Method and device for dynamically evaluating growth vigor of fruit trees

Abstract

The invention provides a dynamic evaluation method and device for fruit tree growth vigor, and the method comprises the steps: obtaining unmanned plane remote sensing data and ground survey data of a target region in the whole annual growth period, and obtaining three-dimensional point cloud data and multispectral orthoimage data based on the unmanned plane remote sensing data; extracting fruit tree phenotype index data in the three-dimensional point cloud data and fruit tree greenness index data in the multispectral orthoimage data; carrying out fusion analysis on the ground survey data, the fruit tree phenotype index data and the fruit tree greenness index data, and constructing a time sequence fruit tree growth comprehensive evaluation model; and inputting the obtained remote sensing data of the to-be-tested unmanned aerial vehicle in the target area into the time sequence fruit tree growth comprehensive evaluation model to obtain a fruit tree growth evaluation result. By combining unmanned aerial vehicle remote sensing data and ground survey data, fruit trees are monitored from two dimensions of spatial structure and spectral information. According to the invention, multi-dimensional evaluation of the growth condition of the fruit tree is realized, and the evaluation accuracy is improved.

Description

Technical Field

[0001] The invention relates to the technical field of fruit tree growth information monitoring, and in particular to a fruit tree growth dynamic evaluation method and device. Background Art

[0002] The growth of fruit trees is a comprehensive reflection of their health status. It is of great significance for guiding the management strategies in the actual planting process, improving the yield and quality of fruit trees, reducing production costs, and increasing the economic benefits of fruit farmers. Traditional fruit tree growth assessment mainly relies on field sampling surveys, which are inefficient, costly, and limited by the distribution and number of sample points, making it difficult to meet the needs of large-scale rapid monitoring. The rapid development of UAV remote sensing platforms has provided new possibilities for rapid monitoring of fruit tree growth. The existing fruit tree growth monitoring methods based on UAV remote sensing platforms have relatively single monitoring methods and cannot comprehensively evaluate the changes in fruit tree spectral images and three-dimensional information, which in turn limits the accurate evaluation of changes in fruit tree growth. Summary of the invention

[0003] The present invention provides a method and device for dynamically evaluating the growth of fruit trees, which is used to solve the defect that the existing technology cannot comprehensively evaluate the spectral image and three-dimensional information changes of fruit trees, and realize accurate evaluation of the growth of fruit trees. The technical solution proposed by the present invention is as follows: In a first aspect, the present invention provides a method for dynamically evaluating the growth of fruit trees, comprising: Acquire UAV remote sensing data and ground survey data of the target area throughout the annual growth period, and acquire three-dimensional point cloud data and multispectral orthophoto data based on the UAV remote sensing data; Extracting fruit tree phenotypic index data based on the three-dimensional point cloud data, and extracting fruit tree greenness index data based on the multispectral orthophoto data; Performing a fusion analysis on the ground survey data, the fruit tree phenotypic index data and the fruit tree greenness index data to construct a time-series comprehensive evaluation model for fruit tree growth; The remote sensing data of the unmanned aerial vehicle to be tested in the target area is obtained, and the remote sensing data of the unmanned aerial vehicle to be tested is input into the time-series fruit tree growth comprehensive evaluation model to obtain the fruit tree growth evaluation result.

[0004] Optionally, extracting fruit tree phenotypic indicator data based on the three-dimensional point cloud data includes: Classifying the three-dimensional point cloud data and extracting the fruit tree point cloud; The point cloud data set is segmented into individual trees to extract fruit phenotype index data of individual fruit trees in the target area.

[0005] Optionally, the fruit tree greenness index data includes each broadband greenness index value; and the step of extracting the fruit tree greenness index data based on the multispectral orthophoto data includes: The multispectral orthophoto data were used to calculate the broadband greenness index in the target area; Obtain canopy area distribution data of individual fruit trees in the target area; According to the broadband greenness index values ​​of individual fruit trees in the target area and the canopy area distribution data of individual fruit trees, the broadband greenness index values ​​of individual fruit trees are calculated; among them, the broadband greenness index includes the normalized difference vegetation index, ratio vegetation index, enhanced vegetation index, visible atmospheric resistance index, leaf chlorophyll index, super green index and green leaf index.

[0006] Optionally, the entire annual growth period includes multiple growth stages, and different growth stages correspond to different time labels; the ground survey data, the fruit tree phenotypic index data and the fruit tree greenness index data are fused and analyzed to construct a time-series fruit tree growth comprehensive evaluation model, including: Dividing the fruit tree phenotypic index data and the fruit tree greenness index data according to time tags; Based on the ground survey data, the fruit tree phenotypic index data and the fruit tree greenness index data of the same time label are fused and analyzed to obtain a fruit tree growth assessment sub-model at the corresponding growth stage; Optimizing the fruit tree growth assessment sub-model using the ground survey data to obtain an optimized fruit tree growth assessment sub-model; The optimized fruit tree growth assessment sub-models at different growth stages during the entire growth period are integrated to obtain the time-series fruit tree growth comprehensive assessment model.

[0007] Optionally, based on the ground survey data, the fruit tree phenotypic index data and the fruit tree greenness index data of the same time label are fused and analyzed to obtain a fruit tree growth assessment sub-model at a corresponding growth stage, including: The fruit tree phenotypic index data and the fruit tree greenness index data are fused using a multi-objective particle swarm optimization algorithm based on the R2 index; wherein the R2 index is used to evaluate the quality of a growth measurement index set obtained by weighting multiple objective functions by a weight vector relative to a corresponding growth reference point; in, represents the R2 index, Represents a set of growth measurement indicators, Represents a set of weight vectors, where The target space is evenly distributed. Indicates the growth reference point, represents the number of weight vectors in the weight vector set, Distributed in The probability of represents the first A weight vector, Indicated in The first The objective function, The number of objective functions, Indicates fruit tree phenotypic index data or fruit tree greenness index data; Through the multi-objective particle swarm optimization algorithm Search for weight vector combinations in the target space to maximize the R2 index and obtain the target index weight combination corresponding to the growth stage; According to the target indicator weight combination and the growth measurement indicator set, a fruit tree growth assessment sub-model corresponding to the growth stage is constructed.

[0008] Optionally, before the fruit tree phenotypic index data and the fruit tree greenness index data of the same time tag are fused and analyzed based on the ground survey data to obtain a fruit tree growth assessment sub-model corresponding to the growth stage, the method further includes: The fruit tree phenotypic index data and the fruit tree greenness index data are continuously interpolated.

