A plant growth data monitoring system based on the Internet of Things

Abstract

The present invention discloses a plant growth data monitoring system based on the Internet of Things, which relates to the field of plant growth monitoring technology. The system obtains spectral characteristic data of each canopy layer to obtain structural characteristic data of the canopy layer. The calculation module substitutes the structural characteristic data and spectral characteristic data into a fusion model, outputs the health convergence of K canopies, and weightedly calculates the health convergence of the K canopies to obtain the overall health index of the plant. The model construction and display module performs sliding window identification on the overall health index of the plant, predicts the growth trend of the plant, and uses LiDAR point cloud data to construct a layered canopy model for display. The monitoring system determines the overall growth health status of the plant after performing multi-level canopy analysis on the plant, and performs time series analysis on the plant to predict the growth trend of the plant, which is conducive to accurately controlling the environmental conditions required by the plant and effectively identifying the early symptoms of the disease.

Description

Technical Field

[0001] The present invention relates to the technical field of plant growth monitoring, and in particular to a plant growth data monitoring system based on the Internet of Things. Background Art

[0002] Plants are affected by environmental factors (such as light, temperature, humidity, soil nutrients, etc.) during their growth process. Accurate monitoring of plant growth status is of great significance to agriculture, forestry, ecological environment protection and scientific research. With the development of science and technology, plant growth monitoring has gradually evolved from traditional manual observation to automation and intelligence, improving the accuracy and efficiency of data collection.

[0003] The existing technology has the following defects:

[0004] Existing monitoring systems usually adopt a unified monitoring method for plant growth areas, including image and environmental data analysis of the growth area to determine the overall growth status of plants in the area. For some high-value plants (such as red sandalwood, Thuja thuja and Phoebe nanmu), since these high-value plants have strict environmental requirements, if a unified monitoring method is adopted for the growth area, it may not be possible to accurately control the environmental conditions required by individuals, and it will be difficult to identify the early symptoms of disease in individual plants in a timely manner. The disease may not be detected until it spreads, increasing the cost and risk of treatment.

[0005] Based on this, the present invention proposes a plant growth data monitoring system based on the Internet of Things. After performing multi-level crown analysis on the plants, the overall growth health status of the plants is judged, and time series analysis of the plants is performed to predict the growth trends of the plants, which is conducive to accurately controlling the environmental conditions required by the plants and effectively identifying the early symptoms of diseases. Summary of the Invention

[0006] The purpose of the present invention is to provide a plant growth data monitoring system based on the Internet of Things to address the shortcomings of the background technology.

[0007] In order to achieve the above-mentioned object, the present invention provides the following technical solutions: a plant growth data monitoring system based on the Internet of Things, comprising an image acquisition module, a spectral analysis module, a structural analysis module, a calculation module, and a model construction and display module;

[0008] Image acquisition module: collects spectral images of plant crowns and divides the plant crowns into K canopy layers from top to bottom;

[0009] Spectral analysis module: After analyzing the spectral characteristics of the canopy based on the spectral image, the spectral characteristic data of each canopy is obtained;

[0010] Structural analysis module: flattens K canopies into a two-dimensional plane, and analyzes the leaf density of the canopy based on the grid method to obtain the structural characteristic data of the canopy;

[0011] Calculation module: Substitute the structural feature data and spectral feature data into the fusion model, output the health convergence of K canopies, and calculate the overall health index of the plant by weighted calculation of the health convergence of K canopies;

[0012] Model construction and display module: Perform sliding window identification on the overall plant health index, predict plant growth trends, and use LiDAR point cloud data to construct a layered canopy model for display.

[0013] In a preferred embodiment, the calculation module obtains the structural characteristic factor and the spectral characteristic factor, and substitutes the structural characteristic factor and the spectral characteristic factor into the fusion model. The model expression is: , For healthy outward restraint, is the spectral characteristic factor, is the structural characteristic factor, 、 is the proportionality coefficient, and the proportionality coefficient 、 Greater than 0.

