Urban air adsorption purification method based on forestry planning

Abstract

The application provides a city air adsorption purification method based on forestry planning, and belongs to the technical field of air purification. The city air adsorption purification method based on forestry planning comprises the following steps: step 1, analyzing the pollution status of city air to determine the main pollutant types, concentration and spatial distribution; step 2, selecting garden plants with strong adsorption and purification capacity according to the pollutant types and concentration; step 3, designing reasonable greening layout and structure according to city climate, wind direction, terrain and building distribution; step 4, formulating a reasonable garden plant planting scheme based on the spatial distribution of pollution sources, including planting location, density and form, to achieve good purification effect; step 5, regularly monitoring and evaluating the concentration of air pollutants and the growth status of garden plants, and adjusting the planting scheme according to the actual situation to maintain the health and vitality of garden plants and improve the efficiency and sustainability of air purification.

Description

Technical Field

[0001] This application relates to the field of air purification technology, specifically a method for urban air adsorption and purification based on forestry planning. Background Technology

[0002] With rapid urbanization, urban air pollution has become increasingly serious, severely impacting residents' health and environmental quality. Traditional air purification methods mainly rely on industrial equipment and pollutant treatment facilities, which are often expensive and unable to comprehensively solve the problem.

[0003] In recent years, plant-based air purification methods have gradually attracted attention. Plants purify the air by absorbing harmful substances and releasing oxygen, providing a sustainable, economical, and eco-friendly solution for cities. However, existing plant-based air purification methods often lack scientific rigor and systematic approach, failing to fully consider the diversity and complexity of urban environments.

[0004] Currently available plant-based air purification systems primarily rely on increasing the number of plants, but lack scientific research on plant species, distribution, and synergistic effects, making it difficult to guarantee purification effectiveness. Furthermore, the lack of accurate predictions of the distribution and changes of urban pollution sources makes it difficult to achieve optimal plant selection and placement.

[0005] Therefore, this application provides an urban air adsorption and purification method based on forestry planning. By fully considering the urban environment, pollution sources, plant characteristics, and data analysis, it provides a scientific, systematic, and sustainable air purification solution, aiming to overcome the shortcomings of existing technologies and provide a more efficient, intelligent, and comprehensive urban air purification method. Summary of the Invention

[0006] In order to overcome a series of defects in the existing technology, the purpose of this patent is to provide an urban air adsorption and purification method based on forestry planning, which includes the following steps:

[0007] Step 1: Analyze the urban air pollution situation to determine the types, concentrations, and spatial distribution of major pollutants;

[0008] Step 2: Select garden plants with strong adsorption and purification capabilities based on the type and concentration of pollutants.

[0009] Step 3: Design a reasonable greening layout and structure based on the city's climate, wind direction, topography, and building distribution;

[0010] Step 4: Based on the spatial distribution of pollution sources, formulate a reasonable planting plan for garden plants, including planting location, density and form, in order to achieve a good purification effect;

[0011] Step 5: Regularly monitor and evaluate the concentration of air pollutants and the growth status of garden plants, adjust the planting plan according to the actual situation, maintain the health and vitality of garden plants, and improve the efficiency and sustainability of air purification.

[0012] Furthermore, step 1 includes the following steps:

[0013] Data on pollutant concentrations, meteorological parameters, and traffic flow are obtained from various monitoring points, and satellite remote sensing data is used as input for spatial data on pollutant distribution.

[0014] Preprocess the collected data to improve its quality and usability;

[0015] Exploratory analysis was conducted on the preprocessed data to understand its basic characteristics and distribution, as well as the main influencing factors and problems of pollutants.

[0016] Based on the data analysis results, a convolutional neural network was selected as the neural network model. Pollutant concentration, meteorological parameters, traffic flow, time series, and spatial coordinates were used as input variables, and spatial data of pollutant distribution were used as output variables to construct a predictive model for pollutant concentration and distribution.