[0009] In a second aspect, the present invention further provides a device for dynamically evaluating the growth of fruit trees, comprising the following modules: A data acquisition module is used to acquire UAV remote sensing data and ground survey data of the target area during the entire annual growth period, and to acquire three-dimensional point cloud data and multi-spectral orthophoto data based on the UAV remote sensing data; A data extraction module, used for extracting fruit tree phenotypic index data based on the three-dimensional point cloud data, and for extracting fruit tree greenness index data based on the multispectral orthophoto data; A model building module is used to integrate and analyze the ground survey data, the fruit tree phenotypic index data and the fruit tree greenness index data to build a time-series comprehensive evaluation model for fruit tree growth; The growth assessment module is used to obtain the remote sensing data of the unmanned aerial vehicle to be tested in the target area, input the remote sensing data of the unmanned aerial vehicle to be tested into the time-series fruit tree growth comprehensive assessment model, and obtain the fruit tree growth assessment result.

[0010] In a third aspect, the present invention further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, the method for dynamically evaluating the growth of fruit trees as described in the first aspect above is implemented.

[0011] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for dynamically evaluating the growth of fruit trees as described in the first aspect above.

[0012] In a fifth aspect, the present invention further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the method for dynamically evaluating the growth of fruit trees as described in the first aspect above.

[0013] Based on the above technical solution, the beneficial effects of the present invention compared with the prior art are as follows: The method and device for dynamic evaluation of fruit tree growth provided by the present invention monitor fruit trees from two dimensions: spatial structure and spectral information, by combining UAV remote sensing data and ground survey data. This fusion of multi-source data not only improves the richness and accuracy of the data, but also can more comprehensively reflect the growth changes of fruit trees. By extracting fruit tree phenotypic indicators from three-dimensional point cloud data and greenness indicators from multispectral orthophoto data, a multi-dimensional evaluation of the growth status of fruit trees is achieved. This joint analysis not only improves the accuracy of the evaluation, but also can reveal potential problems and trends in the growth process of fruit trees.

[0014] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0015] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 This is one of the flow charts of the method for dynamically evaluating the growth of fruit trees provided by the present invention.

[0018] Figure 2 This is the second flow chart of the method for dynamically evaluating the growth of fruit trees provided by the present invention.

[0019] Figure 3 It is a structural schematic diagram of a device for dynamically evaluating the growth of fruit trees provided by the present invention.

[0020] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] Combine the following Figure 1-Figure 3 The invention describes a method and device for dynamically evaluating the growth of fruit trees.

[0023] Reference Figure 1 As shown, the method for dynamic evaluation of fruit tree growth includes the following steps: Step S110, obtaining UAV remote sensing data and ground survey data of the target area throughout the annual growth period, and obtaining three-dimensional point cloud data and multispectral orthophoto data based on the UAV remote sensing data.

[0024] During the entire annual growth period of the fruit tree, the target area is monitored by drones equipped with corresponding sensors to obtain three-dimensional point cloud data and multispectral orthophoto data. These data provide the spatial structure and spectral information of the fruit tree growth. Specifically, multiple key nodes (i.e. growth stages) in the annual growth period of the orchard are selected as the key periods for drone flight monitoring, such as budding, flowering, fruit expansion, fruiting, and dormancy. These key nodes can fully reflect the physiological and ecological characteristics of fruit trees at different growth stages. UAVs are equipped with sensors such as high-resolution cameras and lidar to ensure that the spatial structure and spectral information of fruit tree growth can be captured. According to the preset flight plan and route, the drone conducts flight monitoring of the target area to ensure coverage of the entire orchard area. During the flight, the sensor collects three-dimensional point cloud data and multispectral orthophoto data of the fruit tree in real time. The collected drone remote sensing data is preprocessed, including denoising, registration, correction and other steps to ensure the accuracy and reliability of the data. Subsequently, professional software tools are used to extract three-dimensional point cloud data and multispectral orthophoto data to provide a basis for subsequent analysis and evaluation.

[0025] While the drone remote sensing data is being acquired, a field survey is conducted on the orchards in the target area. The survey content includes the location information of the fruit trees, the date of collection, the growth period, and the growth trend. During the field survey, professional measurement tools are used to measure the original geometric parameters of the fruit trees in situ, including the length, width, and height of the fruit trees. These measurement data provide a basis for the subsequent calculation of the ground truth values ​​of the height and volume of the fruit trees. All data obtained from the field survey are recorded in detail and organized into a spreadsheet or database format to facilitate subsequent data analysis and fusion processing.

[0026] Step S120, extracting fruit tree phenotypic index data based on the three-dimensional point cloud data, and extracting fruit tree greenness index data based on the multispectral orthophoto data.

[0027] The three-dimensional point cloud data acquired by the drone is processed using point cloud processing software to extract phenotypic index data such as the height, crown size, and branch distribution of the fruit tree. Specifically, first, the three-dimensional point cloud data acquired by the drone is preprocessed using point cloud processing software (such as CloudCompare, PCL, etc.). This includes steps such as removing noise points, point cloud filtering, and point cloud registration to ensure the accuracy and integrity of the point cloud data. Secondly, the point cloud of the fruit tree is identified and segmented from the entire point cloud data using the spatial distribution characteristics and density information of the point cloud data through cluster analysis, threshold segmentation, and other methods. Then the phenotypic index is extracted, which includes: calculating the difference between the maximum and minimum values ​​of the point cloud of the fruit tree in the vertical direction to obtain the height of the fruit tree. The horizontal projection area or volume of the crown of the fruit tree is calculated using the convex hull algorithm or grid division method of the point cloud data to characterize the size of the crown. The distribution pattern of the branches of the fruit tree, such as the density, length, and angle of the branches, is analyzed through the spatial distribution characteristics of the point cloud data.