[0014] In a preferred embodiment, after the calculation module obtains the health convergence of K canopies, it performs weighted calculation to obtain the overall health index of the plant, which is expressed as: , where is the overall health index, is the number of plant canopies, is the weight of the i-th canopy, is the health convergence of the i-th canopy.

[0015] In a preferred embodiment, the model building and display module obtains the overall health index of multiple time windows within the monitoring period, and calculates the overall health index mean and overall health index standard deviation of the plant based on the overall health index of multiple time points;

[0016] Predict plant growth trends based on the mean and standard deviation of the overall health index;

[0017] If the mean value of the overall health index is less than the health threshold, and the standard deviation of the overall health index is less than or equal to the standard deviation threshold, it is predicted that the overall health of the plant will show a downward trend and the development speed will be fast;

[0018] If the mean value of the overall health index is less than the health threshold, and the standard deviation of the overall health index is greater than the standard deviation threshold, it is predicted that the overall health of the plant will show a downward trend, but the development speed will be moderate;

[0019] If the mean value of the overall health index is greater than or equal to the health threshold, it is predicted that the overall health of the plant will show an upward trend.

[0020] In a preferred embodiment, the calculation module obtains the photosynthetic efficiency and transpiration rate of the canopy, normalizes the photosynthetic efficiency and transpiration rate so that the value range of the photosynthetic efficiency and transpiration rate is mapped to between [0, 1], obtains the normalized value of the photosynthetic efficiency and the normalized value of the transpiration rate, and sums the normalized value of the photosynthetic efficiency and the normalized value of the transpiration rate to obtain the importance index of the canopy;

[0021] The importance indexes of K canopies are summed to obtain the denominator value, and the weight of the canopy is obtained by dividing the importance index by the denominator value.

[0022] In a preferred embodiment, the structural analysis module uses projection transformation to unfold the three-dimensional canopy into a two-dimensional plane to form K canopy projection images;

[0023] The two-dimensional projection area is divided into n×m grids, each grid represents a local area of the canopy, and the proportion of leaf pixels in each grid is calculated as the leaf density index;

[0024] The mean leaf density of all grids in the two-dimensional projection area is calculated and used as the structural characteristic factor.

[0025] In a preferred embodiment, the leaf pixel ratio in each grid is calculated to define the leaf density, which is expressed as: , where is the leaf density of the grid in row i and column j.

[0026] In a preferred embodiment, after acquiring the spectral image, the spectral analysis module analyzes the data of each band of the spectral image;

[0027] Calculate the normalized vegetation index, chlorophyll index, moisture index and benefit index of each layer of tree canopy based on the band data;

[0028] The normalized vegetation index, chlorophyll index, moisture index and benefit index were substituted into the linear regression algorithm to obtain the spectral characteristic factor of the canopy.

[0029] In a preferred embodiment, the band data includes a blue light band, a green light band, a red light band, a near infrared band, and a short-wave infrared band.

[0030] In a preferred embodiment, after the image acquisition module acquires the spectral image of the plant, it identifies the top and bottom of the plant crown and calculates the crown height, and divides the crown into K layers according to the crown height, wherein: , where Indicates the number of crown division layers, is the crown height, Indicates rounding down.

[0031] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0032] The present invention uses a spectral analysis module to analyze the spectral characteristics of the canopy based on the spectral image, obtaining spectral characteristic data for each canopy. The structural analysis module flattens K canopies into a two-dimensional plane and analyzes the leaf density of the canopy based on a grid method to obtain the structural characteristic data of the canopy. The calculation module substitutes the structural characteristic data and spectral characteristic data into a fusion model and outputs the health convergence of the K canopies. The health convergence of the K canopies is weighted and calculated to obtain the overall health index of the plant. The model construction and display module performs sliding window identification on the overall health index of the plant, predicts plant growth trends, and constructs a layered canopy model using LiDAR point cloud data for display. This monitoring system determines the overall growth health status of the plant after performing multi-level canopy analysis on the plant, and performs time-series analysis on the plant to predict the plant's growth trend. This facilitates precise control of the environmental conditions required by the plant and effectively identifies the early symptoms of the disease. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0034] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0035] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0036] Example 1: Please refer to Figure 1As shown, the plant growth data monitoring system based on the Internet of Things described in this embodiment includes an image acquisition module, a spectrum analysis module, a structure analysis module, a calculation module, and a model construction and display module;

