Device and method for realizing carbon emission reduction prediction based on carbon emission data and support vector machine

Abstract

The invention discloses a carbon emission reduction prediction device and method based on carbon emission data and a support vector machine, and the device comprises a digital twin modeling module which is used for generating a digital terrain model through laser radar scanning and satellite remote sensing data, and constructing a three-dimensional terrain model based on a fractal algorithm; the terrain analysis module is used for inputting a digital terrain model, analyzing terrain features through a machine learning algorithm and outputting vegetation types suitable for planting of the terrain and an air purification species recommendation scheme; the dynamic carbon emission evaluation module is used for collecting real-time environmental parameters and historical data and calculating a carbon emission evaluation value of the current region; and the emission reduction solution generation module is used for generating a plurality of emission reduction schemes in combination with the terrain analysis result when the carbon emission evaluation value does not reach the standard, analyzing the terrain through the detected target area terrain and the support vector machine, and determining a carbon emission scheme through the predicted carbon emission evaluation value, so that the method is more intelligent and reliable.

Description

Technical Field

[0001] The present invention relates to the technical field of computer data processing, and particularly to a carbon emission reduction prediction device and method based on carbon emission data and support vector machines. Background Art

[0002] Although the expansion of the digital industry scale may increase carbon emissions, new technologies and innovative ways can be introduced to reduce carbon emissions. Due to the network effect of the digital economy, there may be a non-linear relationship between the growth of the digital economy and carbon emissions, that is, an inverted "U" shape. The development of the digital economy field, although it will lead to an increase in carbon emissions in the short term, will gradually weaken and eventually stabilize over time. Some scholars have constructed relevant index systems for the digital economy and carbon emissions, and explored how the digital economy can empower carbon emission reduction to promote regional green development from the perspectives of industrial structure and green innovation. In the prior art, carbon emission monitoring is mostly based on fixed sensors and static data visualization platforms, and its limitations are significant. Traditional systems mostly use two-dimensional charts or tables to display carbon emission data, lacking the in-depth integration of geographical space and dynamic environmental parameters, and it is difficult to intuitively reflect the regional carbon footprint distribution and real-time change trend. Existing solutions mostly stay at the data display level, and users cannot verify the effectiveness of emission reduction strategies through interactive operations (such as gesture adjustment, AR simulation). In addition, in order to better determine the carbon emission reduction plan, carbon emission numerical prediction is crucial. In the prior art, such as Chinese Patent Application No. 202410955095.6, with the publication date of November 1, 2024, it discloses a carbon emission information processing method, device and equipment, including: obtaining a target scene scan image; determining a carbon emission monitoring model of the target scene according to the target scene scan image; identifying carbon sources in the carbon emission monitoring model to obtain historical carbon emission data of the carbon sources in the carbon emission monitoring model; training a carbon emission prediction model and carbon emission prediction data of the carbon sources according to the historical carbon emission data of the carbon sources and a preset network model; processing the carbon emission prediction data to obtain the carbon emission value of the carbon source, improving the accuracy of carbon emission data of the carbon source in various environments and stages.

[0003] In the above literature, only the carbon emission prediction data is determined by the target carbon source through the prediction network model, and the amount of emissions that can be absorbed by existing vegetation is not considered when obtaining the carbon emission value, resulting in low accuracy of the predicted carbon emission data. Among them, the carbon emission detection model does not adapt to the terrain according to the data actually obtained in the terrain, and after obtaining the carbon emission value, it is impossible to determine the cost reason according to the obtained vegetation plan and further determine the emission reduction method, which is currently mainly achieved by manual selection, making it impossible to make carbon emission reduction more intelligent and reliable. Summary of the Invention

