Construction method of landscaping street tree growth prediction model based on artificial intelligence

Abstract

The invention discloses a landscaping street tree growth prediction model construction method based on artificial intelligence, and the method comprises the steps: obtaining tree time sequence image data, extracting key point descriptors, finding out matching items, calculating the difference value, analyzing the difference value change mode of the matching items in each growth stage, and carrying out the construction of a landscaping street tree growth prediction model. Constructing a growth stage conversion model to predict growth stage time to obtain residual growth stage prediction time; determining linear and nonlinear correlation functions according to tree historical growth data; calculating environment fitness in the residual growth stage prediction time; judging the growth potential according to tree parameters; and constructing a growth prediction model by using a neural network. The neural network model constructed by the method comprehensively considers various factors, can effectively improve the accuracy and reliability of street tree growth prediction, assists landscaping scientific planning and management, and has good interpretability.

Description

Technical Field

[0001] The present invention relates to the field of digital gardening, and in particular to a method for constructing a growth prediction model for roadside trees in gardening based on artificial intelligence. Background Art

[0002] With the acceleration of urbanization, gardening and greening are playing an increasingly critical role in improving the urban ecological environment and enhancing the quality of life of residents. As an important part of urban greening, the healthy growth and reasonable layout of street trees are of irreplaceable significance for shaping urban landscapes, purifying air, regulating climate, and providing ecological habitats.

[0003] Traditional street tree management methods rely heavily on manual experience and regular field observations, which have many limitations. On the one hand, manual observations make it difficult to achieve real-time, comprehensive monitoring of large-scale street tree groups, and the timeliness and completeness of data acquisition are poor. On the other hand, manual assessments of tree growth conditions often lack accuracy and objectivity, and different observers have different experiences and judgment criteria, resulting in inaccurate and inconsistent evaluations of tree growth conditions, making it difficult to formulate scientific and reasonable maintenance and management strategies.

[0004] In recent years, computer vision technology, sensor technology and artificial intelligence algorithms have made great progress. Integrating and applying these advanced technologies to the field of street tree growth prediction in landscape greening can help improve the efficiency and scientificity of street tree management, reduce maintenance costs, and provide strong data support and decision-making basis for urban greening planning, thus promoting the development of urban greening in the direction of intelligence and sustainability. Summary of the invention

[0005] The purpose of the present invention is to provide a method for constructing a growth prediction model for landscape roadside trees based on artificial intelligence.

[0006] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0007] The first aspect of the present invention provides a method for constructing a growth prediction model for roadside trees in landscape gardening based on artificial intelligence, comprising:

[0008] Obtain image data on the tree time series, extract descriptors of key points, extract matching items from the descriptors, and calculate the difference values ​​of the matching items;

[0009] Analyze the change pattern of the difference value of the matching item at each growth stage of the tree, establish a growth stage conversion model, predict the time of the growth stage, and obtain the predicted time of all remaining growth stages of the tree;

[0010] Obtaining a linear correlation function and a nonlinear correlation function of tree growth based on historical tree growth data, and calculating the environmental adaptability of the tree within the predicted time of all remaining growth stages based on the linear correlation function and the nonlinear correlation function;

[0011] The growth potential of the tree is judged according to the tree parameters, and a growth prediction model is constructed according to the growth potential and the environmental adaptability; the tree parameters include tree height, breast diameter, crown width, root characteristics and tree age.

[0012] As a further method, the method for extracting the descriptor of the key point includes:

[0013] Get the images of the same position of the tree before and after the growth stage, and calculate the dense image features. The expression is:

[0014]

[0015] Where x and y are the horizontal and vertical coordinates of the pixel in the image, respectively; R(x,y), G(x,y) and B(x,y) are the red, green and blue pixel values ​​in the RGB image, respectively; H(x,y), S(x,y) and V(x,y) are the hue, saturation and brightness values ​​in the HSV image, respectively; LBP(x,y) is the local binary pattern feature; HOG(x,y) is the histogram of oriented gradients feature; λ 1 , 2 and λ 3 They are color space conversion weight, local binary model weight and directional gradient histogram weight respectively;

[0016] Extract descriptors using isometric grids and perspective projection, perform differentiable sampling on two dense image features, and obtain descriptors of key points;

[0017] Calculate the similarity function value between two dense image features, and use the gradient descent optimization algorithm to adjust the position of the key points to maximize the similarity function value. The expression of the similarity function is:

[0018]

[0019] Among them, ω i,j is the degree of change of tree characteristics in the area where the descriptor is located, d 1,i and d 2,j are the descriptors extracted from the two images, i and j are the descriptor indexes in the two images, P 1 (x 1,i ,y 1,i ) and P 1 (x 2,j ,y 2,j ) are descriptors d 1,i and d2,j The coordinates of the position in the image after perspective projection, σ 1 and σ 2 are scale parameters used to control the influence of feature point descriptor differences and position differences, respectively.