[0017] Using a portion of the data as a training set, the convolutional neural network model is trained, and the model's parameters and structure are adjusted to improve the model's fitting and generalization abilities.

[0018] Use another set of data as a test set to evaluate the convolutional neural network model;

[0019] Convolutional neural network models are used to predict pollutant concentrations and distributions over a future period, generating heatmaps of the predictions for air quality monitoring, early warning, and management.

[0020] Furthermore, the specific structure of the prediction model for pollutant concentration and distribution includes:

[0021] Input layer: Input data of size 256x256x8, including pollutant concentration, meteorological parameters, traffic flow, time series, spatial coordinates and remote sensing image three bands, which are standardized to make their range between [-1,1];

[0022] Convolutional layer 1: Uses 32 5x5 convolutional kernels with a stride of 1 and padding of the same size to perform convolution operations on the input data and output a feature map of size 256x256x32.

[0023] Activation layer 1: The ReLU function is used as the activation function to perform a non-linear transformation on the output of convolutional layer 1, while keeping the output size unchanged;

[0024] Pooling layer 1: Uses average pooling with a size of 2x2 and a stride of 2. Downsamples the output of activation layer 1 to produce a feature map of size 128x128x32.

[0025] Convolutional layer 2: Uses 64 3x3 convolutional kernels with a stride of 1 and padding of the same. Performs convolution operation on the output of pooling layer 1, and outputs a feature map of size 128x128x64.

[0026] Activation layer 2: The ReLU function is used as the activation function to perform a non-linear transformation on the output of convolutional layer 2, while keeping the output size unchanged;

[0027] Pooling layer 2: Uses average pooling with a size of 2x2 and a stride of 2 to downsample the output of activation layer 2, resulting in a feature map of size 64x64x64.

[0028] Convolutional layer 3: Uses 64 3x3 convolutional kernels with a stride of 1 and padding of the same size. Performs convolution operation on the output of pooling layer 2, and outputs a feature map of size 64x64x128.

[0029] Activation layer 3: The ReLU function is used as the activation function to perform a non-linear transformation on the output of convolutional layer 3, while keeping the output size unchanged;

[0030] Pooling layer 3: Uses average pooling with a size of 2x2 and a stride of 2 to downsample the output of activation layer 3, resulting in a feature map of size 32x32x128.

[0031] Skip connection: The outputs of pooling layer 1, pooling layer 2 and pooling layer 3 are concatenated to obtain a feature map with an output size of 32x32x224;

[0032] Global average pooling layer: Performs global average pooling on the output of the skip connections to obtain a feature vector with an output size of 224;

[0033] Fully connected layer: Uses 1 neuron to multiply the output of the global average pooling layer by a weight matrix, and adds a bias vector to obtain a feature vector with an output size of 1;

[0034] Activation layer: The sigmoid function is used as the activation function to perform a non-linear transformation on the output of the fully connected layer. The output size remains unchanged and is used as the final output of the model.

[0035] Furthermore, step 2 includes the following steps:

[0036] Considering the plant's growth environment, type, characteristics, and resistance and resilience to pollutants, select plants suitable for local climate, soil, and water conditions to ensure their healthy growth and adsorption and purification effects. Also, try to select plants with a high leaf area index and a long growing season to increase their adsorption area and time for pollutants.

[0037] Considering the synergistic effect of plants with other plants, select plant combinations that can complement each other and enhance adsorption and purification capabilities to improve air quality efficiency.

[0038] Furthermore, the selection of garden plants specifically includes the following steps:

[0039] Define objectives and constraints: Based on the urban air pollution situation, the objective function is to maximize the adsorption and purification capacity of plants for major pollutants, and the planting location, density, and form of plants are defined as constraints.

[0040] Generate candidate solutions: Select plants with strong adsorption and purification capabilities, and generate plant combinations based on the synergistic effect of the plants;

[0041] Evaluate candidate solutions: Evaluate and score each candidate solution based on the objective function and constraints;

[0042] Selecting the optimal solution: Based on the evaluation results, select one or more optimal solutions from the candidate solutions as the result of plant selection.

