

[0073] After the feature benchmark is established, it is stored in the benchmark database. In the subsequent identification process, the measured feature vector of the plant to be detected is extracted, and the measured feature vector is compared with the feature benchmark. The category determination result is determined based on the comparison result.

[0074] Step 104: Based on the root and stem size and morphological characteristics, root and stem growth rate characteristics, rhizosphere nutrient physicochemical characteristics, and characteristic criteria of the target plant, determine whether the plant to be tested in the target area is the target plant.

[0075] In one embodiment, the characteristic criteria of the target plant include the root and stem size and morphological characteristics, the root and stem growth rate characteristics, and the rhizosphere nutrient physicochemical characteristics. Step 104 may include the following sub-steps: Sub-step S41: Match the root and stem size and morphological features of the plant to be tested with the root and stem size and morphological features of the target plant to determine the root and stem morphological matching score of the plant to be tested. Specifically, a pre-stored baseline library of rhizome size and morphological features of the target plant (Spartina alterniflora) is invoked. During matching, the cross-sectional structure, internode distribution morphology, number of latent buds, and thickness grade of the plant to be tested are compared item by item with the above baselines. If a feature matches the baseline, that feature receives a matching score; if there is a deviation from the baseline, the score is deducted accordingly based on the degree of deviation; if the deviation exceeds the preset tolerance range, that feature receives no score. The scores of each item are weighted and summed to obtain the rhizome morphology matching score.

[0076] Preferably, since the morphological characteristics of rhizomes (especially cross-sectional structure and thickness grade) are the most direct and stable criteria for distinguishing Spartina alterniflora from native fibrous root plants, the weight of cross-sectional structure and thickness grade is higher than that of internode distribution morphology and number of latent buds in the weighted summation.

[0077] Sub-step S42: Match the root and stem growth rate characteristics of the plant to be tested with the root and stem growth rate characteristic benchmark of the target plant to determine the root and stem dynamic growth rate matching score of the plant to be tested. The growth rate characteristics of the target plant (Spartina alterniflora) were based on the following benchmarks: daily radial growth and daily lateral extension were typically high in spring and summer, while during dormancy (winter), they were typically low but still showed slight growth, exhibiting a dynamic characteristic of continuous growth throughout the year. During matching, the daily radial growth and daily lateral extension of the rhizome to be tested were compared item by item with the benchmark range for the corresponding season. If the measured growth rate fell within the benchmark range, it indicated that the growth activity of the tested rhizome matched that of Spartina alterniflora, and this item received a matching score; if it was below the lower limit of the benchmark range, the score was linearly calculated based on the closeness to the lower limit; if it was significantly below the lower limit (indicating that the rhizome had virtually no growth activity), no score was awarded for this item. The scores for each item were weighted and summed to obtain the rhizome dynamic growth rate matching score.

[0078] Sub-step S43: Match the rhizosphere nutrient physicochemical characteristics with the rhizosphere nutrient physicochemical characteristic benchmark of the target plant to determine the rhizosphere physicochemical fingerprint matching score of the plant to be tested. The rhizosphere physicochemical characteristics benchmark of the target plant (Spartina alterniflora) was established based on a large amount of measured data, including typical fingerprint information such as high salt, high organic matter, and high nitrogen and phosphorus enrichment in the rhizosphere soil, and significantly higher root zone electrical conductivity and soil respiration rate than the surrounding native vegetation areas. During matching, the values ​​of each physicochemical parameter at the test site were compared item by item with the above benchmark. If a parameter value falls within the benchmark range, it indicates that the parameter matches the rhizosphere environmental fingerprint of Spartina alterniflora, and that item receives a matching score; if the measured value deviates from the benchmark range, the score is linearly calculated based on the degree of deviation; if the degree of deviation exceeds the preset tolerance range, that item receives no score. The scores of each item are weighted and summed to obtain the rhizosphere physicochemical fingerprint matching score.

[0079] Sub-step S44: Determine the comprehensive matching score of the plant under test based on the matching score of the root and stem morphology, the matching score of the root and stem dynamic growth rate, and the matching score of the rhizosphere physicochemical fingerprint. The overall matching score is the weighted sum of the three matching scores mentioned above. Among them, the root and stem morphology matching score, as the most stable and intuitive core discrimination criterion, is assigned the highest weight coefficient; the root and stem dynamic growth rate matching score, as a dynamic verification factor reflecting root and stem activity, is assigned the second highest weight coefficient; and the rhizosphere physicochemical fingerprint matching score, as an environmental auxiliary verification factor, is assigned a lower weight coefficient.