[0020] As a further method, the method of extracting matching items from the descriptors and calculating the difference values ​​of the matching items includes:

[0021] For the descriptor set of the image before and after the growth phase, the pairwise dot products between all descriptors are calculated, and the similarity matrix is ​​obtained by summing the products of the descriptors in each dimension;

[0022] Use the similarity matrix to find the maximum value of each row and each column. If the descriptor meets the maximum value of the row and column, it is identified as a match;

[0023] The features of the matching items are normalized and the difference values ​​are calculated using Euclidean distance.

[0024] As a further method, the method of analyzing the change pattern of the difference value of the matching item at each growth stage of the tree and establishing the growth stage conversion model includes:

[0025] Obtain the matching item difference value data calculated from the images before and after each growth stage of historical tree samples, and use the K-Means clustering algorithm to divide the matching item difference value data into different clusters, each cluster represents a growth stage, and determine the difference value range corresponding to each growth stage;

[0026] A growth stage conversion model is constructed using a decision tree algorithm, with the difference value range corresponding to each growth stage and the historical growth stage conversion situation as the training set to train the model;

[0027] The time of the next growth stage is determined by the growth stage transition model, and multiple iterations are performed until the predicted times of all remaining growth stages of the tree are obtained.

[0028] As a further method, the method for obtaining the linear correlation function and the nonlinear correlation function of tree growth based on the historical tree growth data includes:

[0029] Collect historical tree growth data and historical environmental data, use the factors in the historical environmental data as independent variables, and the factors in the historical tree growth data as dependent variables;

[0030] Calculate the Pearson coefficient r between each independent variable and the dependent variable. If |r|≥0.3, there is a linear correlation; if |r|<0.3, there is a nonlinear correlation;

[0031] For linear correlation, if it is a single independent variable linear regression, a single variable linear model is used; if it is a multivariate linear regression, a stepwise regression method is used to select the independent variable combination, and a multivariate linear model is used; for nonlinear correlation, multiple nonlinear models are fitted, and for each nonlinear model, the goodness of fit index is calculated, and the model with the best fit index is selected, where the calculation formula is:

[0032]

[0033] Where P is the number of parameters in the model, l(θ) is the log-likelihood function of the model, n is the number of samples, σ is the error term, and y i is the i-th observation value, f(x i ,θ) is the predicted value of the model, x i is the value of the ith independent variable, and θ is the parameter set of the model.

[0034] As a further method, the method for calculating the environmental adaptability of trees within the predicted time of all remaining growth stages based on the linear correlation function and the nonlinear correlation function includes:

[0035] The predicted time of all remaining growth stages of the tree is input into the pre-trained random forest model to obtain the predicted environmental data, where the random forest model is trained based on the historical time series and historical environmental data;

[0036] Obtain environmental data consistent with historical environmental data variables used in establishing linear and nonlinear correlation functions from the predicted environmental data, and use linear and nonlinear correlation functions to perform average weighted calculations to predict growth value sequences;

[0037] The environmental fitness is calculated based on the predicted growth value sequence, and the expression is:

[0038]

[0039] Among them, ξ is the total number of observed data points, y i is the ith predicted growth value, is the average of all predicted growth values, σ y is the standard deviation of all predicted growth values, x i is the environmental variable vector of the ith observation, μ is the mean value vector of the environmental variable vector, σ x is the standard deviation of the environmental variable vector.

[0040] As a further method, the method for judging the growth potential of a tree based on tree parameters includes:

[0041] Get the tree parameters of the tree at the current time and quantify the growth potential of the tree. The expression is:

[0042]

[0043] Among them, V H is the tree height growth rate, is the average tree height growth rate of the tree species, L a is the effective light ratio that trees can obtain, W f is the wind limiting coefficient on tree height growth, D c is the current breast diameter, D s is the standard breast diameter of the same tree species and age, S r is the root growth space coefficient, V C is the crown width expansion speed, is the average crown expansion speed of the tree species, L i is the ratio of the average light intensity within the crown to the optimal light intensity of the tree species, N a is the ratio of available nutrients in the soil, S d is the root depth in the soil, S q is soil quality, B r is the ratio of root biomass to the aboveground biomass of the tree, B s A is the ratio of the standard root system to the aboveground biomass of the tree species. c is the current age of the tree, is the average age of the tree species, G s is the growth phase coefficient.

[0044] As a further method, the method for constructing a growth prediction model based on the growth potential and the environmental adaptability includes:

[0045] A growth prediction model between growth potential, environmental adaptability and tree growth rate is constructed through a neural network. There are two output layers, which output the predicted tree height and breast diameter in time series respectively.

[0046] Based on the growth potential, a decay function is introduced into the growth prediction model. When the growth potential is smaller, the value of the decay function is larger, which is applied to the hidden layer of the model.