[0028] Spectral analysis is performed on the multispectral orthophoto data to extract the greenness index of fruit trees (such as NDVI, etc.), reflecting the chlorophyll content and photosynthesis intensity of fruit trees. Specifically, the multispectral orthophoto data is corrected and registered to ensure the geometric and radiometric accuracy of the image. As needed, the image is cropped to include only the target fruit tree area. Using the different band information of the multispectral orthophoto, spectral analysis is performed to calculate the ratio or difference between different bands to extract a specific greenness index. Taking the normalized difference vegetation index (NDVI) as an example, it is obtained by calculating the difference and sum ratio of the near-infrared band and the red light band. The higher the NDVI value, the higher the chlorophyll content of the fruit tree and the greater the photosynthesis intensity. According to research needs, other greenness indices can also be extracted, such as the green normalized difference vegetation index (GNDVI), the soil adjusted vegetation index (SAVI), etc., to more comprehensively reflect the greenness of fruit trees.

[0029] Step S130, performing a fusion analysis on the ground survey data, the fruit tree phenotypic index data and the fruit tree greenness index data to construct a time-series comprehensive evaluation model for fruit tree growth.

[0030] Integrate ground survey data (such as tree location information, growth stage, growth, geometric parameters measured in situ, etc.), fruit tree phenotypic index data (such as height, crown size, branch distribution, etc.) and fruit tree greenness index data (such as NDVI, etc.) to form a complete data set. Process missing values ​​and outliers in the data set to ensure the integrity and accuracy of the data. Standardize data from different sources and dimensions for subsequent analysis and modeling. Fusion analysis of ground survey data, fruit tree phenotypic index data and fruit tree greenness index data, considering the combined effects of multiple factors. Use statistical analysis, machine learning and other methods to construct a time-series comprehensive evaluation model for fruit tree growth. This model can reflect the growth status and trends of fruit trees at different growth stages.

[0031] For example, statistical methods (such as Pearson correlation coefficient, Spearman rank correlation coefficient, etc.) are used to analyze the correlation between ground survey data, fruit tree phenotypic index data, and fruit tree greenness index data to identify the key factors affecting the growth of fruit trees. According to the results of the correlation analysis, the features that have a significant impact on the growth of fruit trees are selected for subsequent analysis. If there are a large number of features, feature dimension reduction techniques (such as principal component analysis PCA, linear discriminant analysis LDA, etc.) can be used to reduce the number of features and improve the generalization ability of the model. According to the characteristics of the data and the purpose of analysis, a suitable fusion strategy is selected. For example, ground survey data can be used as prior information and fused with fruit tree phenotypic index data and greenness index data; or different types of data can be weighted and fused to comprehensively consider the influence of multiple factors. Then, time series analysis, regression models in machine learning (such as linear regression, ridge regression, Lasso regression, etc.) or deep learning models (such as recurrent neural network RNN, long short-term memory network LSTM, etc.) are used to establish a comprehensive evaluation model for the growth of time-series fruit trees. The model is trained using the integrated data set, and the performance of the model is optimized by adjusting the parameters and structure of the model. Cross-validation, grid search and other methods can be used to find the optimal model parameters. Use an independent validation data set to validate the model and evaluate the accuracy and generalization ability of the model. The performance of the model can be evaluated using indicators such as accuracy, recall, F1 score, mean square error (MSE), etc. The model is applied to actual scenarios to monitor and predict the growth of fruit trees in real time, providing a scientific basis for fruit tree planting and management.

[0032] The above-mentioned method of standardizing data from different sources and dimensions can be: The Min-Max standardization method was used to standardize the phenotypic index data and greenness index data of various fruit trees to eliminate the influence of the variation dimension and variation range. The formula is: in, is the standardized data, is the original data, is the minimum value of the empirical constant, is the maximum value of the empirical constant.

[0033] Step S140, obtaining the remote sensing data of the drone to be tested in the target area, inputting the remote sensing data of the drone to be tested into the time-series fruit tree growth comprehensive evaluation model, and obtaining the fruit tree growth evaluation result.

[0034] Use drones equipped with corresponding sensors to conduct flight monitoring of the target area according to the preset flight plan and route to obtain the drone remote sensing data to be evaluated. Preprocess the acquired drone remote sensing data, including denoising, registration, correction and other steps to ensure the accuracy and reliability of the data. Extract three-dimensional point cloud data and multispectral orthophoto data from the processed drone remote sensing data.

[0035] The extracted 3D point cloud data was processed using point cloud processing software to identify and segment the fruit tree point cloud. The phenotypic index data of fruit trees, such as the height, crown size, and branch distribution of fruit trees, were extracted by calculating and analyzing the spatial distribution characteristics of the fruit tree point cloud. The extracted multispectral orthophoto data was subjected to spectral analysis to calculate the fruit tree greenness index data (such as NDVI, etc.). The fruit tree greenness index data reflects the chlorophyll content and photosynthesis intensity of fruit trees, and is an important indicator for evaluating the growth of fruit trees.

[0036] The extracted fruit tree phenotypic index data and fruit tree greenness index data are integrated to form a complete data set. The integrated data set is input into the established time-series fruit tree growth comprehensive evaluation model. The model processes and analyzes the input data and outputs the fruit tree growth evaluation results. The evaluation results can include information such as the growth rate, health status, and yield forecast of the fruit trees. Based on the evaluation results, decision support is provided for fruit tree management. For example, according to the growth rate and health status of the fruit trees, a reasonable fertilization and irrigation plan can be formulated; according to the yield forecast results, the planting density and picking strategy of the fruit trees can be adjusted.

[0037] The method for dynamic evaluation of fruit tree growth provided by the present invention combines UAV remote sensing data and ground survey data to monitor fruit trees from two dimensions: spatial structure and spectral information. This fusion of multi-source data not only improves the richness and accuracy of the data, but also can more comprehensively reflect the growth changes of fruit trees. By extracting fruit tree phenotypic indicators from three-dimensional point cloud data and greenness indicators from multispectral orthophoto data, this method realizes a multi-dimensional evaluation of the growth status of fruit trees. This joint analysis not only improves the accuracy of the evaluation, but also can reveal potential problems and trends in the growth process of fruit trees.

[0038] Moreover, this method established a comprehensive time-series fruit tree growth assessment model by integrating ground survey data, three-dimensional point cloud data, and multispectral orthophoto data. This model is not only suitable for the growth monitoring of a single fruit tree, but can also be extended to the growth assessment of a large number of fruit trees, improving the versatility and practicality of monitoring. By adjusting the parameters and variables in the model, this method can adapt to fruit trees of different types and different growth environments, and achieve accurate monitoring of the growth of a large number of fruit trees.

[0039] In an optional embodiment, the step S120 described above of extracting fruit tree phenotypic indicator data based on the three-dimensional point cloud data includes: S1201. Classify the three-dimensional point cloud data and extract the fruit tree point cloud.