[0037] Image acquisition module: The multispectral camera on the drone collects spectral images of plant crowns and divides the plant crowns into K canopy layers from top to bottom. The spectral images are sent to the spectral analysis module and the structural analysis module, and the canopy layer division results are sent to the spectral analysis module and the structural analysis module;

[0038] Spectral analysis module: After analyzing the spectral characteristics of the canopy based on the spectral image, the spectral characteristic data of each canopy is obtained and sent to the calculation module;

[0039] Structural analysis module: flattens K canopies into a two-dimensional plane, and analyzes the leaf density of the canopy based on the grid method to obtain the structural characteristic data of the canopy, which is then sent to the calculation module;

[0040] Calculation module: Substitute the structural feature data and spectral feature data into the fusion model, output the health convergence of K canopies, and calculate the overall health index of the plant by weighted calculation of the health convergence of K canopies. The overall health index is sent to the model construction and display module;

[0041] Model construction and display module: Perform sliding window identification on the overall plant health index, predict plant growth trends, and use LiDAR point cloud data to construct a layered canopy model for display.

[0042] This application uses a spectral analysis module to analyze the spectral characteristics of the canopy based on the spectral image, and then obtains the spectral characteristic data of each canopy. The structural analysis module flattens the K canopies into a two-dimensional plane, and analyzes the leaf density of the canopy based on the grid method to obtain the structural characteristic data of the canopy. The calculation module substitutes the structural characteristic data and the spectral characteristic data into the fusion model, outputs the health convergence of the K canopies, and calculates the health convergence of the K canopies by weighted calculation to obtain the overall health index of the plant. The model construction and display module performs sliding window identification on the overall health index of the plant, predicts the growth trend of the plant, and uses LiDAR point cloud data to construct a layered canopy model for display. This monitoring system determines the overall growth health status of the plant after performing multi-level canopy analysis on the plant, and performs time series analysis on the plant to predict the growth trend of the plant, which is conducive to accurately controlling the environmental conditions required by the plant and effectively identifying the early symptoms of the disease.

[0043] The monitoring system includes the following steps:

[0044] The monitoring system collects spectral images of plant canopies through a multispectral camera carried by a drone, and divides the plant canopy into K canopy layers from top to bottom. After analyzing the spectral characteristics of the canopy based on the spectral image, the spectral characteristic data of each canopy is obtained, the K canopies are flattened into a two-dimensional plane, and the leaf density of the canopy is analyzed based on the grid method to obtain the structural characteristic data of the canopy. The structural characteristic data and the spectral characteristic data are substituted into the fusion model, and the health convergence of the K canopies is output. The health convergence of the K canopies is weighted and calculated to obtain the overall health index of the plant. The overall health index of the plant is identified by sliding window, and the plant growth trend is predicted. The layered canopy model is constructed using LiDAR point cloud data and then displayed.

[0045] Example 2: The image acquisition module collects spectral images of plant crowns using a multispectral camera carried by a drone, and divides the plant crowns into K canopy layers from top to bottom;

[0046] In this application, the drone needs to be equipped with a multispectral camera (or hyperspectral camera) for image acquisition. Common spectral bands include:

[0047] Blue light (B): 450–500 nm (for leaf health analysis).

[0048] Green light (Green, G): 500–570 nm (for chlorophyll reflectance analysis).

[0049] Red light (Red, R): 620–700 nm (used for leaf senescence and disease analysis).

[0050] Near-infrared (NIR): 700–900 nm (used for vegetation health index calculations such as NDVI).

[0051] Shortwave infrared (SWIR): 1000–2500 nm (for moisture content assessment).

[0052] UAV flight parameter settings:

[0053] Flight altitude: Adjusted according to tree height, usually between 30–120m.

[0054] Route planning: Set an automatic cruise path to cover the target area and ensure seamless image stitching.