[0047] The randomness of tree growth and environment is represented in the form of probability density distribution. The probability density distribution is added to the output layer of the growth prediction model, and the probability density distribution is updated using the Bayesian formula. The expression of the Bayesian formula is:

[0048]

[0049] Where Θ is the observed data, G is the tree growth state, E is the environmental factor, P(Θ|G,E) represents the probability of the observed data Θ given the tree growth state G and the environmental factor E, P(G|Θ,E) represents the probability of the tree growth being in state G given the observed data Θ and the environmental factor E, and P(Θ|E) represents the probability distribution of the observed data Θ given the environmental factor E;

[0050] Among them, P(G|Θ,E) obeys multi-source normal distribution, specifically:

[0051]

[0052] Among them, ∑(Θ,E) represents the covariance matrix of the tree growth state G given the observation data Θ and the environmental factor E, μ(Θ,E) represents the mean of the tree growth state G given the observation data Θ and the environmental factor E, and T represents the transposition operation;

[0053] Where P(Θ|E) obeys a mixed distribution, specifically:

[0054]

[0055] Where K is the number of components in the mixed distribution, ω k is the weight of the kth component, Indicates that under environmental factor E, is the conditional probability density function of the normal distribution, where μ k and∑ k is the mean vector and covariance matrix of the kth component;

[0056] The growth prediction model was trained using historical environmental fitness and historical tree height and DBH combinations as training sets.

[0057] In a second aspect, an embodiment of the present application further provides an electronic device, comprising: a processor; and a memory arranged to store computer executable instructions, wherein when the executable instructions are executed, the processor executes the method steps described in the first aspect.

[0058] In a third aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores one or more programs. When the one or more programs are executed by an electronic device including multiple applications, the electronic device executes the method steps described in the first aspect.

[0059] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0060] (1) The present invention can more accurately analyze the growth changes of trees by acquiring image data on the time series of trees and extracting descriptors of key points. This method can improve the accuracy of prediction based on actual image data rather than relying solely on historical statistical data.

[0061] (2) The present invention establishes a growth stage transition model based on the analysis of the difference value change pattern, predicts the time when trees enter the next growth stage, predicts and divides the growth stages of trees, helps to identify the different growth situations of trees in different growth stages, and makes the model construction take into account the time factor, which is more comprehensive.

[0062] (3) The present invention determines the growth potential of trees based on tree parameters, and constructs a growth prediction model based on environmental adaptability. The model introduces a decay function based on the growth potential, taking into account the natural decay of the tree growth potential over time and the long-term impact of environmental factors on tree growth, which helps to more accurately predict the performance of trees at different growth stages. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 The present invention is a flowchart of the steps of a method for constructing a landscaping roadside tree growth prediction model based on artificial intelligence in an embodiment of the present invention.

[0064] Figure 2 It is a schematic diagram of the structure of an electronic device in an embodiment of this specification. DETAILED DESCRIPTION

[0065] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0066] Reference Figure 1 As shown, the present invention provides a method for constructing a growth prediction model for roadside trees in landscaping based on artificial intelligence, comprising:

[0067] S100 obtains image data on a tree time series, extracts descriptors of key points, extracts matching items from the descriptors, and calculates difference values ​​of the matching items;

[0068] It should be explained that this step obtains tree time series images and extracts key point descriptors. It uses multiple attributes and features of pixels, and through specific calculations and processing, accurately depicts the key features of trees, laying the foundation for subsequent comparisons. It finds matching items from descriptors and uses similarity matrices to determine the corresponding features in images at different times, reflecting the continuity and changes of tree growth. These data combined can clearly present the growth trend of trees, providing strong support for growth stage judgment and model construction.

[0069] In a specific embodiment, a group of Sophora japonica trees are newly planted as street trees on a newly built city street. It is planned to use the artificial intelligence-based landscaping street tree growth prediction model of the present invention to predict the future growth status of these Sophora japonica trees, so as to reasonably plan street landscape maintenance, tree maintenance resource allocation and other tasks;

[0070] In the actual evaluation, starting from the first month after the planting of Sophora japonica, at 9 a.m. at the beginning of each month, a high-definition camera was used to take tree images at a distance of 5 meters from the tree and a horizontal angle of 45 degrees. The images of the first and second months were selected for processing. For the pixel coordinates (300, 400) in the image, the RGB values ​​were (100, 80, 60), the HSV values ​​were (35, 0.4, 0.6), the local binary pattern eigenvalue was 0.25, the directional gradient histogram eigenvalue was 0.4, and the color space conversion weight was set to 0.35. The dense image feature value of the pixel was obtained by the dense image feature calculation formula.