[0040] The cloth simulation filtering method in the MATLAB toolbox is used to filter and smooth the 3D point cloud data to reduce the impact of noise and outliers and improve the accuracy and reliability of the data. The improved progressive encryption triangulation filtering algorithm is used to classify the filtered point cloud to distinguish the fruit tree point cloud from other ground point clouds (such as the ground, buildings, other vegetation, etc.). Through classification processing, the fruit tree point cloud in the target area is extracted to provide a basis for subsequent analysis.

[0041] S1202, performing single-tree segmentation on the point cloud data set to extract fruit phenotypic index data of a single fruit tree in the target area.

[0042] The watershed segmentation algorithm was used to segment the fruit tree point cloud to obtain an independent point cloud dataset for each fruit tree. Based on the segmented fruit tree point cloud dataset, phenotypic indicators such as crown area, single tree volume, and number of point clouds were calculated. By calculating the ratio of single tree volume to the number of point clouds, the single tree density was extracted, which reflects the growth status and spatial distribution characteristics of fruit trees. The ground survey data was combined with the visual interpretation results to evaluate the accuracy of the extracted fruit tree phenotypic indicators (including crown area, single tree volume, and single tree density) to ensure the reliability of the extraction results. The fruit trees in the target area were continuously monitored throughout the annual growth period, and the single tree volume and single tree density data at each growth stage were extracted to evaluate the growth trend and health of the fruit trees.

[0043] Three-dimensional point cloud data can accurately reflect the three-dimensional morphological structure of fruit trees, including the extension range of the crown, the density of branches and leaves, and the shape of the trunk. This refined description provides a solid foundation for the accurate measurement of fruit tree phenotypic indicators, making the extraction of key parameters such as crown area and single tree volume more accurate and reliable. The present invention obtains more advanced phenotypic indicators such as single tree density and crown structure complexity by further analyzing and processing the three-dimensional point cloud data. The comprehensive evaluation of these multiple indicators can more comprehensively reflect the growth status and health status of fruit trees, providing a strong basis for the scientific management of fruit trees. The extraction of fruit tree phenotypic indicators based on three-dimensional point cloud data can realize continuous monitoring and long-term tracking of the growth status of fruit trees. By comparing and analyzing data at different time points, the trend and law of fruit tree growth can be revealed, providing important decision-making support for fruit tree management and agricultural production.

[0044] In an optional embodiment, the fruit tree greenness index data includes each broadband greenness index value. The step S120 described above of extracting the fruit tree greenness index data based on the multispectral orthophoto data includes: S210, calculating each broadband greenness index in the target area using the multispectral orthophoto data.

[0045] Acquire multispectral orthophoto data of the target area and ensure good data quality, including appropriate resolution, low noise, and accurate geographic positioning. Use ENVI software to preprocess the multispectral orthophoto data, such as radiation correction, atmospheric correction, and geometric correction, to eliminate atmospheric effects, sensor errors, and terrain effects, and improve data accuracy and reliability.

[0046] Based on the preprocessed multispectral image data and the band information of the multispectral image data, the broadband greenness index in the target area is calculated. These indices are usually calculated based on the reflectance difference between the visible light and near-infrared bands. Each broadband greenness index is an indicator that can easily measure the amount of green vegetation and its growth status. It has a high sensitivity to chlorophyll content, leaf surface canopy and canopy structure. The sensitivity varies for different growth stages of fruit trees. Broadband greenness indices include normalized difference vegetation index (NDVI), ratio vegetation index (SR), enhanced vegetation index (EVI), visible atmospheric resistance index (ARVI), leaf chlorophyll index (such as the chlorophyll content estimate calculated by the reflectance of a specific band), super green index (such as the index obtained by enhancing the reflectance of the green band) and green leaf index (such as the index reflecting the green leaf condition obtained by integrating the reflectance of multiple bands).

[0047] Normalized Differential Vegetation Index (NDVI): .

[0048] Ratio Vegetation Index (RVI): .

[0049] Enhanced Vegetation Index (EVI): .

[0050] Visible Atmospherically Resistance Index (VARI): .

[0051] Leaf Chlorophyll Index (LCI): .

[0052] Super Green Index (SGI): .

[0053] Green Leaf Index (GLI): .

[0054] in, , , , , It is the reflectivity of the blue, green, red, red edge, and near-infrared bands at this location.

[0055] S220. Obtain canopy area distribution data of a single fruit tree in the target area.

[0056] Multispectral image data are segmented and classified using object-oriented or pixel-based classification methods to distinguish between fruit tree canopies and other ground objects (such as soil, buildings, other vegetation, etc.). The boundaries and ranges of the fruit tree canopies are extracted through the classified image data. The canopy area of ​​each fruit tree is calculated based on the extracted boundaries and ranges of the fruit tree canopies. This can be achieved by calculating the area of ​​the polygon enclosed by the canopy boundaries. The canopy area distribution data reflects the spatial distribution and density of fruit trees, and provides a basis for the subsequent calculation of the broadband greenness index value of a single fruit tree.

[0057] S230. Calculate the broadband greenness index values ​​of a single fruit tree according to the broadband greenness index values ​​in the target area and the canopy area distribution data of a single fruit tree.

[0058] For each fruit tree, each fruit tree is matched with its corresponding greenness index according to the extracted crown area and distribution data of the fruit trees. Through matching processing, the broadband greenness index values ​​corresponding to the single fruit tree are extracted, which reflect the key information such as the greenness status, chlorophyll content and health status of the fruit tree. The fruit trees in the target area are continuously monitored throughout the annual growth period, and the broadband greenness index values ​​of each growth stage are extracted. Specifically, the canopy area is taken as the region of interest (ROI), and then the reflectance data of each band in the area are extracted and calculated according to the calculation formula of the broadband greenness index. The broadband greenness index values ​​of the single fruit tree obtained reflect the key information such as the greenness status, chlorophyll content, and health status of the fruit tree.

[0059] The present invention uses multispectral orthophoto data and image processing technology to accurately calculate the broadband greenness index and canopy area distribution data of a single fruit tree in the target area. The extracted fruit tree greenness index data has high accuracy and reliability, and can accurately reflect the growth status and greenness characteristics of the fruit trees. This method is based on remote sensing image data for extraction and analysis, without the need for field sampling or destruction of fruit trees. Therefore, this method can perform multiple monitoring and evaluations without affecting the normal growth of fruit trees. Using remote sensing technology and image processing technology, large-scale and rapid monitoring and analysis of target areas can be achieved. Compared with traditional ground survey methods, this method has higher monitoring efficiency and lower costs.