[0043] Furthermore, step 3 includes the following steps:

[0044] Collect and analyze the city's topography, climate, wind direction, and building distribution to understand the city's natural characteristics, environmental problems, greening needs, and development potential;

[0045] Based on the data analysis results, the goals and indicators of urban greening are determined, including green coverage rate, greening level, greening form and greening effect;

[0046] Based on greening goals and indicators, design a reasonable greening layout and structure, including green space system, green space classification, green space distribution and green space combination;

[0047] Based on the distribution and characteristics of air pollution, and utilizing the principles and methods of urban planning and landscape design, combined with the city's topography, climate, wind direction, and building distribution, we will optimize the layout and structure of greening to improve green coverage, increase greening levels, optimize greening forms, and enhance greening effects.

[0048] The layout and structure of the greening project should be evaluated and optimized to ensure its scientific validity and feasibility.

[0049] Furthermore, step 4 includes the following steps:

[0050] Analyze the type, location, intensity, and emission characteristics of pollution sources to assess the degree and extent of their impact on air quality;

[0051] Based on the garden plants selected in step 2 and the requirements of their growing environment, design reasonable planting locations, densities, and forms.

[0052] Furthermore, step 5 includes the following steps:

[0053] Select appropriate monitoring equipment and indicators, and determine the monitoring frequency and monitoring time according to the types and characteristics of garden plants;

[0054] Monitoring equipment is installed or used on different parts and locations of garden plants to dynamically record monitoring data, which is then uploaded to a data management platform for data analysis and visualization.

[0055] Based on the data analysis results, assess the growth status and air purification effect of garden plants to determine whether the planting plan needs to be adjusted.

[0056] Based on the adjusted planting plan, corresponding management measures will be taken to maintain the health and vitality of garden plants and improve the efficiency and sustainability of air purification.

[0057] Compared with the prior art, this application has at least the following technical effects or advantages.

[0058] This application uses advanced technologies such as satellite remote sensing and neural network models to predict the distribution of urban air pollution. Based on the types and concentrations of pollutants and the absorption capacity of different vegetation, it adopts a data-driven approach to develop personalized landscaping plans, optimizing aspects such as green space selection, layout, and plant selection. This can effectively purify urban air and improve environmental quality. Attached Figure Description

[0059] Figure 1 This is a flowchart of an urban air adsorption and purification method based on forestry planning disclosed in the embodiments of this application.

[0060] Figure 2 This is a structural diagram of the prediction model for pollutant concentration and distribution in this application. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments of this invention will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some embodiments of this invention, but not all embodiments.

[0062] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] The embodiments and directional terms described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0064] Reference Figure 1 A method for urban air adsorption and purification based on forestry planning includes the following steps:

[0065] Step 1: Analyze the urban air pollution situation to determine the types, concentrations, and spatial distribution of major pollutants;

[0066] Step 2: Select garden plants with strong adsorption and purification capabilities based on the type and concentration of pollutants.

[0067] Step 3: Design a reasonable greening layout and structure based on the city's climate, wind direction, topography, and building distribution;

[0068] Step 4: Based on the spatial distribution of pollution sources, formulate a reasonable planting plan for garden plants, including planting location, density and form, in order to achieve a good purification effect;

[0069] Step 5: Regularly monitor and evaluate the concentration of air pollutants and the growth status of garden plants, adjust the planting plan according to the actual situation, maintain the health and vitality of garden plants, and improve the efficiency and sustainability of air purification.