[0004] The present invention provides a carbon emission reduction prediction device and method based on carbon emission data and support vector machines. By detecting the terrain of the target area, analyzing the terrain through support vector machines, and determining the carbon emission reduction plan based on the predicted carbon emission assessment value, it is more intelligent and reliable. On the one hand, the present invention provides a carbon emission reduction prediction method, including: Step S1: Scanning the terrain of the target area through lidar to obtain point cloud data, and synchronously collecting real-time environmental parameters; retrieving vegetation distribution data from the satellite remote sensing database; Step S2: Fusing the point cloud data and the vegetation distribution data, generating a digital terrain model using the fractal algorithm, and annotating the terrain features, where the terrain features include slope, elevation, and soil type; and loading the vegetation carbon absorption rate, growth cycle, and environmental adaptability data into the three-dimensional model through a preset dynamic vegetation parameter library; Step S3: Inputting the digital terrain model into the trained terrain analysis model, dividing the area type through the support vector machine algorithm, and outputting the vegetation types suitable for planting in the terrain and the priority of air purification species; Step S4: The dynamic carbon emission assessment module determines the carbon emission assessment value according to the real-time environmental parameters and historical data, based on the difference between the sum of the energy consumption and traffic emissions in the target area and the carbon emission absorption data of the planting area; Step S5: When the carbon emission assessment value calculated by the system through the dynamic carbon emission assessment unit exceeds the preset threshold, an emission reduction instruction is automatically generated. The emission reduction solution generation module generates multiple solutions according to the emission reduction instruction, determines the cost performance ratio of each solution according to the ratio of the total cost to the emission reduction cost, and determines that the solution with a cost performance ratio higher than the preset cost performance ratio is the optimal solution. The above method constructs a high-precision digital terrain model through lidar and satellite remote sensing, recommends terrain-adapted vegetation solutions in combination with machine learning algorithms, and real-time corrects the carbon emission assessment value based on dynamic environmental parameters. Especially in the calculation process of the carbon emission assessment value, the actual emissions that can be absorbed by vegetation are removed from the industrial emissions and traffic emissions to ensure the reliability of the emissions assessment. After obtaining the carbon emission assessment value, a carbon emission reduction plan is determined, and a high-return plan with high cost performance is determined through the total cost and carbon emission cost, reducing the deviation of manual decision-making, ensuring the intelligence and reliability of machine calculation, and at the same time supporting the data feedback and model iteration after implementation, forming a "prediction - implementation - optimization" closed loop.

[0005] Further, in Step S4: Carbon emission assessment value = ∑(α i ×E i ) + ∑(β j ×M j ) - ∑(γ k ×A k ) where αi is the industrial emission coefficient of the i-th category, E i is the energy consumption, β j is the j traffic emission coefficient of the j category, M is the mileage, γ k is the carbon sink coefficient of the k-th category of vegetation, A k is the planting area.

[0006] Through the above settings, by setting the industrial emission coefficient for various industrial energy consumptions, the traffic emission coefficient before the traffic emissions, and the carbon sink coefficient of the vegetation before the vegetation area, the calculated carbon emissions can be ensured to be more reliable.

[0007] Furthermore, in step S5, calculate the cost performance ratio of each plan: P = ln(Y 总成本 ) / ln(X 减排成本 ), where Y 总成本 is the cost of vegetation planting or equipment procurement, and X 减排成本 is the cost of emission reduction requirements. Select the plans with P > 1.2 to enter the decision-making process.

[0008] Through the above plan, determine the optimal plan by the ratio of the total cost to the emission reduction cost, ensuring that the cost required for emission reduction can reach the emission reduction requirement cost as much as possible and ensuring a certain cost performance ratio.

[0009] Furthermore, after step S5, it also includes: step S6: Display the three-dimensional effect of the plan in an AR overlay manner through the flexible display component, and the user can adjust the parameters with gestures and select the optimal plan.

[0010] Through the above plan, the flexible display component can facilitate the selection of the optimal plan and the display of carbon emission reduction data through the user's gestures.

[0011] Furthermore, the real-time environmental parameters include temperature, humidity, and light intensity. Step S1 also includes retrieving meteorological data from the satellite remote sensing database.

