Method and system for multispectral monitoring and optimization of tea tree growth in tea plantations

Abstract

The present application relates to the technical field of image data processing, and more particularly to a multispectral monitoring and optimization method and system for tea tree growth in a tea garden. The method comprises the following steps: acquiring tea tree growth spectrum data; extracting tea tree growth spectrum features based on the tea tree growth spectrum data to obtain tea tree growth spectrum feature data; calculating a tea tree growth index based on the tea tree growth spectrum feature data to obtain tea tree growth index data; and monitoring the growth state based on the tea tree growth index data to obtain tea tree growth state monitoring data for optimizing tea tree growth cultivation operations. The present application can ensure that the tea tree is always in the best growth environment by continuously monitoring the growth state of the tea tree and dynamically optimizing, thereby maximizing the yield and quality of tea.

Description

Technical Field

[0001] This invention relates to the field of image data processing technology, and in particular to a multispectral monitoring and optimization method and system for tea tree growth in tea gardens. Background Technology

[0002] With the development of agricultural technology, monitoring and managing crop growth using information technology has gradually become an important part of modern agriculture. Traditional crop growth monitoring methods mainly rely on manual observation and simple sensor data collection, which have significant limitations in terms of the breadth, depth, and real-time performance of monitoring. Especially in large-scale crop cultivation environments, how to accurately monitor the growth status of plants (such as tea trees), provide early warnings of abnormal conditions, and conduct scientific cultivation and management has become a major challenge in agricultural production. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention proposes a multispectral monitoring and optimization method and system for tea tree growth in tea gardens, thereby resolving at least one of the aforementioned technical issues.

[0004] This application provides a multispectral monitoring and optimization method for tea tree growth in a tea garden, the method comprising:

[0005] S1. Obtain tea tree growth spectral data;

[0006] S2. Extract tea tree growth spectral features from tea tree growth spectral data to obtain tea tree growth spectral feature data;

[0007] S3. Calculate the tea tree growth index based on the tea tree growth spectral characteristic data to obtain the tea tree growth index data;

[0008] S4. Monitor the growth status of tea trees based on the growth index data to obtain tea tree growth status monitoring data, so as to carry out tea tree growth and cultivation optimization operations.

[0009] This invention utilizes multispectral data (including visible and near-infrared bands) from tea trees to more accurately capture their growth status, reflecting important physiological indicators such as health condition, water content, and chlorophyll content, thus enabling more precise monitoring. The spectral characteristics of tea tree growth can help identify potential problems early, such as leaf diseases, insufficient water, and nutrient imbalances. Compared to traditional manual observation, it can issue early warnings when problems first appear, preventing them from escalating. It allows for fine-grained monitoring of different growth stages and regions of tea trees. For example, individualized monitoring and adjustments can be made for a specific tea tree or a specific region, ensuring that problems are accurately located and addressed promptly.

[0010] Optionally, the spectral data of tea tree growth includes visible light data and near-infrared data of tea tree growth, S1 including:

[0011] Visible light images of tea trees are acquired using visible light acquisition equipment to obtain visible light data on tea tree growth;

[0012] Near-infrared data on tea tree growth were obtained by collecting near-infrared data from tea trees using near-infrared sensors.

[0013] This invention utilizes visible light equipment and near-infrared sensors to cover the visible and near-infrared bands of tea tree growth, acquiring detailed growth information from multiple bands. Visible light data reflects the tea tree's color, morphology, chlorophyll content, etc., while near-infrared data provides deeper physiological information such as leaf internal water content and cell structure. The combination of visible light and near-infrared band data allows the system to simultaneously monitor the external and internal growth status of the tea tree, helping to identify early physiological abnormalities (such as water shortage and disease) and making optimization decisions based on this information.

[0014] Optionally, S2 includes:

[0015] Color and texture features were extracted from the visible light data of tea tree growth to obtain visible light feature data of tea tree growth;

[0016] Near-infrared band features were extracted from near-infrared band data of tea tree growth to obtain near-infrared band feature data of tea tree growth;

[0017] By integrating the visible light characteristic data and the near-infrared band characteristic data of tea tree growth, the spectral characteristic data of tea tree growth are obtained.

[0018] This invention extracts color, texture, and spectral features of tea trees from visible light and near-infrared data, enabling the system to understand the growth status of tea trees from multiple dimensions. Color and texture features reflect the appearance and surface characteristics of tea trees, while near-infrared features reveal their internal physiological information. By extracting features from data in different wavelengths, information can be obtained from both visual and spectral perspectives, helping to reveal the state of tea trees at different growth stages. The combination of multimodal feature data helps to build a more comprehensive tea tree growth model, identify potential problems, and implement more precise management. Through color and texture feature extraction, the system can monitor external manifestations of tea trees, such as color changes, leaf yellowing, and disease patterns. Simultaneously, feature extraction from near-infrared data allows for in-depth analysis of internal physiological indicators such as water content and nitrogen levels. The combination of these two methods provides a more accurate assessment of tea tree health, enabling users to make more precise judgments about the growth status of tea trees. Color and texture feature extraction helps identify external problems of tea trees (such as lesions and leaf withering), while near-infrared feature extraction can detect internal conditions such as water loss and nutrient deficiency. By combining these two feature extraction methods, pests, diseases, or water shortages can be detected earlier and more accurately, allowing for timely responses and preventing the problems from escalating further.

[0019] Optionally, the color texture feature extraction includes:

[0020] The visible light data of tea tree growth is used to segment the region to obtain tea tree region segmentation data;

[0021] Color space transformation is performed on the tea tree region segmentation data to obtain tea tree color space transformed data;

[0022] Color histograms were calculated from the color space transformation data of tea trees to obtain tea tree color histogram data;

[0023] Clustering calculations were performed based on the tea tree color histogram data to obtain the tea tree color histogram clustering data;

[0024] The dissimilarity of the tea tree color histogram clustering data was calculated to obtain the tea tree color histogram clustering dissimilarity data;

[0025] When the cluster difference data of the tea tree color histogram is determined to be greater than the preset difference threshold data, the color feature vector is calculated on the tea tree color histogram data to obtain the tea tree color feature data.