[0055] Shooting angle: You can use vertical or oblique shooting to obtain complete canopy data.

[0056] Image preprocessing:

[0057] Light correction: Use a radiometric calibration plate to eliminate the effects of varying lighting conditions.

[0058] Geometric correction: Corrects image distortion caused by changes in the drone's posture.

[0059] Multispectral band registration: ensures alignment of different spectral bands to improve data accuracy.

[0060] Based on expert knowledge or manual experience, tree crowns are usually divided into the following canopy layers:

[0061] C1 (top canopy): receives the most light and mainly performs photosynthesis.

[0062] C2 (upper middle layer): Light intensity is slightly reduced and water utilization rate is higher.

[0063] C3 (middle layer): limited light, high incidence of disease areas.

[0064] C4 (lower middle layer): The humidity is high and the leaves are aging.

[0065] C5 (bottom layer): close to the tree trunk, with the weakest light and greatly affected by the soil.

[0066] However, the above division method still requires manual on-site inspection before division, which increases labor costs.

[0067] Therefore, in this application, after the image acquisition module obtains the spectral image of the plant, it identifies the top and bottom of the plant crown and calculates the crown height. The crown is divided into K layers according to the crown height, where: , where Indicates the number of crown division layers, is the crown height, Indicates rounding down.

[0068] The spectrum analysis module analyzes the spectral characteristics of the canopy based on the spectral image and obtains the spectral characteristic data of each canopy;

[0069] After acquiring the spectral image, the spectral analysis module analyzes the data of each band of the spectral image, including the blue light band, green light band, red light band, near infrared band and short-wave infrared band;

[0070] Calculate the normalized vegetation index, chlorophyll index, moisture index and benefit index of each layer of tree canopy based on the band data;

[0071] Substituting the normalized vegetation index, chlorophyll index, moisture index and benefit index into the linear regression algorithm, the spectral characteristic factor of the canopy is calculated, and the expression is:

[0072] , where is the spectral characteristic factor, is the regression coefficient of each band data, and the regression coefficients are all greater than 0;

[0073] The present invention uses the logical factors of the spectral characteristic factors as follows: taking the relationship between band data and plant health as an example, the first is the indicator, that is, the factor that causes the change of plant health (the present invention refers to the relationship between band data and plant health); the second is the weight of these indicators, that is, the proportion of each type of band data when it is generated; the third is the operation equation, that is, what kind of mathematical operation process is used to obtain the result, and the spectral characteristic factor is obtained by calculating the indicators with their respective weights through the operation equation.

[0074] The band data obtained from the sample were converted and processed into a data language that can be recognized by computer software. Secondly, these evaluation factors were analyzed by logistic regression using SPSS software to screen out factors and their weights that are significantly correlated with the results. Thirdly, the evaluation factors and weights were brought into the logistic regression equation for calculation to obtain the results, which are as follows:

[0075] First, ensure the integrity of the band data, handle missing values and outliers, and convert the data into a format that SPSS software can recognize. Usually, the data is stored in .csv, .xlsx and other formats, and then imported into SPSS. Open SPSS software, import the processed data file, and transform the variables as needed. For example, for continuous variables, standardize or normalize them. Select the "Analyze" menu, and then select the "Binary Logistic" option under "Regression". In the dialog box, add the dependent variable (outcome) and the independent variable (band data) to the corresponding boxes. SPSS will fit the Logistic regression model based on the selected variables. Type, in the output results, you will see the model's coefficients, standard errors, p-values and other information. Check the coefficients and p-values in the output results to determine which variables have a significant correlation with the results. Usually, a p-value less than 0.05 is considered significant. While fitting the model, use variable selection methods, such as stepwise regression, to help screen the most relevant factors. According to the coefficients of the Logistic regression model, the size of the coefficient reflects the degree of influence of each factor on the result, and the positive and negative signs of the coefficients indicate the direction of the influence. After obtaining the significant factors and their coefficients, the Logistic regression equation is obtained, which is used to calculate the probability of each sample and then predict the results.