[0070] In this embodiment, step 1 requires using modern air monitoring technology to obtain detailed urban air quality data. By analyzing the types, concentrations, and spatial distribution of major pollutants, a comprehensive urban air quality map is created, including the monitoring and analysis of key pollutants such as particulate matter, nitrogen oxides, and volatile organic compounds. Step 2, based on the analysis results of step 1, selects garden plants with good adsorption and purification capabilities for specific pollutants. For example, some plants can effectively adsorb particulate matter, while others may be more effective at adsorbing nitrogen oxides. The adsorption efficiency, growth rate, and adaptability of the plants are considered to ensure their healthy growth and maximum benefit in the urban environment. Step 3 comprehensively considers factors such as urban climate, wind direction, topography, and building distribution to design the most suitable garden structure and layout. For example, under the influence of wind direction, the locations of adsorption plants are rationally set to maximize their adsorption capacity and ensure that airflow is maximized. Based on the spatial distribution of pollution sources identified in Step 1, Step 4 develops a planting plan for garden plants, determining the planting location, density, and form for each area to ensure optimal air purification in key areas. This includes forming adsorption zones or creating green barriers to effectively isolate and mitigate air pollutants. Step 5 establishes a regular monitoring system to track air pollutant concentrations and the growth status of garden plants. Real-time data allows for timely adjustments to the planting plan to adapt to changes in the urban environment and maintain plant health and vitality, contributing to improved air purification efficiency and system sustainability.

[0071] Furthermore, step 1 includes the following steps:

[0072] Data on pollutant concentrations, meteorological parameters, and traffic flow are obtained from various monitoring points, and satellite remote sensing data is used as input for spatial data on pollutant distribution.

[0073] Preprocess the collected data to improve its quality and usability;

[0074] Exploratory analysis was conducted on the preprocessed data to understand its basic characteristics and distribution, as well as the main influencing factors and problems of pollutants.

[0075] Based on the data analysis results, a convolutional neural network was selected as the neural network model. Pollutant concentration, meteorological parameters, traffic flow, time series, and spatial coordinates were used as input variables, and spatial data of pollutant distribution were used as output variables to construct a predictive model for pollutant concentration and distribution.

[0076] Using a portion of the data as a training set, the convolutional neural network model is trained, and the model's parameters and structure are adjusted to improve the model's fitting and generalization abilities.

[0077] Use another set of data as a test set to evaluate the convolutional neural network model;

[0078] Convolutional neural network models are used to predict pollutant concentrations and distributions over a future period, generating heatmaps of the predictions for air quality monitoring, early warning, and management.

[0079] This embodiment provides a scientific basis and effective tools for urban air purification methods through a series of detailed and systematic steps, enabling them to continuously adapt and optimize in the face of ever-changing urban environments. By integrating advanced data analysis techniques and neural network models, this method is expected to improve prediction accuracy and real-time performance while providing urban residents with a cleaner and healthier living environment.

[0080] like Figure 2 As shown, the specific structure of the prediction model for pollutant concentration and distribution further includes:

[0081] Input layer: Input data of size 256x256x8, including pollutant concentration, meteorological parameters, traffic flow, time series, spatial coordinates and remote sensing image three bands, which are standardized to make their range between [-1,1];

[0082] Convolutional layer 1: Uses 32 5x5 convolutional kernels with a stride of 1 and padding of the same size to perform convolution operations on the input data and output a feature map of size 256x256x32.

[0083] Activation layer 1: The ReLU function is used as the activation function to perform a non-linear transformation on the output of convolutional layer 1, while keeping the output size unchanged;

[0084] Pooling layer 1: Uses average pooling with a size of 2x2 and a stride of 2. Downsamples the output of activation layer 1 to produce a feature map of size 128x128x32.

[0085] Convolutional layer 2: Uses 64 3x3 convolutional kernels with a stride of 1 and padding of the same. Performs convolution operation on the output of pooling layer 1, and outputs a feature map of size 128x128x64.

[0086] Activation layer 2: The ReLU function is used as the activation function to perform a non-linear transformation on the output of convolutional layer 2, while keeping the output size unchanged;

[0087] Pooling layer 2: Uses average pooling with a size of 2x2 and a stride of 2 to downsample the output of activation layer 2, resulting in a feature map of size 64x64x64.