[0026] When the cluster difference data of the tea tree color histogram is determined to be less than or equal to the preset difference threshold data, the tea tree color histogram data is aggregated into regions based on the cluster difference data of the tea tree color histogram to obtain the tea tree color histogram region aggregated data.

[0027] Color feature vectors are calculated from the aggregated data of the tea tree color histogram regions to obtain tea tree color feature data;

[0028] Texture features were extracted from visible light data of tea tree growth to obtain tea tree texture feature data;

[0029] Color and texture feature encoding is performed based on tea tree color feature data and tea tree texture feature data to obtain visible light feature data of tea tree growth.

[0030] This invention utilizes regional segmentation of visible light data from tea tree growth to separate different regions of the tea tree (such as leaves, branches, and background). This ensures that color and texture feature extraction focuses solely on the regions of interest within the tea tree, reducing background noise interference and improving the accuracy and precision of feature extraction. By calculating the dissimilarity of color histogram clusters, the method dynamically selects the feature extraction path. When the cluster dissimilarity exceeds a threshold, the system performs more detailed color feature vector calculations for each cluster to capture color details within specific regions. Conversely, when the dissimilarity is low, the system performs regional aggregation on the color histogram data to reduce complexity, dynamically adjust the depth of feature extraction, and balance computational costs, ensuring both efficiency and accuracy. This dissimilarity-driven path selection allows the method to automatically adapt to different growth states and environmental changes, intelligently adjusting analysis strategies to ensure that the system makes optimal feature extraction decisions for different tea tree varieties and growth conditions.

[0031] Optionally, the color texture feature encoding includes:

[0032] Spatial location coding is performed based on tea tree color feature data and tea tree texture feature data to obtain spatially encoded tea tree color feature data and spatially encoded tea tree texture feature data, respectively.

[0033] Based on the spatial coding data of tea tree color features and the spatial coding data of tea tree texture features, multi-head self-attention calculation is performed on the tea tree color feature data and the tea tree texture feature data to obtain the multi-head self-attention feature data of tea tree color and texture.

[0034] Based on the multi-head self-attention feature data of tea tree color and texture, feature weighting and fusion are performed on the tea tree color feature data and tea tree texture feature data to obtain the visible light feature data of tea tree growth.

[0035] This invention utilizes spatial location encoding of tea tree color and texture feature data to capture the spatial context information of each pixel or feature location. Spatial location encoding enhances the expressive power of features, ensuring that the system not only focuses on individual color or texture features but also considers their spatial distribution and arrangement. The morphology and texture of tea leaves often exhibit spatial dependencies, which helps preserve and strengthen these dependencies, thereby improving the accuracy of tea tree growth status monitoring. Through spatial location encoding, the system can handle the complex growth structure of tea trees, such as differences between leaf edges and centers, and variations in local texture, enabling the system to better analyze the relationships between different regions of the tea tree and capture subtle changes in its growth. Through feature weighted fusion, the system can intelligently assign weights to color and texture features. The growth status of tea trees depends on different features at different stages. For example, in some cases, color features may be more important indicators (e.g., changes in leaf color indicate malnutrition or disease), while in others, texture features may better reflect the health status of the tea tree (e.g., subtle structural changes on the leaf surface). Through weighted fusion, the system can dynamically adjust the weights of color and texture features in the comprehensive judgment, ensuring the optimal combination of features.

[0036] Optionally, the near-infrared band feature extraction includes:

[0037] Based on the texture feature data of tea trees, the near-infrared band data of tea tree growth is divided into regions to obtain near-infrared band division data of tea trees.

[0038] Spectral feature points were extracted based on the near-infrared band segmentation data of tea trees to obtain spectral feature point data of tea trees;

[0039] Based on visible light data of tea tree growth and near-infrared band division data of tea tree, band ratio feature extraction was performed to obtain band ratio feature data.

[0040] Spectral envelope removal analysis was performed on the near-infrared band segmentation data of tea plants to obtain envelope removal feature data;

[0041] By integrating the spectral feature point data, band ratio feature data, and envelope removal feature data of tea trees, near-infrared band feature data of tea tree growth is obtained.

[0042] This invention utilizes the texture features of tea trees to regionalize near-infrared data, enabling precise localization of different areas of the tea tree, such as healthy leaves, damaged leaves, and withered areas. Ensuring focused feature extraction concentrates analysis on key parts of the tea tree, avoiding interference from background noise. This not only improves the accuracy of feature extraction but also more accurately identifies problem areas in the tea tree's growth. Through fine-grained regional segmentation, the system can perform more detailed spectral analysis on different regions. For example, healthy areas and potentially diseased areas can be processed separately, ensuring targeted analysis and processing based on the characteristics of different regions. Extraction of spectral feature points (such as absorption peaks, valleys, and slopes) captures the most critical changes in the near-infrared spectrum, allowing the system to accurately locate the health status of tea leaves, such as changes in chemical components like water and nitrogen. Spectral envelope removal eliminates interference from spectral background on absorption features, highlighting absorption peaks and more clearly revealing physiological changes in tea leaves, such as dynamic changes in water and nutrients.

[0043] Optionally, the spectral envelope removal analysis includes:

[0044] Multi-resolution smoothing was performed on the near-infrared band segmentation data of tea plants to obtain smoothed near-infrared band segmentation data.

[0045] Multi-scale spline fitting was performed on the smoothed near-infrared band data to obtain envelope fitting data;

[0046] Dynamic envelope removal was performed based on the near-infrared band segmentation data and envelope fitting data of tea trees to obtain near-infrared band envelope removal data;

[0047] Multi-layer feature decomposition is performed on the near-infrared band envelope removal data to obtain near-infrared band multi-layer feature data;

[0048] Feature fusion is performed on multi-layer feature data in the near-infrared band to obtain envelope-removed feature data.