[0076] The calculation expression of the normalized vegetation index is: NDVI=(NIR-R) / (NIR+R), where NDVI is the normalized vegetation index, NIR is the near-infrared band, and R is the red light band. When the normalized vegetation index is large, it indicates that the vegetation is healthy, with high chlorophyll content and active photosynthesis. When the normalized vegetation index is small, it indicates that the vegetation may be in a state of early stress, drought or disease.

[0077] The calculation expression of the chlorophyll index is: Chl_Index=NIR / G, where Chl_Index is the chlorophyll index, NIR is the near-infrared band, and G is the green light band. A larger chlorophyll index indicates a high chlorophyll content and healthy plants. A smaller chlorophyll index indicates that the cause may be disease, aging, or malnutrition.

[0078] The calculation expression of the moisture index is: NDWI=(NIR-SWIR) / (NIR+SWIR), where NDWI is the moisture index, SWIR is the shortwave infrared band, and NIR is the near infrared band. A larger moisture index indicates that the vegetation has high moisture content and the trees are healthy. A smaller moisture index indicates that the vegetation is water-deficient and may face drought stress.

[0079] The calculation expression of the benefit index is: PRI=(GR) / (G+R), where PRI is the benefit index, G is the green light band, and R is the red light band. When the benefit index is large, it indicates that the vegetation is disease-free and in good health. When the benefit index is small, it indicates that the vegetation may have disease, leaf aging or stress.

[0080] The structural analysis module flattens K canopies into a two-dimensional plane and analyzes the leaf density of the canopy based on the grid method to obtain the structural characteristic data of the canopy;

[0081] The structural analysis module uses projection transformation (including vertical projection and fisheye projection) to unfold the three-dimensional canopy into a two-dimensional plane, forming K canopy projection images to ensure that the morphological characteristics of canopies at different heights are completely preserved. Generally speaking, the top canopy is approximately a circle when unfolded, while the subsequent canopy (i.e., at the bottom of the top canopy) is approximately a ring after unfolding. However, since the center of the ring is a leaf-missing area, directly calculating the leaf density of the two-dimensional ring will lead to inaccurate subsequent calculations (caused by the influence of the leaf-missing area). Therefore, the two-dimensional ring is twice unfolded to approximate a rectangular area before calculation;

[0082] The two-dimensional projection area is divided into n×m grids, each grid represents a local area of the canopy, and the grid size is set to ensure that a single grid can cover a sufficient number of leaves. The proportion of leaf pixels in each grid is calculated as the leaf density indicator;

[0083] Let’s take a tall tree (such as red sandalwood) as an example. Its crown can be divided into K=3 layers, namely:

[0084] Top canopy (outermost layer): approximately circular, with high leaf coverage.

[0085] Middle canopy (second layer): approximately ring-shaped, with the middle area being the area with missing leaves (possibly due to shading by upper leaves).

[0086] Bottom canopy (lowest layer): Approximately a larger ring, but with sparser leaves due to less light.

[0087] Project and expand K=3 crown layers separately:

[0088] Top canopy (K=1) → Directly unfold into a two-dimensional circular projection image.

[0089] The middle canopy (K=2) is approximately annular, but due to the missing leaf area, the density can be incorrectly calculated directly.

[0090] Bottom canopy (K=3) → Approximately expands into a larger ring with sparser leaves.

[0091] For the circular canopy with K = 2 and K = 3 layers, we use quadratic expansion to convert it into a rectangular region and partition it: the converted two-dimensional plane is divided into n × m grids, for example, 100 × 100 small cells, each grid represents a local area.

[0092] Calculate the percentage of leaf pixels in each grid and define the leaf density. The expression is:

[0093] , where The leaf density of the grid in the i-th row and j-th column is calculated, and the mean leaf density of all grids in the two-dimensional projection area is used as the structural characteristic factor.

[0094] The calculation module substitutes the structural feature data and spectral feature data into the fusion model, outputs the health convergence of K canopies, and calculates the overall health index of the plant by weighted calculation of the health convergence of K canopies;

[0095] The calculation module obtains the structural characteristic factors and spectral characteristic factors, and substitutes the structural characteristic factors and spectral characteristic factors into the fusion model. The model expression is: , For healthy outward restraint, is the spectral characteristic factor, is the structural characteristic factor, 、 is the proportionality coefficient, and the proportionality coefficient 、 Greater than 0.