[0071] In the actual evaluation, for the descriptor sets of the images of these two months, the pairwise dot products between all descriptors were calculated to obtain the similarity matrix. The maximum values ​​of each row and column in the similarity matrix were determined, and the matching items were identified. A total of 6 matching items were obtained. After the matching item features were normalized, the Euclidean distance between the feature vectors of two matching items [0.1, 0.18, 0.25] and [0.13, 0.2, 0.28] was calculated, and the difference value was 0.035.

[0072] S200 analyzes the change pattern of the difference value of the matching item in each growth stage of the tree, establishes a growth stage conversion model, predicts the time of the growth stage, and obtains the predicted time of all remaining growth stages of the tree;

[0073] It should be explained that by analyzing the change pattern of the difference value of the matching items, we can gain insight into the evolution rhythm of the characteristics of each growth stage. For example, the difference value in the juvenile stage has small fluctuations, while the fluctuation in the mature stage is large and the trend is obvious. By establishing a growth stage transition model, we can quantify the relationship between this evolution and stage transition in a data-driven way. By predicting the growth stage time through the model, garden managers can plan maintenance strategies in advance and allocate resources reasonably.

[0074] In the actual evaluation, the matching item difference value data calculated from the images before and after the growth stage of Sophora japonica in other areas of the same city in the past 15 years were collected. The K-Means clustering algorithm was used to divide these data into 5 clusters, corresponding to the budding period, seedling period, rapid growth period, stable period and senescence period of Sophora japonica, and the difference value range of each stage was determined. The difference value range of the budding period was [0-0.05]. The decision tree algorithm was used to build a growth stage conversion model. The difference value range and the historical growth stage conversion situation were used as the training set to train the model. The matching item difference value of this batch of Sophora japonica was 0.03. The model determined that it was in the budding period and was expected to enter the seedling period in 6 months. The prediction time of the remaining growth stage is shown in Table 1:

[0075] Table 1 Predicted time of remaining growth stage

[0076] Growth stage Estimated entry time Estimated duration Seedling stage 6 months after planting 12 months Rapid growth period 19 months after planting 24 months Stable period 43 months after implantation 36 months Aging 10 years after planting 20 years

[0077] S300 acquires a linear correlation function and a nonlinear correlation function of tree growth based on historical growth data of the tree, and calculates the environmental adaptability of the tree within the predicted time of all remaining growth stages based on the linear correlation function and the nonlinear correlation function;

[0078] What needs to be explained is that by deeply exploring the intrinsic connection between tree growth and environmental data, building a precise functional relationship to quantify the impact of the environment on tree growth, and then calculating environmental fitness, the significance lies in providing key parameters and basis for the growth prediction model, so that the model can more realistically simulate the growth dynamics of trees in different environments. Through accurate environmental fitness calculation, the model can more accurately predict the growth rate and morphological changes of trees in various growth stages in the future, thereby helping garden managers to formulate targeted maintenance plans in advance and arrange resources reasonably;

[0079] In the actual evaluation, the predicted time of the remaining growth stage of Sophora japonica (based on the growth stage conversion model) was input into the pre-trained random forest model to obtain the predicted environmental data, and the annual light duration, temperature, and soil moisture in the next three years were predicted. The environmental data consistent with the environmental data when the correlation function was established were selected from the predicted environmental data, and the predicted growth value sequence was weighted and calculated using linear and nonlinear correlation functions. Assume that the predicted growth value sequence of tree height in the next three years is [0.5, 0.8, 1.2]; calculate the environmental fitness, the total number of observation data points is 3, and the environmental fitness is calculated according to the formula to be 0.48.

[0080] S400 determines the growth potential of the tree according to the tree parameters, and constructs a growth prediction model according to the growth potential and the environmental adaptability; the tree parameters include tree height, diameter at breast height, crown width, root characteristics and tree age.

[0081] It needs to be explained that through tree parameters, the current growth status of trees can be understood from different dimensions. For example, tree height and breast diameter reflect the vertical and horizontal growth of trees, crown width is related to the range of its photosynthesis, root characteristics affect nutrient absorption, etc., and tree age reflects the growth stage. Combining these can accurately determine how much space and ability the tree still has to grow. Combining the determined growth potential and the previously calculated environmental adaptability to construct a growth prediction model can allow the model to fully consider the adaptation of the tree's own characteristics to the environment in which it is located. The constructed model is no longer a simple speculation of a single factor, but a scientific prediction that comprehensively considers multiple factors, which can more accurately predict the growth rate and morphological changes of trees at different stages in the future.