[0060] In an optional embodiment, the entire annual growth period includes multiple growth stages, and different growth stages correspond to different time labels; the above-mentioned step S130 performs a fusion analysis on the ground survey data, the fruit tree phenotypic index data, and the fruit tree greenness index data to construct a time-series fruit tree growth comprehensive evaluation model, including: S1301. Divide the fruit tree phenotypic index data and the fruit tree greenness index data according to time tags.

[0061] Fruit trees will show different growth characteristics and physiological states at different growth stages. Before building a comprehensive evaluation model for the growth of time-series fruit trees, it is necessary to add different time tags according to the growth stages of fruit trees to divide the collected fruit tree phenotypic index data and fruit tree greenness index data. This step includes determining the growth stage of the fruit tree (such as budding period, flowering period, fruit expansion period, fruiting period, dormancy period, etc.) and time tags (corresponding time of the collection experiment), and classifying the corresponding data into the corresponding time tag data set.

[0062] S1302. Based on the ground survey data, the fruit tree phenotypic index data and the fruit tree greenness index data of the same time label are fused and analyzed to obtain a fruit tree growth assessment sub-model for the corresponding growth stage.

[0063] After determining the growth stage of the fruit tree, it is necessary to perform a fusion analysis on the fruit tree phenotypic index data and the fruit tree greenness index data with the same time label. The model can be constructed using machine learning algorithms (such as support vector machines (SVM), random forests (RF), etc.) or deep learning algorithms (such as convolutional neural networks (CNN), recurrent neural networks (RNN), etc.). Through fusion analysis, a fruit tree growth assessment sub-model for the corresponding growth stage can be established, which can reflect the growth characteristics and physiological state of the fruit tree at this growth stage.

[0064] S1303, optimizing the fruit tree growth assessment sub-model using the ground survey data to obtain an optimized fruit tree growth assessment sub-model.

[0065] The ground survey data includes information such as the growth status of fruit trees, pests and diseases, and soil conditions observed in the field. These data can be used as true values ​​or reference values ​​to optimize the fruit tree growth assessment sub-model. Match the ground survey data with the input data of the fruit tree growth assessment sub-model. Clean and standardize the ground survey data. Use the ground survey data to verify the fruit tree growth assessment sub-model. Adjust the model parameters or weight vectors based on the verification results to improve the accuracy and reliability of the model, and save the optimized fruit tree growth assessment sub-model.

[0066] When verifying the fruit tree growth assessment sub-model, the fruit tree growth assessment results were compared with the results of the field survey, and the performance of the model was comprehensively evaluated using indicators such as accuracy (ACU), precision, recall, and F1-score. Taking accuracy (ACU) as an example: Where: is the number of correctly judged growth potential in the survey sample; It is the number of incorrectly judged growth potential in the survey sample.

[0067] S1304, integrating the optimized fruit tree growth assessment sub-models at different growth stages during the entire annual growth period to obtain the temporal fruit tree growth comprehensive assessment model.

[0068] Collect the optimized fruit tree growth assessment sub-models at different growth stages throughout the growth period. Determine the model fusion strategy according to actual needs, such as weighted average, voting decision, etc. Use the determined fusion strategy to fuse the models at different growth stages throughout the growth period. Use an independent validation data set to verify the fused time-series fruit tree growth comprehensive assessment model, save and output the time-series fruit tree growth comprehensive assessment model. It is also possible to fuse the optimized fruit tree growth assessment sub-models at different growth stages throughout the growth period into a time-series fruit tree growth comprehensive assessment model. In actual application, the fruit tree phenotypic indicator data and the fruit tree greenness indicator data, as well as the corresponding time tags, are input into the time-series fruit tree growth comprehensive assessment model, and the model will call the fruit tree growth assessment sub-model at the corresponding growth stage for growth assessment.

[0069] The present invention can more accurately analyze the growth characteristics and physiological states of fruit trees at different growth stages through data partitioning, and provide reliable data support for the subsequent establishment of a fruit tree growth assessment submodel. Through fusion analysis, the fusion analysis can make full use of the complementarity of fruit tree phenotypic index data and fruit tree greenness index data to improve the accuracy and reliability of the model. At the same time, by constructing a fruit tree growth assessment submodel at different growth stages throughout the growth period, the growth status of fruit trees can be more comprehensively assessed. Through the verification and optimization of ground survey data, the accuracy and reliability of the fruit tree growth assessment submodel can be further improved. At the same time, the optimized model can better adapt to the actual production environment and provide more accurate and reliable decision support for fruit tree management and agricultural production. By integrating the optimized fruit tree growth assessment submodel at different growth stages throughout the growth period, a comprehensive assessment model that can reflect the growth status and physiological state of fruit trees throughout the annual growth period can be constructed. The model can provide more comprehensive and accurate data support for fruit tree management, pest and disease monitoring, nutritional diagnosis, yield prediction, etc. At the same time, the time-series fruit tree growth comprehensive assessment model can also provide a more scientific decision-making basis for agricultural production, and improve agricultural production efficiency and economic benefits.

[0070] In an optional embodiment, the above step S1302, based on the ground survey data, performs a fusion analysis on the fruit tree phenotypic index data and the fruit tree greenness index data of the same time tag to obtain a fruit tree growth assessment sub-model corresponding to the growth stage, including: S13021. Use a multi-objective particle swarm optimization algorithm based on the R2 index to fuse the fruit tree phenotypic index data and the fruit tree greenness index data; wherein the R2 index is used to evaluate the quality of the growth measurement index set obtained by weighting multiple objective functions by a weight vector relative to the corresponding growth reference point; in, represents the R2 indicator, which is the defined risk or performance measure; Represents a set of growth measurement indicators; is a set of weight vectors, representing a set of weight vectors, which are in The target space is evenly distributed, and each weight vector Each objective has a corresponding weight assignment, which is used to balance multiple objectives.

[0071] Indicates the growth reference point; Represents the number of weight vectors in the weight vector set.

[0072] Distributed in The probability on each weight vector The probability of being selected is equal.

[0073] Is a collection An element in the weight vector set A weight vector, which is used to make trade-offs between multiple objectives.