[0080] Preferably, the above weight allocation can be adjusted according to the monitoring scenario in practical applications. For example, in scenarios where the growth rate of Spartina alterniflora is significantly reduced during the dormant period (winter), the weight of the growth rate matching score can be reduced accordingly, while the weight of the morphology matching score can be increased. In scenarios with frequent tidal disturbances and large fluctuations in physicochemical sensor data, the weight of the physicochemical fingerprint matching score can be reduced accordingly, while the weight of the morphology matching score can be increased.

[0081] In sub-step S45, if the comprehensive matching score of the plant to be detected is greater than or equal to a preset threshold, then the plant to be detected is determined to be the target plant.

[0082] The preset threshold is the minimum score required to determine whether the overall matching degree meets the criteria for identifying Spartina alterniflora. When the overall matching degree score is greater than or equal to the threshold, it is determined that there are Spartina alterniflora underground plants or latent patches at the monitoring point, and the identification result of Spartina alterniflora is output, along with the spatial coordinates of the point. When the overall matching degree score is lower than the threshold, it is determined that there are no Spartina alterniflora underground plants at the monitoring point, and the identification result of not having Spartina alterniflora is output.

[0083] Preferably, the preset threshold is determined by statistical analysis of the recognition results of known samples, so as to ensure recognition sensitivity (reducing the false negative rate) while taking into account recognition specificity (reducing the false positive rate).

[0084] In a preferred embodiment of the present invention, the comprehensive matching degree determination and the three-level progressive identification model work together in stages: Level 1 (initial screening based on morphological threshold): Only the root and stem morphology matching score is calculated. If this score is lower than a first preset threshold, it indicates that the plant to be tested does not have a robust creeping root and stem structure, and is directly identified as not *Spartina alterniflora*, terminating the identification process without triggering subsequent data collection and calculation, thus reducing system energy consumption. Level 2 (refined screening based on growth rate): If the root and stem morphology matching score reaches the first preset threshold, the growth rate matching score is further calculated. If both the morphology matching score and the growth rate matching score reach their respective preset thresholds, it indicates that the plant to be tested simultaneously possesses the typical morphological structure and growth activity of *Spartina alterniflora*, and can be identified as *Spartina alterniflora* with high confidence, without needing to proceed to the third level of verification. Level 3 (Rhizosphere Physicochemical Coupling Verification): If the plant under test passes the initial screening at Level 1 but fails to meet the high-confidence criteria at Level 2 (e.g., morphological characteristics deviate from the typical morphology of Spartina alterniflora, or the plant is in a dormant period during monitoring, resulting in a low growth rate), then the rhizosphere physicochemical fingerprint matching score is further calculated, and the three matching scores are weighted and fused to calculate the comprehensive matching degree. The comprehensive matching degree score is compared with a preset threshold to make the final judgment.

[0085] The above-mentioned three-level progressive judgment strategy effectively reduces computational overhead while ensuring recognition accuracy. Approximately 60% of the local plant sites without thick roots are excluded at the first level, eliminating the need to trigger subsequent continuous time-series data acquisition and physicochemical analysis, thus significantly improving the energy efficiency and data processing efficiency of the monitoring system.

[0086] In one embodiment, the method may further include the following steps: Determine the location coordinates of the target plant.

[0087] The spatial location information of each monitoring point can be provided by a positioning module deployed in the monitoring area. In a specific implementation, the monitoring points are deployed in the intertidal zone to be monitored according to a preset grid spacing (e.g., 5m×5m or 10m×10m). During deployment, the positioning module records the spatial coordinates of each monitoring point and physically binds and logically associates them with each sensing unit (layered annular diameter variation sensing unit, micro-root canal imaging unit, gridded rhizosphere physicochemical sensing array) at that point, ensuring that all subsequent multi-source data have clear spatial coordinate attributes.

[0088] In another preferred embodiment, for monitoring areas where fixed monitoring base stations have been established, real-time dynamic positioning (RTK) technology can be used to calibrate the coordinates of each monitoring point to further improve spatial positioning accuracy.