[0049] In this invention, before removing the spectral envelope, the near-infrared band segmentation data of tea plants undergoes multi-resolution smoothing processing, which effectively removes noise from the data while preserving key features of the spectral curve. The multi-scale spline fitting method can fit the envelope at different scales, ensuring the fitting result tightly encloses the local peaks of the spectral curve without overfitting, thus improving the accuracy and flexibility of the envelope fitting. Compared to traditional envelope removal methods, dynamic envelope removal can adaptively adjust the spectral envelope of different regions based on the characteristics of the near-infrared band segmentation data of tea plants. For example, for regions with deep absorption peaks, the envelope removal intensity is weakened to avoid over-removal of features; while for smooth regions, the envelope removal intensity is strengthened to eliminate background influence. By performing multi-level feature decomposition on the spectral data after envelope removal, the spectral signal can be decomposed into multiple frequency levels, each representing different levels of feature information in the spectral curve. For example, the high-frequency layer can reveal subtle changes in the spectrum, while the low-frequency layer can show the overall trend of the spectrum. The feature fusion step effectively integrates information from different scales and levels. By combining high-frequency and low-frequency information obtained from multi-layer decomposition, the system can more comprehensively analyze the spectral characteristics of tea trees and provide accurate data support for monitoring the health status of tea trees.

[0050] Optionally, S3 includes:

[0051] Leaf growth index data were obtained by calculating the leaf growth index from the visible light characteristic data of tea plant growth.

[0052] The moisture content of tea trees was calculated by analyzing the near-infrared band characteristic data of tea tree growth.

[0053] The tea tree growth index data is obtained by calculating the spectral index based on the leaf growth index data and the tea tree moisture content data.

[0054] This invention enables the system to simultaneously capture the external appearance and internal physiological characteristics of tea plants, providing a more comprehensive assessment of their growth status. Visible light data reveals information such as leaf color and texture, while near-infrared data shows moisture and nutrient levels, providing a better reflection of the tea plant's true health condition. Visible light data provides information on the external growth of leaves (such as color and morphological changes), while near-infrared data reveals internal moisture and other physiological characteristics. Through independent analysis and integration of these data, the generated tea plant growth index more accurately reflects the overall health of the tea plant.

[0055] Optionally, S4 includes:

[0056] Based on the tea tree growth index data, the tea tree growth status is classified to obtain tea tree growth status data;

[0057] Based on the growth status data of tea trees, time-series monitoring of growth status is carried out to obtain time-series data of tea tree growth status.

[0058] Based on the time-series data of tea tree growth status, abnormal growth status is detected to obtain tea tree growth status monitoring data, which is used to optimize tea tree growth and cultivation.

[0059] This invention categorizes tea tree growth index data, allowing the system to classify tea tree growth status into different health levels or categories, such as healthy, sub-healthy, water-deficient, and diseased. Time-series monitoring of growth status, by tracking changes in the tea tree's growth status at different points in time, reveals long-term trends in tea tree growth. Anomaly detection helps the system identify abnormal phenomena in tea tree growth, such as sudden leaf yellowing, abnormal water loss, and disease outbreaks. By analyzing time-series data on tea tree growth status, the system can promptly detect abnormal patterns and issue early warnings, helping managers take intervention measures before problems escalate into major issues.

[0060] Optionally, a multispectral monitoring and optimization system for tea tree growth in a tea garden is provided to execute the multispectral monitoring and optimization method for tea tree growth in a tea garden as described above. The multispectral monitoring and optimization system for tea tree growth in a tea garden includes:

[0061] The tea tree growth spectral data acquisition module is used to acquire tea tree growth spectral data;

[0062] The tea tree growth spectral feature extraction module is used to extract tea tree growth spectral features from tea tree growth spectral data to obtain tea tree growth spectral feature data.

[0063] The tea tree growth index calculation module is used to calculate the tea tree growth index based on the tea tree growth spectral characteristic data, and obtain the tea tree growth index data.

[0064] The tea tree growth status monitoring module is used to monitor the growth status of tea trees based on the tea tree growth index data, and obtain tea tree growth status monitoring data for tea tree growth and cultivation optimization.

[0065] The purpose of this invention is:

[0066] 1. The system fully utilizes multispectral data from tea plants, including data fusion from visible and near-infrared bands. By acquiring and processing multispectral data, the system can comprehensively capture the external morphological characteristics (such as leaf color and texture) and internal physiological information (such as moisture content and nitrogen). Precise growth monitoring using multispectral data can improve the overall health of tea plants by 15%-20%. By calculating the leaf growth index and water content data of tea plants, a tea plant growth index is generated, combining external growth performance (such as leaf color changes and leaf thickness) and internal physiological states (such as moisture and nitrogen), thus providing reliable data support for time-series monitoring.

[0067] 2. The time-series monitoring mechanism of this invention allows the system to continuously track the growth status of tea trees, capturing real-time changes in the tea tree's condition at different growth stages. This not only helps managers understand the dynamic trends of tea tree growth but also allows for the rapid detection of potential problems when subtle changes occur in the tea tree's condition. By utilizing visible light data to monitor the appearance characteristics of tea leaves (such as color changes and yellowing) and combining this with near-infrared data analysis of internal physiological characteristics such as moisture and nitrogen, the system can detect potential signals of physiological problems 2-3 weeks earlier than traditional manual detection. Attached Figure Description

[0068] Other features, objects, and advantages of this application will become more apparent from the following detailed description of the non-limiting embodiments, taken with reference to the accompanying drawings:

[0069] Figure 1 A flowchart illustrating the steps of a multispectral monitoring and optimization method for tea tree growth in a tea garden, according to one embodiment, is shown.