[0082] In the actual evaluation, the current height of Sophora japonica trees was measured to be 0.3 meters, the average tree height growth rate of this tree species is 0.45 meters / year, the effective light ratio that trees can obtain is 0.55, the wind restriction coefficient on tree height growth is 0.12, the current breast diameter is 2 centimeters, the standard breast diameter of the same species and age is 1.8 centimeters, the root growth space coefficient is 0.65, the crown expansion speed is 0.15 meters / year, the average crown expansion speed of this tree species is 0.2 meters / year, and the average light intensity inside the crown is The optimal light intensity ratio of this tree species is 0.45, the available nutrient ratio in the soil is 0.35, the root depth in the soil is 0.5 meters, the soil quality is 0.55, the ratio of root biomass to tree aboveground biomass is 0.15, the ratio of standard root biomass to tree aboveground biomass of this tree species is 0.13, the current age of the tree is 0.5 years, the average age of this tree species is 1 year, the growth stage coefficient is 0.7, and the growth potential value is calculated to be 0.52;

[0083] In the actual evaluation, a neural network growth prediction model was constructed, and two output layers were set to predict tree height and breast diameter respectively. Based on the growth potential, an attenuation function was introduced and applied to the hidden layer. The randomness was represented in the form of probability density distribution and added to the output layer, and updated using the Bayesian formula. Specifically, the model was trained using the historical environmental adaptability, tree height and breast diameter data of Sophora japonica in the same city as the training set. After the training was completed, the model predicted that the tree height in the next two years would reach 0.6 meters and 0.9 meters, and the breast diameter would reach 3 centimeters and 4 centimeters, respectively.

[0084] Specifically, through testing the model, it was found that the model's accuracy in predicting tree height was over 94%, and its accuracy in predicting breast diameter was over 87%, with deviation values ​​not exceeding 2% and 3%, respectively, providing an effective decision-making basis for maintenance and management.

[0085] In this embodiment, the method for extracting the descriptor of the key point includes:

[0086] Get the images of the same position of the tree before and after the growth stage, and calculate the dense image features. The expression is:

[0087]

[0088] Where x and y are the horizontal and vertical coordinates of the pixel in the image, respectively; R(x,y), G(x,y) and B(x,y) are the red, green and blue pixel values ​​in the RGB image, respectively; H(x,y), S(x,y) and V(x,y) are the hue, saturation and brightness values ​​in the HSV image, respectively; LBP(x,y) is the local binary pattern feature; HOG(x,y) is the histogram of oriented gradients feature; λ 1 , 2 and λ 3 They are color space conversion weight, local binary model weight and directional gradient histogram weight respectively;

[0089] Extract descriptors using isometric grids and perspective projection, perform differentiable sampling on two dense image features, and obtain descriptors of key points;

[0090] Calculate the similarity function value between two dense image features, and use the gradient descent optimization algorithm to adjust the position of the key points to maximize the similarity function value. The expression of the similarity function is:

[0091]

[0092] Among them, ω i,j is the degree of change of tree characteristics in the area where the descriptor is located, d 1,i and d 2,j are the descriptors extracted from the two images, i and j are the descriptor indexes in the two images, P 1 (x 1,i ,y 1,i ) and P 1 (x 2,j ,y 2,j ) are descriptors d 1,i and d 2,j The coordinates of the position in the image after perspective projection, σ 1 and σ 2 are scale parameters used to control the influence of feature point descriptor differences and position differences, respectively.

[0093] In this embodiment, the method of extracting matching items from descriptors and calculating difference values ​​of the matching items includes:

[0094] For the descriptor set of the image before and after the growth phase, the pairwise dot products between all descriptors are calculated, and the similarity matrix is ​​obtained by summing the products of the descriptors in each dimension;

[0095] Use the similarity matrix to find the maximum value of each row and each column. If the descriptor meets the maximum value of the row and column, it is identified as a match;

[0096] The features of the matching items are normalized and the difference values ​​are calculated using Euclidean distance.

[0097] In this embodiment, the method of analyzing the change pattern of the difference value of the matching item at each growth stage of the tree and establishing the growth stage conversion model includes:

[0098] Obtain the matching item difference value data calculated from the images before and after each growth stage of historical tree samples, and use the K-Means clustering algorithm to divide the matching item difference value data into different clusters, each cluster represents a growth stage, and determine the difference value range corresponding to each growth stage;

[0099] A growth stage conversion model is constructed using a decision tree algorithm, with the difference value range corresponding to each growth stage and the historical growth stage conversion situation as the training set to train the model;

[0100] The time of the next growth stage is determined by the growth stage transition model, and multiple iterations are performed until the predicted times of all remaining growth stages of the tree are obtained.

[0101] In this embodiment, the method for obtaining the linear correlation function and the nonlinear correlation function of tree growth based on the historical growth data of trees includes:

[0102] Collect historical tree growth data and historical environmental data, use the factors in the historical environmental data as independent variables, and the factors in the historical tree growth data as dependent variables;

[0103] Calculate the Pearson coefficient r between each independent variable and the dependent variable. If |r|≥0.3, there is a linear correlation; if |r|<0.3, there is a nonlinear correlation;

[0104] For linear correlation, if it is a single independent variable linear regression, a single variable linear model is used; if it is a multivariate linear regression, a stepwise regression method is used to select the independent variable combination, and a multivariate linear model is used; for nonlinear correlation, multiple nonlinear models are fitted, and for each nonlinear model, the goodness of fit index is calculated, and the model with the best fit index is selected, where the calculation formula is:

[0105]

[0106] Where P is the number of parameters in the model, l(θ) is the log-likelihood function of the model, n is the number of samples, σ is the error term, and y i is the i-th observation value, f(x i ,θ) is the predicted value of the model, x iis the value of the ith independent variable, and θ is the parameter set of the model.