[0012] Through the above settings, more suitable vegetation for planting can be obtained under real-time environmental parameters. Furthermore, "output the vegetation types suitable for planting in the terrain and the priority of air purification species" in step S3 includes: the output result of the terrain analysis model includes recommending fast-growing arbors for areas with a slope < 15° and soil organic matter > 2%, recommending drought-tolerant shrubs for arid areas and associating the cost data of drip irrigation technology, and ranking the air purification species according to the PM2.5 adsorption efficiency and VOC degradation efficiency.

[0013] Through the above settings, the corresponding vegetation types can be obtained according to different terrain conditions, and the optimal air purification species can be determined according to the PM2.5 adsorption efficiency and VOC degradation efficiency.

[0014] Furthermore, the cost performance calculation of the emission reduction solution generation module includes the 5-year cycle rate of return analysis of the planting cost and carbon sink income of the vegetation plan, and the comparison of the unit area cost and emission reduction efficiency between manual spraying and drone spraying in the dust removal plan.

[0015] With the above settings, the cost and emission reduction efficiency in the vegetation plan determine the final solution generation model, ensuring that a more suitable plan can be obtained.

[0016] Furthermore, the execution of the fractal algorithm in step S2 includes: Perform grid processing on the point cloud data and calculate the elevation standard deviation of each grid; Use the Diamond-Square algorithm to generate terrain details and adjust the fractal dimension according to the standard deviation; Map the planting conditions of different species in the vegetation parameter library to the corresponding grids and mark the suitability labels.

[0017] With the above settings, by mapping the point cloud data to the grid and calculating the elevation standard deviation of the network, the data of each network can be balanced, terrain details can be generated, and the fractal dimension can be adjusted by the standard deviation, so that the grid division is more accurate.

[0018] Furthermore, the AR interaction in step S6 includes: After the camera recognizes the real scene, superimpose the vegetation 3D model on the flexible display panel, and the model height is scaled at a real ratio of 1:100; Adjust the vegetation distribution density according to the instruction requirements, and the system calculates the carbon absorption increment in real time and updates the heat map color; If the carbon emission assessment value ≤ the threshold, set the area as a compliant area and trigger the green particle flow special effect.

[0019] With the above settings, the user adjusts the AR simulated vegetation distribution through touch gestures, and the system updates the carbon sink prediction value in real time and triggers the particle special effect feedback, with good user experience and good display effect.

[0020] On the other hand, the present invention also provides a carbon emission reduction prediction device based on carbon emission data and support vector machines, including a flexible display component, a digital twin modeling module, a terrain analysis module, a dynamic carbon emission assessment module, and an emission reduction solution generation module. The flexible display component includes a plurality of sub-display panels and a deformable display frame. The sub-display panels use OLED flexible screens and support curved surface folding and dynamic splicing; The digital twin modeling module is used to generate a digital terrain model DTM through lidar scanning and satellite remote sensing data, and construct a three-dimensional terrain model based on the fractal algorithm; The terrain analysis module inputs a digital terrain model, analyzes terrain features through machine learning algorithms, and outputs recommended vegetation types suitable for planting on the terrain and air purification species recommendation solutions; The dynamic carbon emission assessment module collects real-time environmental parameters and historical data, and calculates the current regional carbon emission assessment value; When the carbon emission assessment value does not meet the standard, the emission reduction solution generation module generates multiple emission reduction solutions in combination with the terrain analysis results, and calculates the cost performance of each solution.

[0021] With the above settings, a high-precision digital terrain model is constructed through lidar and satellite remote sensing, a terrain-adapted vegetation plan is recommended in combination with machine learning algorithms, and the carbon emission assessment value is corrected in real time based on dynamic environmental parameters. Especially during the calculation of the carbon emission assessment value, the actually absorbed emissions by vegetation are removed from industrial emissions and transportation emissions to ensure the reliability of emissions assessment. After obtaining the carbon emission assessment value, a carbon emission reduction plan is determined, and a high-return plan with high cost performance is determined through the total cost and carbon emission cost, reducing manual decision-making bias, ensuring the intelligence and reliability of machine calculation, and at the same time supporting data feedback and model iteration after implementation, forming a "prediction - implementation - optimization" closed loop. Brief Description of the Drawings

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0023] Figure 1 It is a flexible display component diagram of the present invention; Figure 2 It is a connection block diagram between each component module of the present invention.