[0070] Figure 2 A flowchart illustrating the steps of a method for acquiring spectral data of tea tree growth according to an embodiment is shown.

[0071] Figure 3 A flowchart illustrating the steps of a method for extracting spectral features of tea tree growth according to an embodiment is shown.

[0072] Figure 4 A flowchart illustrating the steps of a method for calculating the growth index of tea trees according to an embodiment is shown.

[0073] Figure 5 A flowchart illustrating the steps of a method for monitoring the growth status of tea trees according to one embodiment is shown.

[0074] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0075] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0076] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0077] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0078] Please see Figures 1 to 5 This application provides a multispectral monitoring and optimization method for tea tree growth in a tea garden, the method comprising:

[0079] S1. Obtain tea tree growth spectral data;

[0080] Specifically, multispectral imaging equipment is used to collect spectral data on tea tree growth from different angles, including visible light, near-infrared, and far-infrared bands. Data acquisition methods include multispectral sensors mounted on drones or periodic data collection using ground-based multispectral camera systems. An automated data processing system performs preprocessing operations such as noise reduction and normalization on the collected spectral data to improve data quality.

[0081] Spectral reflectance is calculated for each spectral band to obtain standardized spectral reflectance data. Principal component analysis (PCA) is performed on the spectral data to reduce data dimensionality and extract key spectral components. A spectral segmentation algorithm is applied to the spectral data to distinguish spectral variation regions under different growth conditions.

[0082] S2. Extract tea tree growth spectral features from tea tree growth spectral data to obtain tea tree growth spectral feature data;

[0083] Specifically, based on spectral reflectance data from different spectral bands, key spectral features are extracted, including vegetation indices (such as NDVI and SAVI), moisture indices (such as NDWI), and chlorophyll content indices. Machine learning algorithms (such as random forests or support vector machines) are used to filter the spectral data to ensure that the extracted features are strongly correlated with the tea tree's growth status.

[0084] Common vegetation indices, such as the Normalized Difference Vegetation Index (NDVI), are calculated. The Recursive Feature Elimination (RFE) algorithm is used to reduce the dimensionality of the extracted spectral features, retaining those that best contribute to the growth status. Clustering algorithms (such as K-means or DBSCAN) are then used to classify the spectral features of tea plants, forming different groups based on their growth characteristics.

[0085] S3. Calculate the tea tree growth index based on the tea tree growth spectral characteristic data to obtain the tea tree growth index data;

[0086] Specifically, based on the extracted spectral feature data, mathematical models are used to calculate tea tree growth indices, such as biomass index and leaf area index. Regression analysis is performed using historical spectral data and tea tree growth data to establish a tea tree growth index model.

[0087] A multiple linear regression model is used to fit the tea tree growth index. The input is spectral feature data, and the output is the tea tree growth index data. If a nonlinear relationship exists, nonlinear regression methods such as support vector regression (SVR) or neural networks can be used for fitting. The calculation weights of the growth index are adjusted according to seasonal changes and environmental factors to improve the accuracy of the calculation.

[0088] S4. Monitor the growth status of tea trees based on the growth index data to obtain tea tree growth status monitoring data, so as to carry out tea tree growth and cultivation optimization operations.

[0089] Specifically, tea tree growth index data, combined with environmental parameters (such as temperature, precipitation, and soil moisture), is used to monitor the growth status of tea trees. Based on the monitoring data, cultivation measures for tea trees, such as watering, fertilization, or pruning, are dynamically adjusted to optimize the growth status of the tea trees.

[0090] Time series analysis is used to monitor the changing trends of tea tree growth indices and provide early warnings of abnormal growth. Combined with Bayesian optimization algorithms, cultivation strategies are dynamically adjusted based on real-time monitored growth data to minimize resource usage and maximize growth effectiveness. Control algorithms are used to automatically adjust the frequency and intensity of cultivation operations, such as adjusting water volume and fertilizer application, based on real-time changes in environmental conditions.

[0091] Optionally, the spectral data of tea tree growth includes visible light data and near-infrared band data of tea tree growth, S1 including:

[0092] S11. Use a visible light acquisition device to acquire visible light images of tea trees to obtain visible light data on tea tree growth;

[0093] Specifically, high-resolution visible light cameras are used, mounted on drones or fixed cameras, to capture visible light images of tea trees along pre-defined flight paths or at fixed locations. To ensure the accuracy of the visible light images, parameters such as exposure and white balance are adjusted during image acquisition to address image quality variations under different lighting conditions. After image acquisition, the visible light images undergo preliminary processing, including image denoising, color correction, and geometric correction.

[0094] Image preprocessing algorithms, such as median filtering or Gaussian filtering, are used to denoise the acquired visible light images to eliminate environmental interference. Image registration techniques are used to align images acquired at different times or from different angles to ensure spatiotemporal consistency of the data. Region segmentation is then performed on the preprocessed images, using the watershed algorithm or K-means clustering to distinguish tea tree and background areas, extracting the effective image regions of the tea trees.

[0095] S12. Near-infrared band data of tea tree growth is obtained by collecting near-infrared band data of tea tree growth using a near-infrared sensor.

[0096] Specifically, specialized multispectral or near-infrared sensors are used, mounted on drones or fixed structures, to periodically collect near-infrared data from tea trees. During the acquisition process, the sensor gain and exposure time are adjusted to ensure sufficient contrast and detail in the near-infrared data obtained under different lighting conditions. The acquired near-infrared images are calibrated to compensate for the device's response characteristics and the influence of ambient light on the near-infrared signal.

[0097] A radiometric correction algorithm is used to eliminate band response differences caused by sensor characteristics and external environmental factors, ensuring the accuracy of near-infrared data. A normalization algorithm is employed to standardize the near-infrared band data, enabling consistent comparison of data from different devices or under different acquisition conditions.