[0088] Convolutional layer 3: Uses 64 3x3 convolutional kernels with a stride of 1 and padding of the same size. Performs convolution operation on the output of pooling layer 2, and outputs a feature map of size 64x64x128.

[0089] Activation layer 3: The ReLU function is used as the activation function to perform a non-linear transformation on the output of convolutional layer 3, while keeping the output size unchanged;

[0090] Pooling layer 3: Uses average pooling with a size of 2x2 and a stride of 2 to downsample the output of activation layer 3, resulting in a feature map of size 32x32x128.

[0091] Skip connection: The outputs of pooling layer 1, pooling layer 2 and pooling layer 3 are concatenated to obtain a feature map with an output size of 32x32x224;

[0092] Global average pooling layer: Performs global average pooling on the output of the skip connections to obtain a feature vector with an output size of 224;

[0093] Fully connected layer: Uses 1 neuron to multiply the output of the global average pooling layer by a weight matrix, and adds a bias vector to obtain a feature vector with an output size of 1;

[0094] Activation layer: The sigmoid function is used as the activation function to perform a non-linear transformation on the output of the fully connected layer. The output size remains unchanged and is used as the final output of the model.

[0095] To evaluate the model's classification ability and accuracy, we used confusion matrix, precision, recall, and F1 score. The confusion matrix shows the correspondence between the model's predicted classes and the actual classes, as well as the number of correct and incorrect predictions. Precision is the proportion of samples predicted as belonging to a certain class that actually belong to that class. Recall is the proportion of samples actually belonging to a certain class that the model predicted as belonging to that class. The F1 score is the harmonic mean of precision and recall, used to comprehensively measure the model's classification performance. Generally, higher values ​​for these metrics indicate better classification ability and accuracy. We used the Air Quality Index (AQI) to classify air quality into four categories: Excellent (0-50), Good (51-100), Lightly Polluted (101-150), and Moderately Polluted (151-200). Our test set contains 100 samples, each with one model output value and one actual observation value. The model output value is converted into a discrete class: output values ​​between 0 and 0.25 are classified as "good," values ​​between 0.25 and 0.5 as "good," and so on. We then use a confusion matrix to show the correspondence between the model's predicted and actual classes, as well as the number of correct and incorrect predictions, as shown in Table 1.

[0096] Table 1

[0097] Actual / Forecast excellent good Mild pollution Moderate pollution excellent 23 2 0 0 good 1 22 2 0 Mild pollution 0 1 23 1 Moderate pollution 0 0 1 24

[0098] The classification ability and accuracy of the model are further quantified using metrics such as precision, recall, and F1 score, as shown in Table 2.

[0099] Table 2

[0100] category Accuracy Recall rate F1 score excellent 0.96 0.92 0.94 good 0.88 0.88 0.88 Mild pollution 0.88 0.92 0.90 Moderate pollution 0.96 0.96 0.96

[0101] From Tables 1 and 2, we can see that the confusion matrix and classification evaluation index have good classification ability and accuracy in all categories, with no obvious erroneous predictions, and the values ​​of each index are relatively high, indicating that the model has excellent classification performance.

[0102] Furthermore, step 2 includes the following steps:

[0103] Considering the plant's growth environment, type, characteristics, and resistance and resilience to pollutants, select plants suitable for local climate, soil, and water conditions to ensure their healthy growth and adsorption and purification effects. Also, try to select plants with a high leaf area index and a long growing season to increase their adsorption area and time for pollutants.

[0104] Considering the synergistic effect of plants with other plants, select plant combinations that can complement each other and enhance adsorption and purification capabilities to improve air quality efficiency.

[0105] This embodiment constructs a diverse garden plant community by comprehensively considering the ecological characteristics, growth environment, resistance, resilience, and synergistic effects of plants, maximizing their benefits in urban air purification. This not only helps improve air quality but also provides urban residents with a pleasant and healthy living environment.

[0106] Furthermore, the selection of garden plants specifically includes the following steps:

[0107] Define objectives and constraints: Based on the urban air pollution situation, the objective function is to maximize the adsorption and purification capacity of plants for major pollutants, and the planting location, density, and form of plants are defined as constraints.