[0096] After the calculation module obtains the health convergence of K canopies, it performs weighted calculation to obtain the overall health index of the plant, which is expressed as: , where is the overall health index, is the number of plant canopies, is the weight of the i-th canopy, is the health convergence of the i-th canopy.

[0097] In the existing technology, the weight of the canopy can usually be determined by expert knowledge or manual experience:

[0098] Different canopies have different importance in photosynthesis, transpiration, nutrient distribution, etc., so they can be weighted based on expert knowledge. For example:

[0099] 1) Experience empowerment based on physiological function

[0100] Top canopy (K=1, outermost layer):

[0101] Main function: Receive sunlight for photosynthesis, which is the main source of energy for trees.

[0102] The importance is relatively high, and the weight is usually 0.4~0.5.

[0103] Middle canopy (K=2):

[0104] Main functions: Maintain leaf gas exchange, regulate water and transpiration, and support tree growth.

[0105] The importance is medium, and the weight is usually 0.3~0.4.

[0106] Bottom canopy (K=3, inner layer):

[0107] Main functions: Provide storage function, partially block light, reduce water transpiration, but photosynthesis is weak.

[0108] The importance is relatively low, and the weight is usually 0.1~0.2.

[0109] The model building and display module uses a sliding window to identify the overall health index of plants, predict plant growth trends, and use LiDAR point cloud data to construct a layered canopy model for display.

[0110] 2) Based on expert scoring method (AHP)

[0111] Botanists and forestry experts were asked to rate the importance of different canopies and calculate a weighted average.

[0112] For example:

[0113] Let 10 experts rate:

[0114] K=1 (top): 4.5 / 5;

[0115] K=2 (center): 3.8 / 5;

[0116] K=3 (bottom): 2.2 / 5;

[0117] Calculate the weights after normalization:

[0118] The weight of K1 = 4.5 / (4.5+3.8+2.2)≈0.45;

[0119] The weight of K2 = 3.8 / (4.5+3.8+2.2)≈0.38;

[0120] The weight of K3 = 2.2 / (4.5+3.8+2.2)≈0.17.

[0121] Assigning weights directly based on expert experience is the simplest approach, but it is prone to subjectivity.

[0122] The larger the overall health index of the plant, the better the overall health of the plant.

[0123] The model building and display module uses a sliding window to identify the overall plant health index, predict plant growth trends, and construct a layered canopy model using LiDAR point cloud data for display.

[0124] The model building and display module obtains the overall health index of multiple time windows within the monitoring period, and calculates the mean and standard deviation of the overall health index of the plant based on the overall health index of multiple time points;

[0125] Predict plant growth trends based on the mean and standard deviation of the overall health index;

[0126] If the mean value of the overall health index is less than the health threshold, and the standard deviation of the overall health index is less than or equal to the standard deviation threshold, it is predicted that the overall health of the plant will show a downward trend and the development speed will be fast;

[0127] If the mean value of the overall health index is less than the health threshold, and the standard deviation of the overall health index is greater than the standard deviation threshold, it is predicted that the overall health of the plant will show a downward trend, but the development speed will be moderate;

[0128] If the mean value of the overall health index is greater than or equal to the health threshold, it is predicted that the overall health of the plant will show an upward trend.