[0098] Optionally, S2 includes:

[0099] S21. Extract color and texture features from the visible light data of tea tree growth to obtain visible light feature data of tea tree growth;

[0100] Specifically, color features of tea leaves are extracted from visible light images using methods such as color histograms and color space transformations (e.g., RGB to HSV, Lab) to obtain color feature data. Texture analysis algorithms, such as Gray-Level Co-occurrence Matrix (GLCM), are applied to extract texture features from the image, evaluating the texture information of the tea leaves, including metrics such as contrast, entropy, and uniformity, to obtain texture feature data. Edge detection algorithms (e.g., Canny, Sobel) are applied to extract edge features of the tea leaves, and shape descriptors (e.g., Fourier descriptors) are combined to analyze the shape of the leaves, obtaining shape feature data.

[0101] The RGB image is converted to the HSV color space, and the mean and standard deviation of hue, saturation, and brightness are calculated to generate color distribution data. A color histogram is constructed, and each channel is normalized to obtain standardized color features, resulting in color feature data.

[0102] The gray-level co-occurrence matrix is ​​calculated, and features such as contrast, entropy, correlation, and energy are extracted to describe the surface roughness and texture uniformity of tea leaves. Local binary pattern (LBP) algorithm is used to extract local texture features, generating a texture distribution map of tea leaves, thus obtaining texture feature data.

[0103] The boundary information of tea leaves is obtained by edge detection algorithm, and the geometric features of the leaves, such as aspect ratio, curvature, and boundary complexity, are extracted by combining shape descriptors to obtain shape feature data.

[0104] Spatial multi-head self-attention calculation and feature weighted fusion are performed on color feature data, texture feature data and shape feature data to obtain visible light feature data of tea tree growth.

[0105] S22. Near-infrared band features are extracted from near-infrared band data of tea tree growth to obtain near-infrared band feature data of tea tree growth.

[0106] Specifically, near-infrared data is used to calculate the spectral reflectance of tea leaves at different wavelengths to analyze plant health. Combined with visible light data, common vegetation indices, such as the Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI), are calculated to assess the growth health of tea plants. Water-related indices (such as the Normalized Difference Water Index, NDWI) are calculated to assess the water content of the tea plants.

[0107] A radiometric correction algorithm is used to correct sensor response and ambient light interference, ensuring consistency of reflectance data across different wavelengths. The corrected near-infrared data is then interpolated to obtain a complete spectral curve, and reflectance characteristics at key wavelengths are extracted.

[0108] S23. Integrate the visible light characteristic data and the near-infrared band characteristic data of tea tree growth to obtain the spectral characteristic data of tea tree growth.

[0109] Specifically, data from different sources are normalized to ensure that feature values ​​are within the same range, thus preventing any single feature from having an excessive impact on the results. Z-score normalization or min-max normalization methods are used to process the feature data.

[0110] A multimodal feature fusion algorithm is applied to weighted average the visible light and near-infrared features with different weights to form a comprehensive feature set. Principal component analysis (PCA) is used to extract the principal components from the fused features, reducing data redundancy. The PCA algorithm is then used to select the top few principal components with the largest explained variance, reducing the feature dimensionality. Nonlinear dimensionality reduction methods such as t-SNE can be used to further analyze the internal structure of the high-dimensional data, ensuring that important patterns and feature relationships are preserved.

[0111] Optionally, the color texture feature extraction includes:

[0112] The visible light data of tea tree growth is used to segment the region to obtain tea tree region segmentation data;

[0113] Specifically,

[0114] Color space transformation is performed on the tea tree region segmentation data to obtain tea tree color space transformed data;

[0115] Specifically,

[0116] Color histograms were calculated from the color space transformation data of tea trees to obtain tea tree color histogram data;

[0117] Specifically,

[0118] Clustering calculations were performed based on the tea tree color histogram data to obtain the tea tree color histogram clustering data;

[0119] Specifically,

[0120] The dissimilarity of the tea tree color histogram clustering data was calculated to obtain the tea tree color histogram clustering dissimilarity data;

[0121] Specifically,

[0122] When the cluster difference data of the tea tree color histogram is determined to be greater than the preset difference threshold data, the color feature vector is calculated on the tea tree color histogram data to obtain the tea tree color feature data.

[0123] Specifically,

[0124] When the cluster difference data of the tea tree color histogram is determined to be less than or equal to the preset difference threshold data, the tea tree color histogram data is aggregated into regions based on the cluster difference data of the tea tree color histogram to obtain the tea tree color histogram region aggregated data.

[0125] Specifically,

[0126] Color feature vectors are calculated from the aggregated data of the tea tree color histogram regions to obtain tea tree color feature data;

[0127] Specifically,

[0128] Texture features were extracted from visible light data of tea tree growth to obtain tea tree texture feature data;

[0129] Specifically,

[0130] Color and texture feature encoding is performed based on tea tree color feature data and tea tree texture feature data to obtain visible light feature data of tea tree growth.

[0131] Specifically,

[0132] Optionally, the color texture feature encoding includes:

[0133] Spatial location coding is performed based on tea tree color feature data and tea tree texture feature data to obtain spatially encoded tea tree color feature data and spatially encoded tea tree texture feature data, respectively.

[0134] Specifically, positional encoding is performed on the tea tree color feature data and tea tree texture feature data, embedding the positional information of each pixel into the feature data so that the model can capture spatial positional information. Sine and cosine positional embedding functions are used to transform the spatial positions in the image into high-dimensional vectors, which are then added to the color and texture features.

[0135] Based on the spatial coding data of tea tree color features and the spatial coding data of tea tree texture features, multi-head self-attention calculation is performed on the tea tree color feature data and the tea tree texture feature data to obtain the multi-head self-attention feature data of tea tree color and texture.