[0107] In this embodiment, the method for calculating the environmental adaptability of trees within the predicted time of all remaining growth stages based on the linear correlation function and the nonlinear correlation function includes:

[0108] The predicted time of all remaining growth stages of the tree is input into the pre-trained random forest model to obtain the predicted environmental data, where the random forest model is trained based on the historical time series and historical environmental data;

[0109] Obtain environmental data consistent with historical environmental data variables used in establishing linear and nonlinear correlation functions from the predicted environmental data, and use linear and nonlinear correlation functions to perform average weighted calculations to predict growth value sequences;

[0110] The environmental fitness is calculated based on the predicted growth value sequence, and the expression is:

[0111]

[0112] Among them, ξ is the total number of observed data points, y i is the ith predicted growth value, is the average of all predicted growth values, σ y is the standard deviation of all predicted growth values, x i is the environmental variable vector of the ith observation, μ is the mean value vector of the environmental variable vector, σ x is the standard deviation of the environmental variable vector.

[0113] In this embodiment, the method for judging the growth potential of a tree according to tree parameters includes:

[0114] Get the tree parameters of the tree at the current time and quantify the growth potential of the tree. The expression is:

[0115]

[0116] Among them, V H is the tree height growth rate, is the average tree height growth rate of the tree species, L a is the effective light ratio that trees can obtain, W f is the wind limiting coefficient on tree height growth, D c is the current breast diameter, D s is the standard breast diameter of the same tree species and age, S r is the root growth space coefficient, V C is the crown width expansion speed, is the average crown expansion speed of the tree species, L iis the ratio of the average light intensity within the crown to the optimal light intensity of the tree species, N a is the ratio of available nutrients in the soil, S d is the root depth in the soil, S q is soil quality, B r is the ratio of root biomass to the aboveground biomass of the tree, B s A is the ratio of the standard root system to the aboveground biomass of the tree species. c is the current age of the tree, is the average age of the tree species, G s is the growth phase coefficient.

[0117] In this embodiment, the method for constructing a growth prediction model based on the growth potential and the environmental adaptability includes:

[0118] A growth prediction model between growth potential, environmental adaptability and tree growth rate is constructed through a neural network. There are two output layers, which output the predicted tree height and breast diameter in time series respectively.

[0119] Based on the growth potential, a decay function is introduced into the growth prediction model. When the growth potential is smaller, the value of the decay function is larger, which is applied to the hidden layer of the model.

[0120] The randomness of tree growth and environment is represented in the form of probability density distribution. The probability density distribution is added to the output layer of the growth prediction model, and the probability density distribution is updated using the Bayesian formula. The expression of the Bayesian formula is:

[0121]

[0122] Where Θ is the observed data, G is the tree growth state, E is the environmental factor, P(Θ|G,E) represents the probability of the observed data Θ given the tree growth state G and the environmental factor E, P(G|Θ,E) represents the probability of the tree growth being in state G given the observed data Θ and the environmental factor E, and P(Θ|E) represents the probability distribution of the observed data Θ given the environmental factor E;

[0123] Among them, P(G|Θ,E) obeys multi-source normal distribution, specifically:

[0124]

[0125] Among them, ∑(Θ,E) represents the covariance matrix of the tree growth state G given the observation data Θ and the environmental factor E, μ(Θ,E) represents the mean of the tree growth state G given the observation data Θ and the environmental factor E, and T represents the transposition operation;

[0126] Where P(Θ|E) obeys a mixed distribution, specifically:

[0127]

[0128] Where K is the number of components in the mixed distribution, ω k is the weight of the kth component, Indicates that under environmental factor E, is the conditional probability density function of the normal distribution, where μ k and∑ k is the mean vector and covariance matrix of the kth component;

[0129] The growth prediction model was trained using historical environmental fitness and historical tree height and DBH combinations as training sets.

[0130] Figure 2 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 2 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. The memory may include a memory, such as a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Of course, the electronic device may also include hardware required for other services.

[0131] The processor, network interface and memory can be interconnected through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc.

[0132] The memory is used to store the program. Specifically, the program may include a program code, and the program code includes a computer operation instruction. The memory may include a memory and a non-volatile memory, and provides instructions and data to the processor.

[0133] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a method for constructing a growth prediction model for roadside trees in landscaping based on artificial intelligence at the logical level. The processor executes the program stored in the memory and is specifically used to execute any of the aforementioned methods for constructing a growth prediction model for roadside trees in landscaping based on artificial intelligence.