[0024] Figure 3 It is a terrain display diagram in an embodiment of the present invention. Detailed Description of the Embodiments

[0025] The following will clearly and completely describe the technical solutions of the present invention in combination with the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments of the present invention belong to the protection scope of the present invention.

[0026] As Figure 2As shown in the figure, an embodiment of the present invention provides a carbon emission reduction prediction device based on carbon emission data and support vector machines, including a flexible display component, a digital twin modeling module, a terrain analysis module, a dynamic carbon emission assessment module, a emission reduction solution generation module, and an interactive decision-making module. The flexible display component is composed of a plurality of sub-display panels 100 and a deformable display frame 200. The sub-display panel 100 uses an OLED flexible screen and supports splicing after curved surface folding and unfolding. In this embodiment, the sub-display panel that is spliced after curved surface folding and unfolding is a prior art and will not be elaborated here.

[0027] The digital twin modeling module is used to generate a digital terrain model (DTM) through lidar scanning and satellite remote sensing data, and construct a three-dimensional terrain model based on the fractal algorithm. The terrain analysis module inputs the digital terrain model, analyzes the terrain features through machine learning algorithms, and outputs the recommended vegetation types suitable for planting on the terrain and air purification species. The dynamic carbon emission assessment module collects real-time environmental parameters and historical data. The real-time environmental parameters include temperature, humidity, and light intensity, and calculates the current regional carbon emission assessment value. When the carbon emission assessment value does not meet the standard, that is, when the carbon emission assessment value is greater than the preset standard carbon emission, the emission reduction solution generation module generates multiple emission reduction solutions in combination with the terrain analysis results, and calculates the cost performance of each solution.

[0028] The interactive decision-making module displays the three-dimensional effect of the emission reduction solution in an AR overlay manner through the flexible display component, supports touch adjustment and selection of the optimal solution.

[0029] Specifically, the flexible display component uses a multi-layer OLED flexible substrate with a thickness ≤ 0.3 mm, an adjustable curvature radius range of 5 - 50 mm, supports 180° folding and multi-directional bending. The sub-display panel covered with a hydrophobic coating switches to a high-contrast black-and-white mode under strong light. The sub-screen module corresponding to the high-carbon emission area supports a piezoelectric touch layer with a pressure sensing range of 0.1 - 5 N and a touch error ≤ 0.5 mm. The digital twin modeling module uses a 32-line lidar with a wavelength of 905 nm to scan the target area, generating three-dimensional terrain data with a point cloud density ≥ 100 points / ㎡, an elevation accuracy of ±0.1 m. It calls high-resolution multi-spectral satellite images to extract parameters such as vegetation cover index and surface temperature, uses the Diamond-Square algorithm to recursively subdivide the terrain grid, and dynamically adjusts the fractal dimension according to the elevation standard deviation. The elevation standard deviation is used to measure the accuracy and reliability of DEM data by calculating the difference in elevation between each pixel point and the elevation of its surrounding pixel points. Then, the network with a larger elevation standard deviation can be further divided, such as splitting the network in half, and by calling the geological database preset for the current area, the geological database stores historical data and current data for the current area, the data includes soil type, hydrological data, marking ecological sensitive areas such as wetlands and steep slopes, storing the carbon absorption rate, drought tolerance, and anti-pollution indicators of more than 200 plant species, associating the price and survival rate of nursery stock market data, and supporting cost-benefit analysis. In this embodiment, the Diamond-Square algorithm is an existing algorithm and will not be elaborated here.

[0030] The emission reduction solution generation module pre-sets a variety of general solutions, such as spraying air-purifying substances, vegetation restoration, etc., and combines the terrain analysis results with real-time cost data, such as the price fluctuations in the nursery stock market, to generate customized solutions.