[0136] Specifically, based on spatially encoded color and texture data, multi-head self-attention computation is performed to learn the complex relationship between the color and texture features of tea plants using multiple attention heads. Each attention head independently learns local features, and then the results of each head are combined to obtain stronger expressive power. The query matrix (Q), key matrix (K), and value matrix (V) are calculated; self-attention is computed on the spatially encoded color and texture feature data using multiple independent attention heads, and then the results of each head are concatenated; the outputs of each attention head are weighted and fused to obtain the color and texture multi-head self-attention feature data.

[0137] Based on the multi-head self-attention feature data of tea tree color and texture, feature weighting and fusion are performed on the tea tree color feature data and tea tree texture feature data to obtain the visible light feature data of tea tree growth.

[0138] Specifically, a weight coefficient is assigned to each color and texture feature, and the weights are calculated based on a weighted matrix generated from the multi-head self-attention features. The tea plant color and texture features are weighted and fused, incorporating the multi-head self-attention features to enhance overall expressive power. Weight coefficients are automatically generated through a learning model to ensure that the contributions of color and texture features are adaptive across different growth stages.

[0139] Optionally, the near-infrared band feature extraction includes:

[0140] Based on the texture feature data of tea trees, the near-infrared band data of tea tree growth is divided into regions to obtain near-infrared band division data of tea trees.

[0141] Specifically, texture feature data of tea plants (such as the texture distribution of leaves) is used to guide the regional segmentation of near-infrared band images. K-means, DBSCAN, or texture-based segmentation algorithms (such as SLIC superpixel segmentation) are used to spatially segment the near-infrared band data and extract different growth regions.

[0142] By calculating the local variance of tea tree texture feature data, salient texture regions are identified and used as a segmentation reference for near-infrared band data. Near-infrared band data points are assigned to different cluster centers (representing different regions) to generate near-infrared band segmentation data.

[0143] Spectral feature points were extracted based on the near-infrared band segmentation data of tea trees to obtain spectral feature point data of tea trees;

[0144] Specifically, within each segmented region, spectral feature points at key wavelengths are identified using spectral analysis techniques. Peak and trough detection algorithms are employed to select feature points in specific bands, such as wavelengths with the highest and lowest reflectance. The spectral curves of the segmented regions are then smoothed, for example, using a Savitzky-Golay filter.

[0145] Peaks and troughs are detected by finding the zeros of the first derivative or the extrema of the second derivative. The spectral curves of the divided regions are smoothed, for example, using a Savitzky-Golay filter. Peaks and troughs are detected by finding the zeros of the first derivative or the extrema of the second derivative.

[0146] Based on visible light data of tea tree growth and near-infrared band division data of tea tree, band ratio feature extraction was performed to obtain band ratio feature data.

[0147] Specifically, by combining visible light data (such as the red band) and near-infrared data, the ratio characteristics of different bands are calculated to analyze the health status of vegetation growth. Vegetation indices such as the Normalized Difference Vegetation Index (NDVI) and the Enhanced Vegetation Index (EVI) are extracted through band ratios.

[0148] Spectral envelope removal analysis was performed on the near-infrared band segmentation data of tea plants to obtain envelope removal feature data;

[0149] Specifically, the spectral envelope removal technique is used to eliminate background interference in the spectral reflectance curve to highlight the specific spectral characteristics of the tea plant. After envelope removal, spectral features, such as the depth and width of the absorption bands, are enhanced for further analysis.

[0150] An envelope is constructed by connecting the local maxima on the spectral curve. Dividing the spectral curve by the envelope yields a characteristic curve with the envelope removed. On the spectral curve after envelope removal, the depth, position, and width of the absorption bands are calculated as features of the spectral envelope removal.

[0151] By integrating the spectral feature point data, band ratio feature data, and envelope removal feature data of tea trees, near-infrared band feature data of tea tree growth is obtained.

[0152] Specifically, spectral feature points, band ratio features, and envelope removal features are integrated (or vectorized) to form a comprehensive feature vector containing multiple spectral information.

[0153] Optionally, the spectral envelope removal analysis includes:

[0154] Multi-resolution smoothing was performed on the near-infrared band segmentation data of tea plants to obtain smoothed near-infrared band segmentation data.

[0155] Specifically, multi-resolution filters are used to smooth near-infrared data, eliminating noise and preserving key spectral features. Wavelet transforms or Gaussian smoothing filters are used to smooth spectral data at different scales to improve the stability of spectral curves.

[0156] Wavelet decomposition is used to break down spectral data into details and approximations at different scales. Soft or hard thresholding is applied to the details to remove noise and preserve key features.

[0157] Multi-scale spline fitting was performed on the smoothed near-infrared band data to obtain envelope fitting data;

[0158] Specifically, multi-scale spline fitting technology is used to fit smooth data in the near-infrared band to obtain the envelope. By establishing the spectral envelope through multi-scale spline interpolation, the local extrema of the spectral curve can be accurately captured. Spline curves are constructed by interpolating between the local extrema of the spectral data to form the envelope. By adjusting the scale parameters of the splines, envelope fitting for spectral features at different scales is achieved, resulting in multi-scale envelope fitting data.

[0159] Dynamic envelope removal was performed based on the near-infrared band segmentation data and envelope fitting data of tea trees to obtain near-infrared band envelope removal data;

[0160] Specifically, based on the near-infrared band segmentation data and fitted envelope data of tea plants, the envelope is dynamically removed to obtain processed spectral data. The spectral curve is divided by the corresponding envelope to eliminate baseline drift and preserve absorption band characteristics. The fitting parameters of the envelope are dynamically adjusted, and the envelope removal process is performed on the spectral curve according to different spectral regions and envelope fitting errors. The fitting strength of the envelope is dynamically adjusted according to the characteristics of the spectral region to avoid overfitting or underfitting. By comparing the changes in the spectral curves before and after removal, the effectiveness of the removal is evaluated, and the parameter settings for envelope removal are further optimized.