[0089] When the multi-dimensional coupled recognition model determines that there are Spartina alterniflora underground plants or latent patches at a certain monitoring point (i.e., the comprehensive matching score is ≥ the preset threshold), the spatial coordinate information of that point is bound and associated with the recognition result. Specifically, for each monitoring point identified as the target plant, a record containing the following information is generated: point number, longitude coordinates, latitude coordinates, recognition result ("Spartina alterniflora"), comprehensive matching score, and recognition timestamp.

[0090] In a specific implementation, if a monitoring point is determined to be a non-target plant (i.e., there are no Spartina alterniflora underground plants), the spatial coordinate information of that point will only be used as background data storage and will not be included in the target plant register.

[0091] The associated spatial coordinates of the target plants are output in a preset format to form a visual monitoring log or electronic map marker. Specifically, in one embodiment, the location coordinates are output in tabular form, which includes the latitude and longitude coordinates of each monitoring point identified as a target plant and the corresponding identification confidence information, so that management personnel can conduct on-site verification and targeted treatment based on the coordinate data.

[0092] In another implementation, the location coordinates are output in a Geographic Information System (GIS) compatible format, which can be directly imported into an electronic map platform. The spatial distribution of latent plants under Spartina alterniflora is displayed as marker points on a digital map, providing visual data support for targeted control and delineation of spread boundaries.

[0093] In a preferred embodiment of the present invention, after all monitoring points have been identified, the coordinates of all points identified as target plants are summarized to generate a spatial distribution map of latent underground plants in Spartina alterniflora. The spatial distribution map uses the monitoring area map as a background, and the target plant points are marked with preset markers (such as red dots) at their corresponding coordinate positions. Furthermore, the confidence level can be distinguished by markers of different colors or sizes based on the overall matching score.

[0094] This spatial distribution map can be directly output to mobile terminals or handheld positioning devices, allowing on-site management personnel to navigate to target locations in the tidal flat environment based on coordinates, thus achieving precise location and targeted removal of underground plants in Spartina alterniflora.

[0095] This invention provides a method for plant identification, comprising: acquiring detection data of plants to be detected in a target area; the detection data includes radial dimension data, lateral deformation and displacement data, morphological image data, and rhizosphere soil physicochemical data of underground rhizomes; collecting these four types of heterogeneous data in parallel, simultaneously covering three independent dimensions: static structure, dynamic activity, and microenvironmental influence of the rhizomes, overcoming the limitation of surface monitoring that only sees the above-ground parts and not the underground parts, and bringing underground latent rhizomes into the monitoring range. Based on the detection data, determining the rhizome size and morphological characteristics, rhizome growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plants to be detected; enabling subsequent identification to be based on more stable feature vectors, with the three types of features characterizing the rhizome state from three independent perspectives of structure, activity, and environment, forming multiple verifications. Obtaining the characteristic benchmark of the target plant; by pre-establishing a three-dimensional feature benchmark system for the underground rhizomes of the target plant, qualitative descriptions are transformed into programmable quantitative standards, giving in-situ monitoring data a directly comparable scale, realizing the transformation from manual experience-based judgment to standardized machine comparison. Based on the root and stem size and morphological characteristics, root and stem growth rate characteristics, rhizosphere nutrient physicochemical characteristics, and characteristic benchmarks of the target plant, it is determined whether the plant in the target area is the target plant. Through a three-level progressive judgment, the false judgment rate is reduced compared to single-dimensional judgment, and the judgment does not completely rely on surface features. Even if the target plant is in a dormant state, it can still be accurately identified as long as the underground root and stem are active and match the environmental fingerprint. It can accurately identify dormant plants and ungerminated hidden patches in Spartina alterniflora, and is suitable for early warning and precise control of recurrence after Spartina alterniflora control.