[0031] The interaction decision module identifies the real scene through the built-in camera of the sub-display panel, and superimposes a semi-transparent vegetation 3D model rendered by the Unity engine, with a model scale of 1:100. In this embodiment, the Unity engine rendering is a commonly used rendering tool in the computer.

[0032] The potential interaction includes adjusting the vegetation distribution density according to instructions such as dragging, and calculating the real-time carbon absorption increment; adjusting the thermal layer level (from 1 km² to 100 m² accuracy) by pinching and zooming according to another instruction; popping up a data bubble by long-pressing according to another instruction, and the data bubble includes emissions, costs, implementation progress, etc.; triggering a green particle flow of the UE5 special effect model for the edge of the compliant area; warning with a red pulse in the high-risk area and synchronously driving the piezoelectric vibration membrane.

[0033] As Figure 1 and Figure 3 shown, the present invention provides a carbon emission reduction prediction method based on carbon emission data and support vector machines, including the following steps: Step S1: Data Acquisition and Storage Scan the terrain of the target area by lidar to obtain high-precision point cloud data. The accuracy of the collected point cloud data is ±0.1m, and the real-time environmental parameters including temperature, humidity, and light intensity are collected synchronously; Retrieve historical meteorological data and vegetation distribution data from the satellite remote sensing database and store them in the local edge computing unit; Step S2: Digital Twin Modeling Fuse the point cloud data and satellite images, and use the fractal algorithm to generate a digital terrain model (DTM), and label the terrain features such as slope, elevation, soil type, etc.; After selecting the most cost-effective solution, generate an implementation instruction and synchronize it to the cloud management platform. The fractal algorithm performs grid processing on the point cloud data and calculates the elevation standard deviation of each grid; The elevation standard deviation measures the accuracy and reliability of the DEM data by calculating the difference between the elevation of each pixel point and the elevation of its surrounding pixel points. Then, the grid with a larger elevation standard deviation can be further divided, such as dividing the grid into two, and the Diamond-Square algorithm is used to generate terrain details. Adjust the fractal dimension according to the standard deviation. The standard deviation is the square root of the squared difference between the pixel values of multiple grids and the average value of the entire area. If the standard deviation is greater than the preset standard deviation value, increase the fractal dimension, such as adjusting from one dimension to two dimensions, and further subdivide in the horizontal and vertical directions. The range of the fractal dimension D is 1.2 ≤ D ≤ 2.5; Map the planting conditions of different species in the vegetation parameter library to the corresponding grids and label the suitability tags; Associate with the dynamic vegetation parameter library and load the vegetation carbon absorption rate, growth cycle, and environmental adaptability data into the 3D model; Step S3: Terrain Analysis and Scheme Pre-generation Input the digital terrain model DTM into the trained terrain analysis model, and use the support vector machine (SVM) algorithm to divide the area types, and output the vegetation types suitable for planting on the terrain and the priority of air purification species; The processing of the terrain analysis model by the support vector machine (SVM) algorithm includes dividing the training set and the test set (ratio 7:3) and extracting terrain feature parameters as inputs; The terrain feature parameters include slope, elevation, soil pH value, etc. The support vector machine (SVM) classifier outputs the type labels of the vegetation types suitable for planting on the regional terrain and the priority of air purification species; When the slope < 15° and the soil organic matter > 2%, fast-growing arbors are recommended; When the area is an arid landform, drought-tolerant shrubs are recommended and the drip irrigation cost parameters are bound; And prioritize the air purification species based on the PM2.5 adsorption efficiency and VOC degradation efficiency. The vegetation types corresponding to different terrain feature parameters and the five priority types of air purification can be retrieved through the network or stored locally, and then the type labels that meet the requirements of the prediction model are determined through the training level and the test level.