[0161] Multi-layer feature decomposition is performed on the near-infrared band envelope removal data to obtain near-infrared band multi-layer feature data;

[0162] Specifically, multi-level feature decomposition is performed on the near-infrared band data after envelope removal to extract various spectral features, such as absorption band intensity, width, and location. Empirical Mode Decomposition (EMD) and wavelet transform are used to decompose the spectral signal into different components to extract key features.

[0163] The envelope-removed spectral data is decomposed into a series of intrinsic mode functions (IMFs), each containing spectral features at different scales. Features such as local extrema and band energy are extracted from each IMF to characterize different levels of near-infrared band features. Wavelet decomposition is then applied to the envelope-removed data to extract spectral features of different frequency components, generating multi-layer feature data.

[0164] Feature fusion is performed on multi-layer feature data in the near-infrared band to obtain envelope-removed feature data.

[0165] Specifically, different weights are assigned to the features at each layer based on their importance, and the features at multiple layers are summed in a weighted manner to generate an envelope to remove the feature data.

[0166] Optionally, S3 includes:

[0167] S31. Calculate the leaf growth index from the visible light characteristic data of tea tree growth to obtain leaf growth index data;

[0168] Specifically, information such as color, texture, and shape of tea leaves is extracted from visible light feature data to calculate the leaf growth index. Based on the color, texture, and other characteristics of tea trees, a biomass index (such as leaf area index LAI) is used as a measure of tea leaf growth.

[0169] The color and shape characteristics of tea leaves are used to estimate the leaf coverage area; the edges of each leaf are detected to estimate the area of ​​each leaf, and the total area is calculated. Leaf texture characteristics (such as surface roughness) are combined to supplement the assessment information of leaf growth status, particularly for evaluating health and growth quality.

[0170] S32. Calculate the moisture content of tea trees by analyzing the near-infrared band characteristic data of tea tree growth;

[0171] Specifically, moisture absorption bands (such as those at 950 nm and 1450 nm) are extracted from near-infrared spectral characteristic data to calculate the moisture content of tea plants. The moisture content is then calculated using a moisture index (such as the normalized normalized moisture index NDWI) and evaluated in conjunction with band reflectance data. Based on changes in the moisture index, the moisture content of the tea plants is estimated using an empirical model, and the moisture content of the tea plants is inferred using band ratios and the intensity of characteristic absorption bands.

[0172] S33. Spectral index calculations are performed based on leaf growth index data and tea tree moisture content data to obtain tea tree growth index data.

[0173] Specifically, leaf growth index (such as LAI) and tea tree moisture content index (such as NDWI) are integrated to form a tea tree growth index, which is used to describe the overall growth status of tea trees. The tea tree growth index is generated by combining leaf growth index and moisture content data through a multiple regression model or machine learning algorithm.

[0174] Optionally, S4 includes:

[0175] S41. Classify the growth status of tea trees according to the tea tree growth index data to obtain tea tree growth status data;

[0176] Specifically, based on tea tree growth index data, the growth status of tea trees is divided into different categories (such as normal growth, accelerated growth, slow growth, or abnormal growth). Supervised learning classification algorithms (such as Support Vector Machine (SVM), Random Forest, or KNN) are applied to classify the growth status of tea trees based on historical growth index data.

[0177] Key features (such as leaf growth index and moisture content index) are extracted from tea tree growth index data and trained using a classification algorithm. The trained model is then used to predict and classify the growth status of tea trees. Cross-validation or hold-out methods are used to evaluate the accuracy of the classification model to ensure the reliability of the growth status classification results.

[0178] S42. Based on the tea tree growth status data, perform time-series monitoring of growth status to obtain time-series data of tea tree growth status.

[0179] Specifically, based on the growth status data of tea trees, a time series model is constructed to continuously monitor the status and record the changing trends of growth status. Time series forecasting algorithms (such as ARIMA and LSTM) are applied to analyze the changing patterns of tea tree growth status over time.

[0180] Based on growth status data, a time series model is constructed to capture the status changes of tea trees at different time points. ARIMA (Autoregressive Moving Average) models or Long Short-Term Memory (LSTM) networks are used for time series prediction. Trend analysis is performed on the growth status of tea trees to identify any acceleration, deceleration, or abnormal fluctuations. Moving averages or exponential smoothing methods are used to smooth the time series data, reducing the impact of short-term fluctuations on the monitoring results.

[0181] S43. Based on the time series data of tea tree growth status, abnormal growth status detection is performed to obtain tea tree growth status monitoring data for tea tree growth cultivation optimization.

[0182] Specifically, based on time-series data, anomalies in the growth status of tea trees are identified, such as excessively rapid or slow growth, or drastic changes in the growth index. Statistical methods or machine learning algorithms (such as Isolation Forest, LOF, and PCA) are used for anomaly detection to uncover potential problems that may occur during the growth of tea trees.

[0183] Threshold-based anomaly detection methods are used to monitor the growth status of tea trees and detect deviations from the normal range. Alternatively, Isolation Forest or Local Outlier Factor (LOF) models are used to model the time-series data of tea tree growth status to identify anomalies that do not conform to expected patterns. Once anomalies are detected, feedback is provided through visualization and alarm mechanisms to optimize and adjust the growth status.

[0184] Optionally, a multispectral monitoring and optimization system for tea tree growth in a tea garden is provided to execute the multispectral monitoring and optimization method for tea tree growth in a tea garden as described above. The multispectral monitoring and optimization system for tea tree growth in a tea garden includes:

[0185] The tea tree growth spectral data acquisition module is used to acquire tea tree growth spectral data;

[0186] The tea tree growth spectral feature extraction module is used to extract tea tree growth spectral features from tea tree growth spectral data to obtain tea tree growth spectral feature data.

[0187] The tea tree growth index calculation module is used to calculate the tea tree growth index based on the tea tree growth spectral characteristic data, and obtain the tea tree growth index data.

[0188] The tea tree growth status monitoring module is used to monitor the growth status of tea trees based on the tea tree growth index data, and obtain tea tree growth status monitoring data for tea tree growth and cultivation optimization.