[0134] The present invention can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or an instruction in the form of software. The above processor can be a general-purpose processor, including a central processing unit, a network processor (Network Processor, NP), etc.; it can also be a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor to be executed, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0135] An embodiment of the present application also proposes a computer-readable storage medium, which stores one or more programs, and the one or more programs include instructions. When the instructions are executed by an electronic device including multiple applications, they execute any of the aforementioned methods for constructing a landscaping street tree growth prediction model based on artificial intelligence.

[0136] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. A method for constructing a growth prediction model for roadside trees in landscape gardening based on artificial intelligence, characterized in that: The following steps are involved: Obtain image data on the tree time series, extract descriptors of key points, extract matching items from the descriptors, and calculate the difference values ​​of the matching items; Analyze the change pattern of the difference value of the matching item at each growth stage of the tree, establish a growth stage conversion model, predict the time of the growth stage, and obtain the predicted time of all remaining growth stages of the tree; Obtaining a linear correlation function and a nonlinear correlation function of tree growth based on historical tree growth data, and calculating the environmental adaptability of the tree within the predicted time of all remaining growth stages based on the linear correlation function and the nonlinear correlation function; The growth potential of the tree is judged according to the tree parameters, and a growth prediction model is constructed according to the growth potential and the environmental adaptability; the tree parameters include tree height, breast diameter, crown width, root characteristics and tree age.

2. The method for constructing a growth prediction model for roadside trees in landscaping based on artificial intelligence according to claim 1 is characterized in that: The method for extracting a descriptor of a key point comprises: Calculate dense image features, the expression is: Where x and y are the horizontal and vertical coordinates of the pixel in the image, respectively; R(x,y), G(x,y) and B(x,y) are the red, green and blue pixel values ​​in the RGB image, respectively; H(x,y), S(x,y) and V(x,y) are the hue, saturation and brightness values ​​in the HSV image, respectively; LBP(x,y) is the local binary pattern feature, HOG(x,y) is the histogram of oriented gradients feature, λ1, λ2 and λ3 are the color space conversion weight, local binary model weight and histogram of oriented gradients weight, respectively; Extract descriptors using isometric grids and perspective projection, perform differentiable sampling on two dense image features, and obtain descriptors of key points; Calculate the similarity function value between two dense image features, and use the gradient descent optimization algorithm to adjust the position of the key points to maximize the similarity function value. The expression of the similarity function is: Among them, ω i,j is the degree of change of tree characteristics in the area where the descriptor is located, d 1,i and d 2,j are the descriptors extracted from the two images, i and j are the descriptor indexes in the two images, P1(x 1,i ,y 1,i ) and P1(x 2,j ,y 2,j ) are descriptors d 1,i and d 2,j The coordinates of the position in the image after perspective projection, σ1 and σ2 are scale parameters used to control the influence of feature point descriptor differences and position differences.

3. The method for constructing a growth prediction model for roadside trees in landscaping based on artificial intelligence according to claim 1 is characterized in that: The method of extracting matching items from descriptors and calculating difference values ​​of matching items includes: For the descriptor set of the image before and after the growth phase, the pairwise dot products between all descriptors are calculated, and the similarity matrix is ​​obtained by summing the products of the descriptors in each dimension; Use the similarity matrix to find the maximum value of each row and each column. If the descriptor meets the maximum value of the row and column, it is identified as a match; The features of the matching items are normalized and the difference values ​​are calculated using Euclidean distance.

4. The method for constructing a growth prediction model for roadside trees in landscaping based on artificial intelligence according to claim 1 is characterized in that: The method of analyzing the change pattern of the difference value of the matching item at each growth stage of the tree and establishing a growth stage conversion model includes: Obtain the matching item difference value data calculated from the images before and after each growth stage of historical tree samples, and use the K-Means clustering algorithm to divide the matching item difference value data into different clusters, each cluster represents a growth stage, and determine the difference value range corresponding to each growth stage; A growth stage conversion model is constructed using a decision tree algorithm, with the difference value range corresponding to each growth stage and the historical growth stage conversion situation as the training set to train the model; The time of the next growth stage is determined by the growth stage transition model, and multiple iterations are performed until the predicted times of all remaining growth stages of the tree are obtained.

5. The method for constructing a growth prediction model for roadside trees in landscaping based on artificial intelligence according to claim 1 is characterized in that: The method for obtaining the linear correlation function and the nonlinear correlation function of tree growth based on the historical growth data of trees includes: Collect historical tree growth data and historical environmental data, use the factors in the historical environmental data as independent variables, and the factors in the historical tree growth data as dependent variables; Calculate the Pearson coefficient r between each independent variable and the dependent variable. If |r|≥0.3, there is a linear correlation; if |r|<0.3, there is a nonlinear correlation; For linear correlation, if it is a single independent variable linear regression, a single variable linear model is used; if it is a multivariate linear regression, a stepwise regression method is used to select the independent variable combination, and a multivariate linear model is used; for nonlinear correlation, multiple nonlinear models are fitted, and for each nonlinear model, the goodness of fit index is calculated, and the model with the best fit index is selected, where the calculation formula is: Where P is the number of parameters in the model, l(θ) is the log-likelihood function of the model, n is the number of samples, σ is the error term, and y i is the i-th observation value, f(x i ,θ) is the predicted value of the model, x i is the value of the ith independent variable, and θ is the parameter set of the model.