[0108] Generate candidate solutions: Select plants with strong adsorption and purification capabilities, and generate plant combinations based on the synergistic effect of the plants;

[0109] Evaluate candidate solutions: Evaluate and score each candidate solution based on the objective function and constraints;

[0110] Selecting the optimal solution: Based on the evaluation results, select one or more optimal solutions from the candidate solutions as the result of plant selection.

[0111] This embodiment, through clear objectives, reasonable generation of candidate solutions, comprehensive evaluation process, and scientific selection mechanism, can ensure that the selected garden plant scheme achieves the best air purification effect in actual implementation.

[0112] Furthermore, the objective function and constraints are determined considering pollutant types, vegetation types, pollution monitoring data, and plant resource databases. The process of selecting garden plants is expressed as follows:

[0113] maxf(x) x

[0114] stg i (x)≤0, i=1,2,...,m

[0115] h j (x) = 0, j = 1, 2, ..., n

[0116] Where x represents the selection scheme of candidate plants or plant combinations, f(x) represents the adsorption and purification capacity of the plant, and g i (x) and h j (x) represents some scalar functions, and represents the constraints.

[0117] The goal of the optimization problem in this embodiment is to find a plant selection scheme x such that the objective function f(x) reaches its maximum value while satisfying a series of constraints g. i (x) and h j (x). Solving this optimization problem will provide the optimal selection of garden plants for urban air purification, thereby maximizing air quality.

[0118] Furthermore, step 3 includes the following steps:

[0119] Collect and analyze the city's topography, climate, wind direction, and building distribution to understand the city's natural characteristics, environmental problems, greening needs, and development potential;

[0120] Based on the data analysis results, the goals and indicators of urban greening are determined, including green coverage rate, greening level, greening form and greening effect;

[0121] Based on greening goals and indicators, design a reasonable greening layout and structure, including green space system, green space classification, green space distribution and green space combination;

[0122] Based on the distribution and characteristics of air pollution, and utilizing the principles and methods of urban planning and landscape design, combined with the city's topography, climate, wind direction, and building distribution, we will optimize the layout and structure of greening to improve green coverage, increase greening levels, optimize greening forms, and enhance greening effects.

[0123] The layout and structure of the greening project should be evaluated and optimized to ensure its scientific validity and feasibility.

[0124] This embodiment, through comprehensive urban analysis, fully considers natural features such as topography, climate, wind direction, and building distribution, and combines the results of data analysis to formulate a scientific and reasonable urban greening plan. The plan aims to clarify the goals and indicators of urban greening, including green coverage rate, greening levels, greening forms, and greening effects, to meet the city's environmental improvement needs. The design of a reasonable greening layout and structure includes green space systems, green space classification, green space distribution, and green space combinations to maximize urban green coverage and enhance greening effects through optimized layout and structure. In this process, special attention is paid to the distribution and characteristics of air pollution. Combining urban planning and landscape design principles, the plan optimizes green space settings through reasonable plant selection and layout to address air quality issues. Finally, through evaluation and optimization of the greening layout and structure plan, the scientific validity and implementability of the plan are ensured. This series of steps aims to make urban greening closer to nature and adaptable to urban needs, creating a more livable environment for urban residents.

[0125] Furthermore, step 4 includes the following steps:

[0126] Analyze the type, location, intensity, and emission characteristics of pollution sources to assess the degree and extent of their impact on air quality;

[0127] Based on the garden plants selected in step 2 and the requirements of their growing environment, design reasonable planting locations, densities, and forms.

[0128] By designing and adjusting step 4 in this embodiment, the green space system will more effectively address the impact of pollution sources and improve the overall level of urban air quality.

[0129] Furthermore, step 5 includes the following steps:

[0130] Select appropriate monitoring equipment and indicators, and determine the monitoring frequency and monitoring time according to the types and characteristics of garden plants;

[0131] Monitoring equipment is installed or used on different parts and locations of garden plants to dynamically record monitoring data, which is then uploaded to a data management platform for data analysis and visualization.