[0096] Reference Figure 2The diagram illustrates a flowchart of another plant identification method provided by an embodiment of the present invention. The method may specifically include the following steps: S1. Monitoring point deployment: A layered sensor hardware system is deployed in a grid pattern in the tidal flat area to be monitored, and equipment calibration and data clearing are completed. The hardware adopts a layered modular layout structure, adapting to the three key soil layers of the intertidal zone (10cm, 30cm, and 50cm), covering the main distribution range of Spartina alterniflora rhizomes (15-35cm), while also being compatible with the root distribution depth of native plants, ensuring the accuracy of comparison and identification. S2. Multi-source data synchronous acquisition: synchronously acquire rhizome morphology images, diameter deformation growth data, and rhizosphere soil physicochemical nutrient data around the clock; Layered ring diameter deformation sensing unit: High-sensitivity ring micro-displacement sensors are deployed in three soil layers to fit the soil medium and non-invasively monitor the radial thickness and lateral displacement deformation data of underground rhizomes, capturing the rhizome diameter and lateral spreading growth dynamics. Miniature root canal imaging unit: Equipped with an embedded miniature high-definition multispectral root canal camera, it can acquire images of underground rhizomes in situ and obtain visual information such as rhizome morphology, internode structure, fleshy texture, and sprouting characteristics. Gridded rhizosphere physicochemical sensor array: Simultaneously deploy sensors for salinity, electrical conductivity, soil organic matter, nitrogen and phosphorus nutrients, and soil respiration to collect rhizosphere microenvironment physicochemical data in real time and construct a rhizosphere environment fingerprint database specific to Spartina alterniflora; All hardware features a moisture-proof and corrosion-resistant sealed design, making it suitable for tidal flat environments with repeated tidal submersion and high salt corrosion, and supporting unattended data acquisition around the clock.

[0097] S3. Feature extraction and data preprocessing: Remove abnormal data caused by tides and silt disturbance, and extract three types of core discriminative features; Data was collected through a hardware system, and three types of unique features were extracted to form a differential discrimination standard between Spartina alterniflora and native plants. The specific thresholds are as follows: Rhizome thickness and morphological characteristics (static core discriminant factor): Spartina alterniflora has robust, solid, creeping rhizomes with a stable diameter of 0.6-1.8 cm, distinct segmentation, prominent node buds, and a lateral, continuous distribution; Reed rhizomes are hollow and lignified, with a relatively large diameter but thin walls and a prominent hollow structure; Curcuma longa has short, tuberous rhizomes with a diameter of only 0.2-0.4 cm, much smaller than Spartina alterniflora, with clustered stem bases and extremely short lateral creeping extension, failing to form a large-scale continuous network of roots, short internodes, and a sparse number of latent buds; Suaeda salsa and Scutellaria barbata lack thick creeping rhizomes, with root diameters all <5 mm, exhibiting a fine, fibrous root structure. This dimension accurately acquires rhizome diameter, cross-sectional structure, and distribution morphology through imaging recognition and diameter variation sensing.

[0098] Rhizome growth rate characteristics (dynamic core discriminant factor): *Spartina alterniflora* has highly active fleshy rhizomes, growing continuously throughout the year. In spring and summer, its daily radial growth is 0.2-0.5 mm, and its daily lateral extension is 0.8-2 cm; it still shows slight growth during dormancy. *Phragmites australis* has highly lignified rhizomes with extremely low growth rates, with daily radial growth <0.03 mm and weak lateral spreading ability. Native fibrous-rooted plants such as *Scirpus triqueter* show no obvious rhizome growth deformation characteristics. This dimension calculates dynamic growth parameters using continuous time-series deformation data.

[0099] Rhizosphere nutrient physicochemical characteristics (environmental verification discriminant factors): *Spartina alterniflora* rhizomes secrete specific allelochemical organic acids; its rhizosphere soil is high in salt, organic matter, and nitrogen and phosphorus; and its root zone electrical conductivity and soil respiration rate are significantly higher than those of the surrounding native vegetation. *Phragmites australis* rhizosphere contains abundant lignified residues, and its organic matter composition differs significantly from that of *Spartina alterniflora*. *Scirpus triqueter* prefers low- to medium-salinity tidal flats; its rhizosphere soil has high salt, organic matter, and available nitrogen and phosphorus content, generally falling between that of *Suaeda salsa* and *Phragmites australis*, and it does not secrete characteristic organic acids. *Suaeda salsa* fibrous roots have low nutrient enrichment in the root zone, and its physicochemical indicators fluctuate less. This dimension constructs a rhizosphere environmental fingerprint using a multi-parameter sensor array.