[0074] Indicated in The first objective function; The number of objective functions; It is an input variable or decision variable, representing the fruit tree phenotypic index data or the fruit tree greenness index data.

[0075] This summation symbol represents the sum of the set All weight vectors in Perform the summation.

[0076] Represents the weight vector With the objective function and The dot product of the difference between Relative to threshold The weighted deviation of .

[0077] This maximum operation is in all targets It finds the target with the largest weighted deviation among all targets.

[0078] This minimum operation is performed on all weight vectors It tries to find the weight vector that minimizes the maximum weighted deviation.

[0079] S13022, through the multi-objective particle swarm optimization algorithm The weight vector combination is searched in the target space to maximize the R2 index and obtain the target index weight combination for the corresponding growth stage.

[0080] The multi-objective particle swarm optimization algorithm continuously searches for the optimal solution by simulating the movement of particles in the solution space. In this process, each particle has a position vector and a velocity vector, which represent its position and moving direction in the solution space. The algorithm iteratively updates the position and velocity of the particles to find the weight vector combination that maximizes the R2 index. In the target space, the algorithm searches for a weight vector combination to maximize the R2 index. By continuously iterating and adjusting the position and speed of the particles, the algorithm gradually converges to the optimal solution, which is the weight combination of the target index corresponding to the growth stage.

[0081] Specifically, in The weight vector set is uniformly distributed in the target space . Calculate the R2 index according to the formula to evaluate the quality of the growth measurement index set relative to the growth reference point under the current weight vector combination. Initialize the particle swarm, including the position vector and velocity vector. Iteratively update the position and velocity of the particles, and guide the particles to move towards the optimal solution according to the R2 index. In each iteration, calculate the R2 index of each particle, and update the global optimal solution and individual optimal solution. When the algorithm converges or reaches the preset number of iterations, output the optimal weight vector combination to obtain the above-mentioned target indicator weight combination.

[0082] S13023. Construct a fruit tree growth assessment sub-model for the corresponding growth stage based on the target indicator weight combination and the growth measurement indicator set.

[0083] According to the obtained target index weight combination and growth measurement index set, a fruit tree growth assessment sub-model corresponding to the growth stage is constructed. This model can comprehensively consider the fruit tree phenotypic indicators and greenness indicators to accurately assess the growth of fruit trees.

[0084] The present invention can comprehensively consider the influence of multiple objective functions (i.e., fruit tree phenotypic index and greenness index) on the growth of fruit trees through a multi-objective particle swarm optimization algorithm based on the R2 index, thereby obtaining a more accurate evaluation result. This helps to reduce the error and uncertainty that may be caused by a single indicator evaluation. The constructed fruit tree growth evaluation submodel can be evaluated for a specific growth stage, so it has strong applicability and pertinence. This helps to formulate corresponding management measures and decision support according to the growth characteristics of fruit trees at different growth stages. The weight vector combination obtained by searching the multi-objective particle swarm optimization algorithm can reasonably allocate the weights of different indicators in the evaluation, thereby more objectively reflecting the actual situation of the growth of fruit trees. This helps to avoid the subjectivity and deviation that may be caused by artificial weight allocation. The multi-objective particle swarm optimization algorithm has a faster convergence speed and a stronger global search capability, and can find the optimal solution in a shorter time. This helps to improve the construction efficiency and practical application effect of the evaluation model.

[0085] In an optional embodiment, the method for dynamic evaluation of fruit tree growth aims to construct fruit tree growth evaluation sub-models at different growth stages by integrating fruit tree phenotypic index data and fruit tree greenness index data, and finally integrate to obtain a time-series fruit tree growth comprehensive evaluation model. In order to improve the quality and integrity of the data, before the above-mentioned step S1302, based on the ground survey data, performs a fusion analysis on the fruit tree phenotypic index data and the fruit tree greenness index data of the same time label to obtain the fruit tree growth evaluation sub-model of the corresponding growth stage, the method also includes: S13020. Continuously interpolate the fruit tree phenotypic index data and the fruit tree greenness index data.

[0086] Specifically, collect the fruit tree phenotypic index data and fruit tree greenness index data throughout the annual growth period. Clean the data to remove outliers and missing values. Standardize the data to eliminate the dimensional differences between different indicators. According to the characteristics and needs of the data, select appropriate interpolation methods, such as linear interpolation, polynomial interpolation, spline interpolation, etc. Interpolate the missing values ​​or irregular sampling points in the fruit tree phenotypic index data and fruit tree greenness index data to obtain a continuous and complete data set. Evaluate the accuracy and applicability of the interpolation method by comparing the data before and after interpolation.

[0087] The fruit tree phenotypic index data and fruit tree greenness index data after continuous interpolation are divided according to different time labels, and then the fruit tree phenotypic index data and fruit tree greenness index data of each growth stage are fused and analyzed respectively. The multi-objective particle swarm optimization algorithm based on the R2 index is used to search for the optimal weight vector combination to maximize the R2 index. According to the optimal weight vector combination and the growth measurement index set, the fruit tree growth assessment sub-model corresponding to the growth stage is constructed. The fruit tree growth assessment sub-model is optimized based on the ground survey data, and the model parameters or weight vectors are adjusted. The optimized model is used to accurately evaluate the growth of fruit trees.

[0088] Collect optimized fruit tree growth assessment sub-models at different growth stages. Use a determined fusion strategy to fuse the models at different growth stages to obtain a time-series fruit tree growth comprehensive assessment model. Verify the comprehensive assessment model to ensure its accuracy and reliability.

[0089] The comprehensive evaluation model of time-series fruit tree growth is applied to fruit tree management and agricultural production. According to the evaluation results, targeted management measures and decision support are formulated to improve the production efficiency and quality of fruit trees.

[0090] The present invention can fill in missing values ​​or irregular sampling points in the data through continuous interpolation processing, obtain a continuous and complete data set, and provide a reliable data basis for subsequent fusion analysis and model construction. In the fusion analysis process, a multi-objective particle swarm optimization algorithm based on the R2 indicator is used to comprehensively consider the influence of multiple objective functions and obtain a more accurate weight vector combination and a fruit tree growth assessment sub-model. Optimizing the model based on ground survey data can further improve the accuracy and applicability of the model and make it more in line with actual conditions. The time-series fruit tree growth comprehensive assessment model can comprehensively reflect the growth status of fruit trees at different growth stages, and provide powerful decision support for fruit tree management and agricultural production.