[0129] The following code example reads plant point cloud data (.las, .ply, and other formats), obtains the health convergence values of K canopies, and predicts the overall health of the plant. The following code example generates a layered canopy model using Python tools:

[0130] import numpy as np

[0131] import open3d as o3d

[0132] import matplotlib.pyplot as plt

[0133] # Step 1: Load LiDAR point cloud data

[0134] def load_point_cloud(file_path):

[0135] pcd = o3d.io.read_point_cloud(file_path) # Read point cloud data

[0136] points = np.asarray(pcd.points) # Get point coordinates

[0137] return pcd, points

[0138] # Step 2: Layered crown modeling

[0139] def segment_tree_layers(points, k=5):

[0140] min_z, max_z = np.min(points[:, 2]), np.max(points[:, 2]) # Get the height range

[0141] layers = np.linspace(min_z, max_z, k+1) # Calculate layer height

[0142] segmented_layers = []

[0143] for i in range(k):

[0144] layer_mask = (points[:, 2] >= layers[i]) & (points[:, 2] < layers[i+1])

[0145] segmented_layers.append(points[layer_mask])

[0146] return segmented_layers, layers

[0147] # Step 3: Calculate health convergence (NDVI / NDWI simulation)

[0148] def compute_health_index(layer_points):

[0149] ndvi_values = np.random.uniform(0, 1, len(layer_points)) # Generate simulated NDVI data

[0150] health_index = (ndvi_values - 0.3) 2 # Map to [-1, 1], the higher the health value, the better

[0151] return health_index

[0152] # Step 4: Build a 3D tree crown model (colored by healthy outward convergence)

[0153] def create_colored_point_cloud(layer_points, health_index):

[0154] colors = plt.get_cmap('RdYlGn')(health_index)[:, :3] # Color mapping (red, yellow, green)

[0155] pcd_layer = o3d.geometry.PointCloud()

[0156] pcd_layer.points = o3d.utility.Vector3dVector(layer_points)

[0157] pcd_layer.colors = o3d.utility.Vector3dVector(colors)

[0158] return pcd_layer

[0159] # Step 5: Predict health status (sliding window method)

[0160] def predict_health_trend(health_history, window_size=3):

[0161] Return np.convolve(health_history, np.ones(window_size) / window_size,mode='valid')

[0162] Step 6: Visualize the 3D Model

[0163] def visualize_tree_layers(pcd_layers):

[0164] o3d.visualization.draw_geometry(pcd_layers)

[0165] # Main function

[0166] file_path = "tree_lidar.ply" # Replace with your LiDAR data path

[0167] pcd, points = load_point_cloud(file_path)

[0168] segmented_layers, layers = segment_tree_layers(points, k=5)

[0169] pcd_layers = []

[0170] for_i, layer in enumerate(segmented_layers):

[0171] health_index = compute_health_index(layer)

[0172] pcd_layer = create_colored_point_cloud(layer, health_index)

[0173] pcd_layers.append(pcd_layer)

[0174] #Visualize the health of the tree crown

[0175] visualize_tree_layers(pcd_layers)

[0176] Code analysis:

[0177] The .ply file is read and converted into a NumPy array. The tree crown is divided into K layers by height range. The health convergence is calculated using simulated NDVI data. The point cloud color is adjusted according to the health convergence (red = diseased, green = healthy). The sliding window method is used to predict future health trends. Open3D is used to display the layered tree crown model, and the color reflects the health status.

[0178] Example 3: In Example 2, directly assigning weights based on expert experience is the simplest approach, but it is prone to subjectivity. Therefore, in order to improve the objectivity and accuracy of plant growth analysis, we propose the following solution:

[0179] The calculation module obtains the photosynthesis efficiency and transpiration rate of the canopy, normalizes the photosynthesis efficiency and transpiration rate so that their value ranges are mapped to [0, 1], obtains the normalized value of the photosynthesis efficiency and the normalized value of the transpiration rate, and sums the normalized value of the photosynthesis efficiency and the normalized value of the transpiration rate to obtain the canopy importance index;

[0180] The importance indexes of K canopies are summed to obtain the denominator value, and the weight of the canopy is obtained by dividing the importance index by the denominator value.