[0100] S4. Multi-dimensional model coupling recognition: Input preprocessed data into the trained AI model to complete vegetation category identification and latent patch location; This invention constructs a three-level progressive recognition model to solve the problems of misjudgment of mixed vegetation and missed detection of hidden patches. Specific steps include: Level 1: Initial screening based on morphological thresholds. Based on rhizome thickness, cross-sectional structure, and distribution, plants without thick rhizomes, such as *Suaeda salsa*, *Cynanchum paniculatum*, and *Scirpus triquetrum*, are eliminated, and monitoring sites with thick rhizomes are selected. Level 2: Refined screening based on growth rate. For the initially screened sites, combined with 7-15 days of continuous time-series growth data, the high-speed, continuous growth characteristics of *Spartina alterniflora* are distinguished from the low-speed, steady-state growth characteristics of *Phragmites australis*, completing the initial classification. Level 3: Rhizosphere physicochemical coupling verification. Through multi-indicator fingerprint matching of rhizosphere salinity, organic matter, nutrients, and electrical conductivity, the identification errors of morphology and growth rate are corrected, resolving the cross-interference problem in mixed areas of *Spartina alterniflora* with native plants such as *Phragmites australis* and *Scirpus triquetrum*. Finally, through multi-factor weighted scoring, the identification results are output. When the overall matching degree is ≥92%, it is determined to be an underground plant / latent patch of *Spartina alterniflora*.

[0101] This invention employs a normalized weighted summation of three core feature factors to calculate the comprehensive matching degree. Each factor is scientifically weighted according to its identification weight, and the total weight sum is 1. The formula is as follows:

[0102] in: The score represents the matching score of rhizome morphology (0-100 points), with a weight of 0.4; the score is normalized based on rhizome diameter, solid structure, internode characteristics, and continuous spreading morphology, which are inherent hard features of the species and have the highest priority. The score represents the matching score of the dynamic growth rate of the rhizome (0-100 points), with a weight of 0.35. The score is normalized based on the continuous radial growth over 7-15 days, the persistence of lateral extension, and the activity characteristics during dormancy, and is used to distinguish plant activity from species growth characteristics. The rhizosphere physicochemical fingerprint matching score (0-100 points) is represented by a weight of 0.25. The Euclidean distance similarity matching score based on multidimensional indicators of soil salinity, electrical conductivity, organic matter, and soil respiration is used as an environmental verification anti-interference factor.

[0103] The overall matching score ranges from 0 to 100. An optimal discrimination threshold is set when... If the score is below 92, it is considered an active underground plant / latent dormant patch of Spartina alterniflora; if the score is below 92, it is considered a non-invasive plant such as reed, sea sedge, or native herbs.

[0104] S5. Results Output and Early Warning: Mark active rhizome points and dormant patches of Spartina alterniflora to generate accurate monitoring records and provide coordinate data for targeted control.

[0105] This invention, through multi-dimensional feature coupling verification, reduces the false judgment rate by more than 80% compared to single-sensor monitoring technology, and can accurately distinguish between Spartina alterniflora and various native tidal flat vegetation; it can simultaneously identify surface-grown plants, underground active rhizomes, and dormant residual roots, filling the gap in hidden patch monitoring technology; its modular hardware design makes it easy to deploy and reusable, adaptable to various silty and sandy tidal flat scenarios along the coast, eliminating the need for frequent manual inspections; it can accurately locate invasion points and spread boundaries, providing core data support for zoned targeted treatment, reducing waste of pesticides and manpower, and lowering recurrence rates, thus meeting the needs of ecological and refined wetland management.

[0106] It should be noted that the plant identification method provided in this embodiment of the invention can be executed by a plant identification device or a control module within that plant identification device for executing the method of loading plant identification. This embodiment of the invention uses the execution of the method of loading plant identification by a plant identification device as an example to illustrate the plant identification method provided in this embodiment of the invention.