[0091] Taking the growth monitoring of crisp pear as an example, the following is a specific embodiment and combined with the attached Figure 2 The present invention is further described.

[0092] S310, pear can be divided into five growth stages: budding stage, flowering and fruiting stage, fruit expansion stage, fruit ripening stage, and leaf fall dormancy stage. The entire annual growth period is from March of this year to February of the next year. According to the duration of the growth cycle, 2-3 data collection experiments are carried out in each growth cycle.

[0093] The UAV remote sensing data collection experiment was carried out under the conditions of suitable experimental time (around 10:00-14:00) and experimental weather (clear, cloudy, light wind or no wind), and the study area route was planned by the UAV route planning software, and data was collected in the vertical route shooting mode. The flight altitude of the multispectral UAV was set to 30 meters, the heading overlap was set to 80%, and the lateral overlap was set to 75%. Before takeoff, the 25% and 75% radiation calibration plates were used for calibration.

[0094] Use Pix4D mapper to process drone remote sensing data, import aerial photos, generate point clouds, edit and repair point cloud data, and obtain original three-dimensional point cloud data; obtain multispectral orthophoto data through image stitching, geometric correction, and radiation correction.

[0095] S320, extracting indicators: (1) The original 3D point cloud data was filtered and smoothed using the cloth simulation filtering method in the MATLAB toolbox; the 3D point cloud data was classified using the improved progressive encryption triangulation filtering algorithm to extract the fruit tree point cloud; the fruit tree point cloud was segmented using the watershed segmentation algorithm to obtain the crown area, single tree volume, and number of point clouds; the single tree density was extracted based on the ratio of single tree volume to point cloud number, and the accuracy of the extracted results was evaluated using ground survey data combined with visual interpretation results; the single tree volume and single tree density of fruit trees in the target area throughout the entire growth period were further extracted. Finally, the phenotypic index data of fruit trees throughout the entire growth period were obtained.

[0096] (2) Use ENVI software to process multispectral orthophoto data, extract broadband greenness indexes such as NDVI, RVI, EVI, VARI, LCI, SGI, and GLI based on multispectral data, and extract the broadband greenness index values ​​corresponding to individual fruit trees according to the crown area and distribution data of each fruit tree. Extract the broadband greenness index values ​​of fruit trees throughout the annual growth period, and finally obtain the greenness index data of fruit trees throughout the annual growth period. Continuously interpolate the obtained fruit tree phenotypic index data and fruit tree greenness index data to obtain the growth change curves corresponding to each fruit tree index.

[0097] S330, construct a comprehensive evaluation model for the growth of fruit trees in time series: (1) Use statistical analysis methods and expert judgment to determine the minimum and maximum values ​​of the empirical constants. Based on the Min-Max standardization method, standardize the fruit tree phenotypic index data and fruit tree greenness index data throughout the annual growth period to ensure that the data are compared on the same scale.

[0098] (2) For the data sets with different time labels throughout the annual growth period, the multi-objective particle swarm optimization (MOPSO) algorithm based on the R2 index was used to fuse the fruit tree phenotypic index data and the fruit tree greenness index data of the growth stage to establish a fruit tree growth assessment sub-model. The assessment results were compared with the results of the field survey to complete the accuracy assessment. The fruit tree growth assessment sub-models of each growth stage were integrated to establish a comprehensive time-series fruit tree growth assessment model covering the entire annual growth period.

[0099] S340, conduct fruit tree growth assessment: Collect and extract the drone remote sensing data to be tested in the target area, determine the fruit tree phenotypic index data and the fruit tree greenness index data according to the drone remote sensing data to be tested using the method of step S320 above, input the collection time and various index data (i.e., the fruit tree phenotypic index data and the fruit tree greenness index data) into the time series fruit tree growth comprehensive evaluation model, and obtain the fruit tree growth evaluation result of the current target area.

[0100] The present invention comprehensively considers the growth characteristics of fruit trees in the vegetative growth and reproductive growth stages, and incorporates the phenotypic changes such as volume expansion and increased sparseness caused by the rapid growth of fruit tree branches and leaves, as well as the temporal characteristics of fruit growth and leaf changes during reproductive growth, into the dynamic weight adjustment of phenotypic indicators and greenness indicators. Compared with traditional manual monitoring methods, drone monitoring technology has significant advantages, which can greatly reduce labor costs and effectively reduce the economic burden of agricultural production. At the same time, drones have the ability to quickly cover a wide orchard area, which can achieve efficient and timely monitoring, and provide strong support for timely discovery and resolution of problems in the growth process of fruit trees.

[0101] The present invention relies on the UAV platform to comprehensively evaluate the growth of fruit trees by integrating the coordinated changes of phenotypic and spectral dual indicators. The high-performance high-resolution camera and multispectral sensor carried by the UAV can capture high-precision point cloud data, image information and spectral data in real time. Combined with the analysis method of time series, it can accurately depict the growth status and change process of fruit trees in different growth stages, significantly improving the accuracy and reliability of the evaluation. It can not only accurately evaluate the growth status and potential of fruit trees, but also provide a solid scientific basis for agricultural production. The method of the present invention is widely applicable to various types of fruit trees, and can be flexibly adjusted according to different terrain conditions and fruit tree distribution characteristics. Strong adaptability and flexibility make it an efficient and accurate means of monitoring the growth of fruit trees, which can inject strong impetus into promoting the intelligent and refined development of agricultural production, and lay a solid foundation for the sustainable development of modern agriculture.

[0102] The fruit tree growth dynamic assessment device provided by the present invention is described below. The fruit tree growth dynamic assessment device described below and the fruit tree growth dynamic assessment method described above can be referenced to each other.

[0103] The fruit tree growth dynamic evaluation device provided by the present invention refers to Figure 3 As shown, including: The data acquisition module 410 is used to acquire the UAV remote sensing data and ground survey data of the target area during the entire annual growth period, and acquire the three-dimensional point cloud data and multi-spectral orthophoto data based on the UAV remote sensing data; A data extraction module 420 is used to extract fruit tree phenotypic index data based on the three-dimensional point cloud data, and to extract fruit tree greenness index data based on the multispectral orthophoto data; A model building module 430 is used to perform a fusion analysis on the ground survey data, the fruit tree phenotypic index data and the fruit tree greenness index data to construct a time-series fruit tree growth comprehensive evaluation model; The growth assessment module 440 is used to obtain the remote sensing data of the drone to be tested in the target area, input the remote sensing data of the drone to be tested into the time-series fruit tree growth comprehensive assessment model, and obtain the fruit tree growth assessment result.