6. The method for constructing a growth prediction model for roadside trees in landscaping based on artificial intelligence according to claim 1 is characterized in that: The method for calculating the environmental adaptability of trees within the predicted time of all remaining growth stages based on the linear correlation function and the nonlinear correlation function comprises: The predicted time of all remaining growth stages of the tree is input into the pre-trained random forest model to obtain the predicted environmental data, where the random forest model is trained based on the historical time series and historical environmental data; Obtain environmental data consistent with historical environmental data variables used in establishing linear and nonlinear correlation functions from the predicted environmental data, and use linear and nonlinear correlation functions to perform average weighted calculations to predict growth value sequences; The environmental fitness is calculated based on the predicted growth value sequence, and the expression is: Where ξ is the number of data points in the predicted growth value sequence, y i is the ith predicted growth value, is the average value of all predicted growth values ​​in the growth stage, σ y is the standard deviation of all predicted growth values ​​within the growth stage, x i is the environmental data vector corresponding to the i-th growth value, μ is the average vector of all environmental data vectors, σ x is the standard deviation vector of all environmental data vectors.

7. The method for constructing a growth prediction model for roadside trees in landscaping based on artificial intelligence according to claim 1 is characterized in that: The method for judging the growth potential of a tree based on tree parameters comprises: Get the tree parameters of the tree at the current time and quantify the growth potential of the tree. The expression is: Among them, V H is the tree height growth rate, is the average tree height growth rate of the tree species, L a is the effective light ratio that trees can obtain, W f is the wind limiting coefficient on tree height growth, D c is the current breast diameter, D s is the standard breast diameter of the same tree species and age, S r is the root growth space coefficient, V C is the crown width expansion speed, is the average crown expansion speed of the tree species, L i is the ratio of the average light intensity within the crown to the optimal light intensity of the tree species, N a is the ratio of available nutrients in the soil, S d is the root depth in the soil, S q is soil quality, B r is the ratio of root biomass to the aboveground biomass of the tree, B s A is the ratio of the standard root system to the aboveground biomass of the tree species. c is the current age of the tree, is the average age of the tree species, G s is the growth phase coefficient.

8. The method for constructing a growth prediction model for roadside trees in landscaping based on artificial intelligence according to claim 1 is characterized in that: The method for constructing a growth prediction model according to the growth potential and the environmental adaptability comprises: A growth prediction model between growth potential, environmental adaptability and tree growth rate is constructed through a neural network. There are two output layers, which output the predicted tree height and breast diameter in time series respectively. Based on the growth potential, a decay function is introduced into the growth prediction model. When the growth potential is smaller, the value of the decay function is larger, which is applied to the hidden layer of the model. The randomness of tree growth and environment is represented in the form of probability density distribution. The probability density distribution is added to the output layer of the growth prediction model, and the probability density distribution is updated using the Bayesian formula. The expression of the Bayesian formula is: Where Θ is the observed data, G is the tree growth state, E is the environmental factor, P(Θ|G,E) represents the probability of the observed data Θ given the tree growth state G and the environmental factor E, P(G|Θ,E) represents the probability of the tree growth being in state G given the observed data Θ and the environmental factor E, and P(Θ|E) represents the probability distribution of the observed data Θ given the environmental factor E; Among them, P(G|Θ,E) obeys multi-source normal distribution, specifically: Among them, ∑(Θ,E) represents the covariance matrix of the tree growth state G given the observation data Θ and the environmental factor E, μ(Θ,E) represents the mean of the tree growth state G given the observation data Θ and the environmental factor E, and T represents the transposition operation; Where P(Θ|E) obeys a mixed distribution, specifically: Where K is the number of components in the mixed distribution, ω k is the weight of the kth component, Indicates that under environmental factor E, is the conditional probability density function of the normal distribution, where μ k and∑ k is the mean vector and covariance matrix of the kth component; The growth prediction model was trained using historical environmental fitness and historical tree height and DBH combinations as training sets.

9. An electronic device, comprising: processor; as well as A memory arranged to store computer executable instructions, which when executed cause the processor to perform the method according to any one of claims 1 to 8.

10. A computer-readable storage medium storing one or more programs, which, when executed by an electronic device including a plurality of application programs, enables the electronic device to execute the method according to any one of claims 1 to 8.

Images