[0107] Reference Figure 3 The diagram shows a structural block diagram of a plant identification device provided in an embodiment of the present invention, which may specifically include the following modules: The detection data acquisition module 301 is used to acquire detection data of the plants to be detected in the target area; the detection data includes radial dimension data of underground rhizomes, lateral deformation displacement data, morphological image data and rhizosphere soil physicochemical data; The feature determination module 302 is used to determine the root size and morphological characteristics, root growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plant to be detected in the target area based on the detection data. The feature reference acquisition module 303 is used to acquire the feature reference of the target plant. The target plant determination module 304 is used to determine whether the plant to be detected in the target area is the target plant based on the root and stem size and morphological characteristics, root and stem growth rate characteristics, rhizosphere nutrient physicochemical characteristics, and the characteristic criteria of the target plant.

[0108] In one embodiment, the characteristic criteria of the target plant include the root and stem size and morphological characteristics, the root and stem growth rate characteristics, and the rhizosphere nutrient physicochemical characteristics. The target plant identification module includes: The root and stem morphology matching score determination submodule is used to match the root and stem size and morphology features of the plant to be tested with the root and stem size and morphology feature benchmark of the target plant to determine the root and stem morphology matching score of the plant to be tested. The root and stem dynamic growth rate matching score determination submodule is used to match the root and stem growth rate characteristics of the plant to be tested with the root and stem growth rate characteristic benchmark of the target plant to determine the root and stem dynamic growth rate matching score of the plant to be tested. The rhizosphere physicochemical fingerprint matching score determination submodule is used to match the rhizosphere nutrient physicochemical characteristics with the rhizosphere nutrient physicochemical characteristic benchmark of the target plant to determine the rhizosphere physicochemical fingerprint matching score of the plant to be tested. The comprehensive matching score determination submodule is used to determine the comprehensive matching score of the plant under test based on the matching score of the root and stem morphology, the matching score of the root and stem dynamic growth rate, and the matching score of the rhizosphere physicochemical fingerprint. The target plant determination submodule is used to determine the plant to be detected as the target plant if the comprehensive matching score of the plant to be detected is greater than or equal to a preset threshold.

[0109] In one embodiment, the radial dimension data and the lateral deformation displacement data are acquired by a layered annular radial deformation sensor installed in the soil of the target area; The morphological image data was acquired by a miniature root canal imaging unit installed in the soil of the target area. The rhizosphere soil physicochemical data were collected by a gridded rhizosphere physicochemical sensor array set in the soil of the target area.

[0110] In one embodiment, the feature determination module includes: The first determining submodule is used to determine the apparent diameter, cross-sectional structure, internode distribution morphology, and number of latent buds of the underground rhizome of the plant to be detected based on the morphological image data. The second determining submodule is used to determine the thickness grade of the underground rhizome to be tested based on the radial dimension data, the apparent diameter, and the cross-sectional structure. The rootstock size and morphology characteristic determination submodule is used to determine the cross-sectional structure, the internode distribution pattern, the number of latent buds, and the thickness grade as the rootstock size and morphology characteristics of the underground rootstock of the plant to be tested.

[0111] In one embodiment, it further includes: The third determining submodule is used to determine the average daily radial growth of the underground rhizome of the plant to be tested based on the radial dimension data of the underground rhizome in a continuous time series. The fourth determination submodule is used to determine the average daily lateral extension of the underground rhizomes of the plant under test based on the lateral deformation displacement data in a continuous time series. The root and stem growth rate characteristic determination submodule is used to determine the radial average daily growth and the lateral average daily extension as the root and stem growth rate characteristics of the underground root and stem of the plant to be tested.

[0112] In one embodiment, the rhizosphere soil physicochemical data includes at least one of the following: salt content, organic matter content, nitrogen and phosphorus content, electrical conductivity, and soil respiration rate. The feature determination module further includes: The fifth determination submodule is used to determine the nutrient enrichment coefficient and rhizosphere environmental fingerprint of the underground rhizome rhizosphere region to be tested based on at least one of the salt content, organic matter content, nitrogen and phosphorus content, electrical conductivity and soil respiration rate in the rhizosphere soil physicochemical data. The rhizosphere nutrient physicochemical characteristics determination submodule is used to determine the nutrient enrichment coefficient and the rhizosphere environmental fingerprint as the rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plant to be tested.

[0113] In one embodiment, it further includes: The anomaly removal module is used to remove abnormal data caused by tidal and silt disturbances from the detection data.

[0114] In one embodiment, after determining whether the plant to be detected in the target area is a target plant, the method further includes: The coordinate determination module is used to determine the position coordinates of the target plant.