[0034] Step S4: Dynamic carbon emission assessment Calculate the carbon emission assessment value based on real-time environmental parameters and historical data: Carbon emission assessment value = ∑(α i ×E i ) + ∑(β j ×M j ) − ∑(γ k ×A k ); where α i is the industrial emission coefficient of the i-th category, E i is the energy consumption, β j is the traffic emission coefficient of the j -th category, M j is the mileage, γ k is the vegetation carbon sink coefficient of the k-th category, A k is the planting area; α i , β j , γ k are pre-set coefficients. For example, for the industrial emission energy consumption of a chemical plant, it is obtained from the historical data of the same period of the plant. The industrial emission coefficient is the largest. For the mileage of traffic emissions, it is determined according to the length of the roads in the current area. Different emission coefficients are preset for different types of transportation tools. For the planting area, it can be determined by the area of the detected vegetation area currently.

[0035] If the assessment value exceeds the preset threshold, trigger an emission reduction instruction; otherwise, do not trigger an emission reduction instruction.

[0036] Step S5: Generation and optimization of emission reduction solutions According to the emission reduction instruction, call the terrain analysis results to generate multiple solutions, including vegetation planting solutions and dust removal equipment deployment solutions; Calculate the cost performance of each solution: P = ln(Y 总成本 ) / ln(X 减排成本 ) where Y 总成本 is the vegetation planting or equipment procurement cost, and X 减排成本 is the emission reduction demand cost; ln is the natural logarithm. If it is a vegetation planting solution, determine the vegetation planting cost as the price and quantity required to purchase vegetation seedlings, etc. If it is a dust removal equipment deployment solution, determine the type of dust removal equipment and the price and quantity required to purchase air purifying substances sprayed by the dust removal equipment, etc. Select the solutions with P > 1.2 to enter the decision-making process; discard the solutions with P less than or equal to 1.2 and re-plan the solutions. Step S6: Interactive solution selection and implementation The three-dimensional effect of the solution is displayed in an AR overlay manner through a flexible display component; the split display panel is unfolded into a tablet form, and after the camera recognizes the real scene, a virtual vegetation model is overlaid; the implementation process of displaying the three-dimensional effect of the solution in an AR overlay manner is prior art and will not be elaborated here.

[0037] The solution parameters are adjusted through user gesture instructions, and the user gesture instructions include dragging. The system updates the carbon emission prediction value and the heat map in real time, and the display of the heat map on the three-dimensional map is prior art.

[0038] The above is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. A carbon emission reduction prediction method based on carbon emission data and support vector machines, characterized in that, Including: Step S1: Scan the terrain of the target area through lidar to obtain point cloud data, and synchronously collect real-time environmental parameters; retrieve vegetation distribution data from the satellite remote sensing database. Step S2: Integrate the point cloud data and the vegetation distribution data, use the fractal algorithm to generate a digital terrain model, and label the terrain features, where the terrain features include slope, elevation, and soil type; and load the vegetation carbon absorption rate, growth cycle, and environmental adaptability data into the 3D model through a preset dynamic vegetation parameter library. Step S3: Input the digital terrain model into the trained terrain analysis model, divide the area types through the support vector machine algorithm, and output the vegetation types suitable for planting in the terrain and the priority of air purification species. Step S4: The dynamic carbon emission assessment module determines the carbon emission assessment value according to the real-time environmental parameters and historical data, based on the difference between the sum of energy consumption and traffic emissions in the target area and the carbon emission adsorption data of the planting area. Step S5: When the carbon emission assessment value calculated by the system through the dynamic carbon emission assessment unit exceeds the preset threshold, an emission reduction instruction is automatically generated. The emission reduction solution generation module generates multiple solutions according to the emission reduction instruction, determines the cost performance ratio of each solution according to the ratio of the total cost to the emission reduction cost, and determines that the solution with a cost performance ratio higher than the preset cost performance ratio is the optimal solution.

2. The carbon emission reduction prediction method based on carbon emission data and support vector machine according to claim 1, characterized in that, In step S4: Carbon emission assessment value = ∑(α i × E i ) + ∑(β j × M j ) - ∑(γ k × A k ) Among them, α i is the industrial emission coefficient of the i-th category, E i is the energy consumption, β j is the traffic emission coefficient of the j th category, M j is the mileage, γ k is the vegetation carbon sink coefficient of the k-th category, A k is the planting area.