[0132] Based on the data analysis results, assess the growth status and air purification effect of garden plants to determine whether the planting plan needs to be adjusted.

[0133] Based on the adjusted planting plan, corresponding management measures will be taken to maintain the health and vitality of garden plants and improve the efficiency and sustainability of air purification.

[0134] In this embodiment, by regularly monitoring, evaluating and adjusting, problems can be identified and corresponding improvements can be made in a timely manner, and the configuration of garden plants can be optimized to improve the overall efficiency of the air purification system. This process is a key step in continuously improving and optimizing urban air quality.

Claims

1. A method for urban air adsorption and purification based on forestry planning, characterized in that, Includes the following steps: Step 1: Analyze the urban air pollution situation to determine the types, concentrations, and spatial distribution of major pollutants; Step 2: Select garden plants with strong adsorption and purification capabilities based on the type and concentration of pollutants. Step 3: Design a reasonable greening layout and structure based on the city's climate, wind direction, topography, and building distribution; Step 4: Based on the spatial distribution of pollution sources, formulate a reasonable planting plan for garden plants, including planting location, density and form, in order to achieve a good purification effect; Step 5: Regularly monitor and evaluate the concentration of air pollutants and the growth status of garden plants, adjust the planting plan according to the actual situation, maintain the health and vitality of garden plants, and improve the efficiency and sustainability of air purification. Step 1 includes the following steps: Data on pollutant concentrations, meteorological parameters, and traffic flow are obtained from various monitoring points, and satellite remote sensing data is used as input for spatial data on pollutant distribution. Preprocess the collected data to improve its quality and usability; Exploratory analysis was conducted on the preprocessed data to understand its basic characteristics and distribution, as well as the main influencing factors and problems of pollutants. Based on the data analysis results, a convolutional neural network was selected as the neural network model. Pollutant concentration, meteorological parameters, traffic flow, time series, and spatial coordinates were used as input variables, and spatial data of pollutant distribution were used as output variables to construct a predictive model for pollutant concentration and distribution. Using a portion of the data as a training set, the convolutional neural network model is trained, and the model's parameters and structure are adjusted to improve the model's fitting and generalization abilities. Use another set of data as a test set to evaluate the convolutional neural network model; Convolutional neural network models are used to predict pollutant concentrations and distributions over a future period, generating heatmaps of the predictions for air quality monitoring, early warning, and management. The selection of garden plants specifically includes the following steps: Define objectives and constraints: Based on the urban air pollution situation, the objective function is to maximize the adsorption and purification capacity of plants for major pollutants, and the planting location, density, and form of plants are defined as constraints. Generate candidate solutions: Select plants with strong adsorption and purification capabilities, and generate plant combinations based on the synergistic effect of the plants; Evaluate candidate solutions: Evaluate and score each candidate solution based on the objective function and constraints; Selecting the optimal solution: Based on the evaluation results, select one or more optimal solutions from the candidate solutions as the result of plant selection.