[0181] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0182] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0183] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A plant growth data monitoring system based on the Internet of Things, characterized by: It includes image acquisition module, spectrum analysis module, structure analysis module, calculation module, model building and display module; Image acquisition module: collects spectral images of plant crowns and divides the plant crowns into K canopy layers from top to bottom; Spectral analysis module: After analyzing the spectral characteristics of the canopy based on the spectral image, the spectral characteristic data of each canopy is obtained; Structural analysis module: flattens K canopies into a two-dimensional plane, and analyzes the leaf density of the canopy based on the grid method to obtain the structural characteristic data of the canopy; Calculation module: Substitute the structural feature data and spectral feature data into the fusion model, output the health convergence of K canopies, and calculate the overall health index of the plant by weighted calculation of the health convergence of K canopies; Model building and display module: This module uses sliding windows to identify the overall plant health index, predict plant growth trends, and construct a layered canopy model using LiDAR point cloud data for display. The calculation module obtains the photosynthesis efficiency and transpiration rate of the canopy, normalizes the photosynthesis efficiency and transpiration rate so that the value range of the photosynthesis efficiency and transpiration rate is mapped to [0, 1], obtains the normalized value of the photosynthesis efficiency and the normalized value of the transpiration rate, and sums the normalized value of the photosynthesis efficiency and the normalized value of the transpiration rate to obtain the importance index of the canopy; The importance indexes of K canopies are summed to obtain the denominator value, and the weight of the canopy is obtained by dividing the importance index by the denominator value; The structural analysis module uses projection transformation to unfold the three-dimensional canopy into a two-dimensional plane to form K canopy projection images; The two-dimensional projection area is divided into n×m grids, each grid represents a local area of the canopy, and the proportion of leaf pixels in each grid is calculated as the leaf density index; Calculate the mean leaf density of all grids in the two-dimensional projection area and use the mean leaf density as the structural characteristic factor; After acquiring the spectral image, the spectral analysis module analyzes the data of each band of the spectral image; Calculate the normalized vegetation index, chlorophyll index, moisture index and benefit index of each layer of tree canopy based on the band data; Substitute the normalized vegetation index, chlorophyll index, moisture index and benefit index into the linear regression algorithm to calculate the spectral characteristic factor of the canopy; After the image acquisition module acquires the spectral image of the plant, it identifies the top and bottom of the plant crown and calculates the crown height. The crown is divided into K layers according to the crown height, where: , where Indicates the number of crown division layers, is the crown height, Indicates rounding down.

2. The plant growth data monitoring system based on the Internet of Things according to claim 1, characterized in that: The calculation module obtains the structural characteristic factors and the spectral characteristic factors, and substitutes the structural characteristic factors and the spectral characteristic factors into the fusion model. The model expression is: , For healthy outward restraint, is the spectral characteristic factor, is the structural characteristic factor, 、 is the proportionality coefficient, and the proportionality coefficient 、 Greater than 0.

3. The plant growth data monitoring system based on the Internet of Things according to claim 2, characterized in that: After the calculation module obtains the health convergence of K canopies, it performs weighted calculation to obtain the overall health index of the plant, which is expressed as: , where is the overall health index, is the number of plant canopies, is the weight of the i-th canopy, is the health convergence of the i-th canopy.

4. The plant growth data monitoring system based on the Internet of Things according to claim 3, characterized in that: The model building and display module obtains the overall health index of multiple time windows within the monitoring period, and calculates the overall health index mean and overall health index standard deviation of the plant based on the overall health index of multiple time points; Predict plant growth trends based on the mean and standard deviation of the overall health index; If the mean value of the overall health index is less than the health threshold, and the standard deviation of the overall health index is less than or equal to the standard deviation threshold, it is predicted that the overall health of the plant will show a downward trend and the development speed will be fast; If the mean value of the overall health index is less than the health threshold, and the standard deviation of the overall health index is greater than the standard deviation threshold, it is predicted that the overall health of the plant will show a downward trend, but the development speed will be moderate; If the mean value of the overall health index is greater than or equal to the health threshold, it is predicted that the overall health of the plant will show an upward trend.

5. The plant growth data monitoring system based on the Internet of Things according to claim 4, characterized in that: Calculate the percentage of leaf pixels in each grid and define the leaf density. The expression is: , where is the leaf density of the grid in row i and column j.

6. The plant growth data monitoring system based on the Internet of Things according to claim 5, characterized in that: The band data includes a blue light band, a green light band, a red light band, a near infrared band, and a short-wave infrared band.

Images