3. The carbon emission reduction prediction method based on carbon emission data and support vector machine according to claim 1, characterized in that Calculate the cost performance of each solution in step S5: P = ln(Y 总成本 ) / ln(X 减排成本 ), and select the solutions with P > 1.2 to enter the decision-making process.

4. The carbon emission reduction prediction method based on carbon emission data and support vector machine according to claim 1, characterized in that, After step S5, it further includes: Step S6: Display the 3D effect of the solution in an AR overlay manner through the flexible display component, and the user adjusts the parameters with gestures and selects the optimal solution.

5. The carbon emission reduction prediction method based on carbon emission data and support vector machine according to claim 1, characterized in that, The real-time environmental parameters include temperature, humidity, and light intensity. Step S1 further includes retrieving meteorological data from the satellite remote sensing database.

6. The carbon emission reduction prediction method based on carbon emission data and support vector machine according to claim 1, characterized in that, In step S3, "outputting the vegetation types suitable for planting in the terrain and the priority of air purification species" includes: the output result of the terrain analysis model includes recommending fast-growing arbors for areas with a slope <15° and soil organic matter > 2%, recommending drought-tolerant shrubs for arid areas and associating the cost data of drip irrigation technology, and ranking the air purification species according to the PM2.5 adsorption efficiency and VOC degradation efficiency.

7. The carbon emission reduction prediction method based on carbon emission data and support vector machine according to claim 1, characterized in that The cost performance ratio calculation of the emission reduction solution generation module includes the 5-year cycle return rate analysis of the planting cost and carbon sink income of the vegetation plan, and the comparison of the unit area cost and emission reduction efficiency of artificial spraying and drone spraying in the dust removal plan.

8. The carbon emission reduction prediction method based on carbon emission data and support vector machine according to claim 1, wherein The execution of the fractal algorithm in step S2 includes: Performing grid processing on the point cloud data and calculating the elevation standard deviation of each grid. Using the Diamond-Square algorithm to generate terrain details and adjusting the fractal dimension according to the standard deviation. Mapping the planting conditions of different species in the vegetation parameter library to the corresponding grids and labeling the suitability labels.

9. The carbon emission reduction prediction method based on carbon emission data and support vector machine according to claim 2, characterized in that, The AR interaction in step S6 includes: After the camera recognizes the real scene, overlay the 3D vegetation model on the flexible display panel, and the model height is scaled at a real ratio of 1:

100. Adjust the vegetation distribution density according to the instruction requirements, and the system calculates the carbon absorption increment in real time and updates the color of the heat map. If the carbon emission assessment value ≤ the threshold, set this area as a compliant area, trigger the green particle flow special effect, and at the same time the piezoelectric vibration membrane generates a short pulse feedback.

10. The prediction device for the carbon emission reduction prediction method based on carbon emission data and support vector machine according to claim 1, characterized in that, It includes a flexible display component, a digital twin modeling module, a terrain analysis module, a dynamic carbon emission assessment module, and a carbon emission reduction solution generation module. The flexible display component includes multiple sub-display panels and a deformable display frame. The sub-display panels use OLED flexible screens and support curved folding and dynamic splicing. The digital twin modeling module is used to generate a digital terrain model (DTM) through lidar scanning and satellite remote sensing data, and construct a three-dimensional terrain model based on the fractal algorithm. The terrain analysis module inputs the digital terrain model, analyzes the terrain features through machine learning algorithms, and outputs the recommended vegetation types suitable for planting on the terrain and the recommended air purification species. The dynamic carbon emission assessment module collects real-time environmental parameters and historical data, and determines the carbon emission assessment value according to the difference between the sum of energy consumption and traffic emissions in the target area and the carbon emission absorption data of the planting area. When the carbon emission assessment value does not meet the standard, the carbon emission reduction solution generation module generates multiple carbon emission reduction solutions in combination with the terrain analysis results, determines the cost performance of each solution according to the ratio of the total cost to the carbon emission reduction cost, and determines that the solution with a cost performance higher than the preset cost performance is the optimal solution.

Images