[0104] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4 As shown, the electronic device may include: a processor 510, a communications interface 520, a memory 530 and a communication bus 540, wherein the processor 510, the communications interface 520 and the memory 530 communicate with each other via the communication bus 540. The processor 510 may call the logic instructions in the memory 530 to execute the method for dynamically evaluating the growth of fruit trees.

[0105] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0106] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the dynamic evaluation method of fruit tree growth provided by the above-mentioned methods.

[0107] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the method for dynamically evaluating the growth of fruit trees provided by the above-mentioned methods.

[0108] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0109] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for dynamically evaluating the growth of fruit trees, characterized in that: include: Acquire UAV remote sensing data and ground survey data of the target area throughout the annual growth period, and acquire three-dimensional point cloud data and multispectral orthophoto data based on the UAV remote sensing data; Extracting fruit tree phenotypic index data based on the three-dimensional point cloud data, and extracting fruit tree greenness index data based on the multispectral orthophoto data; Performing a fusion analysis on the ground survey data, the fruit tree phenotypic index data and the fruit tree greenness index data to construct a time-series comprehensive evaluation model for fruit tree growth; The remote sensing data of the unmanned aerial vehicle to be tested in the target area is obtained, and the remote sensing data of the unmanned aerial vehicle to be tested is input into the time-series fruit tree growth comprehensive evaluation model to obtain the fruit tree growth evaluation result.

2. The method for dynamic evaluation of fruit tree growth according to claim 1, characterized in that: The extracting of fruit tree phenotypic indicator data based on the three-dimensional point cloud data comprises: Classifying the three-dimensional point cloud data and extracting the fruit tree point cloud; The point cloud data set is segmented into individual trees to extract fruit phenotype index data of individual fruit trees in the target area.

3. The method for dynamic evaluation of fruit tree growth according to claim 1, characterized in that: The fruit tree greenness index data includes each broadband greenness index value; the fruit tree greenness index data extracted based on the multispectral orthophoto data includes: The multispectral orthophoto data were used to calculate the broadband greenness index in the target area; Obtain canopy area distribution data of individual fruit trees in the target area; According to the broadband greenness index values ​​of individual fruit trees in the target area and the canopy area distribution data of individual fruit trees, the broadband greenness index values ​​of individual fruit trees are calculated; among them, the broadband greenness index includes the normalized difference vegetation index, ratio vegetation index, enhanced vegetation index, visible atmospheric resistance index, leaf chlorophyll index, super green index and green leaf index.

4. The method for dynamic evaluation of fruit tree growth according to claim 1, characterized in that: The entire annual growth period includes multiple growth stages, and different growth stages correspond to different time labels; the ground survey data, the fruit tree phenotypic index data and the fruit tree greenness index data are integrated and analyzed to construct a time-series fruit tree growth comprehensive evaluation model, including: Dividing the fruit tree phenotypic index data and the fruit tree greenness index data according to time tags; Based on the ground survey data, the fruit tree phenotypic index data and the fruit tree greenness index data of the same time label are fused and analyzed to obtain a fruit tree growth assessment sub-model at the corresponding growth stage; Optimizing the fruit tree growth assessment sub-model using the ground survey data to obtain an optimized fruit tree growth assessment sub-model; The optimized fruit tree growth assessment sub-models at different growth stages during the entire annual growth period are integrated to obtain the time-series fruit tree growth comprehensive assessment model.

5. The method for dynamic evaluation of fruit tree growth according to claim 4, characterized in that: Based on the ground survey data, the fruit tree phenotypic index data and the fruit tree greenness index data of the same time label are fused and analyzed to obtain a fruit tree growth assessment sub-model at the corresponding growth stage, including: The fruit tree phenotypic index data and the fruit tree greenness index data are fused using a multi-objective particle swarm optimization algorithm based on the R2 index; wherein the R2 index is used to evaluate the quality of a growth measurement index set obtained by weighting multiple objective functions by a weight vector relative to a corresponding growth reference point; in, represents the R2 index, Represents a set of growth measurement indicators, Represents a set of weight vectors, where The target space is evenly distributed. Indicates the growth reference point, represents the number of weight vectors in the weight vector set, Distributed in The probability of represents the first A weight vector, Indicated in The first The objective function, The number of objective functions, Indicates fruit tree phenotypic index data or fruit tree greenness index data; Through the multi-objective particle swarm optimization algorithm Search for weight vector combinations in the target space to maximize the R2 index and obtain the target index weight combination corresponding to the growth stage; According to the target indicator weight combination and the growth measurement indicator set, a fruit tree growth assessment sub-model corresponding to the growth stage is constructed.

6. The method for dynamic evaluation of fruit tree growth according to claim 4, characterized in that: Before the fruit tree phenotypic index data and the fruit tree greenness index data of the same time label are fused and analyzed based on the ground survey data to obtain a fruit tree growth assessment sub-model corresponding to the growth stage, the method further includes: The fruit tree phenotypic index data and the fruit tree greenness index data are continuously interpolated.

7. A device for dynamically evaluating the growth of fruit trees, characterized in that: include: A data acquisition module is used to acquire UAV remote sensing data and ground survey data of the target area during the entire annual growth period, and to acquire three-dimensional point cloud data and multi-spectral orthophoto data based on the UAV remote sensing data; A data extraction module, used for extracting fruit tree phenotypic index data based on the three-dimensional point cloud data, and for extracting fruit tree greenness index data based on the multispectral orthophoto data; A model building module is used to integrate and analyze the ground survey data, the fruit tree phenotypic index data and the fruit tree greenness index data to build a time-series comprehensive evaluation model for fruit tree growth; The growth assessment module is used to obtain the remote sensing data of the unmanned aerial vehicle to be tested in the target area, input the remote sensing data of the unmanned aerial vehicle to be tested into the time-series fruit tree growth comprehensive assessment model, and obtain the fruit tree growth assessment result.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the method for dynamically evaluating the growth of fruit trees as described in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for dynamically evaluating the growth of fruit trees as described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for dynamically evaluating the growth of fruit trees as described in any one of claims 1 to 6 is implemented.

Images