[0189] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for multispectral monitoring and optimization of tea plant growth in a tea garden, characterized in that, The method comprises: S1, acquiring visible light image of tea tree by visible light acquisition device to obtain visible light data of tea tree growth; acquiring near-infrared wave band of tea tree by near-infrared sensor to obtain near-infrared wave band data of tea tree growth; S2, color texture feature extraction is carried out on the visible light data of tea tree growth to obtain visible light feature data of tea tree growth; near-infrared wave band feature extraction is carried out on the near-infrared wave band data of tea tree growth to obtain near-infrared wave band feature data of tea tree growth; the visible light feature data of tea tree growth and the near-infrared wave band feature data of tea tree growth are integrated to obtain the spectral feature data of tea tree growth; S3, calculating tea tree growth index according to tea tree growth spectral feature data to obtain tea tree growth index data, wherein the tea tree growth spectral data comprises tea tree growth visible light data and tea tree growth near-infrared wave band data; S4, monitoring growth state according to tea tree growth index data to obtain tea tree growth state monitoring data for tea tree growth cultivation optimization operation; The color texture feature extraction comprises: region segmentation is carried out on the visible light data of tea tree growth to obtain tea tree region segmentation data; color space conversion is carried out on the tea tree region segmentation data to obtain tea tree color space conversion data; color histogram calculation is carried out on the tea tree color space conversion data to obtain tea tree color histogram data; cluster calculation is carried out according to the tea tree color histogram data to obtain tea tree color histogram cluster data; difference degree calculation is carried out on the tea tree color histogram cluster data to obtain tea tree color histogram cluster difference degree data; when it is determined that the tea tree color histogram cluster difference degree data is greater than the preset difference degree threshold data, color feature vector calculation is carried out on the tea tree color histogram data to obtain tea tree color feature data; when it is determined that the tea tree color histogram cluster difference degree data is less than or equal to the preset difference degree threshold data, region aggregation is carried out on the tea tree color histogram data according to the tea tree color histogram cluster difference degree data to obtain tea tree color histogram region aggregation data; color feature vector calculation is carried out on the tea tree color histogram region aggregation data to obtain tea tree color feature data; texture feature extraction is carried out on the visible light data of tea tree growth to obtain tea tree texture feature data; color texture feature coding is carried out according to the tea tree color feature data and the tea tree texture feature data to obtain the visible light feature data of tea tree growth.

2. The method of claim 1, wherein, The color texture feature coding comprises: space position coding is carried out according to the tea tree color feature data and the tea tree texture feature data to obtain tea tree color feature space coding data and tea tree texture feature space coding data respectively; multi-head self-attention calculation is carried out on the tea tree color feature data and the tea tree texture feature data according to the tea tree color feature space coding data and the tea tree texture feature space coding data to obtain tea tree color texture multi-head self-attention feature data; feature weighted fusion is carried out on the tea tree color feature data and the tea tree texture feature data according to the tea tree color texture multi-head self-attention feature data to obtain the visible light feature data of tea tree growth.

3. The method of claim 1, wherein, The near-infrared wave band feature extraction comprises: According to the tea tree texture feature data, the tea tree growth near-infrared band data is divided into regions to obtain tea tree near-infrared band division data; According to the tea tree near-infrared band division data, spectral feature points are extracted to obtain tea tree spectral feature point data; According to the tea tree growth visible light data and the tea tree near-infrared band division data, band ratio features are extracted to obtain band ratio feature data; The tea tree near-infrared band division data is analyzed by removing the spectral envelope to obtain envelope removal feature data; The tea tree spectral feature point data, band ratio feature data and envelope removal feature data are integrated to obtain tea tree growth near-infrared band feature data.

4. The method of claim 3, wherein, The spectral envelope removal analysis includes: The tea tree near-infrared band division data is processed by multi-resolution smoothing to obtain near-infrared band division smoothing data; The near-infrared band division smoothing data is fitted by multi-scale spline to obtain envelope fitting data; According to the tea tree near-infrared band division data and the envelope fitting data, dynamic envelope removal is performed to obtain near-infrared band envelope removal data; The near-infrared band envelope removal data is decomposed by multi-layer feature to obtain near-infrared band multi-layer feature data; The near-infrared band multi-layer feature data is fused to obtain envelope removal feature data.

5. The method of claim 1, wherein S3 It includes: The tea tree growth visible light feature data is calculated by leaf growth index to obtain leaf growth index data; The tea tree growth near-infrared band feature data is calculated by tea tree water content to obtain tea tree water content data; According to the leaf growth index data and the tea tree water content data, the spectral index is calculated to obtain the tea tree growth index data.

6. The method according to claim 1, characterized by S4 It includes: According to the tea tree growth index data, the tea tree growth state classification is performed to obtain tea tree growth state data; According to the tea tree growth state data, the growth state time series monitoring is performed to obtain tea tree growth state time series data; According to the tea tree growth state time series data, the growth state anomaly detection is performed to obtain tea tree growth state monitoring data for tea tree growth cultivation optimization operation.

7. A multispectral monitoring and optimization system for tea plant growth in a tea garden, characterized in that, The tea tree growth in the tea garden is used to perform the multi-spectral monitoring and optimization method as claimed in claim 1, and the multi-spectral monitoring and optimization system for tea tree growth in the tea garden includes: A tea tree growth spectrum data acquisition module is used to acquire tea tree growth spectrum data; A tea tree growth spectrum feature extraction module is used to extract tea tree growth spectrum features according to tea tree growth spectrum data to obtain tea tree growth spectrum feature data; A tea tree growth index calculation module is used to calculate tea tree growth index according to tea tree growth spectrum feature data to obtain tea tree growth index data; A tea tree growth state monitoring module is used to monitor the growth state according to the tea tree growth index data to obtain tea tree growth state monitoring data for tea tree growth cultivation optimization operation.

Images