2. The urban air adsorption and purification method based on forestry planning according to claim 1, characterized in that, The specific structure of the predictive model for pollutant concentration and distribution includes: Input layer: Input data of size 256x256x8, including pollutant concentration, meteorological parameters, traffic flow, time series, spatial coordinates and remote sensing image three bands, which are standardized to make their range between [-1,1]; Convolutional layer 1: Uses 32 5x5 convolutional kernels with a stride of 1 and padding of the same size to perform convolution operations on the input data and output a feature map of size 256x256x32. Activation layer 1: The ReLU function is used as the activation function to perform a non-linear transformation on the output of convolutional layer 1, while keeping the output size unchanged; Pooling layer 1: Uses average pooling with a size of 2x2 and a stride of 2. Downsamples the output of activation layer 1 to produce a feature map of size 128x128x32. Convolutional layer 2: Uses 64 3x3 convolutional kernels with a stride of 1 and padding of the same. Performs convolution operation on the output of pooling layer 1, and outputs a feature map of size 128x128x64. Activation layer 2: The ReLU function is used as the activation function to perform a non-linear transformation on the output of convolutional layer 2, while keeping the output size unchanged; Pooling layer 2: Uses average pooling with a size of 2x2 and a stride of 2 to downsample the output of activation layer 2, resulting in a feature map of size 64x64x64. Convolutional layer 3: Uses 64 3x3 convolutional kernels with a stride of 1 and padding of the same size. Performs convolution operation on the output of pooling layer 2, and outputs a feature map of size 64x64x128. Activation layer 3: The ReLU function is used as the activation function to perform a non-linear transformation on the output of convolutional layer 3, while keeping the output size unchanged; Pooling layer 3: Uses average pooling with a size of 2x2 and a stride of 2 to downsample the output of activation layer 3, resulting in a feature map of size 32x32x128. Skip connection: The outputs of pooling layer 1, pooling layer 2 and pooling layer 3 are concatenated to obtain a feature map with an output size of 32x32x224; Global average pooling layer: Performs global average pooling on the output of the skip connections to obtain a feature vector with an output size of 224; Fully connected layer: Uses 1 neuron to multiply the output of the global average pooling layer by a weight matrix, and adds a bias vector to obtain a feature vector with an output size of 1; Activation layer: The sigmoid function is used as the activation function to perform a non-linear transformation on the output of the fully connected layer. The output size remains unchanged and is used as the final output of the model.

3. The urban air adsorption and purification method based on forestry planning according to claim 1, characterized in that, Step 2 includes the following steps: Considering the plant's growth environment, type, characteristics, and resistance and resilience to pollutants, select plants suitable for local climate, soil, and water conditions to ensure their healthy growth and adsorption and purification effects. Also, select plants with a high leaf area index and a long growing season to increase their adsorption area and time for pollutants. Considering the synergistic effect of plants with other plants, select plant combinations that can complement each other and enhance adsorption and purification capabilities to improve air quality efficiency.

4. The urban air adsorption and purification method based on forestry planning according to claim 1, characterized in that, Step 3 includes the following steps: Collect and analyze the city's topography, climate, wind direction, and building distribution to understand the city's natural characteristics, environmental problems, greening needs, and development potential; Based on the data analysis results, the goals and indicators of urban greening are determined, including green coverage rate, greening level, greening form and greening effect; Based on greening goals and indicators, design a reasonable greening layout and structure, including green space system, green space classification, green space distribution and green space combination; Based on the distribution and characteristics of air pollution, and utilizing the principles and methods of urban planning and landscape design, combined with the city's topography, climate, wind direction, and building distribution, we will optimize the layout and structure of greening to improve green coverage, increase greening levels, optimize greening forms, and enhance greening effects. The layout and structure of the greening project should be evaluated and optimized to ensure its scientific validity and feasibility.

5. The urban air adsorption and purification method based on forestry planning according to claim 1, characterized in that, Step 4 includes the following steps: Analyze the type, location, intensity, and emission characteristics of pollution sources to assess the degree and extent of their impact on air quality; Based on the garden plants selected in step 2 and the requirements of their growing environment, design reasonable planting locations, densities, and forms.

6. The urban air adsorption and purification method based on forestry planning according to claim 1, characterized in that, Step 5 includes the following steps: Select appropriate monitoring equipment and indicators, and determine the monitoring frequency and monitoring time according to the types and characteristics of garden plants; Monitoring equipment is installed or used on different parts and locations of garden plants to dynamically record monitoring data, which is then uploaded to a data management platform for data analysis and visualization. Based on the data analysis results, assess the growth status and air purification effect of garden plants to determine whether the planting plan needs to be adjusted. Based on the adjusted planting plan, corresponding management measures will be taken to maintain the health and vitality of garden plants and improve the efficiency and sustainability of air purification.

Images