[0115] The plant identification device in this embodiment of the invention can be a device, or a component, integrated circuit, or chip in a terminal. The device can be a mobile electronic device or a non-mobile electronic device. For example, mobile electronic devices can be mobile phones, tablets, laptops, PDAs, in-vehicle electronic devices, wearable devices, ultra-mobile personal computers (UMPCs), netbooks, or personal digital assistants (PDAs), etc., while non-mobile electronic devices can be servers, network attached storage (NAS), personal computers (PCs), televisions (TVs), ATMs, or self-service machines, etc. This embodiment of the invention does not impose specific limitations.

[0116] The plant identification device in this embodiment of the invention can be a device with an operating system. This operating system can be Android, iOS, or other possible operating systems; this embodiment of the invention does not impose specific limitations.

[0117] The plant identification device provided in this embodiment of the invention can achieve Figures 1 to 3 The various processes implemented by the plant identification device in the method embodiment will not be described again here to avoid repetition.

[0118] In this embodiment of the invention, a plant identification method is introduced, comprising: a detection data acquisition module, used to acquire detection data of the plant to be detected in a target area; the detection data includes radial dimension data, lateral deformation and displacement data, morphological image data, and rhizosphere soil physicochemical data of underground rhizomes; the four types of heterogeneous data are collected in parallel, simultaneously covering three independent dimensions of root static structure, dynamic activity, and microenvironmental influence, breaking through the limitation of surface monitoring that only sees the above-ground parts and not the underground parts, so that the underground latent rhizomes are included in the monitoring range. A feature determination module is used to determine the root size and morphological characteristics, root growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plant to be detected based on the detection data; so that subsequent identification is based on more stable feature vectors, and the three types of features characterize the root state from three independent perspectives of structure, activity, and environment, forming multiple verifications. The feature benchmark acquisition module is used to acquire the feature benchmarks of target plants. By pre-establishing a three-dimensional feature benchmark system for the underground rhizomes of target plants, qualitative descriptions are transformed into programmable quantitative standards, providing a directly comparable benchmark for in-situ monitoring data and realizing the transformation from manual experience-based judgment to standardized machine comparison. The target plant identification module is used to determine whether the plants to be detected in the target area are target plants based on the size and morphological characteristics of the rhizomes, the growth rate of the rhizomes, the physicochemical characteristics of the rhizosphere nutrients, and the feature benchmarks of the target plants. Through a three-level progressive judgment, the false judgment rate is reduced compared to single-dimensional judgment, and the judgment does not completely depend on surface features. Even if the target plant is in a dormant state, it can still be accurately identified as long as the underground rhizomes are active and match the environmental fingerprint. It can accurately identify underground latent plants and ungerminated hidden patches in Spartina alterniflora, and is suitable for early warning and precise control of recurrence after Spartina alterniflora control.

[0119] This invention also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described plant identification method embodiments and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0120] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0121] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of the present invention is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0122] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0123] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of the present invention.

Claims

1. A method for identifying a plant, characterized in that, include: Acquire detection data of the plants to be tested in the target area; the detection data includes radial dimension data, lateral deformation and displacement data, morphological image data, and rhizosphere soil physicochemical data of the underground rhizomes; Based on the detection data, the root size and morphological characteristics, root growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plants to be tested in the target area are determined. Obtain the characteristic benchmarks of the target plant; Based on the root and stem size and morphological characteristics, root and stem growth rate characteristics, rhizosphere nutrient physicochemical characteristics, and characteristic criteria of the target plant, it is determined whether the plant to be tested in the target area is the target plant.

2. The plant identification method according to claim 1, characterized in that, The characteristic criteria of the target plant include the root and stem size and morphology characteristics, root and stem growth rate characteristics, and rhizosphere nutrient physicochemical characteristics. The step of determining whether a plant in the target area is a target plant based on the rootstock size and morphological characteristics, rootstock growth rate characteristics, rhizosphere nutrient physicochemical characteristics, and characteristic criteria of the target plant includes: The root and stem size and morphological characteristics of the plant to be tested are matched with the root and stem size and morphological characteristics of the target plant to determine the root and stem morphological matching score of the plant to be tested. The root and stem growth rate characteristics of the plant to be tested are matched with the root and stem growth rate characteristics benchmark of the target plant to determine the root and stem dynamic growth rate matching score of the plant to be tested. The rhizosphere nutrient physicochemical characteristics are matched with the rhizosphere nutrient physicochemical characteristics benchmark of the target plant to determine the rhizosphere physicochemical fingerprint matching score of the plant to be tested. The comprehensive matching score of the plant under test is determined based on the matching score of the root and stem morphology, the matching score of the root and stem dynamic growth rate, and the matching score of the rhizosphere physicochemical fingerprint. If the overall matching score of the plant to be tested is greater than or equal to a preset threshold, then the plant to be tested is determined to be the target plant.

3. The plant identification method according to claim 1, characterized in that, The radial dimension data and the lateral deformation displacement data were acquired by a layered annular radial deformation sensor installed in the soil of the target area. The morphological image data was acquired by a miniature root canal imaging unit installed in the soil of the target area. The rhizosphere soil physicochemical data were collected by a gridded rhizosphere physicochemical sensor array set in the soil of the target area.

4. The plant identification method according to claim 1, characterized in that, The step of determining the rootstock size and morphological characteristics, rootstock growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plants to be tested in the target area based on the detection data includes: Based on the morphological image data, the apparent diameter, cross-sectional structure, internode distribution morphology, and number of latent buds of the underground rhizome of the plant to be tested are determined. Based on the radial dimension data, the apparent diameter, and the cross-sectional structure, the thickness grade of the underground rhizome to be tested is determined; The cross-sectional structure, the internode distribution pattern, the number of latent buds, and the thickness grade are determined as the root and stem size and morphological characteristics of the underground rhizome of the plant to be tested.

5. The plant identification method according to claim 1, characterized in that, The step of determining the rootstock size and morphological characteristics, rootstock growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plants to be tested in the target area based on the detection data also includes: Based on the radial dimension data of the underground rhizomes within a continuous time series, the average daily radial growth of the underground rhizomes of the plant under test is determined. Based on the lateral deformation displacement data within a continuous time series, the average daily lateral extension of the underground rhizomes of the plant under test is determined. The radial daily average growth and the lateral daily average extension are determined as the root growth rate characteristics of the underground rhizomes of the plant under test.

6. The plant identification method according to claim 1, characterized in that, The rhizosphere soil physicochemical data include at least one of the following: salt content, organic matter content, nitrogen and phosphorus content, electrical conductivity, and soil respiration rate. The step of determining the rootstock size and morphological characteristics, rootstock growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plants to be tested in the target area based on the detection data also includes: Based on at least one of the following in the rhizosphere soil physicochemical data: salt content, organic matter content, nitrogen and phosphorus content, electrical conductivity, and soil respiration rate, determine the nutrient enrichment coefficient and rhizosphere environmental fingerprint of the underground rhizome rhizosphere region to be tested; The nutrient enrichment coefficient and the rhizosphere environmental fingerprint are used to determine the rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plant under test.

7. The plant identification method according to claim 1, characterized in that, Before determining the rootstock size and morphological characteristics, rootstock growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the target plant in the target area based on the detection data, the method further includes: Abnormal data caused by tidal and silt disturbances are removed from the detection data.

8. The plant identification method according to claim 1, characterized in that, After determining whether the plant to be detected in the target area is the target plant, the method further includes: Determine the location coordinates of the target plant.

9. A plant identification device, characterized in that, include: The detection data acquisition module is used to acquire detection data of the plants to be tested in the target area; the detection data includes radial dimension data of underground rhizomes, lateral deformation and displacement data, morphological image data, and rhizosphere soil physicochemical data; The feature determination module is used to determine the root size and morphological characteristics, root growth rate characteristics, and rhizosphere nutrient physicochemical characteristics of the underground rhizomes of the plant to be detected in the target area based on the detection data. The feature benchmark acquisition module is used to acquire the feature benchmarks of the target plant. The target plant determination module is used to determine whether the plant to be detected in the target area is the target plant based on the root and stem size and morphological characteristics, root and stem growth rate characteristics, rhizosphere nutrient physicochemical characteristics, and the characteristic criteria of the target plant.

10. An electronic device, characterized in that, It includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the plant identification method as described in claims 1-8.

11. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the plant identification method as described in claims 1-8.

Images